H088

Jrg Knblein (Ed.
)
Modern Biopharmaceuticals
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ISBN: 3-527-31184-X
Further Titles of Interest
Gary Walsh
Biopharmaceuticals
Biochemistry and Biotechnology
2003
ISBN 0-470-84326-8
Gary Walsh
Proteins
Biochemistry and Biotechnology
2001
ISBN 0-471-89907-0
Rodney J. Y. Ho, Milo Gibaldi
Biotechnology
and Biopharmaceuticals
Transforming Proteins and Genes into Drugs
2003
ISBN 0-471-20690-3
Chi-Huey Wong (Ed.)
Carbohydrate-based
Drug Discovery
2003
ISBN 3-527-30632-3
Oliver Kayser, Rainer H. Mller (Eds.)
Pharmaceutical Biotechnology
Drug Discovery and Clinical Applications
2004
ISBN 3-527-30554-8
Rainer Fischer, Stefan Schillberg (Eds.)
Molecular Farming
Plant-made Pharmaceuticals and Technical
Proteins
2004
ISBN 3-527-30786-9
Martin Schleef (Ed.)
DNA-Pharmaceuticals
Formulation and Delivery in Gene Therapy
and DNA Vaccination
2005
ISBN 3-527-31187-4
Rolf D. Schmid, Ruth Hammelehle
Pocket Guide to Biotechnology
and Genetic Engineering
2003
ISBN 3-527-30895-4
Volume
Design, Development and Optimization
Edited by
Jrg Knblein
Editor
Dr. Jrg Knblein
Head Microbiological Chemistry
Schering AG
Mllerstrae 178
13342 Berlin
Germany
Library of Congress Card No.:
applied for
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the
British Library
Bibliographic information published by
Die Deutsche Bibliothek
Die Deutsche Bibliothek lists this publication in the
Deutsche Nationalbibliografie; detailed bibliographic data
is available in the Internet at http://dnb.ddb.de.
2005 WILEY-VCH Verlag GmbH & Co. KGaA,
Weinheim
All rights reserved (including those of translation into
other languages). No part of this book may be repro-
duced in any form nor transmitted or translated into
machine language without written permission from the
publishers. Registered names, trademarks, etc. used in
this book, even when not specifically marked as such,
are not to be considered unprotected by law.
Printed in the Federal Republic of Germany
Printed on acid-free paper
Cover Tim Fonseca, www.fonsecatim.com
Typsetting K+V Fotosatz GmbH, Beerfelden
Printing betz-druck GmbH, Darmstadt
Bookbinding J. Schffer GmbH, Grnstadt
ISBN-13 978-3-527-31184-2
ISBN-10 3-527-31184-X
n All books published by Wiley-VCH are carefully pro-
duced. Nevertheless, authors, editors, and publisher do
not warrant the information contained in these books,
including this book, to be free of errors. Readers are
advised to keep in mind that statements, data, illustrati-
ons, procedural details or other items may inadvertently
be inaccurate.
Volume 1
Prologue XXV
Dedication XXIX
Foreword XXXI
Foreword XXXV
Quotes XXXVII
Executive Summary XLI
List of Contributors CXXIII
Introduction
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 1
Gary Walsh
1 What are Biopharmaceuticals? 2
2 A Global Snapshot 2
3 Upstream and Downstream Processing 3
4 Trends in Approvals 6
5 Declining Number of Approvals 8
6 Products Approved for Human Use 9
7 Products Approved for Veterinary Use 25
8 Likely Future Directions 27
9 Concluding Remarks 33
V
Contents
ISBN: 3-527-31184-X
Part I Biopharmaceuticals Used in Molecular Medicine
From Genome to Clinic Correlation Between Genes, Diseases and Biopharmaceuticals 37
1 Beginning to Understand the End of the Chromosome 37
Thomas R. Cech
1.1 Introduction 37
1.2 Telomere Terminal Transferase 38
1.3 Telomerase Contains an Essential RNA 38
1.4 Finally, the Protein: Telomerase Reverse Transcriptase 39
1.5 Current Picture of Telomerase 40
1.6 Regulation of Telomerase 42
1.7 Cellular Immortality 44
1.8 Cancer 44
2 The Role of Pharmacogenetics/Pharmacogenomics
in Drug Development and Regulatory Review: Current Status 49
Shiew-Mei Huang and Lawrence J. Lesko
2.1 Introduction 50
2.2 Variability in Drug Response 50
2.3 Drug-metabolizing Enzymes and Transporters 52
2.4 Applications of Pharmacogenetics and Pharmacogenomics in Drug Development
and Regulatory Review 54
2.5 Determination of Different Genotype Groups based on Known Valid
and Probable Valid Biomarkers 56
2.6 Drug Interactions 60
2.7 Voluntary versus Required Submissions 60
2.8 Labeling Implications 63
2.9 Conclusion 64
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 71
Joerg Geistlinger and Peter Ahnert
3.1 Genetic Variation, Disease Susceptibility and Drug Response 73
3.2 Pharmacogenetics and Pharmacogenomics 74
3.3 Personalized Medicine 76
3.4 SNPs in Clinical Applications 78
3.5 Strategies in SNP Discovery 80
3.6 SNP Technologies 83
3.7 Polydimensional SNP-Chips: The Array-On Technology 88
3.8 Outlook 93
4 A Systems Biology Approach to Target Identification and Validation
for Human Chronic Disease Drug Discovery 99
Bonnie E. Gould Rothberg, Carol E. A. Pena, and Jonathan M. Rothberg
4.1 Limitations in the Chronic Disease Drug Discovery Process 100
Contents VI
4.2 Creating the Pharmaceutically Tractable Genome 104
4.3 Integrated Systems Biology Approaches to Drug Target Validation for Specific
Clinical Indications 110
4.4 Conclusion 123
5 The Development of Herceptin
:
Paving the Way for Individualized Cancer Therapy 127
Thorsten S. Gutjahr and Carsten Reinhardt
5.1 Introduction 128
5.2 HER2 129
5.3 Herceptin Mechanism of Action and Effects on Cellular Processes 130
5.4 Preclinical Evidence 131
5.5 HER2 Testing as a Prerequisite for Herceptin Therapy: Development
of Commercially Available and Validated Testing Methodologies 133
5.6 HER2 Testing Algorithm 135
5.7 Herceptin in Clinical Use 136
5.8 Future Prospects for Herceptin and other Targeted Therapies 143
5.9 Herceptin in Early Breast Cancer 143
5.10 Herceptin Adjuvant Trials 143
5.11 Conclusion 145
siRNA the Magic Bullet and Other Gene Therapeutical Approaches 151
6 Adenovirus-based Gene Therapy: Therapeutic Angiogenesis
with Adenovirus 5 Fibroblast Growth Factor-4 (Ad5FGF-4) in Patients
with Chronic Myocardial Ischemia 151
Michael McCaman, Francisco J. Castillo, Farah Fawaz, Yasushi Ogawa, Erik Whiteley,
Elisabeth Lehmberg, Mei Tan, Jacob Kung, Bruce Mann, Erno Pungor Jr.,
and Gabor M. Rubanyi
6.2 Therapeutic Angiogenesis and the Importance of Collateral Vessels 153
6.3 Designing an Intervention Suitable for Therapeutic Angiogenesis 153
6.4 Production and Characterization of the Ad5FGF-4 Vector 156
6.5 Pre-clinical Efficacy and Safety of Ad5FGF-4 in Pigs 172
6.6 Clinical Studies 175
6.7 Summary and Conclusions 178
7 MIDGE Vectors and dSLIM Immunomodulators:
DNA-based Molecules for Gene Therapeutic Strategies 183
Manuel Schmidt, Barbara Volz, and Burghardt Wittig
7.1 Vectors for Gene Therapy 184
7.2 Immunomodulatory Molecules 193
7.3 Application of MIDGE Vectors and dSLIM Immunomodulators 198
Contents VII
8 Nonprotein-coding RNAs and their Potential as Biopharmaceuticals 213
Maciej Szymanski, Jan Barciszewski and Volker A. Erdmann
8.2 The Contents of the Genomes 214
8.3 npcRNAs 215
8.4 Functions of npcRNAs 217
8.5 npcRNAs and Human Diseases 219
8.6 miRNAs 222
8.7 Future Prospects 223
9 Double-stranded Decoy Oligonucleotides as new Biopharmaceuticals 229
Andreas H. Wagner and Heiko E. von der Leyen
9.2 Therapeutic Decoy ODN Application 232
10 Rational siRNA Design for RNA Interference:
Optimizations for Therapeutic Use and Current Applications 243
Anastasia Khvorova, Queta Boese, and William S. Marshall
10.1 RNAi: History and Mechanism 244
10.2 Early siRNA Design Parameters 248
10.3 Current siRNA Design Considerations 251
10.4 Therapeutic Applications of RNAi 259
10.5 Summary: The Future of RNAi in Biopharmaceutical Development 264
Mobilis in Mobile Human Embryonic Stem Cells and Other Sources for Cell Therapy 269
11 The First Cloned Human Embryo: An Unlimited Source of Stem Cells
for Therapeutic Cloning 269
Woo Suk Hwang, Byeong Chun Lee, Sung Keun Kang, and Shin Yong Moon
11.2 Human Somatic Cell Nuclear Transfer (SCNT) 270
11.3 Establishment and Characterization of Human SCNT ES Cells 276
11.4 Reprogramming Adult Cells into an Embryonic State 277
11.5 Discussion and Conclusion 279
12 Myocardial Regeneration Strategies using Human Embryonic Stem Cells 283
Izhak Kehat, Oren Caspi, and Lior Gepstein
12.2 Derivation of Human Embryonic Stem Cells 286
12.3 Cardiomyocyte Differentiation of ES Cells 289
12.4 Possible Research and Clinical Applications of the hES-derived
Cardiomyocytes 293
12.5 Early Cardiac Lineage Differentiation 293
12.6 Myocardial Regeneration Strategies using hES-derived Cardiomyocytes 295
Contents VIII
12.7 Functional Integration of the Cell Grafts 296
12.8 Cardiomyocyte Enrichment, Purification, and Up-scaling Strategies 298
12.9 Prevention of Immunological Rejection 299
12.10 Conclusions 300
13 Gene and Cell-based Therapies for Cardiovascular Disease 305
Abeel A. Mangi
13.2 Gene Therapy as Novel Drug Delivery 306
13.3 Cell-based Gene Therapy and Regenerative Cardiovascular Medicine 319
13.4 Future Directions and Challenges 321
14 Spheramine
: A Cell Therapeutic Approach to Parkinsons Disease 325

Elke Reissig, Hermann Graf, and Friedrich-Joachim Kapp
14.2 PD 326
14.3 Spheramine 334
14.4 Randomized, Double-blind, Placebo-controlled Multicenter Study of the Safety,
Tolerability and Efficacy of Spheramine Implanted Bilaterally
into the Postcommissural Putamen of Patients with Advanced PD 343
14.5 Summary and Outlook 348
15 Applying Human Cells to Organogenesis and Transplantation 353
Benjamin Dekel and Yair Reisner
15.1 Growing Demands for Kidney Allograft Transplantation 354
15.2 Alternative Sources for Human Renal Allografts 354
Volume 2
Part II Biopharmaceuticals and Their Mode of Action
Quid pro Quo Lysis vs. Coagulation in the Fine-tuned Balance of the Clotting Cascade 377
1 Mechanisms of Serine Proteinase Activation: Insights for the Development
of Biopharmaceuticals for Coagulation and Fibrinolysis 377
Rainer Friedrich
1.2 Bacterial Activators of Host Zymogens 381
1.3 Some Remarks on Nonproteolytic Activators 388
2 Application of the Principle of Polyvalency to Protease Inhibition 395
Luis Moroder
Contents IX
2.2 Thermodynamic Model of Bivalent Ligand Binding 396
2.3 Homo- and Heterobivalent Inhibitors of the Yeast 20S Proteasome 398
2.4 Bivalent Inhibition of Mast Cell b-Tryptase 405
2.5 Heterobivalent Inhibition of Thrombin 411
2.6 Perspectives 414
3 A New Technology Standard for Safety and Efficacy in Factor VIII Replacement Therapy:
Designing an Advanced Category rFVIII Concentrate 419
Norbert Riedel and Friedrich Dorner
3.2 Development of rFVIII 428
3.3 Production of rFVIII 430
3.4 Pathogen Safety 433
3.5 Quality Control 435
3.6 Purity and Potency 435
3.7 Preclinical Studies 436
3.9 Summary 447
Errare Humanum Est What Causes Cancer and How to Selectively Fight Tumors 451
4 Biopharmaceutical Drugs from Natural Sources 451
David J. Newman, Gordon M. Cragg, and Barry R. OKeefe
4.1 Biotechnologically Produced Proteins and Peptides as Approved Drugs 452
4.2 Potential Agents from Non-mammalian Sources as Leads
to Novel Therapies 481
4.3 Overall Concluding Comments 488
5 Biopharmaceuticals as Targeting Vehicles for In situ Radiotherapy
of Malignancies 497
Raymond M. Reilly
5.2 Principles of Targeted In situ Radiotherapy of Malignancies 499
5.3 RIT of Non-Hodgkins B-Cell Lymphomas: The Pre-eminent Success Story 500
5.4 Other Strategies for In situ Radiotherapy of Non-Hodgkins Lymphoma 505
5.5 Radioimmunotherapy of AML: Success but not Cure 505
5.6 RIT of Solid Tumors: Encouraging Results n Minimal Residual Disease 507
5.7 Pre-Targeting Strategies: Improving the Therapeutic Index of RIT 511
5.8 Peptide-Directed In situ Radiotherapy: Targeting Somatostatin Receptors 516
5.9 Auger Electron Radiotherapy: Anti-tumor Effects at the Single Cell Level 519
5.10 a-Particle RIT: Anti-tumor Effects at the Multi-cell Level 525
5.11 Conclusion 526
Contents X
6 New Directions in Tumor Therapy
Amino Acid Deptetion with GlutaDON
as Treatment for Cancer 537

Rolf Kalhammer and Natarajan Sethuraman
6.1 Rationale for GlutaDON
Therapy 537
6.3 PEGylation and Protection from Inactivation 541
6.4 Toxicology 545
6.5 Clinical Trial 545
Mundus Vult Decipi High Mutation Rates of HIV and New Paradigms for Treatment 549
7 AIDS Gene Therapy: A Vector Selectively Able to Destroy Latently HIV-1-infected
Cells 549
Francisco Luque Vzquez and Ricardo Oya
7.1 The Genes and Life Cycle of HIV-1 551
7.2 Gene Therapy of AIDS 553
7.3 Viral Latency: the Real Challenge 557
7.4 A Vector Able Selectively to Destroy Latently Infected Cells 559
8 Combinatorial RNA-based Therapies for HIV-1 569
Kevin V. Morris and John J. Rossi
8.2 RNA-based Antiviral Agents 570
8.3 RNAi: Diversity of Viral Targets 571
8.4 Delivery of siRNAs to Target Cells 573
8.5 Challenges for RNA-based Therapies 577
8.6 Summary and Conclusion 577
Part III Improving the Development of Biopharmaceuticals
Citius, Altius, Fortius Acceleration by High Throughput and Ultra-HT 583
1 Design of Modern Biopharmaceuticals by Ultra-high-throughput Screening
and Directed Evolution 583
Markus Rarbach, Wayne M. Coco, Andre Koltermann, Ulrich Kettling,
and Manfred Eigen
1.1 Modern Biopharmaceuticals 584
1.2 Directed Evolution Fundamentals 585
1.3 Generation of Protein Diversity 586
1.4 Selection Strategies 593
1.5 High-throughput and High-content Screening of Protein Libraries 594
1.6 Directed Evolution of Biopharmaceuticals 598
1.7 Conclusions 601
Contents XI
2 Learning from Viruses: High-throughput Cloning using the Gateway
System
to Transfer Genes without Restriction Enzymes 605
Jonathan D. Chesnut
2.2 Background 606
2.3 Engineering the Lambda System to Create Gateway 609
2.4 The Gateway Reactions 610
2.5 Creating Gateway Entry Clones 611
2.6 Gateway Destination Vectors 613
2.7 Applications Enabled by Gateway Cloning 614
2.8 HTP Expression Analysis in Mammalian Cells 614
2.9 HTP Cloning and Expression in a Baculovirus System 615
2.10 Multisite Gateway 616
2.11 Creation of Entry Vectors and Three-fragment Multisite Assembly Reaction 618
2.12 Perspective 621
In Vivo Veritas Early Target Validation in Knock-out Mice and More 621
3 Target Validation: An Important Early Step in the Development
of Novel Biopharmaceuticals in the Post-genomic Era 621
Christoph P. Bagowski
3.2 RNA- and DNA-based Techniques for Post-transcriptional Regulation
of Molecular Targets, and their Potential as Biopharmaceutical Drugs 624
3.3 Peptide and Protein-based Approaches 636
3.4 Protein Kinases as Targets for Drug Development 639
3.5 Cell-based Assays for In vitro Target Validation in the Drug Discovery
Process 640
3.6 Animal Models as the Ultimate Target Validation 645
4 Genetically Modified Mice in Medical and Pharmaceutical Research 649
Cord Brakebusch
4.1 Disease-oriented Research in Genetically Modified Mice 649
4.2 Generation of Genetically Modified Mice by Gene Targeting 651
4.3 Analysis of Genetically Modified Mice 659
4.4 Alternative Methods 659
5 An NIH Model Organism for Biopharmaceutical and Biomedical Research:
The Lower Eukaryote Dictyostelium discoideum 661
Thomas Winckler, Ilse Zndorf, and Theodor Dingermann
5.2 The Gene Discovery Tool Box or Dictyostelium Research 665
5.3 Production of Recombinant Proteins in D. discoideum 672
Contents XII
5.4 Dictyostelium discoideum in Biomedical Research 685
5.5 Conclusions 689
Revolution by Evolution Rational Design for Desire and Scientific Art of Optimization 695
6 Releasing the Spring: Cofactor- and Substrate-assisted Activation of Factor IXa 695
Hans Brandstetter and Katrin Sichler
6.2 The Zymogen Form of fIX is Fully Inactive 697
6.3 Relevance of Tyr99 on the Stability of the 99-loop 697
6.4 Lys98 Hinders Substrate Binding to fIXa both Sterically and Electrostatically 698
6.5 Tyr177 Locks the 99-loop in an Inactive Conformation, which is Released
by Cofactor fVIIIa and Modified by the Physiologic Substrate fX 699
6.6 S1 Site Mutations Decrease the Activity of fIXa 699
6.7 Evolutionary Relation of fIXa and fXa is Reflected in the Dependence
of Activity Changes on Arg/Lys Substrates 700
6.8 By Binding at the 60-loop Ethylene Glycol Indirectly Reorganizes the 99-loop
and Allosterically Stimulates the Activity of fIXa 700
7 Accelerating Diagnostic Product Development Process with Molecular Rational Design
Harald Sobek, Rainer Schmuck, and Zhixin Shao
7.2 Strategies for Optimizing Diagnostic Proteins 705
7.3 Examples 709
7.4 Summary 717
Volume 3
Part IV Production of Biopharmaceuticals
The Industrys Workhorses Mammalian Expression Systems 723
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian
Cells in Bioreactors 723
Florian M. Wurm
1.2 Vectors, Transfections, and Cell Line Generation 727
1.3 Host Cell Engineering 731
1.4 Gene Transfer and Gene Amplification in Mammalian Cells 733
1.5 Production Principles for Mammalian Cells: Anchorage-dependent Cultures and
Suspension Cultures 737
1.6 Large-scale Transient Expression 744
Contents XIII
1.7 Regulatory Issues 745
1.8 Concluding Remarks 751
2 Alternative Strategies and New Cell Lines for High-level Production
of Biopharmaceuticals 761
Thomas Rose, Karsten Winkler, Elisabeth Brundke, Ingo Jordan and Volker Sandig
2.1 Mammalian Cells as a Workhorse to Produce Protein-based
Biopharmaceuticals 761
2.2 The Cell Line of Choice 762
2.3 Pushing Expression Levels Impact of Vector Design and Cell Clone
Selection 764
2.4 A Single CHO High-producer Clone for Multiple Products 766
2.5 The G-line: Use of the Immunoglobulin Locus of a Human/Mouse
Heterohybridoma for Heterologous Gene Expression 769
2.6 Human Designer Cell Lines 774
3 PER.C6
Cells for the Manufacture of Biopharmaceutical Proteins 779

Chris Yallop, John Crowley, Johanne Cote, Kirsten Hegmans-Brouwer, Fija Lagerwerf,
Rodney Gagne, Jose Coco Martin, Nico Oosterhuis, Dirk-Jan Opstelten,
and Abraham Bout
3.2 Generation of PER.C6 Cells 782
3.3 PER.C6 Cells for the Manufacture of Recombinant Proteins 784
3.4 Fed-batch Process Development 789
3.5 Operation of PER.C6 Cells in Continuous Perfusion 794
3.6 Characterization of Antibodies Produced by PER.C6 Cells 797
3.7 Conclusion 803
4 Use of the Glutamine Synthetase (GS) Expression System for the Rapid Development
of Highly Productive Mammalian Cell Processes 809
John R. Birch, David O. Mainwaring, and Andrew J. Racher
4.2 Cell Line Construction and Selection 810
4.3 Cell Line Stability 818
4.4 Cell Engineering to Increase Productivity 819
4.5 Selection of Useful Cell Sub-populations 822
4.6 Process Development 823
4.7 Summary 830
Vivat, Crescat, Floreat A Ripe and Blooming Market for Transgenic Animals and Plants 833
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 833
Julio Baez
Contents XIV
5.2 Advantages and Disadvantages of Transgenic Systems for the Production
5.3 Commercial Biopharmaceuticals with Human Clinical Experience
for Therapeutic, Immunoprophylactic, and Medical Device Use
derived from Transgenic Systems 852
5.4 Conclusions 873
6 Production of Recombinant Proteins in Plants 893
Victor Klimyuk, Sylvestre Marillonnet, Jrg Knblein, Michael McCaman,
and Yuri Gleba
6.2 Plant-based Expression Systems 894
6.3 Plant-made Recombinant Proteins available Commercially,
and under Development 903
6.4 Comparative Analysis of the Expression Systems and Production Platforms 907
7 Humanized Glycosylation: Production of Biopharmaceuticals
in a Moss Bioreactor 919
Gilbert Gorr and Sabrina Wagner
7.2 Mosses: Some General Aspects 920
7.3 Cell Culture 922
7.4 Recombinant Expression 923
7.5 N-Glycosylation 924
7.6 Conclusions and Outlook 927
8 ExpressTec: High-level Expression of Biopharmaceuticals in Cereal Grains 931
Ning Huang and Daichang Yang
8.2 Development of ExpressTec for High-level Expression of Recombinant Proteins
in Cereal Grains 932
8.3 High-level Expression of Biopharmaceuticals in Cereal Grain
using ExpressTec 938
8.4 Impact of Expression Level on the Cost of Goods 945
8.5 Perspectives of Expressing Biopharmaceuticals in High Plants 946
9 Biopharmaceutical Production in Cultured Plant Cells 949
Stefan Schillberg, Richard M. Twyman, and Rainer Fischer
9.2 Recombinant Proteins Produced in Plant Cell Suspension Cultures 951
9.3 Challenges and Solutions for the Production of Recombinant Proteins 954
9.4 Process Engineering 958
9.5 Downstream Processing 959
Contents XV
9.6 Regulatory Considerations 960
9.7 Conclusions 961
10 Producing Biopharmaceuticals in the Desert: Building an Abiotic Stress Tolerance
in Plants for Salt, Heat, and Drought 967
Shimon Gepstein, Anil Grover, and Eduardo Blumwald
10.1 General Comments on Abiotic Stresses 968
10.2 Drought and Salt Tolerance 969
10.3 High-temperature Stress 981
10.4 Conclusions and Perspectives 989
11 The First Biopharmaceutical from Transgenic Animals: ATryn
995
Yann Echelard, Harry M. Meade, and Carol A. Ziomek
11.2 Recombinant Production of AT 998
11.3 Characterization of rhAT 1003
11.5 Clinical Trials with rhAT 1011
Alea Non Iacta Est Improving Established Expression Systems 1021
12 Producing Modern Biopharmaceuticals: The Bayer HealthCare Pharma Experience
with a Range of Expression Systems 1021
Heiner Apeler
12.1 The Escherichia coli Expression Platform 1022
12.2 The Saccharomyces cerevisiae Expression Platform 1027
12.3 The HKB11 Expression Platform 1029
12.4 Outlook and Conclusion 1031
13 Advanced Expression of Biopharmaceuticals in Yeast at Industrial Scale:
The Insulin Success Story 1033
Asser Sloth Andersen and Ivan Diers
13.2 Design and Optimization of the Insulin Precursor Molecule 1036
13.3 Production of Insulin 1041
13.4 Conclusions and Future Aspects 1042
14 Baculovirus-based Production of Biopharmaceuticals using Insect Cell Culture
Processes 1045
Wilfried Weber and Martin Fussenegger
14.2 Molecular Tools for the Construction of Transgenic Baculoviruses 1046
14.3 Insect Cell Culture 1047
14.4 Insect Cell Glycosylation and Glycoengineering 1047
Contents XVI
14.5 Nutrient and Kinetic Considerations for Optimized BEVS-based Protein
Production 1048
14.6 Scaling-up Baculovirus-based Protein Production 1050
14.7 Generic Protocol of Optimized Protein Production 1050
14.8 Case study: Rapid Optimization of Expression Conditions and Large-scale
Production of a Brutons Tyrosine Kinase Variant (BTK) 1053
14.9 Conclusion 1058
15 Robust and Cost-effective Cell-free Expression of Biopharmaceuticals:
Escherichia Coli and Wheat Embryo 1063
Luke Anthony Miles
15.2 Transcription 1066
15.3 Translational 1068
15.4 Treatment of Extracts for Synthesis of Disulfide-bonded Proteins 1072
15.5 ATP Regeneration Systems 1074
15.6 Reaction Conditions 1075
When Success Raises its Ugly Head Outsourcing to Uncork the Capacity Bottleneck 1083
16 Contract Manufacturing of Biopharmaceuticals Including Antibodies
or Antibody Fragments 1083
J. Carsten Hempel and Philipp N. Hess
16.2 Expression Systems and Manufacturing Procedures 1085
16.3 Outsourcing and Contract Manufacturing 1089
Part V Biopharmaceuticals used for Diagnositics and Imaging
From Hunter to Craftsman Engineering Antibodies with Natures Universal Toolbox 1105
1 Thirty Years of Monoclonal Antibodies:
A Long Way to Pharmaceutical and Commercial Success 1105
Uwe Gottschalk and Kirsten Mundt
1.2 Making Monoclonal Antibodies 1109
1.3 Other Antibody Formats: Antibody Fragments 1113
1.4 Medical Application Areas for MAbs 1116
1.5 From Initial Failure to Success: Getting the Target Right 1117
1.6 The Market Perspective 1119
1.7 Drug Targeting: The Next Generation in Cancer Treatment 1122
1.8 Developing a Manufacturing Process for MAbs 1126
1.9 Routine Manufacture of MAbs 1127
Contents XVII
1.10 Glycosylation and Other Post-translational Modifications 1132
1.11 Emerging Issues in MAb Production 1134
1.12 The Future of MAbs 1136
2 Modern Antibody Technology: The Impact on Drug Development 1147
Simon Moroney and Andreas Plckthun
2.2 Immunogenicity 1148
2.3 Technology 1153
2.4 Reaching the Target: The Importance of Specificity, Affinity and Format 1163
2.5 Exerting an Effect at the Target 1168
2.6 Antibodies in their Natural Habitat: Infectious Diseases 1175
2.7 Opportunities for New Therapeutic Applications Provided
by Synthetic Antibodies 1176
2.8 Future Directions and Concluding Statements 1177
3 Molecular Characterization of Autoantibody Responses in Autoimmune Diseases:
Implications for Diagnosis and Understanding of Autoimmunity 1187
Constanze Breithaupt
3.1 Autoantibodies in Autoimmune Diseases 1188
3.2 Autoantibody Epitopes 1190
3.3 Visualization of Epitopes 1195
3.4 Structural Characterization of AutoantibodyAutoantigen Complexes 1199
Find, Fight, and Follow Target-specific Troika from Mother Natures Pharmacopoiea 1211
4 Molecular Imaging and Applications for Pharmaceutical R&D 1211
Joke G. Orsel and Tobias Schaeffter
4.2 Imaging Modalities and Contrast Agents 1213
4.3 Molecular Imaging 1225
4.4 Molecular Imaging for Drug Discovery and Development 1230
5 Design and Development of Probes for In vivo Molecular and Functional Imaging
of Cancer and Cancer Therapies by Positron Emission Tomography (PET) 1243
Eric O. Aboagye
5.1 What is Positron Emission Tomography? 1244
5.2 Radiochemistry Considerations 1246
5.3 Pharmacological Objectives in Oncology Imaging Studies 1249
5.4 The Use of Radiolabeled Drugs to Image Tumor and Normal Tissue
Pharmacokinetics 1250
Contents XVIII
5.5 Pharmacodynamic Studies 1254
6 Ligand-based Targeting of Disease:
From Antibodies to Small Organic (Synthetic) Ligands 1271
Michela Silacci and Dario Neri
6.2 Ligands 1273
6.3 Classes of Diseases 1276
6.4 From a Ligand to a Product 1288
7 Ultrasound Theranostics: Antibody-based Microbubble Conjugates as Targeted In vivo
Contrast Agents and Advanced Drug Delivery Systems 1301
Andreas Briel, Michael Reinhardt, Mathias Murer, and Peter Hauff
7.1 Motivation: Find, Fight and Follow! 1302
7.2 Ultrasound: Hear the Symptoms 1304
7.3 Ultrasound Contrast: Tiny Bubbles 1305
7.4 The Perfect Modality: Sensitive Particle Acoustic Quantification (SPAQ) 1308
7.5 Targeting and Molecular Imaging: The Sound of an Antibody 1309
7.6 Drug Delivery: The Magic Bullet 1315
7.7 Ultrasound, Microbubbles and Gene Delivery:
Noninvasive Micro-Gene Guns 1318
7.8 Summary: Ultrasound Theranostics
Building a Bridge between Therapy and Diagnosis 1320
Getting Insight Sense the Urgency for Early Diagnostics 1325
8 Development of Multi-marker-based Diagnostic Assays with the ProteinChip
System 1325
Andreas Wiesner
8.1 The Urgency of Earlier Diagnosis 1326
8.2 Proteins are Best Choice Again 1327
8.3 Current Tools for Protein Biomarker Detection 1328
8.4 The ProteinChip
System at a Glance 1329

8.5 Distinctions of the SELDI Process 1333
8.6 The Pattern Track
TM
Process:
From Biomarker Discovery to Assay Development 1334
8.7 Protein Variants as Disease Markers 1337
8.8 Conclusion and Outlook 1338
9 Early Detection of Lung Cancer: Metabolic Profiling of Human Breath with Ion Mobility
Spectrometers 1343
Jrg Ingo Baumbach, Wolfgang Vautz, Vera Ruzsanyi, and Lutz Freitag
Contents XIX
9.2 Material and Methods: IMS 1345
9.3 Results and Discussion 1347
9.4 Clinical Study 1349
Volume 4
Part VI Advanced Application Routes for Biopharmaceuticals
Getting Inside Quest for the Best and How to Improve Delivery 1361
1 Advanced Drug Delivery Systems for Biopharmaceuticals 1361
Gesine E. Hildebrand and Stephan Harnisch
1.2 Challenges for the Administration of Biopharmaceuticals 1363
1.3 Drug Delivery Strategies 1366
1.4 Outlook 1384
Pathfinder New Ways for Peptides, Proteins and Co 1393
2 Poly(ethylene) Glycol Conjugates of Biopharmaceuticals in Drug Delivery 1393
Michael D. Bentley, Mary J. Bossard, Kevin W. Burton, and Tacey X. Viegas
2.2 The Polymer 1394
2.3 Safety and Disposition of PEG 1396
2.4 PEG Reagents and Conjugation 1397
2.5 Biopharmaceutical Conjugates 1400
2.6 PEGylation of Peptides 1407
2.7 Formulations of PEGylated Biopharmaceuticals 1408
2.8 Analysis of PEG-conjugates 1411
2.9 Summary and Future Outlook 1415
3 Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems 1419
Karen Lingnau, Christoph Klade, Michael Buschle, and Alexander von Gabain
3.1 Vaccines and their Importance in the Fight against Human Diseases 1420
3.2 Adjuvants: An Overview 1423
3.3 Cationic Peptides as Novel Vaccine Adjuvants 1426
3.4 Cationic Antimicrobial Peptides (CAMP) as Novel Adjuvants 1433
3.5 Cationic Peptide Delivery Systems in Combination with Other Adjuvants 1437
3.6 The Development of IC31 and Future Prospects 1440
Contents XX
4 The Evolving Role of Oralin
TM
(Oral Spray Insulin) in the Treatment of Diabetes
using a Novel RapidMist
TM
Diabetes Management System 1445
Pankaj Modi
4.2 Rationale for Oralin
TM
Development 1446
4.3 The Benefits of Oralin
TM
1447
4.4 The Preparation and Pharmaceutical Properties of Oralin
TM
1448
4.5 Phase II, Long-term Safety and Efficacy Study 1457
5 Improvement of Intestinal Absorption of Peptide and Protein Biopharmaceuticals
by Various Approaches 1463
Akira Yamamoto
5.1 Improvement of Peptide and Protein Absorption 1464
5.2 Use of Protease Inhibitors 1467
5.3 Chemical Modification of Peptide and Protein Biopharmaceuticals 1472
5.4 Chitosan Capsules for the Colon-specific Delivery of Insulin 1480
5.3 Conclusion 1484
Via Mala the Stoney Road of DNA Delivery: Back-pack, Feed-back, and Pay-back 1487
6 DNA Vaccine Delivery from Poly(ortho ester) Microspheres 1487
Chun Wang, Herman N. Eisen, Robert Langer, and Jorge Heller
6.2 Poly(Ortho Esters) 1494
6.3 Preparation and Characterization of Microspheres 1496
6.4 In vivo Evaluation of Immune Responses 1500
7 Liposomal In vivo Gene Delivery 1507
Shigeru Kawakami, Fumiyoshi Yamashita, and Mitsuru Hashida
7.1 Cationic Charge-mediated In vivo Gene Transfer to the Lung 1510
7.2 Asialoglycoprotein Receptor-mediated In vivo Gene Transfer to Hepatocytes 1512
7.3 Mannose Receptor-mediated In vivo Gene Transfer to Macrophages 1513
7.4 Folate Receptor-mediated In vivo Gene Transfer to Cancer Cells 1515
7.5 Transferrin Receptor-mediated In vivo Gene Transfer to Brain 1517
8 Programmed Packaging:
A New Drug Delivery System and its Application to Gene Therapy 1521
Kentaro Kogure, Hidetaka Akita, Hiroyuki Kamiya, and Hideyoshi Harashima
8.1 New Concept for Gene Delivery 1521
8.2 Controlled Intracellular Trafficking 1525
Contents XXI
8.3 Transgene Expression and Gene Correction 1531
8.4 Towards Clinical Applications of Transgene Expression
and Gene Correction 1534
Getting Beyond Rocket Science vs. Science Fiction 1537
9 Bionanotechnology and its Role to Improve Biopharmaceuticals 1537
Oliver Kayser
9.2 Drug and Gene Delivery 1539
9.3 Gene Delivery 1543
9.4 Biosensors 1544
9.5 Implants and Tissue Engineering 1546
9.8 Safety Aspects 1548
9.7 Conclusions and Future Trends 1550
Part VII From Transcription to Prescription of Biopharmaceuticals
Dosis Facit Venenum The Therapeutic Window between Systemic Toxicity
and Lack of Efficacy 1557
1 Analytics in Quality Control and In vivo 1557
Michael Hildebrand
1.3 Classes of Biopharmaceuticals 1560
1.4 Analytical Methods and Specifications 1560
1.5 International Guidelines on Quality Control 1571
1.6 Analytics In vivo 1573
2 Design, Development and Optimization: Crystal Structures
of Microsomal Cytochromes P450 1581
Dijana Matak Vinkovic, Sheena Whyte, Harren Jhoti, Jose Cosme,
and Pamela A. Williams
2.1 P450: The Background 1581
2.2 Importance of P450s for Drug Development 1582
2.3 Variability and Drug Metabolism 1583
2.4 The Structure of Cytochrome P450 1584
3 Mettox
TM
: A Suite of Predictive In silico and In vitro Assays for Metabolic
and Genotoxicological Profiling of Preclinical Drug Candidates 1603
Michael Murray
3.1 Issues and Economics of Early ADMET (Absorption, Distribution, Metabolism,
Excretion and Toxicity) Assessment 1604
Contents XXII
3.2 Phase I Metabolism Prediction: Computational Approaches 1608
3.3 Phase I Metabolism Prediction: In vitro Techniques 1613
3.4 Genotoxicity Prediction 1624
Happy End: Claim to Fame and Approval 1637
4 Considerations for Developing Biopharmaceuticals: FDA Perspective 1637
Kurt Brorson, Patrick G. Swann, Janice Brown, Barbara Wilcox,
and Marjorie A. Shapiro
4.2 Regulatory Authority 1639
4.3 Overview of Product Development: CMC Perspective 1643
4.4 Chemistry, Manufacturing and Controls Considerations 1645
4.5 Quality Control and Assurance 1647
4.6 Microbial Issues Specific to Biopharmaceuticals 1650
4.7 Process Validation 1653
4.8 Inspectional Considerations 1653
4.9 Biotech Development:
Lessons Learned and Issues Overcome by Industry and FDA 1654
4.10 FDA Initiatives to Improve the Pharmaceutical and Biopharmaceutical
Development Process 1661
5 The Regulatory Environment for Biopharmaceuticals in the EU 1669
Axel F. Wenzel and Carina E. A. Sonnega
5.2 History and Background 1673
5.3 The Competent Regulatory Bodies 1676
5.4 What is the EU Authorities Definition of a Biotechnological Product? 1681
5.5 The Regulatory Framework 1682
5.6 CP: The Biotech Procedure 1683
5.7 From Transcription to Prescription:
What is Different for Biotechnological Drugs? 1688
5.8 Biogenerics 1700
Part VIII From Bench to Bedside The Aftermaths
Think Big and Dealmaking for Growth Global Changes in the Health-care Sector 1711
1 Healthcare Trends and their Impact on the Biopharmaceutical Industry:
Biopharmaceuticals Come of Age 1711
Alexander Moscho, Markus A. Schfer, and Kristin Yarema
1.2 Despite Robust Demand the Industry Faces Severe Challenges 1713
Contents XXIII
1.3 Why Biopharmaceuticals can Succeed in Rougher Markets 1724
1.4 Biopharmaceutical Players Will Need to Adapt their Portfolios
and Business Models 1728
News and Views Quo Vadis, Biopharmaceuticals? 1741
2 mondoBIOTECH: The Swiss biotech BOUTIQUE 1741
Dorian Bevec and Fabio Cavalli
2.1 Introduction
2.2 Product Platforms 1742
2.3 Interferon-c + Genechip 1750
2.4 Bacteriophages 1751
2.5 Outlook for the Company 1752
3 G-CSF and Bioequivalence: The Emergence of Healthcare Economics 1755
James Harris, III
3.2 Biogenerics and Bioequivalence 1756
Light at the End of the Tunnel or Back to the Roots? 1771
4 Bioinformatics: From Peptides to Profiled Leads 1771
Paul Wrede and Matthias Filter
4.2 Basic Concepts of Virtual Drug Discovery 1773
4.3 Pep2Lead Concept 1778
4.4 ADMETox Profiling 1785
4.5 Outlook 1798
5 Engineering and Overproduction of Polyketide Natural Products 1803
Martha Lovato Tse and Chaitan Khosla
5.2 Polyketide Synthases 1806
5.3 Engineering PKSs to Produce Novel Polyketides 1815
5.4 Development of Scalable Production Processes 1820
Epilog 1833
More about the Editor 1835
Supplement CD-ROM 1837
Subject Index 1841
Contents XXIV
Volume 1
Prologue XXV
Dedication XXIX
Foreword XXXI
Foreword XXXV
Quotes XXXVII
Introduction
Gary Walsh
Thomas R. Cech
V
Contents
ISBN: 3-527-31184-X
:
Abeel A. Mangi
Contents VI
14 Spheramine

Volume 2
Rainer Friedrich
Luis Moroder
2.6 Perspectives 414
3.9 Summary 447
Contents VII
4.2 Potential Agents from Non-mammalian Sources as Leads
to Novel Therapies 481
4.3 Overall Concluding Comments 488
of Malignancies 497
Raymond M. Reilly
5.4 Other Strategies for In situ Radiotherapy of Non-Hodgkins Lymphoma 505
5.6 RIT of Solid Tumors: Encouraging Results n Minimal Residual Disease 507
5.11 Conclusion 526

6.1 Rationale for GlutaDON
Therapy 537
6.4 Toxicology 545
Cells 549
Contents VIII
8.2 RNA-based Antiviral Agents 570
8.5 Challenges for RNA-based Therapies 577
and Manfred Eigen
1.1 Modern Biopharmaceuticals 584
1.7 Conclusions 601
System
Jonathan D. Chesnut
2.2 Background 606
2.4 The Gateway Reactions 610
2.7 Applications Enabled by Gateway Cloning 614
2.8 HTP Expression Analysis in Mammalian Cells 614
2.11 Creation of Entry Vectors and Three-fragment Multisite Assembly Reaction 618
2.12 Perspective 621
Contents IX
3.2 RNA- and DNA-based Techniques for Post-transcriptional Regulation
of Molecular Targets, and their Potential as Biopharmaceutical Drugs 624
3.5 Cell-based Assays for In vitro Target Validation in the Drug Discovery
Process 640
3.6 Animal Models as the Ultimate Target Validation 645
Cord Brakebusch
4.1 Disease-oriented Research in Genetically Modified Mice 649
4.3 Analysis of Genetically Modified Mice 659
4.4 Alternative Methods 659
5.2 The Gene Discovery Tool Box or Dictyostelium Research 665
5.5 Conclusions 689
6.2 The Zymogen Form of fIX is Fully Inactive 697
6.4 Lys98 Hinders Substrate Binding to fIXa both Sterically and Electrostatically 698
6.5 Tyr177 Locks the 99-loop in an Inactive Conformation, which is Released
by Cofactor fVIIIa and Modified by the Physiologic Substrate fX 699
6.7 Evolutionary Relation of fIXa and fXa is Reflected in the Dependence
of Activity Changes on Arg/Lys Substrates 700
Contents X
6.8 By Binding at the 60-loop Ethylene Glycol Indirectly Reorganizes the 99-loop
and Allosterically Stimulates the Activity of fIXa 700
7.3 Examples 709
7.4 Summary 717
Volume 3
Florian M. Wurm
3 PER.C6

and Abraham Bout
Julio Baez
and Yuri Gleba
Contents XI
995
Heiner Apeler
Processes 1045
Luke Anthony Miles
Contents XII
Eric O. Aboagye
System 1325
Andreas Wiesner
Spectrometers 1343
Contents XIII
Volume 4
TM
TM
Pankaj Modi
Akira Yamamoto
Oliver Kayser
Contents XIV
Michael Hildebrand
3 Mettox
TM
Michael Murray
James Harris, III
Contents XV
Martha Lovato Tse and Chaitan Khoslat
Epilog 1833
Subject Index 1841
Contents XVI
Volume 1
Prologue XXV
Dedication XXIX
Foreword XXXI
Foreword XXXV
Quotes XXXVII
Introduction
Gary Walsh
Thomas R. Cech
V
Contents
ISBN: 3-527-31184-X
:
Abeel A. Mangi
Contents VI
14 Spheramine

Volume 2
Rainer Friedrich
Luis Moroder
of Malignancies 497
Raymond M. Reilly

Cells 549
Contents VII
and Manfred Eigen
System
Jonathan D. Chesnut
Cord Brakebusch
Volume 3
Florian M. Wurm
Contents VIII
1.5 Production Principles for Mammalian Cells: Anchorage-dependent Cultures and
Suspension Cultures 737
1.6 Large-scale Transient Expression 744
2.1 Mammalian Cells as a Workhorse to Produce Protein-based
Biopharmaceuticals 761
2.3 Pushing Expression Levels Impact of Vector Design and Cell Clone
Selection 764
2.5 The G-line: Use of the Immunoglobulin Locus of a Human/Mouse
Heterohybridoma for Heterologous Gene Expression 769
3 PER.C6

and Abraham Bout
3.7 Conclusion 803
4.3 Cell Line Stability 818
4.5 Selection of Useful Cell Sub-populations 822
4.7 Summary 830
Contents IX
Julio Baez
5.2 Advantages and Disadvantages of Transgenic Systems for the Production
5.3 Commercial Biopharmaceuticals with Human Clinical Experience
for Therapeutic, Immunoprophylactic, and Medical Device Use
derived from Transgenic Systems 852
5.4 Conclusions 873
and Yuri Gleba
6.3 Plant-made Recombinant Proteins available Commercially,
and under Development 903
7.3 Cell Culture 922
8.2 Development of ExpressTec for High-level Expression of Recombinant Proteins
in Cereal Grains 932
8.3 High-level Expression of Biopharmaceuticals in Cereal Grain
using ExpressTec 938
Contents X
9.4 Process Engineering 958
9.6 Regulatory Considerations 960
9.7 Conclusions 961
10.1 General Comments on Abiotic Stresses 968
995
Heiner Apeler
12.4 Outlook and Conclusion 1031
13.4 Conclusions and Future Aspects 1042
Contents XI
Processes 1045
14.2 Molecular Tools for the Construction of Transgenic Baculoviruses 1046
14.3 Insect Cell Culture 1047
14.5 Nutrient and Kinetic Considerations for Optimized BEVS-based Protein
Production 1048
14.6 Scaling-up Baculovirus-based Protein Production 1050
14.8 Case study: Rapid Optimization of Expression Conditions and Large-scale
Production of a Brutons Tyrosine Kinase Variant (BTK) 1053
Luke Anthony Miles
15.5 ATP Regeneration Systems 1074
Contents XII
1.4 Medical Application Areas for MAbs 1116
1.6 The Market Perspective 1119
1.8 Developing a Manufacturing Process for MAbs 1126
1.12 The Future of MAbs 1136
2.3 Technology 1153
2.7 Opportunities for New Therapeutic Applications Provided
by Synthetic Antibodies 1176
Contents XIII
Eric O. Aboagye
5.4 The Use of Radiolabeled Drugs to Image Tumor and Normal Tissue
Pharmacokinetics 1250
6.2 Ligands 1273
6.4 From a Ligand to a Product 1288
7.2 Ultrasound: Hear the Symptoms 1304
7.4 The Perfect Modality: Sensitive Particle Acoustic Quantification (SPAQ) 1308
7.7 Ultrasound, Microbubbles and Gene Delivery:
Noninvasive Micro-Gene Guns 1318
7.8 Summary: Ultrasound Theranostics
Building a Bridge between Therapy and Diagnosis 1320
System 1325
Andreas Wiesner
8.1 The Urgency of Earlier Diagnosis 1326
8.3 Current Tools for Protein Biomarker Detection 1328
8.4 The ProteinChip

Contents XIV
TM
Process:
From Biomarker Discovery to Assay Development 1334
8.8 Conclusion and Outlook 1338
Spectrometers 1343
Volume 4
TM
TM
Pankaj Modi
Akira Yamamoto
Contents XV
Oliver Kayser
Michael Hildebrand
3 Mettox
TM
Michael Murray
Contents XVI
James Harris, III
Epilog 1833
Subject Index 1841
Contents XVII
Volume 1
Prologue XXV
Dedication XXIX
Foreword XXXI
Foreword XXXV
Quotes XXXVII
Introduction
Gary Walsh
Thomas R. Cech
V
Contents
ISBN: 3-527-31184-X
:
Abeel A. Mangi
Contents VI
14 Spheramine

Volume 2
Rainer Friedrich
Luis Moroder
of Malignancies 497
Raymond M. Reilly

Cells 549
Contents VII
and Manfred Eigen
System
Jonathan D. Chesnut
Cord Brakebusch
Volume 3
Florian M. Wurm
Contents VIII
3 PER.C6

and Abraham Bout
Julio Baez
and Yuri Gleba
995
Heiner Apeler
Contents IX
Processes 1045
Luke Anthony Miles
Eric O. Aboagye
Contents X
System 1325
Andreas Wiesner
Spectrometers 1343
Volume 4
1.4 Outlook 1384
2.3 Safety and Disposition of PEG 1396
2.9 Summary and Future Outlook 1415
Contents XI
3.6 The Development of IC31 and Future Prospects 1440
TM
TM
Pankaj Modi
4.2 Rationale for Oralin
TM
Development 1446
TM
1447
TM
1448
Akira Yamamoto
5.2 Use of Protease Inhibitors 1467
5.3 Chemical Modification of Peptide and Protein Biopharmaceuticals 1472
5.4 Chitosan Capsules for the Colon-specific Delivery of Insulin 1480
5.3 Conclusion 1484
6.2 Poly(Ortho Esters) 1494
6.3 Preparation and Characterization of Microspheres 1496
6.4 In vivo Evaluation of Immune Responses 1500
7.1 Cationic Charge-mediated In vivo Gene Transfer to the Lung 1510
7.2 Asialoglycoprotein Receptor-mediated In vivo Gene Transfer to Hepatocytes 1512
7.3 Mannose Receptor-mediated In vivo Gene Transfer to Macrophages 1513
7.4 Folate Receptor-mediated In vivo Gene Transfer to Cancer Cells 1515
7.5 Transferrin Receptor-mediated In vivo Gene Transfer to Brain 1517
Contents XII
8.1 New Concept for Gene Delivery 1521
8.2 Controlled Intracellular Trafficking 1525
8.3 Transgene Expression and Gene Correction 1531
8.4 Towards Clinical Applications of Transgene Expression
and Gene Correction 1534
Oliver Kayser
9.2 Drug and Gene Delivery 1539
9.3 Gene Delivery 1543
9.4 Biosensors 1544
9.5 Implants and Tissue Engineering 1546
9.8 Safety Aspects 1548
9.7 Conclusions and Future Trends 1550
Michael Hildebrand
1.3 Classes of Biopharmaceuticals 1560
1.4 Analytical Methods and Specifications 1560
1.5 International Guidelines on Quality Control 1571
1.6 Analytics In vivo 1573
2.1 P450: The Background 1581
2.2 Importance of P450s for Drug Development 1582
2.3 Variability and Drug Metabolism 1583
2.4 The Structure of Cytochrome P450 1584
Contents XIII
3 Mettox
TM
Michael Murray
3.1 Issues and Economics of Early ADMET (Absorption, Distribution, Metabolism,
Excretion and Toxicity) Assessment 1604
3.2 Phase I Metabolism Prediction: Computational Approaches 1608
3.3 Phase I Metabolism Prediction: In vitro Techniques 1613
3.4 Genotoxicity Prediction 1624
4.2 Regulatory Authority 1639
4.3 Overview of Product Development: CMC Perspective 1643
4.4 Chemistry, Manufacturing and Controls Considerations 1645
4.5 Quality Control and Assurance 1647
4.6 Microbial Issues Specific to Biopharmaceuticals 1650
4.7 Process Validation 1653
4.8 Inspectional Considerations 1653
4.9 Biotech Development:
Lessons Learned and Issues Overcome by Industry and FDA 1654
4.10 FDA Initiatives to Improve the Pharmaceutical and Biopharmaceutical
Development Process 1661
5.2 History and Background 1673
5.3 The Competent Regulatory Bodies 1676
5.4 What is the EU Authorities Definition of a Biotechnological Product? 1681
5.5 The Regulatory Framework 1682
5.6 CP: The Biotech Procedure 1683
5.7 From Transcription to Prescription:
What is Different for Biotechnological Drugs? 1688
5.8 Biogenerics 1700
Contents XIV
1.2 Despite Robust Demand the Industry Faces Severe Challenges 1713
1.3 Why Biopharmaceuticals can Succeed in Rougher Markets 1724
1.4 Biopharmaceutical Players Will Need to Adapt their Portfolios
and Business Models 1728
2.1 Introduction
2.2 Product Platforms 1742
2.3 Interferon-c + Genechip 1750
2.4 Bacteriophages 1751
2.5 Outlook for the Company 1752
James Harris, III
3.2 Biogenerics and Bioequivalence 1756
4.2 Basic Concepts of Virtual Drug Discovery 1773
4.3 Pep2Lead Concept 1778
4.4 ADMETox Profiling 1785
4.5 Outlook 1798
5.2 Polyketide Synthases 1806
Contents XV
5.3 Engineering PKSs to Produce Novel Polyketides 1815
5.4 Development of Scalable Production Processes 1820
Epilog 1833
Subject Index 1841
Contents XVI
Mens Sana in Corpore Sana
Rationale for Modern Biopharmaceuticals
I have a dream . . .. Once, on an early
Sunday morning in 2003, the 50th anni-
versary year of DNA discovery, I woke up
and had the idea to bring together all the
world-renowned leaders from biotech aca-
demia and industry, in order to publish a
comprehensive book on modern biophar-
maceuticals. As learned from nature, some
things happen best if at all sponta-
neously. So, I contacted some of my
friends, presented the idea and discussed
with them the current hot topics in the
LifeSciences arena. Very quickly a list with
topics and authors emerged, which I pre-
sented to Wiley-VCH and they sponta-
neously agreed to publish this book.
From my past career I knew a number
of highly educated scientists and managers
in the LifeSciences: first, when I studied
biotechnology and did my diploma thesis
at the GBF (Gesellschaft fr Biotechnolo-
gische Forschung), then when I worked in
the biotech industry with Professor Nor-
bert Riedel, before I went on to also study
biochemistry and do my PhD at the Max
Planck Institute (MPI) with Professor Ro-
bert Huber. This was also a spontaneous
move: I remember quite well, when I
stopped by at the MPI on my way back
from a snowboard trip in the Munich
mountains. Quite naively, I asked if I
could talk to the Nobel prize laureate Pro-
fessor Huber asking for the opportunity
to work in one of the most famous labora-
tories in the world without even having an
appointment. It was an incredible honor
that he accepted. Now as my teacher and
co-founder of our biotech start-up he en-
couraged me to write this book and was
also willing to contribute to this endeavor.
I am also pleased that a colleague from
this start-up company is contributing with
a chapter on genetically engineered factor
IXa with 7000 times increased activity.
This proves that we had the right concept
for the company; needless to say that I am
glad that as well as the scientific success
this company is continuously growing,
whereas most of the companies founded
at the same time no longer exist.
After this entrepreneurial exercise I
switched gear and started working in a
high-tech consultancy with a focus on bio-
technology, before I started with Schering
AG. Heading the Department of Microbial
Chemistry again involved a number of
state-of-the-art biotechnologies (from ex-
pression system design and fermentation
process development to downstream pro-
cessing, Good Manufacturing Practice and
analytics). Obviously, over the years I was
exposed to a huge variety of different com-
panies, people and biopharmaceutical en-
XXV
Prologue
ISBN: 3-527-31184-X
vironments, and it was a great honor for
me when I was elected to the Executive
Board of the European Association of
Pharmaceutical Biotechnology (EAPB, and
very recently as its designated president)
and to the Editorial Board of the European
Journal of Pharmaceutics and Biopharmaceu-
ticals (EJPB). Altogether, in my past career,
I had the pleasure to meet a vast number
of brilliant scientists from world-class uni-
versities and academic institutes as well as
business leaders from major pharmaceuti-
cal companies.
I am very grateful for these various op-
portunities, which inspired the book pro-
ject Modern Biopharmaceuticals and pro-
vided me with the required large number
of excellent contacts at the same time,
and, I am happy to say, that most of my
contacts have spontaneously agreed to
provide a chapter for this book collea-
gues from academia and industry, from
regulatory authorities, and from consulting
business.
I hope that the reader will agree that
this book is the first of its kind, introduc-
ing a comprehensive set of technologies
recently developed, showing their impact
on drug development, discussing para-
digm shifts in the healthcare system and
also reflecting these changes in industrial
research. Compiling this wealth of infor-
mation in a sophisticated manner was
only possible if all chapters were written
by the experts themselves, and most of
them are working in academic institutes
and (often in their own) biotech compa-
nies at the same time. The authors come
from some of the worlds most famous
academic institutes, and biotech compa-
nies, such as CalTech, Cambridge, Charit,
ETH Zrich, Fraunhofer-Institute, Har-
vard, Johns Hopkins, Karolinska, Kyoto
University, London Imperial College, Max-
Planck-Institute, MIT, Moscow and Polish
Academy of Sciences, NCI, NIH, Oxford,
Princeton, Scripps, Seoul University, Stan-
ford, Technion, Weizmann, Yale. They are
CEOs, Board Members or Global R & D
Heads of world-class companies, e.g. Am-
gen, Bayer, Baxter, Berlex, Crucell, DSM,
DuPont Merck, Genentech, Genzyme, In-
vitrogen, Lonza, McKinsey, Mologen, Mon-
santo, MorphoSys, Novartis, Novo Nordisk,
Philips, Roche, SmithKline, Schering
and from FDA.
This profound and balanced mixture of
academia and industry was intended to
make the book equally appealing to scien-
tists at research institutes, physicians at
hospitals, students at universities and labo-
ratory technical staff from areas like medi-
cine, all different areas of LifeSciences, as
well as other healthcare professionals. It is
my hope that it will serve as an inspiration
for all professionals in the field, since it of-
fers a very good framework for under-
standing the complex nature of biophar-
maceuticals, the mainstay of modern med-
icine.
Of course, some of the breakthrough
technologies described need to be treated
with caution, e.g., human cloning. The
chapter by Hwang et al. impressively dem-
onstrates that we are now able to clone a
human being, but this ground-breaking
scientific success also has various ethical
implications, depending on how it will be
applied. These pluripotent embryonic stem
cells from somatic cell nuclear transfer
of reprogrammed human adult cells can
be grown to have an unlimited source
of autologous cells for transplantation
medicine (some striking examples on
organogenesis and organ repair are pre-
sented in the section cell therapeutical
approaches). This is very exciting, because
for the first time transplant rejection can
be overcome, since the transplant is built
from the patients own cells with the pa-
Prologue XXVI
tients own genetic setup. However, this
breakthrough for therapeutic cloning could
also be misused for evil purposes (as with
nuclear power) and this reminds me of a
quote from Jurassic Park: Biotechnology is
the most powerful force which was ever
on the planet. But you play with it like a
child, who just found his fathers gun. So,
whether these powerful biotechnologies
are solely (and exclusively) applied in a
positive way that is beneficial for mankind
really depends on how respectful we as
scientists deal with them!
We all know that since the remarkable
debut of modern biopharmaceuticals, the
field of pharmaceutical biotechnology has
evolved tremendously. By comparison,
when I follow how quickly (life) sciences
advance, it would make Newtons apple ap-
pear to fall in slow motion. I am very hap-
py that people who contributed most to
this fast and exciting development of bio-
pharmaceuticals, who helped to usher in a
golden age of molecular biology, also con-
tributed to Modern Biopharmaceuticals
Design, Development and Optimization. I
would like to take the opportunity to thank
all of the authors for their excellent contri-
butions and hope that the reader will en-
joy this fantastic collection of scientific art.
I also wish to thank the publisher Wiley-
VCH for making this project happen espe-
cially Andrea Pillmann and Waltraud Wst.
Both were very supportive from the begin-
ning of this exciting book project not just
in managing the publication itself, but also
for managing my ideas. And I guess some-
times it was a tough job to stop my creativ-
ity in generating new ideas. Another idea
which both were in favor of was to also pro-
vide a supplementary CD-ROM with a
PowerPoint presentation that I have as-
sembled over the years. This I use for edu-
cational purposes when I share with stu-
dents the fascination of (20 000 years of)
biotechnology. The CD-ROM also includes
some fantastic video animations, e.g., show-
ing the whole process from DNA unwind-
ing in the nucleus through transcription
into mRNA to the expression of a biophar-
maceutical. By focusing on key aspects,
these animations tremendously help in the
understanding of such complex processes.
Or, as a homage to Albert Einstein (whose
theory of relativity would have its 100th an-
niversary this year): make it simpler, but
not simple.
Having said that, if you have any valuable
educational video animations that I could
incorporate into the presentation on the
supplementary CD-ROM, I would be very
grateful if you could contact me. Also, if
you identify any areas, topics or technolo-
gies which you feel are not yet captured,
please let me know. In addition, I would ap-
preciate any comments which will help to
keep the topics/content up to date and
make the next edition of Modern Biopharma-
ceuticals (which is in preparation already)
even more comprehensive. Pleast visit our
biotechnology hub at www.get-gps.net to
discuss current trends with the respective
Global Pharma Specialist from our world-
wide competence network. And this will
help all of us in the LifeScience community,
because as we all know: knowledge is power
shared knowledge is success.
Thank you very much in advance and
enjoy reading Modern Biopharmaceuticals!
Jrg Knblein
Scientific Advisor
Executive Boardmember
and designated President
of European Association
of Pharmaceutical Biotechnology
Berlin
May 2005
Prologue XXVII
Modern Biopharmaceuticals Design, Devel-
opment and Optimization is dedicated to
the man who made all this possible:
Francis Crick (19162004).
Francis Harry Compton Crick was born on
8 June 1916 in Northampton, England. He
studied physics at University College, Lon-
don, where he obtained a BSc in 1937. He
then started his PhD in physics, which
was interrupted in 1939 by the outbreak of
World War II. Crick worked as a scientist
for the British Admiralty until he left in
1947 to study biology in Cambridge, where
he worked at the Strangeways Research
Laboratory.
Two years later, he joined the Medical
Research Council Unit at Cambridge Uni-
versitys Cavendish Laboratory, headed by
Max F. Perutz, to study X-ray diffraction
by the helix. The work of Perutz laid the
groundwork for his interest in protein
structures (as it did for my teacher, Robert
Huber, and later for myself as well). Thus,
he became a research student for the sec-
ond time and was accepted in 1950 as a
member of Caius College, Cambridge.
Cricks career was then critically influ-
enced by his friendship with James D.
Watson, which in April 1953 led to the
ground-breaking proposal of the double-
helical structure for DNA and its mecha-
nism of replication published in Nature:
We wish to suggest a structure for the
salt of deoxyribose nucleic acid (DNA).
This structure has novel features which
are of considerable biological interest.
Crick obtained a PhD in 1954 on his
thesis entitled X-ray diffraction: polypep-
tides and proteins and in 1959 became a
Fellow of the Royal Society for his work
on DNA as well as for his study of the
structure of proteins. Finally, in 1962, Wat-
son and Crick, along with their colleague
Maurice Wilkins, were awarded the Nobel
Prize in Physiology or Medicine.
After his work on the double helix,
which changed the face of modern-day
XXIX
Dedication
ISBN: 3-527-31184-X
Francis Crick (courtesy of Marc Lieberman,
Salk Institute, La Jolla, CA)
My family and loved ones for their continuous support and patience
science and medicine, Crick collaborated
with Sydney Brenner at Cambridge to de-
velop the adaptor hypotheses about protein
synthesis and the genetic code. Between
1966 and 1976, Crick worked on embryol-
ogy until he moved to the Salk Institute
for Biological Studies in San Diego, CA.
There, he began to work on the under-
standing of the brain and neural correlates
of consciousness, which he continued for
the rest of his career.
Darwin has interested us in the history
of natures technology (Karl Marx), and
Watson and Crick paved the ground for
modern biotechnologies: all the work de-
scribed in this book only became possible
through the elucidation of the structure of
DNA the greatest scientific accomplish-
ment of the 20th century. This fascinating
work on DNA revolutionized science, and
enabled molecular genetics, biotechnology
and the development of modern biophar-
maceuticals.
When I was asking Francis Crick for his
contribution to my book, on 4 November
2003, he replied that he very much appre-
ciated this endeavor of creating such a
comprehensive book: . . . Nice of you to
ask me to contribute to your book on bio-
pharmaceuticals . . . Unfortunately I am in
very poor health so do please excuse me.
Apologies, Francis Crick.
Francis Crick, the DNA code-breaker,
died after a long battle with cancer at the
age of 88 years on 29 July 2004, at Thorn-
ton Hospital of the University of California
in La Jolla.
I will always remember Francis for his
extraordinarily focused intelligence and for
the many ways he showed me kindness
and developed my self-confidence, says
his long-time colleague and friend James
D. Watson, whom I had the honor of
spending an evening with in October 2004
during the Faculty Meeting at Charit,
Berlin. I am very grateful that I had the
opportunity to discuss recent trends in
biotechnology and his view on Modern
Biopharmaceuticals. He treated me as
though I were a member of his family,
Watson says. Being with him for 2 years
in a small room in Cambridge was truly a
privilege. I always looked forward to being
with him and speaking to him, up until
the moment of his death. He will be sorely
missed.
Francis Crick made an enormous contri-
bution to science, and our understanding
of biology and the health of mankind. His
death is a sad loss to science and espe-
cially to modern biotechnology.
Dr. Jrg Knblein
Head of Microbiological Chemistry
Schering AG
Berlin
May 2005
Dedication XXX
James Watson and Jrg Knblein, October 2004
during Watsons visit at the Charit in Berlin,
Germany
History of Modern Biopharmaceuticals:
Where Did We Come From and Where
Will We Go
It is a pleasure to write the Foreword to this
unusual and excellent biotech book! Modern
Biopharmaceuticals Design, Development
and Optimization gives a comprehensive
overview of the status of pharmaceutical
biotechnology today, but also looks ahead
and shows future trends with an outstand-
ing collection of very recent results. It pre-
sents a comprehensive overview on break-
through achievements with state-of-the-art
biotechnologies, demonstrating that Life-
Sciences are nowadays shaped by amazing
and ground-breaking discoveries.
When James D. Watson and Francis
Crick elucidated the structure of the mol-
ecule of life in 1953, it was compelling in
its sheer beauty. However, more impor-
tantly, the three-dimensional structure of
DNA led to the mechanisms of replication
and was the first three-dimensional Xerox
machine (Kenneth Boulding). Three years
later, Arthur Kornberg isolated the enzyme
that synthesizes the molecule of life: DNA
polymerase.
Around that time, scientists were also
working on the more complex structures
of proteins: John C. Kendrew and co-
workers described the structure of myoglo-
bin in Nature in 1958 (A three-dimen-
sional model of the myoglobin molecule
obtained by X-ray analysis, Nature 181,
6626), and shortly after that Max F. Pe-
rutz and coworkers described the structure
of hemoglobin (A three-dimensional Four-
ier synthesis of reduced human haemoglo-
bin at 5.5 resolution, Nature 199, 6338).
Both shared the Nobel Prize in Chemistry
in 1962 for their studies on the structures
of globular proteins.
In the field of DNA, one important dis-
covery was followed by the next: the first
plasmid was isolated in 1959, 1 year later
Franois Jacob and Jacques Monod defined
mRNA as the carrier for the blueprint of
the entire protein, and in 1961 Marshall
W. Nirenberg started decoding the genetic
alphabet by identifying that at the mRNA
level the codon UUU encodes the amino
acid phenylalanine. The respective phenyl-
analyl-t-RNA was later discovered by Aaron
Klug (see his quote for Modern Biophar-
maceuticals). Now the mystery of tran-
scription and even of translation is solved,
and, in 1962, Watson (see his quote for
Modern Biopharmaceuticals) and Crick,
along with their colleague Maurice Wil-
kins, were awarded the Nobel Prize in
Physiology or Medicine. In 1968, gene
scissors, discovered by Werner Arber, re-
volutionized molecular biology, since these
restriction enzymes are capable of specifi-
cally cutting bacterial DNA. This enabled
XXXI
Foreword
ISBN: 3-527-31184-X
scientists for the first time to prepare re-
combinant DNA. Two years later the cen-
tral dogma of biochemistry, i.e., that the
genetic flow is unidirectional from DNA
via mRNA to protein, was proven wrong:
Howard Temin and David Baltimore dis-
covered the enzyme reverse transcriptase,
synthesizing cDNA from mRNA. This
breakthrough discovery eventually allowed
the expression of eukaryotic genes, as the
untranslated segments in the genome are
spliced out by this process.
At Brookhaven National Laboratory, New
York, the Protein Data Bank (PDB) was es-
tablished in 1971, and has become a repos-
itory for protein coordinates which are
shared between scientists worldwide. The
PDB is a very important tool and the basis
for rational, structure-based drug design
a prerequisite for the development of mod-
ern biopharmaceuticals.
In 1973, a new era in biotechnology
started with the advent of gene technology,
when Allan Maxam and Walter Gilbert
(Harvard) and Frederick Sanger (Cam-
bridge) developed a DNA sequencing
method. Combining all these fascinating
findings, Stanley Cohen (see his quote for
Modern Biopharmaceuticals) and Herbert
Boyer re-combined in vitro DNA pieces to
form a new gene for the first time.
At the same time, Georges J. F. Khler
and Csar Milstein were working together
at the Medical Research Council Laborato-
ry of Molecular Biology in Cambridge,
where in 1975 they discovered a technique
to produce monoclonal antibodies. Pre-
viously, to prepare substantial quantities of
antibodies, scientists had to inject an anti-
gen into an animal, wait for antibodies to
form, draw blood from the animal and iso-
late (a mixture of different types of) anti-
bodies. The only way to obtain monoclonal
antibodies was to clone lymphocytes, se-
creting one form of antibody molecules.
Lymphocytes, however, are short-lived and
cannot be cultivated easily. By fusing lym-
phocytes with myeloma cells, Khler and
Milstein obtained hybrid cells synthesizing
a single species of antibody while perpetu-
ating themselves indefinitely. Together with
Niels K. Jerne, they received the Nobel Prize
in Physiology or Medicine 1984. Most pres-
ent biopharmaceuticals (i.e., therapeutic
and diagnostic proteins) are antibody-based
molecules, and this is why the development
of monoclonal antibodies revolutionized
medicine and paved the way for new, tar-
get-specific approaches, where pure, uni-
form and highly sensitive protein molecules
can be used as biopharmaceuticals for diag-
nosis and therapy.
The recombinant DNA technology of
Cohen and Boyer enabled them to gener-
ate the first commercial product in 1978:
human insulin expressed in Escherichia
coli. These efforts also led to the first bio-
tech company: on 15 October 1980 Genen-
tech went public on the New York Stock
Exchange. Fascination about this modern
biopharmaceutical and the huge potential
of the new biotechnology caused the stock
price to jump from US$ 35 to 89 in the
first 20 minutes; by the evening of the
same day, the market capitalization was
US$ 66 million!
The year 1984 is another landmark: the
first transmembrane protein, the photo-
synthetic reaction center (RC) from Rhodo-
pseudomonas viridis, was solved. The chal-
lenge in solving the structure of this huge
(150 kDa) protein was that it consists of 11
membrane-spanning, hydrophobic a-he-
lices. Solving the RC structure was a ma-
jor breakthrough, since many of the most
interesting drug targets are membrane-
bound proteins. In 1988, my colleagues
Johann Deisenhofer and Hartmut Michel
and myself were awarded the Nobel Prize
for Chemistry for this work.
Foreword XXXII
Then there was the advent of a surpris-
ingly simple tool that readily revolution-
ized molecular biology and heavily influ-
enced modern biotechnology. In 1983,
Kary Mullis invented a process he called
the polymerase chain reaction (PCR),
which solved a core problem in molecular
genetics, i.e., gene amplification. In other
words: how to make copies of a strand of
DNA that you are interested in? PCR turns
the job over to the very biomolecules that
nature uses for copying DNA as well. Two
primers flag the beginning and end of
the DNA stretch to be copied and an en-
zyme called polymerase walks along the
segment of DNA, reading its code and
assembling a copy. To complete the PCR
cocktail, a pile of DNA building blocks is
added, which the polymerase needs to
make that DNA copy in vitro. Kary Mullis
won the 1993 Nobel Prize in Chemistry
for this discovery (see his quote for Mod-
ern Biopharmaceuticals). Exactly 10 years
later another breakthrough for modern
medicine was awarded: Paul Lauterbur re-
ceived the Nobel Prize for his pioneering
work in imaging technique. This enabled
early diagnostic and enhanced earlier treat-
ment leading to high success rates (see his
quote for Modern Biopharmaceuticals).
Another quantum leap for modern bio-
technology was the first cloned mammal
by Ian Willmut in 1996 (see his quote for
Modern Biopharmaceuticals) by means
of somatic cell nuclear transfer (SCNT)
the sheep Dolly. Then, in 2004, the first
human embryo was cloned by a team led
by Woo Suk Hwang, who was able to ob-
tain pluripotent embryonic stem cells by
SCNT of reprogrammed human adult
cells. The highly differentiated genetic pro-
gram of the nucleus from an adult cell
can be completely reprogrammed after
being introduced into an enucleated oocyte
from a donor, so that these embryonic
stem cells can be grown to produce an un-
limited source of autologous cells for
transplantation medicine. This experiment,
together with some striking examples of
how one can apply this new source of
stem cells for organogenesis and organ re-
pair, is also presented in this book.
An unusual feature of Modern Biopharma-
ceuticals Design, Development and Optimi-
zation is that, for a book with so many facts,
it is a delight to read. Whilst being easy to
read, it is a guide to both broad surveys
and key papers, which are provided in con-
venient, but at the same time comprehen-
sive, reference lists and Internet links. Im-
plementing this structure, the reader can
easily begin to explore the very extensive lit-
erature on all the relevant topics and also
has a guide to navigate through the World
Wide Web as well. On top of this comes a
very educational CD-ROM with impressive
video material.
I am convinced that my student Jrg Kn-
blein has done a great job in compiling a
cutting edge and comprehensive book on
modern biopharmaceuticals written by
knowledgeable experts from academia and
industry. I wish this extraordinary book a
numerous and broad readership, and I hope
the reader will enjoy this collection of scien-
tific art as much as I did!
Professor Robert Huber
(Nobelprize in Chemistry, 1988)
Max-Planck-Institute for Biochemistry
Martinsried
May 2005
Foreword XXXIII
Modern Biopharmaceuticals A Primer:
Stem Cell Research and Very Recent
Breakthroughs
The opportunities and challenges of an
aging population make it mandatory to re-
think our attitude towards medicine. With
the advent of genomics and proteomics
and with the entire new field of molecular
medicine, we now have the means in our
hands to cope with the challenges lying
ahead of us.
More diseases than ever are likely to be
treated with innovative and completely
new therapeutic approaches. And above all
early precise diagnostic procedures at the
molecular level will allow us to get a better
understanding of the underlying pro-
cesses. Today, the so-called molecular
imaging allows already for a functional di-
agnosis as well as for diagnostic measures
at the molecular level.
At the same time the refinement in our
ability to diagnose in vitro genetic and pro-
tein alterations will eventually lead to a
much better understanding of our predis-
position to develop certain diseases or will
give us a clue as to what types of treat-
ment will be most appropriate for certain
groups of individual patients.
Pharmaceutical and biopharmaceutical
research hence is a perfect application ori-
ented continuation of what is going on in
laboratories of biomolecular research. Prob-
ably for the first time in biomedical re-
search there is not only an opportunity but
also an irrevocable necessity for public pri-
vate partnerships in order to fully exploit
the potentialities in the field of biomedi-
cine. If predisposition towards genetically
influenced diseases can be detected early
and if specific molecular imaging diagnos-
tics allowing for a precise and early detec-
tion of diseases becomes feasible, it is more
than likely that the borderline between sec-
ondary and primary prevention will be
shifted towards earlier timepoints of pre-
ventive intervention. Primary prevention
either with lifestyle changes or with phar-
maceutical means, will become a routine
measure in many forms of early and even
very early disease states as well as in those
cases where only statistical probability gives
a hint towards upcoming diseases.
Taken together, this will allow for a
much better and more rational employ-
ment of preventive medicine than today.
Of course, this can be regarded as utopia,
probably also wishful thinking, but there
is no doubt that in certain areas and under
certain circumstances this will become
reality provided that the ways in which we
educate and inform patients and potential
patients properly change adequately.
Cell biology including stem cell research
has become a very exciting new field of
XXXV
Foreword
ISBN: 3-527-31184-X
molecular biology, since this new science
offers a much better understanding of ele-
mentary processes holding out the pro-
spect of understanding why certain cells
e.g. become tumor cells. Making use of
the knowledge gained with this research
and the cells being produced has another
dimension: the so-called Regenerative
Medicine. Human embryonic stem cells
and adult stem cells are of prime interest
for these new fields of biomedical re-
search.
For me there is no doubt that to fully un-
derstand the possibilities of stem cells, the
mechanism of cell differentiation and hence
a basic mechanism of life, we must not con-
centrate on adult stem cells only, but rather
include embryonic stem cells, too.
The question to what extent stem cells
be it embryonic or be it adult will be
used and have to be used in medical and
clinical practice later on is still a very open
one. There is at least hope that the in vivo
activation of existing adult stem cells could
be a fascinating dimension of research
work, resulting from the current work in
embryonic and adult stem cells. It is
obvious that we still have different legal
frameworks for working with embryonic
stem cells. It is also true that we are in
the middle of a major ethical debate. How-
ever, for me it is equally important that
the progress to be expected and the advan-
tages to be envisaged from research in
stem cells are so fundamental that even-
tually we will find a regulatory framework
within which this type of research can be
performed worldwide. If this turns out not
to be the case, it could easily happen that
the major breakthrough for this type of
modern biopharmaceutical research will
be achieved in Asian countries, which are
already at the forefront in this field.
In summary, modern biopharmaceutical
technologies offer enormous opportunities
and are a great intellectual challenge for
our imagination and for our daily research
work. The field is highly dynamic, it is ex-
panding, and it offers great opportunities
for enthusiastic young people. And above
all, this exciting field of new science is giv-
ing and will give us a much better basis
for understanding human life and for add-
ing a new quality of life.
The present book Modern Biopharma-
ceuticals is a good primer to interest
young scientists in this multidisciplinary
field of modern biotechnology. It is com-
prehensive, touches most of the currently
available modern trends in LifeSciences,
and it is written for a broad and cross-dis-
ciplined audience. Jrg Knbleins Mod-
ern Biopharmaceuticals provides the win-
dow into biopharmaceutical research, de-
velopments and applications of today and
tomorrow, and I hope that readers from
academia as well as from industry will ap-
preciate this collection of outstanding con-
tributions from world class researchers
working in both fields.
Professor Gnter Stock
Senator of German Research
Foundation (DFG)
Board Member and Head of
Research of Schering AG
Berlin
May 2005
Foreword XXXVI
The making of pharmaceutical and diag-
nostic agents in cells has moved from
edge to the center of their respective com-
mercial development.
With Modern Biopharmaceuticals, Jrg
presents an outstanding collection of arti-
cles from groundbreaking scientists, com-
prehensively describing the many novel
ways cells so are being deployed toward
human good.
Professor James D. Watson,
DNA code-breaker and Nobel Prize laureate
(Physiology or Medicine, 1962)
COLD SPRING HARBOR LABARATORY,
New York
The new book Modern Biopharmaceuti-
cals has an impressive list of authors
drawn both from world-renowned aca-
demic research laboratories and also from
the worlds leading biotech and pharma-
ceutical companies. The experts from this
coalition of world-class companies, insti-
tutes and universities have direct experi-
ence of the cutting edge technologies de-
scribed and understand the various needs,
met and unmet. This fantastic line up of
authors make it a truly world class book
a four-volume educational platform cover-
ing the full spectrum of science from dis-
covery to applications.
It is hoped, that there will also follow
(an inexpensive) student edition, which
would be more widely accessible.
Professor Sir Aaron Klug,
Discoverer of the Phenylalanyl-t-RNA
and Nobel Prize laureate (Chemistry, 1982)
MRC LABORATORY
OF MOLECULAR BIOLOGY
Cambridge, United Kingdom
XXXVII
Quotes
ISBN: 3-527-31184-X
The comprehensive coverage provided in
Modern Biopharmaceuticals by eminent in-
vestigators should stir the imagination of
all scientists interested in possible medical
applications of their own research. I wish
you best of luck in your endeavors with
this excellent biotech book.
Professor Stanley Cohen,
Designer of the first cloning vector
and Nobel Prize laureate
VANDERBILT UNIVERSITY
SCHOOL OF MEDICINE
Nashville, Tennesee
We always seem to be right on the edge
of solving all our health problems, just like
we always seem to be on the verge of ulti-
mately discovering the physical mysteries
of the universe. It does seem like we are
about to understand cancer, genetic dis-
eases, infectious diseases all the things
that bring us discomfort on the personal
level.
Gunther Stent decided in the late Six-
ties, in his wonderful lectures at Berkeley
entitled the Rise and Fall of Molecular
Biology, that all the interesting stuff in
molecular biology had already been fig-
ured out. Only the boring details remained
just then biotechnology exploded. Our
latest shocking advance, the ease of read-
ing and manipulating DNA, is what is re-
sponsible, I suppose for our latest bout of
thinking we know almost everything im-
portant. It turns out though, that there are
always new things to discover.
You need to keep up on what is known
already and you always need to know
whats already known. So, read this book
Modern Biopharmaceuticals and you will
get a very good overview of what is cur-
rently known in the exciting field of Life-
Sciences.
Professor Kary Mullis,
Inventor of PCR and Nobel Prize laureate
(Chemistry 1993)
Newport Beach, California
Quotes XXXVIII
It is not easy to obtain a wide overview of
the developing impact of new knowledge
in the basic pharmaceutical sciences on
medicine. This book, Modern Biopharma-
ceuticals, is an admirable attempt to meet
that need, for all experts in this field, as
well as for students who need an orienta-
tion for possibilities in academia, industry,
and medicine.
Professor Paul Lauterbur,
Pioneer of MRI and Nobel Prize laureate
UNIVERSITY OF ILLINOIS
Department of Chemistry
The new biopharmaceuticals that are
being developed at present will provide im-
portant new opportunities in therapy and
diagnosis that cannot be met in any other
way. Jrg Knblein has assembled in these
four new volumes a unique collection of
reports by the world leaders in their fields.
They describe the present state of their
field and the requirement for further re-
search. Modern Biopharmaceuticals will be
an important resource for students and re-
searchers alike.
Professor Ian Wilmut,
Clone-father of sheep Dolly
ROSLIN INSTITUTE, Scotland
Department of Gene function
and Development
Quotes XXXIX
The explosion of biological products as
novel human therapeutic agents is based
on remarkable advances in the enabling
sciences that comprise modern biotechnol-
ogy. Modern Biopharmaceuticals provides a
broad, up to date analysis of the many fa-
cets of discovery and development re-
quired to successfully generate biopharma-
ceuticals. The scope is all encompassing,
the chapters are authored by the who is
who of biotechnology experts, and the cov-
erage is admirable. The Knblein should
be a unique resource.
Professor Chris Walsh,
HARVARD MEDICAL SCHOOL,
Department of Biological Chemistry
and Molecular Pharmacology
The Charit has achieved its international
reputation by the close association of basic
research with its diagnostic and therapeu-
tic application. Outstanding examples of
this are Robert Koch, Paul Ehrlich and
Emil von Behring. Jrg Knblein con-
tinues this great tradition in Berlin with
his book Modern Biopharmaceuticals De-
sign, Development and Optimization. It pro-
mises to be a great success.
Professor Detlev Ganten, CEO and President
CHARIT UNIVERSITY MEDICINE
Berlin, Germany
Quotes XL
A New Era in the New Millennium
Modern Biopharmaceuticals Design, Devel-
opment and Optimization is an attempt to
give a broad overview on this exciting field
of LifeSciences. It is a real challenge (and
a priori an impossible task) to capture all
of the trends which currently shape the
landscape of modern biotechnology. Ac-
cording to Ralph Waldo Emerson (1803
1882) Men love to wonder, and that is the
seed of science, and this is especially true
for modern biopharmaceuticals and the ex-
ploding number of new biotechnologies
that have developed over the last few years
in the LifeSciences. Obviously, as there are
always more and more new topics appear-
ing on the horizon (and although always
interesting) I had to stop at some point
the only question was where and when to
make a clear cut. As always, the guiding
principle came from nature: iridium (an
element in the precious metal group of
the periodic table) is spatially surrounded
by 77 electrons (and theoretically completely
wrapped by an electron shell) and the GPS
system Iridium with its planned 77 low-
earth orbiting satellites will enable 100%
coverage of our planet. Thus, with my glo-
bal biopharmaceutical surveillance I was
hoping to provide at least a representative,
although far from complete, truly global
snapshot. Therefore, this book now con-
sists of 77 chapters divided in Volumes I
IV (like the consecutive clinical phases of
a successful biopharmaceutical, including
postmarket studies in phase IV) from all
different areas of biopharmaceutical re-
search and development, and from all dif-
ferent parts of the world again, by its
very nature, incomplete.
To compensate for this incomplete-
ness, we have implemented a web page
Global Pharma Specialists (www.get-
gps.net) where we will continuously pro-
vide and discuss recent achievements in
the biotech arena that could not be covered
in the first four volumes. One can also
download animations and other content
from the supplementary CD-ROM sup-
plied with Modern Biopharmaceuticals. Hav-
ing said that, I would like to encourage
everybody to visit this biotechnology hub
at www.get-gps.net and get an insight into
latest trends in modern biotechnologies.
You can also seek advice from the entire
global network of the most knowledgeable
experts from academia and pharmaceutical
industry: the Global Pharma Specialists.
In addition, I would appreciate any com-
ments and submission of latest develop-
ments at www.get-gps.net, which will help
to keep the topics/content up to date and
XLI
Executive Summary
ISBN: 3-527-31184-X
make the next edition of Modern Biopharma-
ceuticals (which is in preparation already)
even more comprehensive, and therefore
more valuable, because as we all know:
knowledge is power shared knowledge
is success
The cover of Modern Biopharmaceuticals
was inspired by Fantastic Voyage, a science
fiction novel from the quintessential
author Isaac Asimov (19201992), pub-
lished in the same year I was born. The
picture shows a jet-powered nanosubmar-
ine flying past several red cells as it navi-
gates its way through a human blood ves-
sel. Brightly backlit endothelial cells pro-
tectively coat the blood vessel walls and a
large vascular bifurcation looms ahead.
This scenario nicely describes the
approach Find, Fight and Follow, e.g., to
hook up the biopharmaceutical to a specif-
ic delivery system, selectively bring the
drug to the desired target (without harm-
ing the rest of the body) and, finally, moni-
tor the therapeutic success. Such a troika
can be realized with, for example, the
same antibody loaded with different active
substances: diagnostic or therapeutic
early diagnostic (find) with a weak imaging
radionuclide, therapy with high radiation
energy of a strong emitter (fight) and thera-
py control again with the weak radionu-
clide (follow). In general, targeting moie-
ties (ligands) can be converted into imag-
ing or therapeutic agents by modification
with suitable radionuclides, fluorophores,
drugs, enzymes, cytokines or other bioac-
tive molecules. This sounds like the vision
of Paul Ehrlichs (18541915) magic bullet.
However, as we will see in Modern Bio-
pharmaceuticals, this is no longer a vision
we are almost there.
In addition to that, some examples from
the field of bionanotechnology (which is
judged as the key technology of the 21st
century) will be presented with the main
focus on fabrication and miniaturization.
We will see the pharmaceutical use of
smart drug delivery systems like micronee-
dles and biosensing microchips. We will
also learn about a pill-sized video device,
which can be swallowed by the patient and
will show the doctor the intestinal tract,
for example very similar to Asimovs na-
nosubmarine on the cover.
Modern Biopharmaceuticals starts with
the fair and very essential question: What
are biopharmaceuticals? The term origi-
nated in the 1980s for a class of pharma-
ceuticals produced by modern biotechnolog-
ical techniques mainly protein-based
drugs produced by genetic engineering.
Over the years, the class of biopharmaceuti-
cals was extended, first with monoclonal
antibodies (mAbs) from hybridoma cells,
then with DNA-based drugs like antisense
technologies and gene therapy, and very
recently with siRNAs (small interfering
RNAs) and stem cells. All these aspects
are covered in an excellent Introductory
chapter from Gary Walsh, Professor at
Limerick University, Ireland. Gary answers
the question in a comprehensive way, and
also gives an overview on the history of
biopharmaceuticals and their current ap-
proval status. Gary and myself are Execu-
tive Board members of the European Asso-
ciation of Pharmaceutical Biotechnology
(EAPB) and he was my number one
choice for this chapter: he is very experi-
enced with a sound background on bio-
pharmaceuticals, gives lectures on this top-
ic and has also published several articles
in Nature Biotechnology. In addition, he is
author of some comprehensive textbooks
including the landmark bestseller Biophar-
maceuticals, Biochemistry and Biotechnology.
In his excellent Introduction he describes
how rapidly the biopharmaceutical sector
has matured and illustrates the medical as
well as commercial importance of these
Executive Summary XLII
drugs. The biopharmaceutical industry,
although only 20 years old, already has a
turnover of more than US$ 30 billion per
year. With a dramatic growth rate of 20%,
this is obviously a highly lucrative and val-
uable market, and it is estimated that it
will grow to over US$ 90 billion by 2010.
Approximately a quarter of all genuinely
new drugs currently coming on the mar-
ket are biopharmaceuticals already and
some 250 million people worldwide have
been treated to date with this class of
drugs. As we will see, continued advances
in developing new and exciting modern
biopharmaceuticals will fuel the growth of
this new drug class at the dawn of this
new millennium.
Consequently, the next very substantial
question is: Where did we come from
and where will we go?, and is answered
by my teacher, friend and co-founder of
our joint biotech company, Professor Ro-
bert Huber, from the Max-Planck-Institute
for Biochemistry in Martinsried near Mu-
nich. I cannot imagine anybody in the
world who could give this answer in a
more sophisticated way than Robert does:
he has an extremely sound background in
biochemistry, and solved the X-ray struc-
ture of many proteins and large biological
assemblies, which in turn paved the way
for developing a number of modern bio-
pharmaceuticals. Because of his various
scientific achievements, he has been
elected as a Member and Honorary Mem-
ber to more than 20 societies, e.g., Na-
tional Academy of Sciences USA, the Roy-
al Society, London, and the European
Academy of Arts, Sciences and Humani-
ties. On top of that, he has received more
than 30 honors, including the Max Tishler
Prize (Harvard University) and The Grand
Decoration of Honor with Star for Services
to the Republic of Germany. A hallmark
was in 1984 when he solved the structure
of the first transmembrane protein, the
photosynthetic reaction center (RC) from
Rhodopseudomonas viridis. Solving the RC
structure (a huge 150-kDa protein consist-
ing of 11 membrane-spanning, hydropho-
bic a-helices) was a major breakthrough,
since many of the most interesting drug
targets are membrane-bound proteins.
Some examples with their working mecha-
nism and mode of action are shown in an
impressive video animation on the supple-
mentary CD-ROM. For his pioneering
work on biopharmaceuticals and outstand-
ing achievements in the entire field, Pro-
fessor Huber was awarded the Nobel Prize
in 1988. I am very grateful that I had the
chance to be his student, and I am also
very thankful that he encouraged me and
supported Modern Biopharmaceuticals.
Part I: Biopharmaceuticals Used
in Molecular Medicine
From Genome to Clinic Correlation
between Genes, Diseases and
Biopharmaceuticals
Part I, Biopharmaceuticals Used in Mo-
lecular Medicine, again starts with a con-
tribution from a Nobel Prize laureate. Tho-
mas R. Cech from the Howard Hughes
Medical Institute received the Nobel Prize
for Chemistry in 1989 for the discovery of
catalytic properties of RNA and, as we will
see, this also led to the development of in-
teresting biopharmaceutical approaches. In
his chapter he lays the ground for under-
standing the molecular biology of cells
and their chromosomes, and also describes
the molecular mechanisms leading to cel-
lular immortality and cancer. He then, in
an easy-to-follow way, explains on a molec-
ular level and on the basis of the ge-
nome how to identify new and attractive
Executive Summary XLIII
targets for the development of modern bio-
pharmaceuticals. Subsequently, this nice
introduction to pharmacogenetics and
pharmacogenomics is followed by a chap-
ter from Dr. Shiew-Mei Huang, Deputy Of-
fice Director for Clinical Pharmacology
and Biopharmaceutics at the US Food and
Drug Administration (FDA). Her expertise
in this field stems from her entire aca-
demic and industrial career: before joining
the FDA, she was Director of the Pre-Clin-
ical ADME Group at DuPont Merck Phar-
maceutical Company. Her colleague, Pro-
fessor Lawrence J. Lesko, chairs the FDA
Pharmacogenetics Working Group and
currently serves as a Regent of the Ameri-
can College of Clinical Pharmacology. Both
focus on the different genetic setups of in-
dividual patients and how this is reflected
in their respective response (or nonre-
sponse) to certain biopharmaceuticals. The
current state of knowledge with regard to
DNA-based differences [e.g., single nucleo-
tide polymorphism (SNP)] in pharmacoki-
netics and pharmacodynamics of medica-
tions is discussed as well as the impact on
individualized medicine, in which a spe-
cific drug is developed for a certain cohort
of patients with the same genetic makeup.
As part of the FDAs strategic action plan
and critical path initiative, the Agency is
developing standards to effectively handle
emerging technologies, especially in the
areas of pharmacogenomics, in order to
provide efficient and rapid translation of
new scientific developments and break-
throughs into safe and effective biophar-
maceuticals. Therefore, in their chapter
they also provide an update on how phar-
macogenomic information is being applied
and reviewed in Investigational New Drug
(IND) and New Drug Application (NDA)
submissions. Critical issues in the FDA
regulatory review and labeling implications
of pharmacogenomic data, together with
various challenges in the effective transla-
tion of pharmacogenomic information to
clinical practice, are also discussed.
How large-scale detection of genetic vari-
ation works and how this technically leads
to personalized medicine is explained by
Dr. Jrg Geistlinger, founder of Array-On
Microarray Technologies, and Dr. Peter
Ahnert from Ohio State University, now at
the University of Leipzig, Institute for
Clinical Immunology and Transfusion
Medicine. Human genomes are more than
99% identical and less than 1% variation
determines the genetic differences be-
tween individuals. Over 80% of this varia-
tion is due to SNPs and a major cause of
different individual responses to drugs
and environmental substances; further-
more, this determines the presence of sus-
ceptibility alleles and, therefore, genetic
predisposition for hereditary diseases. Both
colleagues describe how to determine the
large amount of base variation in the hu-
man genome in a reproducible, fast and
economical manner with an extremely pre-
cise high-throughput DNA chip-based
technology for SNP typing.
Obviously, with the human genome se-
quence now available, new strategies are
necessary to efficiently mine the genome
for the set of human drug targets. One
such class of drug targets are GPCRs (G-
protein coupled receptors): they present
the largest single protein family in our
pharmaceutically tractable genome (PTG).
A modern approach to identify such fami-
lies is called homology/orthology min-
ing, which seeks to identify novel genes
with a similar sequence to known genes
or proteins. This approach uses Basic Lo-
cal Alignment Search Tool (BLAST) algo-
rithms where a known gene or protein is
used as the seed with which to search
the novel sequence space. The query ele-
ment can be a sequence from the same or-
Executive Summary XLIV
ganism (i.e., homology mining) or from a
different organism (i.e., orthology mining).
This approach identifies genes likely to be
family members of the query gene, and
thus extends the number of genes belong-
ing to protein families previously proven
to be drug targets (e.g., GPCRs). The
working of GPCRs (and the underlying
mechanism of signal transduction, which
makes them drug targets) is shown in an
impressive video animation on the supple-
mentary CD-ROM.
CuraGen is one of the first companies
to apply a systems biology approach to sys-
tematically identify 6273 potential drug tar-
gets, defining for the first time a complete
PTG. Professor Jonathan M. Rothberg
from Yale University founded CuraGen
Corporation, headquartered in New Haven,
in 1993. He was named an Ernst and
Young Entrepreneur of the Year, and in
2004 also elected to the US National Aca-
demy of Engineering for his pioneering
work in mining the human genome. To-
gether with Dr. Carol Pea from Yale and
Johns Hopkins and Dr. Bonnie E. Gould
Rothberg from Yale, who is also Director
of Clinical Development for CuraGens
lead compound, they describe their inte-
grated systems biology approach that uses
high-throughput genomic, transcriptomic
and proteomic technologies to anchor a
drug development process, which can ef-
fectively and efficiently nominate, prosec-
ute, validate and ultimately develop targets
and their relevant biopharmaceutical
drugs. This is why, in 2003, CuraGen be-
came one of the first companies to go di-
rectly from the genome to the clinic, with
a drug candidate derived from their inte-
grated systems biology approach.
A very striking example where knowledge
of the genetic setup from an individual pa-
tient is converted into the benefit for this
patient by means of individualized medi-
cine is Herceptin
from Genentech/Roche.
Cancer development is the result of cumu-
lating genetic alterations. A recent major
advance in the treatment of cancer is the
emergence of therapies aimed specifically
at altered gene products or distorted gene
expression, often referred to as targeted
therapy. As these modifications are not
present in normal cells, these new antican-
cer drugs very specifically target the tumor
cells and, to a large extent, avoid damage
to normal cells. Reliable detection of the al-
tered gene or its protein product to identify
patients that may benefit from these tar-
geted therapies is therefore often indicated.
Herceptin (trastuzumab) is one of the few
targeted therapeutics, which is directed
against the human epidermal growth factor
receptor-2 (HER-2) of breast cancer cells.
HER-2 was originally identified by Robert
Weinbergs group in 1984, and it was later
discovered by Axel Ullrich and Denis Sla-
mon that the HER2/neu gene was amplified
in breast cancers.
Herceptin is a humanized mAb targeted
to HER-2 overexpressed in 2030% of hu-
man breast cancers because only a certain
patient population expresses this specific
biomarker, only these women are suscepti-
ble for treatment with Herceptin. Applying
one round of diagnostic before the start of
the therapy reveals those individuals who
exhibit the HER-2 receptor and can there-
fore successfully be treated against breast
cancer with Herceptin. The working mech-
anism of Herceptin and the selection of pa-
tients is nicely shown on a movie on the
supplement CD-ROM. This success story
is described in a chapter provided by Dr.
Thorsten Gutjahr and Dr. Carsten Rein-
hardt from Hoffmann-La Roche. Before
joining the company, Thorsten Gutjahr did
postdoctoral studies at the University of Ca-
lifornia and is currently leading the Hercep-
tin program at Roche. Carsten Reinhardt is
Executive Summary XLV
a Medical Doctor and did his practical year
at Mount Sinai Hospital, New York, before
he became Head of a research group at
the Max-Planck-Institute. After a position
as Head of Clinical Development at Frese-
nius, he is now working as International
Medical Leader for Roches Strategic Mar-
keting Department. Their contribution out-
lines the clinical development of Herceptin
and the parallel development of diagnostic
tests that identify those patients that are
most likely to benefit from Herceptin-based
therapy. The manufacture of Herceptin is il-
lustrated in a video on the supplementary
CD-ROM. [Note in press. Two current phase
III studies using Herceptin to fight an ag-
gressive form of early-stage breast cancer
after surgery have been halted early, be-
cause data showed that Herceptin cuts the
rate of disease recurrence by 52%. Follow-
ing these positive results, analysts from
Merrill Lynch increased their sales esti-
mates for 2009 from 1.6 up to 2.7 billion
Euros]. The colleagues from Roche explain
why the co-development (drug and diagnos-
tic) serves as a prime example for individua-
lized cancer therapy, sets a new standard for
drug development in oncology and paves
the way for a new generation of modern bio-
pharmaceuticals.
siRNA The Magic Bullet and other Gene
Therapeutical Approaches
Once a certain disease is diagnosed, gene
therapy is becoming a powerful tool for its
treatment. Therapeutic angiogenesis is a
novel and first-of-its-kind treatment strat-
egy for patients with chronic myocardial
ischemia (stable angina). As nicely out-
lined by my colleagues and friends (I had
the opportunity to work with them for half
a year in sunny California) Gabor Rubanyi,
Michael McCaman, Frank Castillo, Jacob
E. Kung, Yasushi Ogawa and Farah S. Fa-
waz and Erik Whiteley, Elisabeth Lemberg,
Mei Tan, Bruce Mann, Enno Pungor, who
complete the team of experts from Berlex.
Erno Pungor received his PhD from MIT
where he is still active in giving lectures.
He is also an Editorial Member of the
journal Current Pharmaceutical Biotechnol-
ogy, has extensively published in this filed
and holds several patients. The goal of
their therapeutic angiogenesis approach is
to stimulate the formation of collateral ves-
sels to restore blood flow to ischemic re-
gions of the heart. They all have extensive
expertise in this field: Gabor, for example,
was Associate Professor at Mayo Clinic
Medical School before becoming Head of
the Gene Therapy Department at Berlex
and Professor at the University of Califor-
nia, Davis. He is the author of several
books and founder of the biomedical jour-
nal Endothelium. Michael has broad exper-
tise in process and analytical development
for protein, viral and cell therapy products,
and is currently heading the Bio-analytical
Department at Berlex, where he is develop-
ing novel cell therapies. Frank was founder
and Head of the Fermentation Laborato-
ries at the Venezuelan Institute of Scientif-
ic Research in Caracas before becoming
Head of Fermentation and Cell Culture
Development at Berlex. His current work
involves process development for the Good
Manufacturing Practice (GMP) production
of mAbs, recombinant proteins and vec-
tors for gene therapy. Jacob studied at Uni-
versity of California, Berkeley and success-
fully developed an HIV urine antibody de-
tection ELISA (enzyme linked immuno
sorbent assay) before joining Berlex as
Manager of Quality Control. Farah studied
at the American University of Beirut, Leba-
non, where she received the Presidents
Honor List Award before moving on to the
University of Michigan, Ann Arbor, the
University of California, San Francisco and
Executive Summary XLVI
then finally to Berlex. Here she was re-
sponsible for the development of potency
assays for the adenovirus gene therapy. Ya-
sushi, who is Director of Quality Control
and responsible for GMP issues, did his
PhD in Biochemistry at the University of
California, Los Angeles. Before joining
Berlex, he was Director of Pharmaceutical
Development at Celtrix, where he worked
on the characterization of growth factors
(as also published in Science), and commu-
nication with the FDA on protein analysis
and Chemistry, Manufacturing and Con-
trols (CMC) issues. Now these colleagues
share with us their excitement in this
promising gene therapeutic approach:
Ad5FGF-4. Preclinical studies in pigs with
myocardial ischemia showed that intracor-
onary injection of Ad5FGF-4 (consisting of
the gene coding for angiogenic growth fac-
tor FGF-4 and an adenovirus-based deliv-
ery vehicle) is well tolerated and effective.
Intracoronary infusion of Ad5FGF-4 in-
creased myocardial perfusion to control
levels and restored normal heart wall mo-
tion. Based on these data, development of
this unique biopharmaceutical and clinical
trials in patients with stable angina were
initiated. Two clinical trials have been
completed employing intracoronary gene
transfer of Ad5FGF-4: the angiogenic gene
therapy (AGENT) trial was the first multi-
center, randomized, double-blind, placebo-
controlled, dose-escalation trial of intracor-
onary gene therapy on 79 chronic stable
angina patients. A subsequent trial
(AGENT-2) of intracoronary Ad5FGF-4 at a
single dose of 10
10
virus particles was un-
dertaken to study the effect on myocardial
perfusion on 52 patients with chronic
stable angina. Safety, efficacy and pharma-
cokinetic data of these clinical trials are
presented and discussed.
Another promising gene therapeutic
approach is DNA vaccination with the
MIDGE
and dSLIM
technologies from
MoloGen. The mechanism behind this
technology is shown on the supplementary
CD-ROM. From my studies at the Free
University of Berlin, I know Professor
Burghardt Wittig, founder and CEO of this
innovative biotech company, who is Profes-
sor of Molecular Biology and Bioinfor-
matics at the Free University. Burghardt
was the supervisor of my Diploma thesis
and we share not only our interest in gene
cloning, but also a passion for snow board-
ing. He has a dual background (medicine
and physics), because in addition to his
studies in medicine he attended courses in
physics at the California Institute of Tech-
nology (CalTech), and was Visiting Profes-
sor from the Weizmann Institute of
Science and at MIT, where he worked with
one of the most influential scientists in
the LifeSciences, Professor Alexander
Rich. The foundation of Professor Wittigs
chair, combining Molecular Biology and
Bioinformatics, in 1989 has been ground-
breaking few would have thought that
10 years later molecular biological infor-
matics would be essential for projects in
genomics all over the world.
MoloGen AG was founded in 1998 in
the same year the company had its initial
public offering (IPO) in the Berlin Stock
Exchange and in July 2001 it stepped up
to the Frankfurt Stock Exchange. It was
the first German start up solely financed
by the stock market and the first German
biotech company to go public in the year
of its foundation. For his courage and in-
novative spirits in forming Germanys first
PrivatePublic Partnership between Molo-
Gen AG and his academic alma mater, the
Free University, Professor Wittig was
granted the prestigious Entrepreneur of
the Year award in 1999.
As the inventor of MIDGE he describes
the concept of Minimalistic Immunogeni-
Executive Summary XLVII
cally Defined Gene Expression. MIDGE
vectors are innovative, nonviral expression
constructs that are characterized by unique
features, particularly by their high safety.
MIDGE expression vectors consist only of
a minimal genetic content: the cytomega-
lovirus (CMV) promoter, the transgene
and a polyadenylation site, leading to the
characteristic small size of these vectors.
As a prominent feature, the nucleotides of
the loops are ideally suited to covalently at-
tach various molecules, leading to defined
properties of the vector such as tissue- and
cell-specific targeting, increasing the rate
of transfection of cells or expression of
proteins. Modification of MIDGE vectors
with T
H
1 peptide results in MIDGET
H
1
vectors with increased potency. In addition
to this, they induce a T
H
1-type of immune
response and, thus, are particularly suit-
able for DNA vaccination. Convincing re-
sults have been achieved with MIDGE
T
H
1 vectors to prevent leishmaniasis in
dogs. Furthermore, MIDGET
H
1 coding
for HBsAg (antigen from hepatitis B virus)
induces high titers of antibodies in mice
and now needs to be evaluated in clinical
trials. The emphasis upon hepatitis B no
doubt reflects the global significance of
this condition. Two billion people are in-
fected worldwide, with 350 million individ-
uals suffering from lifelong chronic infec-
tions. In excess of a million sufferers die
each year from liver cancer and/or cirrho-
sis triggered by hepatitis B. In tumor ther-
apy, either combined effects of MIDGE
vectors and immunomodulators dSLIM
(double-stem loop immunomodulator) or a
combination of dSLIM with chemotherapy
are used to increase the patients immuno-
logical response against tumor cells. In
mice, tumor diseases could be prevented
by immunization with ex vivo transfected
tumor cells (for cell-based therapy) or by
DNA vaccination with a tumor-associated
antigen (TAA) in both strategies dSLIM
was used as immunomodulatory mole-
cules. In a human clinical trial, the appli-
cation of a therapeutic vaccine using
MIDGE vectors for ex vivo transfection of
autologous tumor cells and combination of
this cell-based vaccine with dSLIM re-
sulted in a clinical response of 50% of the
patients. In addition, another clinical can-
cer trial showed the safety, benefit and im-
munomodulatory potential of dSLIM in
combination with chemotherapy.
A new class of potential biopharmaceuti-
cal drugs that can be designed to specifi-
cally target transcription factors involved
in the pathogenesis of a given disease is
the double-stranded decoy oligodeoxynu-
cleotides (ODNs). There has been an ex-
plosion in the use of transcription factor
decoys as tools for studying gene regula-
tion and as experimental therapy to treat a
variety of pathological conditions. Ongoing
preclinical and clinical development pro-
grams at various emerging biotech compa-
nies as well as academic research institu-
tions are currently elucidating the poten-
tial of this promising new drug class. My
colleague from EAPB, Professor Heiko E.
von der Leyen was at the Department of
Cardiovascular Medicine, Stanford Univer-
sity with research focusing on cardiovascu-
lar gene therapy and oligonucleotide deliv-
ery, before he was appointed CSO of
AVONTEC, developing decoy ODNs. Ini-
tial clinical studies employing AVONTECs
STAT-1 (signal transducer and activator of
transcription) decoy ODN in allergic asth-
ma as well as psoriasis have shown no
side-effects (early phase I trials) and unre-
markable tolerability. Another successful
clinical trial applying decoy ODNs is de-
scribed as well: binding of E2F (transcrip-
tion factor) to decoy ODN prevents it from
transactivating the gene expression of cell
cycle regulatory proteins like PCNA (prolif-
Executive Summary XLVIII
erating cell nuclear antigen), cdk2 (cyclin-
dependent kinase 2) and c-myc (oncogene),
thereby inhibiting vascular smooth muscle
cell proliferation and subsequent neointi-
ma formation. Clinical therapeutic effects
in a clinical development program (PRE-
VENT) with a pressure device-mediated ex
vivo application of E2F decoy ODN in vas-
cular grafts of patients with late-stage pe-
ripheral artery disease were first demon-
strated at Harvard University. The E2F de-
coy ODN has been shown to be effective
in phase I/II and IIb trials, and is currently
being evaluated in two phase III clinical
trials. The peripheral artery bypass study
(PREVENT 3) is testing edifoligide
TM
(E2F
decoy) in 1400 patients who have under-
gone peripheral artery bypass surgery at
approximately 80 medical centers through-
out the US. PREVENT 4 is evaluating edifo-
ligide in 2400 patients who have undergone
surgery at more than 100 US medical cen-
ters. The FDA has granted edifoligide a
fast track status for both coronary and pe-
ripheral indications due to the unmet med-
ical needs the product may address (bypass
atherosclerosis). Enrolment for both studies
has been completed and data are expected
to be presented in 2005.
Christoph Bagowski, another friend from
the Max-Planck-Institute, was working on a
related topic at Stanford University. He
shares with us his several years of first-hand
experience with antisense RNA and siRNA.
Christoph did his PhD together with Axel
Ullrich, who discovered that the HER2/
neu gene was amplified in breast cancers.
As described above, this finding and the
consequent development of the successful
targeting approach eventually led to Her-
ceptin. Currently, Christoph is Assistant
Professor at the Institute of Biology at the
University of Leiden, and describes how in
recent years the chemical structures of oli-
gonucleotides have been optimized and
the half-lives of the antisense molecules
have been improved. These next-generation
compounds have recently recaptured the in-
terest of the pharmaceutical industry in
antisense technology. Vitravene
TM
(fomivir-
sen), which treats a condition called CMV
retinitis in people with AIDS, is the first
FDA-approved antisense drug. Vitravene
has been developed by Isis Pharmaceuticals
(Carlsbad), which licensed the worldwide
commercial rights to Novartis. Another suc-
cessful example from Isis Pharmaceuticals
is presented as well: they established a drug
development collaboration programme with
OncoGenexTechnologies (Vancouver) in
2001 to develop and commercialize OGX-
011. OGX-011 (ISIS 112989) is an antican-
cer antisense drug, which inhibits clusterin,
and has currently entered phase II clinical
trials for patients with prostate cancer and
other solid tumors. Another promising ex-
ample is LY2181308 (ISIS 23722), a sec-
ond-generation antisense drug which was
licensed to Eli Lilly. In preclinical studies,
LY2181308 demonstrated activity in multi-
ple in vivo models of cancer, hence in No-
vember 2004 Lilly initiated phase I clinical
trials in cancer patients. LY2181308 targets
survivin, a molecule that allows the survival
of cells that would normally undergo pro-
grammed cell death or apoptosis.
After a comprehensive review of anti-
sense RNA, Christoph then switches to
siRNA. In comparison to antisense RNA,
RNAi (RNA interference) is just a little over
5 years old and the basis of this gene sup-
pression in plants and fungi only became
clear in 1998 when Fire and colleagues pub-
lished their seminal work in Nature describ-
ing gene silencing in Caenorhabditis elegans
by the artificial introduction of double-
stranded RNA (dsRNA). RNAi reigns
among the most significant scientific dis-
coveries at the turn of the 21st century, both
for its impact on fundamental genetic re-
Executive Summary XLIX
search and on biotechnology and the devel-
opment of biopharmaceuticals. RNAi has
been shown to inhibit gene expression post-
transcriptional via cytoplasmic mRNA de-
gradation; it has been successfully utilized
for tissue-specific gene knockdown in
mice, thus proving its function in a whole
animal. Now well-documented in mam-
mals, RNAi results in gene suppression by
cleavage or translational attenuation of tar-
get mRNA using siRNA or short hairpin
RNA (shRNA), respectively, as the func-
tional intermediates in a highly coordinated
protein: RNA complex known as RISC
(RNA-induced silencing complex). Fortui-
tously, these regulatory RNA molecules are
readily synthesized and when artificially in-
troduced in vitro or in vivo effect mRNA tar-
get-specific suppression. Coupled with the
ease of producing the siRNAs (and related
shRNAs), RNAi-mediated gene silencing
has now emerged as an extremely valuable
technology to reduce or knockdown expres-
sion of specific genes and allow for assess-
ment of gene function. In the laboratory,
RNAi is routinely used to reveal the genetic
secrets of development, intracellular signal-
ing, cancer, infection and a full range of
other phenomena. However, can the phe-
nomenon hailed by Science as the Break-
through of the Year in 2002 break out of
the laboratory and lead to novel therapies
as well? Pharmaceutical giants are hoping
so and several biotech companies have bet
their futures on it; however, not everyone
is so optimistic about the future of RNAi
therapy. At the heart of its promise as a
powerful biopharmaceutical drug lies the
exquisite selectivity of RNAi like the fa-
bled magic bullet, an RNAi sequence
seeks out and destroys its target without
affecting other genes (similar as for anti-
bodies, as described by Uwe Gottschalk
elsewhere in this book). Indeed, the discov-
ery of RNAi has, in a very short time, ini-
tiated a revolution in molecular biology
and the study of gene expression.
The challenges facing siRNA are similar
to those that any potential drug candidate
would face: (a) ensuring highly potent tar-
get inhibition, (b) achieving appropriate
target specificity, (c) assuring stability of
the active drug in biological fluids, (d) di-
recting distribution to the appropriate tar-
get organ and (e) minimizing target- or
chemical class-based toxicity. Strategies to
address these issues include rational
siRNA sequence selection (based on bioin-
formatics and sophisticated design algo-
rithms) and the use of chemical modifica-
tions of, and conjugation to, siRNA that
enhance serum stability, pharmacokinetics
and biodistribution. It is exactly this that
has been successfully performed by Dhar-
macon Inc., where an interdisciplinary
approach with contributions from bioinfor-
matics, molecular and cell biology, chemis-
try, and pharmacology is followed to de-
sign highly active, stable and specific
siRNAs. Anastasia Khvorova from Dhar-
macon Inc., shares her tremendous experi-
ence on the development of stable and se-
lective siRNA with us: she has extensively
published on this topic only over the last
2 years in Nature, Nature Biotechnology, Cell
and Proceedings of the National Academy of
Sciences of the USA. Another successful ap-
plication of siRNA, namely the treatment
of HIV, will be discussed later by John
Rossi from the City of Hope hospital.
After learning about SNPs, gene thera-
py, antisense, decoy oligonucleotides and
siRNA, we will now focus on another type
of RNA, which was only discovered re-
cently when large-scale sequencing and
data-mining technologies became available
in the postgenomic era. In the last decade,
there were a growing number of reports
concerning novel genes which produced
transcripts without protein-coding capacity.
Executive Summary L
Such RNAs, named noncoding or nonpro-
tein-coding RNAs (npcRNAs), play impor-
tant roles in many regulatory processes in
all organisms. In eukaryotes, they are in-
volved in cell differentiation and develop-
ment. Many of the mammalian npcRNAs
are localized within chromosomal regions,
which are linked to certain diseases, in-
cluding neurobehavioral and developmen-
tal disorders and cancer. The understand-
ing of npcRNAs biology may open new
perspectives for molecular diagnostics and
biopharmaceuticals.
Professor Volker A. Erdmann from the
Free University at Berlin was always an
RNA pioneer and close friend to James
D. Watson. During my studies at the Free
University, I learned a lot from him about
all types of ribonucleic acids and, in addi-
tion, he introduced me to his friend the
Father of DNA. Thus, I had the privilege
to get to know James D. Watson and also
to meet with him in person. During the Fac-
ulty Meeting at Charit in October 2004, I
had the opportunity to spend an entire
evening with him to discuss recent develop-
ments in biotechnology and his view on
modern biopharmaceuticals which was
of course a very valuable input for this book
(see his quote for Modern Biopharmaceuti-
cals). Unfortunately, I did not have the same
luck with the other DNA code-breaker: I
only had E-mail correspondence with Fran-
cis Crick when I invited him to contribute to
this book. He replied that he very much ap-
preciated the endeavor of creating such a
comprehensive book: ... Nice of you to
ask me to contribute to your book on bio-
pharmaceuticals . . . Unfortunately I am in
very poor health so do please excuse me.
Apologies, Francis Crick Half a year later
he died and I deeply regret that I never
had the chance to meet him in person. Sir
Francis Harry Compton Crick made an en-
ormous contribution to science, our under-
standing of biology and the health of man-
kind. His death is a sad loss to science,
especially modern biotechnology he will
be sorely missed!
Coming back to RNAs, I was very grate-
ful when Professor Erdmann agreed to
provide a chapter on npcRNAs. Professor
Volker Erdmann received his PhD from
the Max-Planck-Institute for Experimental
Medicine and moved, after postdoctoral
studies at the University of Wisconsin,
back to Berlin to work at the Max-Planck-
Institute for Molecular Genetics. Since
1980 he has held the Chair of Biochemis-
try and Molecular Biology at the Free Uni-
versity of Berlin and is also Director of the
RNA-Network. He is recipient of the pres-
tigious Leibniz Award from the German
Research Council (DFG) and member of
the Academy of Sciences. Together with
his colleagues Professor Jan Barciszewski
and Dr. Maciej Szymanski from Polish
Academy of Science, they provide an excel-
lent overview on npcRNAs and their po-
tential as biopharmaceuticals.
For a very long time, it has been as-
sumed that the regulation of gene expres-
sion essentially depends on the activity of
specific proteins, i.e., transcription factors,
responsible for switching genes on and
off. This is nicely shown for the Gene-
Switch
experiment (turning on Lucifer-

ase-light in the mouse hind limb) on the
supplement CD-ROM. In general, the ex-
pression of a particular gene can be regu-
lated on three different levels. The struc-
ture of chromatin and/or epigenetic fac-
tors (e.g., methylation) establishes the pre-
transcriptional level which determines if
the gene can be transcribed (active) or not
(inactive). The expression of active genes
is regulated on the transcriptional level by
a number of trans- and cis-acting factors
which govern the timing and efficiency of
transcription. Finally, at the posttranscrip-
Executive Summary LI
tional level, the amount of the protein to
be produced is determined. The posttran-
scriptional regulatory mechanisms can op-
erate on different steps of the pathway
from pre-mRNA splicing to translation on
the ribosome. The last decade has seen an
unprecedented growth of biological data
resulting mostly from nearly industrial-
scale sequencing projects of genomes from
a variety of organisms, including humans.
The results of these efforts demonstrated,
beyond all doubt, that the views concern-
ing many aspects of molecular biology
which were obvious in the pregenomic era
need to be revised in particular, this ap-
plies to the role of RNA in the cell. In the
context of recent discoveries in this field, it
is evident that the functions of RNAs can
no longer be treated as accessory to pro-
teins. It seems that the correct gene ex-
pression patterns are governed by the in-
tricate network of RNAs which control the
flow of information in the cell.
The most striking finding revealed by the
analysis of mammalian genomes is the rel-
atively small number of protein genes. In
the initial analysis of the draft on human
genome, presented by the Human Genome
Sequencing Consortium, it was estimated
that the human genome contains 26500
29000 protein-coding genes. Similar num-
bers (2700030500) were revealed for the
mouse genome, and more recent analyses,
using a gene prediction program based on
comparative analysis of human and mouse
genomes, yielded 44242 and 44770 protein
genes, respectively. The protein-coding part
or the open reading frames (ORFs) of these
genes account for less than 2% of the geno-
mic DNA. ORFs, together with the untrans-
lated regions (5'- and 3'-UTRs) and introns
represent around 2527% of the genome.
The rest of the nonprotein-coding portions
are composed primarily of repetitive se-
quences, which make up approximately
46% of the genome. Although the function
of the remaining quarter of genomic DNA
is largely unknown, one can assume that
at least some of its portions are responsible
for the spatial and temporal coordination of
gene expression.
In prokaryotes, where intergenic and un-
translated regions are short and splicing is
an exception rather than the rule, the
ORFs of protein-coding genes account for
over 90% of genomic DNA. In simple eu-
karyotes, the nonprotein-coding DNA con-
stitutes 1040%. In invertebrates, the non-
coding part accounts for around 7090%
of the genome and in mammals as much
as around 98%. Therefore, these nonpro-
tein-coding regions of genomic DNA may
be responsible for the regulation of com-
plex mechanisms which underlie develop-
ment and differentiation by means of con-
trolling the expression of proteins which
play a role in the cells hardware.
Obviously, our current knowledge of the
nature of npcRNAs and RNA-mediated
regulation of cellular processes is still very
superficial, and only a small fraction of
npcRNAs identified to date have been
characterized in terms of function or ex-
pression patterns. However, there is no
doubt that npcRNAs hold the key to un-
derstanding the functioning, development
and evolution of complex biological sys-
tems, and as demonstrated by several ex-
amples, npcRNAs are potentially good
markers for human diseases including
early detection of certain forms of cancer.
Systematic studies on npcRNA expression
profiles will lead to the development of
highly accurate molecular diagnostic tools
which could be applied not only for detec-
tion, but also for prognosis and the design
of modern biopharmaceuticals.
Executive Summary LII
Mobilis in Mobile Human Embryonic Stem
(hES) Cells and other Sources for Cell Therapy
After discussing the current status and fu-
ture trends in nucleotide-based biopharma-
ceuticals, cell therapy and transplantation
medicine is another very powerful and
promising field. Although organ transplan-
tation has been one of the major medical
advances of the past 30 years, it is becom-
ing increasingly apparent that the supply
of organs is limited and will not improve
with current medical practice. In addition
to the technical hurdles, one reason for
this dilemma is that in Berlin, for exam-
ple, only 3% of the (potential) donors have
previously provided written consent that
their organs can be used for transplanta-
tion. For Charit hospital alone (one of the
more than 50 hospitals in Berlin), about
1000 patients are waiting for a new kidney.
Also, it cannot be ruled out completely
that contamination with viruses or other
adventitious agents will occur in the
course of transplantation as just hap-
pened in Germany early 2005. Six patients
who received organ transplantation were
infected with rabies, because the female
donor was infected with this RNA virus,
which belongs to the Rhabdoviridae family.
The recipients who were lucky in the
first place to receive an organ at all even-
tually died, because along with the organ
they also received this fatal disease.
Although the normal prevalence of this
animal virus in Europe and the US is lit-
erally zero, in countries like India about
50000 people die each year from rabies.
Several questions arise from such an inci-
dent, e.g., How complex does organ need
testing to be and what procedures can
completely exclude the risk of infection?,
bearing in mind that 70000 transplanta-
tions have been successfully performed in
Germany so far without any transmitted
infection. Another observation comes into
these considerations: some infections have
long latencies, e.g., for rabies it can take
up to 7 years until the first symptoms of
the disease for slow viruses even longer!
Organogenesis represents a welcome al-
ternative to combat organ shortage and
also to prevent the recipient becoming in-
fected through donor contamination. Or-
ganogenesis of complex tissues, such as
the kidney, requires a coordinated sequen-
tial transformation process, with individual
stages involving time-dependent expres-
sion of cellcell, cellmatrix and cellsig-
nal interactions in three dimensions. Em-
bryonic precursor tissues are composed of
functionally diverse stem/progenitor cell
types that are organized in spatially com-
plex arrangements. The theme of tempor-
alspatial patterning of progenitor cell in-
teractions is programmed in precursor tis-
sues leading to their growth and develop-
ment. Indeed, recent data presented in
Transplantation, Trends in Biotechnology and
Blood by Professor Yair Reisner from Weiz-
mann Institute pinpoint a window of time
in human kidney organogenesis that may
be optimal for transplantation into mature
recipients. Professor Reisner explains in
his chapter how window transplants are
defined by their remarkable ability to grow,
differentiate and undergo vascularization,
achieving successful organogenesis of
urine-producing miniature kidneys with
no evidence of transdifferentiation into
nonrenal cell types. Also, they lack tumori-
genicity and have reduced immunogenic-
ity, compared to adult counterparts. Reis-
ner and coworkers were able to show that
transplanted early human kidney precur-
sors develop as chimeric organs in which
blood vessels are of both donor and host
origin, whereas external vessels required
for graft maintenance are mostly host
derived. In contrast, nonwindow trans-
Executive Summary LIII
plants (earlier or later in gestation) can, re-
spectively, form teratomas or are more
prone to immune rejection and are there-
fore less suitable for organogenesis.
Despite this success, there are obviously
still some drawbacks, and the greatest hur-
dle for cells, and transplants in general, is
the rejection reaction from the recipient,
i.e., patient. As the genetic setup from the
donor is different, the donor cells are recog-
nized as foreign and these newly received
cells trigger the patients immune system.
To overcome rejection reactions in part, or
at least reduce them, immune suppressants
are given to patients receiving donor cells,
tissues or other transplants. This approach
does not solve the problem per se, since in
most cases the immune response is only re-
duced, not eliminated, and some patients
have severe side-effects or this approach
does not work at all. To completely circum-
vent these issues, the patient needs to re-
ceive his or her own cells and this can
be achieved by the use of stem cells.
Since 1998 stem cell research has be-
come a major and central part of biomedi-
cal research. James Thomson at the Uni-
versity of Wisconsin and John Gearhart at
Johns Hopkins University were the first
who were able to isolate and propagate hu-
man stem cells, hence making them avail-
able for research purposes. Human em-
bryonic stem cells (hES) are extremely val-
uable, because they can differentiate into
all of the more than 200 cell types of the
human body and are therefore very pro-
mising biopharmaceuticals. However, bio-
ethical aspects influence worldwide re-
search activities with hES cells and limit
their access at least in some regions.
This topic is very controversially discussed
and the intention, as well as the pros and
cons, can range from political, over ethical
and commercial to personal reasons. Cur-
rent US President George W. Bush, for ex-
ample, states that scientists funded by the
federal government could not derive fresh
human stem cells because, he said, it is
immoral. Nancy Reagan announced that
she supports ES cell research to cure, for
example, Alzheimers disease, which
claimed the life of her husband, former
President Ronald Reagan. In addition, bi-
partisan groups of Congressmen and Se-
nators each wrote to Bush to ask him to
change his policy. Some States, such as
California and New Jersey, have even
passed their own measures to support hES
research. The debate on stem cell research
remains a hot topic, and during the US
President election campaign, John Kerry
replied to Natures questionnaire on
science issues that he would lift the re-
strictions that Bush has placed on stem
cell research, while ensuring ethical lines
would not be crossed. [Note in press. After
a long battle, the United Nations accepted
a declaration forced by the US to com-
pletely ban human cloning. According to
this declaration, not only reproductive clon-
ing, but also therapeutic cloning is forbid-
den. However, again, this declaration is
not binding, because every country makes
its national laws, hence every country can
legally perform cloning if they want. For
example, in Europe, France and the UK
voted against the declaration, whereas Ger-
many voted for it and this obviously re-
presents the current status of the respec-
tive national laws. Also in the US, follow-
ing the new results from Professor Hwang
regarding patient-specific cloning of hES
cells, the topic is being very controversially
discussed again: The US House of Repre-
sentatives has decided after a very emo-
tional debate to publicy finance hES cell
research and reverse the restriction that
was implemented by Bush. In Washing-
ton, 238 representatives voted for this new
law and 194 against it and this despite
Executive Summary LIV
President Bushs warning. Since he is
strictly against this research, Bush claimed
that he for the first time during his four-
and-a-half years Presidential period
would make use of its right to place a veto
against this bill if this new law were to
pass the Senate. Even that would not erase
the significance of this victory in the Re-
publican-controlled House for advocates of
hES cell research: 50 Republican members
also voted for the change in policy under-
lining that the pendulum of public opin-
ion is swinging strongly for change. This
again is the reason why the Democratic
Senator Ted Kennedy says that Bush
shows one more time, that he has no un-
derstanding for the priorities of American
people. Indeed, according to a Gallup
study, 60% of the US citizens support hES
cell research].
Friedrich Nietzsche (18441900) once
raised the question: Do you believe that
sciences would ever have arisen and be-
come great, if there had not beforehand
been magicians, alchemists, astrologers,
and wizards who thirsted and hungered
after secrets and forbidden powers? As we
have seen, the topic of human cloning is
probably the most controversially dis-
cussed topic ever, because this technology
makes us in a sense God-like, in that we
are now empowered to create human
beings and also may materialize mans
dream of being forever young. In particu-
lar, since 1996 when Ian Wilmut (see his
quote for Modern Biopharmaceuticals) from
the Roslin Institute in Scotland cloned the
sheep Dolly, people from all over the
world have warned against cloning hu-
mans, and also Wilmut himself claims in
a 2001 Science article Dont Clone Hu-
mans. Other researchers are less restric-
tive when they talk about therapeutic clon-
ing, for example, and James D. Watson
himself is not strictly against human clon-
ing. As he explained to me his personal
view in an exciting discussion during the
Faculty Meeting at Charit in October
2004, this always depends on the aim of
the study and the design of a certain ex-
periment. Obviously, we need to be critical,
we need to have clear regulations, and we
need to discuss and decide cloning experi-
ments on a case-by-case basis but for me
the key question remains: quis custodiet
ipsos custodes (who controls the decision
makers)?
For this purpose, UNESCO established
an International Bioethics Committee
(IBC), which in April 2001 issued a report
on The use of embryonic stem cells in
therapeutic research to discuss hES cell
research in the light of ethical aspects.
However, despite this initiative, bioethical
laws differ somewhat in different coun-
tries and only approximately 30 countries
actually have regulations regarding the use
of human stem cells. Some of these coun-
tries are not allowed to generate hES cells
while others are only allowed to import
such cells for research purposes. Again,
most countries do not have any regulations
at all or they have changed them over
the years. In the UK, for example, it was
forbidden, but regulations have now
changed so that the ban on work with hES
cells no longer exist. In 2004, the British
Human Fertilization and Embryology
Authority (HFEA) has even relaxed their
rules on so-called designer babies to help
sick siblings: the new rules would let cou-
ples with a sick child test embryos con-
ceived during fertility treatment and pick
one for implantation that matches the old-
er childs tissue type. Babies created to
match the tissue of an ill brother or sister
could, for example, donate stem cells to
their sibling to weed out disorders.
From an opposite standpoint, in Ger-
many, generation and even the import of
Executive Summary LV
hES cells is forbidden by law due to a deci-
sion of the Parliament (Deutscher Bundes-
tag). Only as an exception, and only in a
very important case, and only if the cells
were generated before 1 January 2002,
might clearance for import be granted for
a certain research project. These legal cir-
cumstances obviously lead to an imbalance
in freedom to operate, and hence in re-
search opportunities for different institutes
and companies in different countries. This
situation is the reason why in the meantime
the German Ministry for Research has
voted for rethinking these restrictive laws.
The Head of Germanys ruling social party,
SPD, states, that who ever wants social pro-
gress, also needs to have the courage to
approach new frontiers: We cannot let
others perform research abroad, gain profit
from their results and wash our hands in
innocence. And to do so in the technically
and ethically contentious arena of stem cells
is playing with fire. Taking these controver-
sies together, it is a delicate situation and I
had the opportunity to discuss this with
Professor Gnter Stock, Vice-president of
The Max-Planck-Society, Senator of The
German Research Foundation (DFG), as
well as Board Member and Head of Re-
search of Schering AG, and I am very grate-
ful that he contributed a foreword to this
book, also addressing this topic.
Another similar restrictive and limiting
decision from the Bundestag was taken in
November 2004 when they decided to im-
plement an even more restrictive law for
genetic engineering. Despite warnings
from the Max-Planck-Society and the Ger-
man Research Foundation that ... research
would become impossible..., the German
legislature implemented this new law,
which also (down)regulates the growth of
transgenic plants. At the same time, when
German politicians from the Bundestag
decided against the advice from these two
major scientific societies, the Swiss popu-
lation decided with an overwhelming ma-
jority to allow stem cell research in Swit-
zerland. This again shows the different
perceptions, and the impact on research
and development opportunities legisla-
tion in this case is hampering science and
progress rather than supporting it.
A side-track discussion is going on re-
garding the different potencies of hES
cells to differentiate into specific cells (e.g.,
neurons, muscle, hair, skin, heart, liver,
blood cells, etc.) and their respective no-
menclature. The nomenclature according
to the American National Bioethics Advi-
sory Commission (NBAC 1999) is as fol-
lows: pluripotent cells (plures = many) can
differentiate into many different cells and
omnipotent cells (omnia = all) can differ-
entiate into all cell types. The same is true
for totipotent cells (totus = complete), but
these cells in addition to forming all cell
types are able to harmonically assemble
along the system of axis in a proper way
(self-assembly): they are capable of pro-
longed in vitro proliferation and self-renew-
al, but also retain the ability to differenti-
ate into derivatives of all three germ layers
both in vitro and in vivo. Following cultiva-
tion in suspension, these hES cells tend to
spontaneously create three-dimensional ag-
gregates of differentiating tissue known as
embryoid bodies (EBs). Theoretically, if
continuously grown and propagated, these
EBs would lead to the generation of a
complete human being therefore they
are considered totipotent.
According to a definition by Edwards
and Beard (1997), totipotency is the abil-
ity of a cell other than an oocyte to develop
into an entire offspring including the
germ line. The dilemma is that to show
omnipotency and totipotency (that all cell
types are generated, or in addition assem-
ble properly), one would need to grow the
Executive Summary LVI
hES cells to complete (cloned) viable indi-
viduals (and in the best case also the off-
spring to show that germ cells are also
generated and function properly). For
other species, totipotency was shown a
long time ago. Hans Spemann (1869
1941), Nobel Prize laureate of 1935, al-
ready performed some crucial experiments
in 1914, when he mechanically separated
blastomers from the two-cell stage of the
sea urchin (Parechinus) just by vigorous
shaking. Both completely separated blasto-
mers grew to intact adult twins. In 1952,
Friedrich Seidel (18971992) and collea-
gues were able to successfully perform a
similar offspring experiment with mam-
mals: from a two-cell stage, they were able
to grow two normally developed rabbits
(Oryctolagus cuniculus). However, methods
nowadays are much more sophisticated
and it is probably experimentally feasible
to perform the experimentum crucis for toti-
potency in humans; this is referred to as
reproductive cloning, which is strictly forbid-
den for humans. Therefore, showing om-
nipotency and totipotency is practically im-
possible for hES cells.
Thus, if there are any concerns with hu-
man embryonic stem cells, why not use
adult stem cells? Adult stem cells, derived
from bone marrow, for example, would ob-
viously make the entire ethical discussion
about hES cells obsolete. The problem is
that despite 15 years of intense research
adult stem cells cannot be sufficiently
propagated ex vivo to yield interesting
amounts of stem cells. This is why such
adult stem cells are not yet an alternative.
In general, pluripotency of ES cells can be
established traditionally using three differ-
ent approaches. Mouse ES cells can be re-
transferred into early mouse embryos
where they eventually give rise to all so-
matic cells of the chimeric embryo, includ-
ing the germ cells. Such a test cannot be
applied to human ES for obvious ethical rea-
sons. The second approach relates to the
demonstration that ES cells can differenti-
ate to generate derivatives of all three germ
layers in vivo. When hES cells are injected
into immunodeficient mice, they form be-
nign teratomas containing advanced differ-
entiated tissue types representing all three
germ layers. The third and most exciting
approach establishes ES pluripotency dur-
ing in vitro differentiation. Both mouse
and human ES cells, when removed from
the mouse embryonic fibroblast (MEF)
feeder layer and allowed to differentiate,
can form three-dimensional cell aggregates,
EBs, that contain tissue derivatives of endo-
dermal, ectodermal and mesodermal origin.
Indeed, the ability of hES cells to generate a
variety of mature somatic cell types was im-
pressively demonstrated using both sponta-
neous and directed in vitro differentiation
systems. Hence, since the initial report of
the derivation of hES cells, they were shown
to be able to differentiate into cardiac tissue,
neuronal tissue including dopaminergic
cells, islet pancreatic cells, hematopoietic
progenitors, keratinocytes, bone tissue and
endothelial cells.
A real breakthrough with human ES
cells was just recently achieved by the Kor-
ean scientist Professor Woo Suk Hwang
from Seoul University, using enucleated
adult cells as the starting material
Hwang and coworkers were able to obtain
pluripotent human embryonic stem cells
from somatic cell nuclear transfer (SCNT)
of reprogrammed human adult cells as
they recently described in Science. Luckily,
they agreed to provide a chapter for Mod-
ern Biopharmaceuticals on their impressive
results: the highly differentiated genetic
program of the nucleus from an adult cell
was completely reprogrammed after being
introduced into the enucleated oocyte from
a donor. After fusion, a (totipotent) hES
Executive Summary LVII
cell was generated, which (like a fertilized
oocyte) can develop into a blastocyst. (See
also the video kindly provided by Professor
Hwang on the supplementary CD-ROM.)
This major success in the production of a
human SCNT-ES cell line was attributed
to optimization of several factors including
the donor cell type, reprogramming time,
activation protocol and use of sequential
culture system with newly developed in vi-
tro culture medium. One factor of utmost
importance was the use of a less-invasive
enucleation method the squeezing meth-
od (this method is also shown in a video
on the supplementary CD-ROM, kindly
provided by Professor Hwang). Here, the
oocytes were squeezed using a glass pi-
pette so that the DNAspindle complex is
extruded through a small hole in the zona
pellucida. The same small hole is used to
subsequently introduce the somatic nu-
cleus containing the donor DNA before it
is sealed applying an electric stimulus
(this method is also shown in a video on
the supplementary CD-ROM, kindly pro-
vided by Professor Hwang). Applying this
squeezing method rather than aspirating the
DNAspindle complex with a glass pipette
(as others have done so far), the success
rate could be increased tremendously. Con-
trary to that, with use of an aspiration
method, others have recently reported de-
fective mitotic spindles after SCNT in non-
human primate embryos, most likely re-
sulting from the depletion of microtubule
motor and centrosome proteins lost to the
meiotic spindle after aspiration enuclea-
tion. After continuous proliferation for
more than 70 passages, Hwang could
show that SCNT-hES (recipient) cells main-
tained normal karyotypes and were geneti-
cally identical to the somatic nuclear donor
cells. (See also the video kindly provided by
Professor Hwang on the supplementary
CD-ROM.)
[Note in press. Professor Hwang and col-
leagues were able to clone hES cell lines
from 11 patients ranging from 2 to 56
years in age with an unprecedented suc-
cess rate. Whereas in his previous experi-
ments 248 human oocytes were needed to
successfully generate one SCNT-hES cell
line, now only less than 20 oocytes were
needed. This is why famous stem cell re-
searchers like George Daley from Harvard
University and Gerald Schatten from Pitts-
burgh University say that they are im-
pressed by these results which they had
not anticipated even within the next 10
years. Even Ian Wilmut, the pioneer in
cloning, asked Hwang for support and ad-
vice in human cloning the student be-
comes the teacher. Indeed, this method is
ideally suited for therapeutic cloning (e.g.,
generating organs and other transplants),
because the patient is at the same time the
donor (of the somatic nucleus, which is re-
programmed and used to generate the
SCNT-hES cells) and recipient (of the
SCNT-hES cells and/or transplants there-
of). Therefore, these cells have the same
genetic makeup, and will not be recog-
nized as foreign and not stimulate any
immune response this means that there
is no risk of rejection. In addition, these
hES cells can be differentiated into any
specific cells (e.g., neurons, heart, liver or
blood cells). Hence, this methodology is a
quantum leap a major step towards the
use of stem cells in the treatment of dis-
ease and for the first time yields an unlim-
ited source of autologous cells for trans-
plantation medicine!
In August 2004, the British HFEA ap-
proved Ian Wilmuts and other researchers
request to use the same method as their
Korean colleague Hwang to prepare hES
cells for therapeutic cloning. Incidentally,
the same day that Hwang published his
groundbreaking new results, Miodrag Stoj-
Executive Summary LVIII
kovic and colleagues from the International
Centre for Life at Newcastle University (UK)
were able to clone a human embryo apply-
ing the Dolly method, although they could
not yet achieve any hES cells. Stojkovics aim
is to develop insulin-producing islets cells
by means of SCNT-hES from the patients
own cells. Diabetes patients could then be
cured by this method without any risk of
rejection, because these cells have the
patients genetic setup. Stojkovic, a German
veterinary doctor from Munich who moved
to the UK 2 years ago, says that in other
EU countries he would have gone to prison
for his experiments although the 36 oo-
cytes he used came from a clinic specializ-
ing in in vitro fertilization and would have
dumped these oocytes otherwise. Why is
embryonic stem cell research dangerous?,
he asks. In the UK, this research is per-
formed to find out why stem cells are good
and how they can be used. And exactly this
difference in the mental setting is directly
reflected in the different national laws. (Ir-
onically, Professor Ian Wilmut is the recipi-
ent of the German Paul Ehrlich Prize 2005
for his groundbreaking research leading
to the first cloned mammal).
Of course, recent impressive results
(and oxymorons) again fuel the contro-
versy about human cloning. Although the
spokesman of Germanys ruling social
party, SPD, states that Chancellor Gerhard
Schrder is not seeking for a change in
current restrictive laws, he admits that this
might be due in 2 years time. Also, the
German Ministry of Research (SPD) ad-
mits that one needs to rethink current
laws and the Chief of the German FDP
even claims that therapeutic cloning must
be allowed without any restrictions. The
Head of the German National Ethics Com-
mittee, Professor Spiros Simitis, states that
due to the continuous progress in biotech-
nology one needs to re-evaluate current
laws on an ongoing basis and adjust these
as needed.]
Gene and cell-based strategies have
evolved into powerful therapeutic plat-
forms capable of influencing the patho-
physiology of complex, acquired, polyge-
netic diseases. Cardiovascular disease, for
example, is the leading cause of lost pro-
ductivity, morbidity and mortality in indus-
trialized nations, and it is projected that by
2020 cardiovascular disease will be the
leading contributor to the worldwide bur-
den of disease. The past several decades
have witnessed significant advances in car-
diovascular therapeutics the invention of
new interventional and electrophysiological
devices, the development of minimally in-
vasive surgical techniques, and the discov-
ery of new and effective drugs. All this has
altered the natural history of cardiovascu-
lar disease, but ironically, the increased
survival resulting from these treatment
modalities has produced a growing popula-
tion with chronic cardiovascular diseases
who are, or will be, failing current state-
of-the-art therapies. Accordingly, there is
an increasing need for the development
of advanced and innovative therapeutics.
Recent advances in gene- and cell-based
approaches provide unprecedented oppor-
tunities for the discovery of novel therapies
to address this pressing demand.
How SCNT-hES cells can be applied for
myocardial regeneration and transplanta-
tion is impressively demonstrated by Pro-
fessor Lior Gepstein from Technion Insti-
tute in Haifa, Israel. His recent break-
through achievements were another hall-
mark in stem cell research and were
highlighted by the international press. In
addition to his scientific results, the Tech-
nion Institute was recently in the scientific
spotlight when his colleagues Aaron Chie-
chanover and Avram Hershko were
awarded the Nobel Prize in 2004. As first
Executive Summary LIX
described in Nature Biotechnology, Lior and
colleagues used a special scheme to gener-
ate reproducible spontaneous cardiomyo-
cytes differentiating in a hES cell system
by applying undifferentiated hES cells of a
single-cell clone, which was propagated on
top of the MEF feeder layer. The hES cells
were then removed from the feeder layer,
dissociated into small clumps of 320 cells
and grown in suspension for 710 days
where they formed EBs. The EBs were then
plated on gelatin-coated culture dishes and
observed microscopically for the appearance
of spontaneous contraction. Rhythmically
contracting areas appeared at 422 days
after plating in about 10% of the EBs (as im-
pressively shown in a movie on the supple-
mentary CD-ROM). In addition to the mo-
lecular and structural studies described
above, several functional assays including
extracellular and intracellular electrophys-
iological recordings, calcium imaging, and
pharmacological studies clearly demon-
strated that the contracting areas within
the EBs also displayed physiological prop-
erties consistent with an early-stage human
cardiac phenotype. Hence, all the compo-
nents of normal cardiac excitationcontrac-
tion coupling were demonstrated to be pres-
ent within this tissue, including the typical
electrical activation, increase in [Ca
2+
]
i
and
the resulting contraction. Similar to pre-
vious studies in a mouse ES model, whole-
cell patch-clamp studies demonstrated that
these human ES cell-derived cardiomyocytes
also displayed all cardiac-specific action
potential morphologies and ion currents.
To demonstrate the ability of these hES
cell-derived cardiomyocytes to survive,
function and integrate in an in vivo heart,
Professor Gepstein and colleagues as-
sessed the ability to function as a biological
pacemaker in an animal model of a slow
heart rate and successfully paced its heart
to normal rates. The animal model of
complete atrioventricular (AV) block was
generated in pigs by ablating their AV
node, the major electrical conduction path-
way between the atria and the ventricles.
This resulted in complete dissociation be-
tween the atrial and ventricular electrical
activities, and generation of a slow ventric-
ular rate, mimicking the clinical scenario
of patients suffering from complete AV
block, usually requiring the implantation
of an electronic pacemaker. Following crea-
tion of an AV block in these animals, they
injected the spontaneously contracting EBs
into the posterolateral left ventricular wall
and monitored their electrocardiogram.
Following cell grafting, a new ectopic ven-
tricular rhythm was detected in 11 out of
13 animals studied, in six of which it was
characterized by sustained and long-term
activity. Three-dimensional electrophysio-
logical mapping revealed that this ectopic
ventricular rhythm originated from the
area of hES cell-derived cardiomyocyte
transplantation, and pathological studies
validated the presence and integration of
the grafted cardiomyocytes at the site of
cell transplantation!
This exciting breakthrough described by
Professor Gepstein, i.e., the development
of hES lines and their ability to differenti-
ate into cardiomyocyte tissue, holds great
promise for several cardiovascular research
and clinical areas. Research based on the
cells may help to elucidate the mecha-
nisms involved in early human cardiac
lineage commitment, differentiation and
maturation. Moreover, this research may
promote the discovery of novel growth and
transcriptional factors using gene trapping
techniques, functional genomics and pro-
teomics, as well as providing a novel in vi-
tro model for drug development and test-
ing.
In summary, the ability to generate for
the first time human cardiac tissue from
Executive Summary LX
hES cells in vitro provides an exciting and
promising cell source for the emerging
discipline of regenerative medicine, tissue
engineering and myocardial repair: various
differentiated derivatives of specific hES
cell lines could be safely transplanted with-
out the risk of immune rejection. One of
the most promising strategies in doing so
is really based on the generation of iso-
genic hES cell lines tailored specifically for
each patient, as described above by Profes-
sor Hwang from Seoul University. Accord-
ing to Professor McKay from the NIH, cell
cultures produced from stem cells could
also help make predictions about the true
effect of a certain drug in a specific patient
(one step further to individualized medi-
cine). As we will see in other chapters, this
information would provide a lot more an-
swers and would be more predictive than
currently used animal models.
Another exciting approach to treat car-
diovascular disease also uses stem cells,
but mesenchymal rather than embryonic
ones, to treat infarcted myocardium. Pro-
fessor Abeel Mangi recently published his
impressive results from Harvard Medical
School in Nature Medicine and now shares
his excitement with us. He describes how
to repair infarcted myocardium by geneti-
cally enhanced mesenchymal stem cells
transduced by ex vivo genetic manipulation
to express the prosurvival Akt protein (pro-
tein kinase B). Together with his collea-
gues from Harvard, they characterized a
highly purified population of mesenchy-
mal stem cells harvested from the bone
marrow of adult animals that is easily ex-
pandable and scalable, and induces recov-
ery of cardiac function after myocardial in-
farction by differentiating into cardiomyo-
cytes in vivo.
Transplantation of adult bone marrow-
derived mesenchymal stem cells has been
proposed as a strategy for cardiac repair
following earlier myocardial damage. How-
ever, poor cell viability with transplantation
has limited the reparative capacity of these
cells in vivo. In the study, Professor Mangi
and coworkers genetically engineered rat
mesenchymal stem cells using ex vivo ret-
roviral transduction to overexpress the
gene AKT1, which encodes the prosurvival
Akt protein. Transplantation of 510
6
cells
overexpressing Akt into the ischemic rat
myocardium inhibited the process of cardi-
ac remodeling by reducing intramyocardial
inflammation, collagen deposition and car-
diac myocyte hypertrophy, regenerated 80
90% of lost myocardial volume, and com-
pletely normalized systolic and diastolic
cardiac function. These observed effects
were dose (cell number) dependent. Mes-
enchymal stem cells transduced with Akt1
restored 4-fold greater myocardial volume
than equal numbers of cells transduced
with the reporter gene LacZ. In this study
it could be demonstrated that mesenchy-
mal stem cells genetically enhanced with
Akt1 can repair infarcted myocardium,
prevent remodeling and nearly normalize
cardiac performance.
Recent developments in cell biology per
se also enable regenerative medicine for
other diseases, including the use of retina
cells to treat neurodegenerative diseases,
e.g., Parkinsons disease. Neurodegenera-
tive disease can be regarded as the greatest
challenge in neurology: Alzheimers dis-
ease, multiple sclerosis and Parkinsons
disease are the most frequent diseases af-
fecting the human brain. Individuals suf-
fering from one of these neurodegenera-
tive conditions form the vast majority of
neurology patients both in clinics and doc-
tors offices, and as in-patients. Although
the decade of the brain (19902000) has
produced a tremendous quantity of re-
search results elucidating mechanisms of
neuronal decay and cell degeneration, in
Executive Summary LXI
general, no realistic chance of a cure is in
sight. For Parkinsons disease, a number
of symptomatic treatment approaches have
already existed since 1961, when high
doses of l-DOPA were given to the pa-
tients. Since then, l-DOPA has been the
gold standard in Parkinsons disease ther-
apy. As dopamine is metabolically sensitive
and is too polar to easily cross the blood
brain barrier, its precursor l-DOPA must
instead be used. Decarboxylase inhibitors
administered simultaneously, almost al-
ways in combinatory formulations, prevent
immediate transformation into dopamine
in the periphery and allow it to be built up
to therapeutic concentrations in the brain.
Despite this trick and the development of
slow-release formulations, the half-life of
l-DOPA is still very short and multiple
doses over the day are required in ad-
vanced Parkinsons disease patients.
A promising new therapy is beginning
to emerge, and is described by my collea-
gues Elke Reissig, Joachim-Friedrich Kapp
and Hermann Graf from Schering AG in
the next chapter: Spheramine
: A Cell
Therapeutic Approach to Parkinsons Dis-
ease. Elke Reissig is a medical doctor spe-
cializing in neurology and psychiatry, and
a Core Clinician in clinical trials for multi-
ple sclerosis and Parkinsons disease. She
has certified training in neurology, e.g.,
magnetic resonance techniques, and
stereotactic neurosurgery, and is currently
acting as a Senior Medical Advisor. Joa-
chim-Friedrich Kapp is a medical doctor,
too, and currently Head of Global Busi-
ness Unit Therapeutics. Before joining
Schering, he was working in several lead-
ing positions in the pharmaceutical indus-
try, including VP and Head, International
Clinical Development at Warner Lambert.
The third colleague is Hermann Graf, a
biochemist and biologist who was Head of
the Cell Biology Department at Metreon
before joining Schering AG. Currently, as
Head of Cell Therapy, he is responsible for
the Spheramine cooperation with Titan
Pharmaceuticals. Spheramine is a biologi-
cal product composed of human retinal
pigment epithelial (hRPE) cells placed on
microcarriers, for implantation into the
human brain. The retina cells are placed
on microcarriers of crosslinked porcine
gelatin to enhance their survival. The phar-
macologically active parts of Spheramine
are the human retinal pigment epithelial
cells which produce l-DOPA, which may
provide an approach to support local dopa-
mine generation in the brain. As is known
from experimental work, continuous dopa-
minergic stimulation likely represents a
more physiologic presentation of the neu-
rotransmitter lacking in Parkinsons dis-
ease than achievable by fluctuations
achieved with oral l-DOPA therapy.
Spheramine is administered in one ses-
sion of stereotactic neurosurgery and no
follow-up operations or tuning sessions
are required, as is the case for deep brain
stimulation. The immediate risk of surgery
is thought to be the same as for deep
brain stimulation, but the cumulative risk,
including that caused by hardware remain-
ing in the brain and repeated surgery in
the case of electric stimulator implanta-
tion, may be lower.
At this point in time, promising efficacy
results from a pilot study in six patients
followed over more than 36 months have
been observed with no safety complica-
tions. Preliminary safety results from an
ongoing double-blind, placebo-controlled
(STEPS) study in 68 patients are encourag-
ing, while efficacy data from this trial can
be expected in the first or second quarter
2006.
As a conclusion for Part I, we have seen
by some impressive examples that oligo-
nucleotide-based (antisense DNA, decoy
Executive Summary LXII
oligonucleotides, siRNA) and cell-based
approaches (hES cells, mesenchymal stem
cells, retina cells) are becoming more and
more prominent, and I am convinced that
we will see some more exciting break-
throughs in this field over the next couple
of years.
Part II: Biopharmaceuticals and their Mode
of Action
Quid Pro Quo Lysis versus Coagulation in
the Fine-tuned Balance of the Clotting Cascade
Before we can move on to the improve-
ment of biopharmaceuticals, we should
first gain more insight into the different
modes of action for the very diverse kinds
of biopharmaceuticals. Three different ex-
amples/classes are discussed each of
them ranking high on the list of fatal dis-
eases: bleeding disorders (hemophilia),
cancer and AIDS.
We start with Zymogen activation, which
is a central regulation mechanism in
many important biological processes in-
cluding blood coagulation, fibrinolysis and
the complement system. As recently pub-
lished in Nature by another friend from
the Max-Planck-Institute, Rainer Friedrich,
most trypsin-like serine proteinases are
synthesized as inactive precursors (proen-
zymes or zymogens) that must either bind
to a specific cofactor to develop substantial
catalytic activity and/or be activated by lim-
ited proteolytic processing which induces a
conformational change creating a func-
tional catalytic machinery, which removes
an N-terminal peptide or entire N-terminal
domains. This structural rearrangement
can be seen in an animation on the sup-
plementary CD-ROM. As a consequence of
the activation, catalytic activity of the en-
zyme is usually enhanced by several orders
of magnitude. The so-called zymogenic-
ity of the proenzyme is a measure for the
increase in catalytic efficiency after activa-
tion. While the precursors of the digestive
enzymes trypsin or chymotrypsin are al-
most completely inactive, they obtain a
10
4
- to 10
6
-fold activity increase due to
their activation!
In cascades like coagulation, fibrinolysis
and complement activation, a given pro-
teinase activates another pro-proteinase in
an amplification cascade, so that even
these drastic activation steps are (in addi-
tion and on top) again amplified many
fold. Therefore, a number of regulatory
steps are required to fine tune such com-
plex processes in which a precise balance
is mandatory to guarantee proper physio-
logical and life-saving functions. One such
example is the blood coagulation cascade:
in response to vascular injury, the body
must tightly seal the leakage while pre-
venting unrestrained intravascular clot de-
velopment and vessel occlusion. The coag-
ulation process is a complex interplay of
the blood vessel wall, platelets and other
blood cells, as well as many soluble plas-
ma proteins (coagulation factors). In the
ultimate step of the coagulation cascade,
the trypsin-like serine proteinase thrombin
[ factor (F) IIa] is released into the blood
stream, where it performs several essential
pro-coagulant functions. Free a-thrombin
(the active form) converts soluble fibrino-
gen to fibrin, which spontaneously poly-
merizes to form the fibrillar matrix of the
blood clot. Thrombin also activates FXIII,
a transglutaminase which thereafter cova-
lently crosslinks fibrin monomers, form-
ing an insoluble clot. Binding of thrombin
to its receptor thrombomodulin leads to a
dramatic change in the substrate specifici-
ty of thrombin, converting it from a pro-
coagulant to an anticoagulant and antifi-
brinolytic agent. An excellent video anima-
Executive Summary LXIII
tion showing this complex interplay in a
very educational manner is available on
the supplementary CD-ROM.
A method for the design of inhibitors
for such proteases was published in Pro-
ceedings of the National Academy of Sciences
of the USA by another colleague from the
Max-Planck-Institute. Professor Luis Moro-
der, who is also Professor of Biochemistry
and Biotechnology and the Technical Uni-
versity Munich, presents the principle of
polyvalency by structure-based design of
mono- and bivalent inhibitors for tryptase,
proteasome and serine proteases, e.g.,
FXa.
Hemophilia A is caused by the absence
or severe deficiency of FVIII, a protein in
human blood critical for proper blood
coagulation. More than 350000 people
worldwide have hemophilia approxi-
mately 80% of them have hemophilia A.
As this congenital bleeding disorder re-
sults from insufficient levels of FVIII coag-
ulation activity, it is characterized by a pro-
longed clotting time. Patients suffering
from this impaired blood coagulation can
experience spontaneous, uncontrolled in-
ternal bleeding that often is associated
with pain, debilitation, chronic joint de-
struction and, if left untreated, the risk of
death. Because the FVIII gene that codes
for the FVIII protein is located on the X
chromosome, virtually all clinically af-
fected individuals are male. Here we pres-
ent a case study for the development of a
modern biopharmaceutical against this
most abundant bleeding disorder: antihe-
mophilic FVIII (AHF). This blood clotting
factor is deficient or absent in individuals
with classic hemophilia A. The develop-
ment of AHF exemplifies the complex
mode of action for drugs targeting the very
sensitive coagulation cascade.
I am happy to say that I know from my
previous work experience probably the two
most knowledgeable experts in this area:
Friedrich Dorner and Norbert G. Riedel.
Only they are able to provide such a com-
prehensive and balanced overview on the
development of a modern biopharmaceuti-
cal against this prominent disease, because
both have extensive academic as well as in-
dustrial experience. Professor Friedrich
Dorner, who spent 4 years at Harvard Uni-
versity, is now Executive Board Member
and President of Global R & D of Baxter.
In addition, he was elected to the WHO
Advisory Board for Recombinant DNA
Technology, is member of the Austrian
Academy of Science and was awarded the
Grand Decoration of Honor in Gold for
Services to the Republic of Austria. He has
almost 200 publications in peer-reviewed
journals and holds more than 50 patents.
Professor Norbert G. Riedel spent 8 years
of his career at Harvard University, MIT
and Boston University School of Medicine,
before he became Head of Global Biotech-
nology of Hoechst Marion Roussel, and
currently serves as CSO and Senior Vice
President at Baxter. In addition to his cur-
rent position, Norbert serves on Scientific
Advisory Boards and Boards of Directors
of German and US-based biotechnology
companies, is a member of the Board of
Management of the German Association
of Biotechnology Companies, and a mem-
ber of the Board of Directors of the Bio-
technology Industry Organization (BIO).
Both colleagues are obviously experts in
this field and their chapter starts with a
historical overview of transfusion therapy
for hemophilia, which was first proposed
in the mid-19th Century and then began
with whole-blood transfusion early in the
20th Century. The large volumes of blood
or citrated plasma replacement required to
achieve hemostasis following major bleed-
ing episodes evolved over time to more
manageable amounts of cryoprecipitate, to
Executive Summary LXIV
highly purified plasma-derived FVIII
(pdFVIII) concentrates and finally to re-
combinant human FVIII concentrates
(rFVIII), which offer the advantages of
lower risk for blood-borne pathogen trans-
mission, reduced impact on the immune
system and supply that is independent of
plasma availability.
However, all previously developed rFVIII
concentrates incorporate human- or ani-
mal-derived proteins at some point in pro-
cessing; thus, concerns remain within the
hemophilia community regarding possible
pathogen transmission through these
additives. Here, both colleagues describe
the development, production and clinical
study programme of a novel full-length
protein-free rFVIII preparation for the
treatment of hemophilia A. ADVATE
was initially approved by the European

Commission in March 2004 and is the
first FVIII to be processed using a
plasma/albumin-free method. Hence, this
approach is a breakthrough and provides a
new standard of pathogen safety for hemo-
philia A patients!
Errare Humanum Est What Causes Cancer
and How to Selectively Fight Tumors
After gaining insight into the complex
mode of action of the coagulation cascade,
the next section deals with the pathophysi-
ology of cancer, and provides recent trends
and achievements on how to selectively
monitor and fight cancer cells. Cancer is a
major health issue worldwide. The most
common solid tumors are breast, colorec-
tal, ovarian, prostate and lung cancer,
which account for more than 3.2 million
new cases annually and 1.7 million deaths
each year. In addition, large numbers of
individuals are diagnosed with, and die
each year from, hematological malignan-
cies such as lymphomas (around 166000
new cases and 93000 deaths, respectively)
or leukemias (144000 new cases and
109000 deaths, respectively). Although, as
we will see later in the Diagnostic and
Imaging section, early detection com-
bined with advances in surgery and exter-
nal radiotherapy have improved the prog-
nosis for many patients with solid tumors
(in which the disease is confined to the
primary anatomical site), the outlook for
patients with advanced disseminated can-
cer remains poor.
Thus, the goal is to develop as many po-
tent biopharmaceuticals for as many types
of cancer as soon as possible. To achieve
this ambitious goal, utilizing natures re-
pertoire is one approach, which is followed
by David J. Newman and Gordon M.
Cragg from the NCI. David worked in a
number of US-based pharmaceutical com-
panies such as Smith Kline before joining
the NCI. In 2003, David, who is also Pro-
fessor at the Center of Marine Biotechnol-
ogy, University of Maryland, was awarded
the NIH Merit Award for development of
anti cancer agents. Gordon did his PhD at
Oxford University and joined the NCI in
1985. His interests lie in cancer and AIDS,
and he also received an NIH Merit Award
for his contributions to the development
of taxol. Both explain natures role in de-
veloping potent biopharmaceuticals and
start with an investigation of drugs ap-
proved worldwide since 1981. A substan-
tial proportion of them indeed fall into the
category of biological agents and, on
further subdivision, a significant number
of these are in fact compounds that either
are natural products (small to medium
sized polypeptides) or are derivatives of
such natural materials. Some of them have
been produced via recombinant or bio-
technologies, and have been expressed via
fermentative processes in both prokaryotic
and eukaryotic systems.
Executive Summary LXV
Almost 400 biotech medicines are cur-
rently undergoing trials in the US, with
the majority (close to 50%) being directed
against cancer in one or more of its many
manifestations, and a significant number
being directed towards infectious disease,
autoimmune disease and HIV. A signifi-
cant proportion of all materials in the can-
cer and infectious disease areas are vac-
cines of one type or another, and the sec-
ond largest class of biopharmaceuticals un-
der development are mAbs, with the
largest numbers being directed towards
cancer and autoimmune diseases. The
authors demonstrate the current and fu-
ture potential of the search in nature for
biologically active peptides and proteins,
and the ability to express these agents in
homologous and/or heterologous hosts.
The ability to manipulate the gene se-
quences in order to produce subtle modifi-
cations of existing active agents and the
potential for semisynthesis to modify the
basic properties of the initial agents is am-
ply demonstrated in the discussions of
these compounds. David concludes that
for a very large number of disease entities
not only in man, but in veterinary medi-
cine and perhaps as importantly in crop
protection, the potential for large-scale pro-
duction of these agents by biotechnological
means is limited only by imagination: the
basic compounds are there and working
with Mother Natures Pharmacopoeia has
several advantages!
As discussed earlier, the outlook for pa-
tients with advanced disseminated cancer
remains poor and this is why research
really needs to focus on this topic. Tar-
geted in situ radiotherapy, described by
Professor Raymond M. Reilly from the
General Hospital, University of Toronto, is
one strategy intended to selectively eradi-
cate disseminated cancer while sparing
normal tissue. As a member of the Society
of Nuclear Medicine and as Professor at
the Leslie Dan Faculty, Raymond has es-
tablished the Laboratory of Molecular
Imaging and Targeted Radiotherapeutics at
this University. The idea is to selectively
guide biomolecules to the tumor cells
hooked up to radionuclides that emit a-
particles, b-particles or Auger electrons.
The field is more than 20 years old in con-
cept and the success in treating non-Hodg-
kins B cell lymphoma (NHL) using anti-
CD20 mAbs conjugated to iodine-131
([
131
I]tositumomab; Bexxar
, BMS) or yt-
trium-90 ([
90
Y]ibritumomab tiuxetan; Zeva-
lin
, Schering AG) has reinvigorated the

search for other biologically targeted radio-
therapeutic agents. One is the application
of short-range Auger electron-emitters or
a-emitters for targeted radiotherapy of
cancer, which is described for human epi-
dermal growth factor (hEGF) conjugated
to diethylenetriamine-pentaacetic acid
(DTPA) and labeled with
111
In ([
111
In]-
DTPAhEGF) for treatment of EGF recep-
tor (EGFR)-overexpressing breast cancer.
EGFRs are overexpressed up to 100-fold
compared to most normal epithelial tis-
sues in almost all estrogen receptor (ER)-
negative, hormone-resistant and poor prog-
nosis breast cancers. Professor Reilly de-
scribes his work on exploring a type of
Trojan Horse targeted radiotherapeutic
strategy for breast cancer that exploits the
internalization and nuclear translocation
pathway of hEGF following binding to its
cell surface receptor. Indeed, only very re-
cently it was shown that the EGF/EGFR
complex may have a novel role as a nucle-
ar transcription factor for the cyclin D
1
gene, particularly in rapidly dividing cells
(e.g., cancer cells). A differential nuclear
versus cytoplasmic distribution of EGF/
EGFR in malignant versus normal cells
has profound implications for Auger elec-
tron-emitting radiotherapeutic agents tar-
Executive Summary LXVI
geted at EGFR expression, since the radia-
tion absorbed dose deposited in the nu-
cleus is about 15 times higher when
111
In
or
125
I decays in the nucleus compared to
decay in the cytoplasm, and 30 times high-
er than when the decay occurs on the cell
surface. The differential nuclear versus cy-
toplasmic uptake could provide a second
level of selectivity in protecting EGFR-posi-
tive normal cells, in addition to their lower
level of EGFR expression compared to ma-
lignant cells. [
111
In]DTPA-hEGF was rapid-
ly and efficiently internalized by EGFR-
overexpressing human breast cancer cells
and the emission of Auger electrons in
close proximity to DNA reduced the sur-
viving fraction of cells to less than 5%.
This means that [
111
In]DTPA-hEGF is up
to 300-fold more toxic on a molar level
than selected chemotherapeutic agents
commonly used for breast cancer such as
paclitaxel, methotrexate or doxorubicin (pi-
comolar concentrations provided cytotoxic
effects equivalent to those of 4 Gy of exter-
nal c-radiation!). This excellent chapter
concludes with the exciting future of the
field, exemplified by research demonstrat-
ing surgical cleavage of specific gene se-
quences in cancer cells using triplex-form-
ing oligonucleotides conjugated to Auger
electron-emitters, also referred to as anti-
gene radiotherapy.
In contrast to chemotherapy, which
leads to systemic toxicity and hence is
very harmful to the patient, selectivity is
the goal for cancer therapy. Therefore,
after the approach of selectively killing
cancer cells by target-specific in situ radio-
therapy, we will now focus on another very
promising strategy: selectively starving
tumor cells. The core technology and prod-
ucts of the company Medical Enzymes are
based on a proven approach using amino
acid-depleting enzymes as therapeutic
agents. This technique capitalizes on the
dependence of cancer cells on particular
amino acids relative to normal cells. The
use of amino acid-depleting enzymes pro-
vides a way to starve cancer cells by effi-
ciently reducing the concentration of the
selected amino acid. The proof-of-concept
for the amino acid-depleting strategy was
demonstrated with asparaginase (EL-
SPAR
, Merck), a chemotherapeutic agent

for leukemia. Joseph Roberts, Professor at
University of South Carolina, Columbia,
pioneered the development of asparagi-
nase and his accomplishments in advanc-
ing the field of therapeutic enzymes form
the basis of Medical Enzymes core tech-
nology: GlutaDON
. The principle here is

the depletion of glutamine rather than as-
paragine, because glutamine is the most
abundant circulating amino acid in the
body, and hence the major respiratory fuel
for tumor cells.
GlutaDON
is a combination therapy
for the treatment of cancer consisting of
PEGylated glutaminase, which breaks
down the glutamine and the glutamine
analog 6-diazo-5-oxo-l-norleucine (DON).
The rationale for using glutamine antago-
nists in combination with the enzyme glu-
taminase is based on the premise that the
effectiveness of the antagonist will be dras-
tically enhanced when the available pool of
glutamine is depleted by the enzyme. Tu-
mor cells are avid glutamine consumers
due to their decreased expression of gluta-
mine synthetase and the need for gluta-
mine as a substrate in nucleotide and pro-
tein biosynthesis, for energy production,
and for the generation of key metabolic in-
termediates. To cope with the demand for
glutamine, tumor cells express highly effi-
cient transporters to ensure that substrate
availability does not become rate limiting:
human hepatoma cells transport gluta-
mine at a rate up to 20 times faster than
normal hepatocytes do.
Executive Summary LXVII
DON is an antitumor antibiotic and as a
structural analog of l-glutamine it functions
as antagonist, interfering with several key
biochemical reactions, such as inhibition
of DNA replication and protein synthesis
resulting in inhibition of tumor growth.
DON has been shown to possess promising
antineoplastic activity against a variety of
animal tumors and human tumor xeno-
grafts (in nude mice), including colon,
breast and lung carcinomas, but has limited
potential when used as a single agent in the
treatment of cancer in humans, because of
severe toxicity that prevents dose escalation
into the required therapeutic range. How-
ever, by depleting glutamine in the blood-
stream through the combined glutaminase
activity, DON is much more rapidly taken
up by the tumor cells, because they possess
an enhanced transport mechanism for glu-
tamine. This increased efficiency of DON
uptake by tumor cells allows for much low-
er dosing levels. Another advantage in the
combination therapeutic regimen could be
achieved by the development of an econom-
ically feasible method for producing and
purifying a new and PEGylated form of
the glutaminase, which showed promising
anticancer activity with little host toxicity
in preclinical studies utilizing human lung,
breast, colorectal and ovarian tumor xeno-
grafts growing in athymic nude mice. Pre-
clinical toxicology studies in animals were
completed and numerous studies demon-
strated that tumors do not develop resis-
tance to glutaminase treatment as they do
to most anticancer therapies. These fantas-
tic results led to a multicenter phase IIIa
trial with principle investigator Professor
Clemens Unger from the Clinic for tumor
biology in Freiburg, Germany. In these pro-
mising clinical trials, GlutaDON targets
lung, breast, ovarian, colorectal and prostate
cancers the predominant cancers in the
Western World.
Mundus Vult Decipi High Mutation Rates
of HIV and New Paradigms for Treatment
Insight into the latest development of bio-
pharmaceuticals against cancer is followed
by the latest therapies against HIV-1,
which is the causative agent of AIDS. The
first cases were reported in 1981 already
and Robert Gallo (who discovered the virus
two years later) talked about a hybris, be-
cause medical scientists claimed that they
would have previously eradicated infec-
tious diseases at least in the wealthy re-
gions of the industrialized world. But this
virus is currently spread worldwide and af-
fects millions of individuals; the present
situation is especially critical in sub-Sahar-
an countries (roughly 70% of the world
cases), where the virus affects more than
30 million people. Altogether the global
AIDS pandemic has killed more than 28
million people so far and infected more
than 42 million it is estimated that 45
million new cases of HIV infections will
occur by 2010. The impact of this virus in
health and social relationships has been
tremendous in many countries around the
world since it was discovered in the early
1980s and is despite massive informa-
tion and protection campaigns still a rap-
idly spreading disease (14000 new cases
per day!). The lifetime treatment cost for a
person with HIV is estimated by the US
Centers for Disease Control and Preven-
tion to be US$ 155000 this translates
into an annual amount of more than
US$ 6 billion.
HIV-1 is a member of the retrovirus
family and belongs to the lentivirus genus.
Due to the complexity of the HIV-1 infec-
tion and natures strategies for efficient
evolutionary adaptation, it was difficult to
find a safe and efficient therapy in the past
two decades: in natural evolution, the enti-
ties with the fastest adaptation to changing
Executive Summary LXVIII
environmental conditions are viruses. Pre-
sumably, viruses developed this extreme
adaptability to escape eradication by the of-
ten quickly changing defense mechanisms
of their hosts. Looking at the molecular
mechanisms that confer this ability, it be-
comes apparent that replication of the viral
genome operates at very high mutation
rates. Hence it is difficult to find and fight
the virus, because it is continuously chang-
ing its envelope and therefore escapes rec-
ognition and treatment. This mechanism
of recognition and binding is shown on
the supplementary CD-ROM. Therefore,
another approach (rather than at the pro-
tein level) has to be followed, and impor-
tant efforts have been made to develop
gene therapy approaches aimed at inhibit-
ing viral replication and making cells resis-
tant to the virus or eliminating the in-
fected cells. At present, virus replication is
quite efficiently blocked by conventional
highly active antiretroviral therapy
(HAART). However, HAART does have its
limitations, including the emergence of
drug-resistance mutants, patient noncom-
pliance and the overall cost of multidrug
combination therapy. In addition, the exis-
tence of long-lasting latently HIV-1-in-
fected cells in the patient does not allow
the eradication of the virus. In fact, the vi-
ral reservoirs of latently infected cells are
not affected by HAART and have become
the most problematic area in HIV-1 thera-
py.
For this reason, Francisco Luque who,
since his EMBO Fellowship, heads the An-
dalusian Research System Molecular
Studies of Human Pathologies, is develop-
ing gene therapeutic strategies that are not
inhibiting viral replication, but are destroy-
ing the viral reservoirs. The genetic con-
structs contain an externally inducible sys-
tem that promotes the expression of any
latent HIV-1 provirus without affecting the
cell cycle state by the expression of the po-
tent viral transactivation TAT protein. A
second genetic system included in the vec-
tor allows the expression of a suicide gene
in response to the presence of the essen-
tial viral REV protein, which in turn in-
duces a quick death of the cell by apopto-
sis due to the overexpression of p53 in re-
sponse to the presence of HIV-1 provirus.
The vector can be packaged into HIV-1
viral particles to deliver the genetic con-
structs in any cell susceptible to HIV-1.
This biopharmaceutical is designed such
that any alteration of the normal cell func-
tion is prevented, so that uninfected trans-
duced cells remain unaffected and fully
functional. Data submitted to the Journal
of Molecular Medicine show that this sys-
tem permits a very efficient and specific
destruction of any HIV-1-infected cell, even
those that contain a silent provirus.
Lentiviral vector delivery of RNA-derived
modalities offers another valid and poten-
tially efficacious gene therapy-based
approach to augment current anti-HIV-1
therapeutics. I am very thankful that the
next contribution again comes from one of
the most knowledgeable experts in this
field: John J. Rossi, Chair of Beckman Re-
search Institute of the City of Hope and
Professor at the City of Hope National
Medical Center. Due to his credibility in
this field, John serves as Principal Investi-
gator and Program Director for several
NIH studies for the treatment of HIV with
siRNA, HIV ribozymes and other RNA-
based therapeutics. For his achievements
he was the recipient of the very prestigious
Merit Award, Division of AIDS, from the
National Institute of Allergy and Infectious
Diseases. His colleague Kevin V. Morris
worked with Monsanto before he did his
PhD at the University of California, Davis
and then moved on to the Center for AIDS
at the University of California, San Diego.
Executive Summary LXIX
Both have published extensively in Nature
Biotechnology and Science, and share with
us their experience with the many lentivi-
ral vector systems that have been studied,
including those based on feline immuno-
deficiency virus (FIV), HIV-1 and HIV-2/
SIV (simian immunodeficiency virus) as
well as replication incompetent, self-inacti-
vating versus conditionally replicating (mo-
bilizable) vectors. A major limitation in
utilizing these lentiviral vectors and a gene
therapy-based approach in treating HIV-1
infection has been the relative lack of an
efficacious therapeutic modality. Recently,
siRNAs have been described, and shown
to potently and specifically suppress HIV-1
by applying virus-based vectors to deliver
certain anti-HIV-1 genes to the respective
target cells. The problem, however, is that
resistance to siRNA occurs rather rapidly
and is only contingent on a single nucleo-
tide substitution recently HIV-1 demon-
strated an ability to elude siRNA targeting
by the evolution of alternative splice vari-
ants for the siRNA targeted transcripts.
Therefore, to avoid the emergence of HIV-
1 resistance, a multitargeting approach
should be taken. Importantly, anti-HIV-1
ribozymes could be incorporated along
with multiple siRNAs targeting the most
conserved regions of HIV-1 as well as
splice junctions. Indeed lentiviral vectors
can accommodate a roughly 6.5-kb payload
offering a promising delivery vehicle for
RNA-based modalities such as siRNAs and
ribozymes. Different combinations of each,
including a collection of their respective
variants, can be developed as future bio-
pharmaceuticals for the treatment of HIV-
1 infections.
In summary, Part II has shown exam-
ples of three different types of diseases
(hemophilia, cancer and AIDS), their un-
derlying pathogenic mechanisms and how
these are reflected in the development of
different classes of biopharmaceuticals.
Although a number of breakthroughs have
already been achieved, as highlighted and
explained in depth by the authors, we still
have some development work ahead of us.
Biopharmaceuticals suffer from a wide-
spread imbalance between perception and
facts, and some perceive biopharmaceuti-
cals as new, exotic and not well under-
stood, despite the fact that the first recom-
binant therapeutic protein was approved
by the FDA more than two decades ago
(Eli Lillys insulin Humulin, approved in
1982). Indeed, several biologics have now
been released as second- and even third-
generation products that exhibit improved
efficacy, fewer side-effects and better pro-
duction efficiency; however, much effort
was spent in terms of time and investment
to get there. However, in the light of in-
creasing competition, biopharmaceuticals
coming off patent, generic markets, cuts
in the healthcare systems and smaller R &
D budgets, one needs to speed up the
development process, and reduce time to
market and overall cost. Some feasible
approaches, options and technologies to
do so will be presented in the next section.
Part III: Improving the Development
of Biopharmaceuticals
Citius, Altius, Fortius Acceleration by
High-throughput and Ultra-high-throughput
Techniques
After learning about three different classes
of disease and how their very diverse
mode of action impacts the development
of specific biopharmaceuticals, we will
now discuss how we can improve the de-
velopment process of biopharmaceuticals
per se independent of their indication
and mode of action. Speeding up the de-
Executive Summary LXX
velopment process and screening more
candidates in parallel at the same time is
key. This means that automation, paralleli-
zation, miniaturization, etc. are the param-
eters to tune the process or in one
phrase: high-throughput (or even ultra-
high-throughput). We will see some im-
pressive results and some new technology
trends which really do speed up biophar-
maceutical development. It is worth noting
at this point that all of these technologies
(and almost all the work described in this
book) are based on, or became only possi-
ble through, a fascinating invention which
revolutionized molecular biology, human
genetics, biotechnology and of course the
development of biopharmaceuticals: the
polymerase chain reaction (PCR). The
PCR was invented by Kary Mullis, who in
turn received the Nobel Prize in 1993 (see
his quote for Modern Biopharmaceuticals). I
see PCR as one of the greatest scientific
accomplishments and enabling technolo-
gies of the 20th Century, and we will now
see some striking examples of its utiliza-
tion in biopharmaceutical development.
The next section again starts with a con-
tribution from a Nobel Prize laureate:
Manfred Eigen from the Max-Planck-Insti-
tute in Gttingen, Germany. I am very
pleased to have known Professor Eigen for
quite a while. I remember well when I
met him for the first time: the PhD stu-
dents from Manfred Eigens lab and Ro-
bert Hubers lab were having meetings
once a year in Klosters, Switzerland. This
was an excellent opportunity to exchange
ideas between these two famous groups
and very stimulating discussions emerged
on how to improve biopharmaceutical re-
search. Because these meetings were
planned in a similar manner to the Gor-
don conferences, there was also some time
to enjoy the fantastic scenery of the gla-
ciers and mountains and to do some
snow boarding as well. Anyway, Professor
Manfred Eigen is the inventor of the in-
triguing principle of directed evolution
and I had the honor as well as pleasure to
learn first-hand experience from this gen-
tleman. He also explained to me why the
choice of the library strategy is one of the
most crucial decision points in any direct-
ed evolution project. The library strategy
determines the composition of protein
variants present in the libraries, as well as
the number of possible mutants and muta-
tions. The number of possible mutants is
termed the complexity of a library and it
becomes obvious that even for simple pro-
tein libraries its complexity easily surpass
by orders of magnitudes the number of
variants that can be generated and techni-
cally evaluated. This imbalance is known
as the complexity problem in protein evolu-
tion. A simple calculation demonstrates
this imminent and intrinsic problem. The
number of all possible variants of a 200
amino acid protein is approximately 10
260
(theoretical required sequence space). The
enormous dimension of this theoretical
number can be better grasped when com-
paring it with real figures. For example, if
the mass of the entire universe were to
consist of solely such proteins, this would
amount to only approximately 10
75
mole-
cules (maximal possible sequence space).
These figures show that the difference be-
tween the theoretical required and maximal
possible (assuming the universe would con-
sist of exclusively the protein variants we
are interested in) sequence space is ob-
viously huge: a 10 with 185 zeros! Even
applying a brute force method (exploring
the entire sequence space with a trial-and-
error approach) one would never be suc-
cessful, because the required material is
simply not available.
Therefore, nature works differently nat-
ural evolution works by selecting variants
Executive Summary LXXI
(mutants) with improved fitness from smal-
ler populations of closely related variants
within a species. Selection pressure favors
a subset of variants of such a quasispecies
over competing variants and confers a high-
er replication rate to these variants, which
eventually leads to a shift of the quasispe-
cies towards a distinct phenotype. Our un-
derstanding that mutations can improve
an organisms phenotype gained increasing
acceptance in the decades following the
work of Johann Gregor Mendel (1822
1884). However, in vitro mutagenesis of pro-
teins (through their respective genes) was
only made possible after the structure of
DNA was discovered, the genetic code deter-
mined and basic DNA manipulation tech-
nologies became available.
From computer simulations as well as
experimental data it has been derived that
the fastest evolving species in nature are
those that replicate their genome with a
mutation rate near the error threshold.
This has been shown for RNA viruses,
which are known for their fast adaptation
to changing environmental conditions, and
this finding can be transferred to directed
evolution given the technical capability of
screening sufficiently large populations.
However, employing common mutagenic
methods like error-prone PCR can be in-
sufficient to efficiently search the fitness
landscape of a given protein for a variant
that serves a given task. It is the combina-
tion with other DNA variation techniques
like cassette mutagenesis and DNA shuf-
fling that offers the possibility to explore
sequence space with reasonable resources.
Such approaches are, however, limited by
the inherent enormous complexity when
generating protein or enzyme variants.
The complexity that has to be handled
when including, for example, only single,
double and triple mutants is already be-
yond classical screening capabilities.
Current approaches in directed evolution
of biopharmaceuticals therefore tackle both
sides of the problem: (a) increasing
throughput in the therapeutic evaluation
of protein variants while maintaining phar-
macologically relevant conditions and (b)
limiting the complexity of protein libraries
while preserving or even increasing the
chances to contain improved variants of
the specific biopharmaceutical. Different
approaches can be employed to select im-
proved variants of, for example, an enzyme
from any pool; among these are selection
by growth (or survival) of a producing or-
ganism, selection by binding to a target
substance (e.g., antibody) and selection by
screening. The methods differ in their
ability to address specific targets and their
combination forms a cyclic process able to
produce optimized protein variants even
for ambitious tasks. Together with his col-
leagues from DIREVO Biotech AG in Co-
logne, Manfred Eigen as co-founder exactly
describes this excellent approach and
shows some striking examples. Directed
evolution makes the profound difference
between early and modern biopharmaceu-
ticals we are no longer forced to discover
a compound with a particular activity, but
are instead able to intentionally develop it.
As published in Nature and Proceedings of
the National Academy of Sciences of the
USA, this approach is so compelling be-
cause it harnesses natures fundamental
tools to optimize molecules and even such
complex systems as entire organisms for
their respective (and ever-changing) envi-
ronment!
The relevance of certain fundamental
parameters is impressively demonstrated
with examples, in which application-rele-
vant characteristics (such as evolution of a
proteases substrate affinity or specificity to
hydrolyze a pharmacologically relevant tar-
get sequence) have been successfully opti-
Executive Summary LXXII
mized. As also described in their excellent
chapter, directed evolution optimization of
proteins has proven its potential in an en-
ormous number of studies targeting var-
ious proteins, including enzymes, antibod-
ies, peptide hormones and cytokines, to
name a few, and aiming for a broad variety
of optimization goals such as binding af-
finity, catalytic activity, thermostability, pH
stability, expression yield and many others.
In terms of biopharmaceuticals, directed
evolution is mainly employed to improve
the characteristics of biopharmaceutical
drugs that are under development, for the
engineering of marketed drugs, i.e., for
the generation of second- and third-gen-
eration products, or for the engineering of
follow-on biologics, i.e., unrelated proteins
that have the same functionality as mar-
keted biopharmaceuticals.
After learning about the exciting pos-
sibilities of directed evolution to develop or
even design the desired biopharmaceutical,
we will now focus on cloning and expres-
sion. The huge number of enzyme
variants which can be obtained from direc-
ted evolution and ultra-high-throughput
screening obviously also need to be cloned
and expressed. To avoid any potential bot-
tleneck in this step, and to identify the
most suitable protein species, it is manda-
tory to apply also high-throughput tech-
niques for cloning and expression. How-
ever, cloning genes by standard restriction
enzyme and ligase methods is not suitable
for this demanding task: a researcher must
choose the correct restriction enzymes to
allow isolation of the fragments, cut and
purify the vector, and insert and assemble
them in a ligation reaction in the proper
ratios. Each gene fragment must be con-
sidered on a case-by-case basis given that
the particular restriction enzymes needed
to isolate the fragment may not be com-
patible with the vector for which it is in-
tended. Also, the gene itself might contain
the same restriction sites internally that
match those needed for subcloning; hence,
the gene of interest would be destroyed.
To mitigate these types of problems, the
gene is commonly amplified by PCR with
additional restriction sites encoded in the
primers. For a small number of fragments,
this approach (including appropriate plan-
ning of the cloning strategy, restriction en-
zymes, cleavage sites, etc.) can be used to
efficiently subclone them to the appropri-
ate expression vector. Working at modern
scales, however, high-throughput cloning
calls for transfer of hundreds, thousands
or even millions of genes from platform to
platform. The ideal system suitable for
such a task would have maximum compat-
ibility and flexibility, would be useful with
a minimum amount of planning, and
would maintain the orientation and read-
ing frame of the ORFs transferred within
the system. It would also be rapid, need-
ing no restriction enzymes, no gel purifi-
cation of DNA fragments and no lengthy
ligation steps. Although that is obviously
quite a challenge, Jonathan D. Chesnut
from Invitrogen is one of the inventors of
such a high-throughput cloning platform.
Before joining Invitrogen 10 years ago,
Jon studied at the University of California,
San Diego, the University of California,
Davis and was working with Hybritech on
antibody engineering. At Invitrogen, he
was developing several topoisomerase-
mediated cloning technologies, as well as
strategies for stem cell engineering. More
recently he was leading the Gateway
R&D. The Gateway recombinational clon-

ing system represents a new paradigm in
molecular biology by improving speed and
efficiency beyond traditional restriction
cloning methods. Harnessing the recombi-
nation mechanism of bacteriophage k, it
provides an efficient method for manipu-
Executive Summary LXXIII
lating DNA elements and accelerates the
discovery process. It does this by allowing
efficient, high-throughput and automatable
cloning and transfer of DNA elements for
applications such as high-level protein pro-
duction for structural analysis, generation
of antibodies or other detection reagents,
functional analysis studies such as protein
interactions, subcellular localization, post-
translational modification and RNA inter-
ference of translation.
Gateway facilitates efficient gene trans-
fer via a re-engineered phage recombinase
system from the double-stranded DNA
bacteriophage k, which usually uses the
bacterial host cellular machinery to propa-
gate itself. To accomplish this, it first in-
fects the cell and inserts its entire genome
into the host chromosome. In this conser-
vative and site-specific reaction, a 25-bp re-
gion in the bacterial chromosome (termed
the attB site, bacterial attachment) is rec-
ognized by the phage recombination com-
plex and aligns with a 24-bp sequence in
the phage genome (attP, phage attach-
ment). In the presence of a dimeric bacte-
rial protein, integration host factor (IHF)
and phage-encoded Integrase (Int), the
DNA align and exchange strands effec-
tively inserting the phage genome into the
bacterial chromosome at the attB site. The
reaction that integrates the phage DNA
into the bacterial genome creates two new
sites that now flank the phage DNA
termed attL (for left) and attR (for right).
The phage DNA is replicated along with
the host DNA and can stay integrated in
the host genome for an extended period of
time. At some point, possibly after some
sort of stress or other insult to the bacteri-
um, the phage DNA is excised from the
host chromosome and is re-encapsulated
allowing it to move on to infect another,
healthy bacterium. The excision reaction is
essentially the reverse of the integration
reaction described above. Int and IHF are
again involved, but are joined by an addi-
tional phage-encoded protein, Excisionase
(Xis). These enzymes functionally join the
attL and attR sites leading to strand cleav-
age and exchange, and creation of attB
and attP sites on the two respective circu-
lar chromosomes. In the Gateway system,
the attP and attB sites remain on their
own respective (donor and expression)
plasmids, while the attL and attR sites ex-
ist on different (entry and destination)
plasmids, respectively; hence the att sites
can be used in pairs and flipped in orien-
tation to allow transfer DNA cassettes be-
tween any Gateway compatible vectors.
However, the heart of the Gateway sys-
tem is the entry clone. In this plasmid, attL
sites flank the DNA fragment of interest.
Once a gene, ORF or other DNA element
has been inserted between these sites to
create the entry clone, it can be transferred
to any compatible destination vector (e.g., a
battery of expression vectors for various
different hosts for parallel expression
screening). This transfer maintains the ori-
entation and reading frame (if the DNA is
an ORF) and is accomplished regardless of
the sequence of the DNA. Since there is
no specific sequence requirement that the
element must have (apart from containing
flanking att sites), entire pools or highly
diverse libraries of DNA fragments can be
captured and transferred with the same
high efficiency, and in a single reaction. In
order to do so, the entry vector is mixed
with a destination vector that carries attR
sites (or a collection of destination vectors
for multiple screening purposes), then
clonase (Int, IHF, and Xis) is added to cat-
alyze recombination between specific attL
and attR sites. The products of this reac-
tion consist of a donor vector that contains
the (toxic) ccdB gene flanked by attP sites
and the expression clone that carries the
Executive Summary LXXIV
DNA of interest flanked by attB sites (i.e.,
an ORF now downstream of a promoter
sequence). Given the appropriate amounts
of reactants, the efficiency of the recombi-
nation reaction (the amount of entry clone
converted to expression clone) is approxi-
mately 70%. An aliquot of this reaction
mix is used to transform competent Escher-
ichia coli and the desired expression clone is
selected by two means. First, any cell car-
rying either a donor vector or an unreacted
destination vector is selected against by the
presence of the ccdB gene. The product of
this gene is extremely toxic to most strains
of E. coli and therefore any cell containing
this gene is efficiently killed. Secondly, the
transformation mixture is plated on the
appropriate selective antibiotic that effec-
tively selects against cells transformed
with an unreacted entry vector. The recom-
binational specificity and the stringency of
the positive and negative selection leads to
an overall efficiency (the proper clones ex-
isting as colonies on the agar plate) of
greater than 95%! This high cloning effi-
ciency is a key attribute to the Gateway
system and the strong selection system of-
ten makes it unnecessary to screen multi-
ple colonies in order to select the desired
clone, thereby facilitating many high-
throughput gene cloning and expression
applications. For example, it can be used
to express the same protein in hundreds
of different hosts (applying only one entry
clone) to obtain the highest yield or to ex-
press different constructs of one protein to
identify the optimal domain boundaries or
best linker between them.
In summary, the Gateway system offers
an efficient mechanism to streamline clon-
ing without concern for the specific se-
quence of an ORF. Since it does not use re-
striction enzymes, ORFs are transferred to
an expression vector without the require-
ment for amplification by PCR or subse-
quent concern for PCR-induced mutations.
Gateway is an example of a new genre of
recombination-based cloning systems that
facilitate subcloning of genetic material
by site-specific recombination, and enable
transfer and assembly of ultra-large num-
bers of DNA elements using a single, stan-
dard mechanism. This system can speed up
tremendously (at least the cloning part of)
biopharmaceutical development.
In Vivo Veritas Early Target Validation in
Knockout Mice and More
Now we know how to utilize natures im-
pressive toolbox to optimize and accelerate
the development of biopharmaceuticals for
their intended application. However, if we
are also able to functionally optimize, for ex-
ample, an enzyme for its use as active ingre-
dient, we might still be far away from using
it as a biopharmaceutical drug in the phys-
iological setting. Target validation is the key
word here an important early step in the
development of novel biopharmaceuticals.
About a decade ago, the biopharmaceutical
industry was very eager to identify novel tar-
gets and use them in their drug discovery
programs. Today, the incentive is more on
conserving money by focusing on more va-
lidated targets. With the Human Genome
Project finished, the identification of targets
is easier and, in principle, all of the roughly
30000 genes identified could be used as po-
tential targets for drug discovery programs.
More realistic estimates range from a few
thousand to 10000 putative drug discovery
targets in humans. Since no biopharmaceu-
tical company has the resources to investi-
gate more than a handful of targets at a
time, assessing the quality of these targets
as early as possible, or validating these tar-
gets, is a crucial step in the drug discovery
process.
Executive Summary LXXV
Several target databases have been estab-
lished and are helpful tools for future ad-
vances in drug discovery. Examples are the
NIH target database, which includes data
from the worldwide structural genomic
and proteomic project, and the therapeutic
target database of the National University
of Singapore. However, the key questions
remain: What is a target and what makes
it a valid one? Here, companies generally
do differ in their requirements for target
validation. In general terms, a target is a
molecule (often a protein) that is instru-
mental to a disease process, although it
might not be directly involved. Target valida-
tion verifies that the target molecule is an
essential element in the disease process
and that it constitutes a potential point of
therapeutic intervention under physiologi-
cal conditions. During the target identifica-
tion process, targets are linked to the gen-
eration, the progression or the symptoms
of a disease. A potential target becomes a va-
lidated target when it is convincingly dem-
onstrated that altering (increasing or inhib-
iting) its biological activity leads to improve-
ment in a respective disease model.
The functional validation of a target prior
to the start of a drug discovery program is a
very critical early step fast and accurate
target validation technologies are the foun-
dation for future successful drug discovery.
However, equally important at this early
stage can be the less pleasant target invali-
dation. The early investment in target vali-
dation technologies can definitely pay off
at later stages or at least save a lot of money
later spent. Many different technologies in-
cluding targeted gene silencing, protein in-
hibition, cellular assays, chemical genetics
and combinatorial biology are used for the
target validation process. In most cases, a
combination of tools, including gene/pro-
tein modifications and model system ap-
proaches, along with expression and proteo-
mic data, are utilized to validate and priori-
tize targets of interest.
One very experienced expert in the field
of target validation is my colleague Chris-
toph Bagowski, who I have already intro-
duced along with the topic of RNAi. Chris-
toph first worked at Stanford University
on this emerging field and later also in a
biotech company in the Bay area focusing
on the development of target validation
technologies. Christoph shares with us his
broad expertise and discusses several, but
not all, state-of-the-art molecular and cellu-
lar biology methods some of which bear
the potential to turn into biopharmaceuti-
cal drugs themselves. Several techniques
for the validation of members of the pro-
tein kinase family are described as well as
two novel biopharmaceutical drugs on the
market which target cellular kinases.
We will also learn that a very crucial as-
pect of the target validation process is to
understand the systems biology and, if
possible, to use a good model system for
the respective disease. To help understand
systems biology it is necessary to be famil-
iar with the cellular signal transduction
networks. This encompasses aspects of
qualitative and quantitative biology, and in-
volves timing and localization of signaling
complexes as well as their regulation.
Many intracellular signaling pathways
have been described and, interestingly,
many of these are very highly conserved
from yeast, to worms to flies and humans.
From the more than 500 protein kinases
that have recently been described in the
human genome, one good example pre-
sented here is the EGFR pathway, which
leads to activation of mitogen-activated
protein kinases (MAPK). The EGFR path-
way is conserved from worms (C. elegans),
where it plays a role in vulva development,
to flies (Drosophila melanogaster), where it
is involved in the development of the eye,
Executive Summary LXXVI
to humans, where it is important for cell
growth and differentiation. In order for
the pharmaceutical industry to benefit
from the huge progress made in geno-
mics, proteomics and the associated bioin-
formatics, it remains crucial to understand
and test the systems biology.
Undoubtedly, understanding systems
biology can provide a means for identify-
ing pathways that are critical to disease
and it is important to not only understand
qualitative, but also quantitative, aspects of
systems biology. It should also be kept in
mind that, for example, two-dimensional
models of signaling networks will only
partially help understanding systems biol-
ogy, and bear the potential danger of mis-
leading and oversimplifying systems biol-
ogy properties. Therefore, it is essential to
verify these (preliminary) findings in a
complex disease model such as knockout
mice genetically modified mice that al-
low us to model disease-related genetic al-
terations to study the exact mechanistic
consequences of the mutations in vivo,
and to design and test new therapeutic
strategies to fight a certain disease.
Again, I am very happy that another col-
league from the Max-Planck-Institute with
several years of hands-on experience is
covering this critical topic. Cord Brake-
busch describes how genetically modified
mice are generated and used in medical
and pharmaceutical research. As we all
know, many diseases are caused or facili-
tated by genomic alterations; however,
studies with human patients are often dif-
ficult to perform and analyze due to differ-
ences in genetic background, age, disease
history and living conditions, the limited
amount of material available for histologi-
cal and biochemical analysis, and often
also low numbers of patients. Mouse mod-
els for human diseases have the advantage
that large numbers of genetically identical
animals of the same age and gender can
be handled, that the biology of mice is rel-
atively close to humans in comparison to
the aforementioned fishes, flies or worms,
and that they can be genetically modified.
This is why mice serve as the most impor-
tant models to study the mechanism of in-
herited and acquired diseases, and are ex-
tremely useful to test the efficacy of new
biopharmaceuticals and to validate poten-
tial drug targets.
Once we have generated mice with a
phenotype similar to the human disease,
these are analyzed in depth to identify
pathways that are critical to the disease,
and to understand qualitative and quantita-
tive aspects of the systems biology. With
these transgenic mice, new disease thera-
pies can be tested and evaluated in detail,
and primary cells might be used for high-
throughput screening for new drugs as
well.
In addition to that, genetically modified
mice can also be used as therapy models.
In that case a potential drug target is al-
tered in its activity by mutation in a simi-
lar way as the potential drug would do.
Gene targeting allows us to generate mice
that either lack this molecule (correspond-
ing to 100% inhibition) or express instead
a constitutively active form of it (corre-
sponding to 100% activation). Since inhibi-
tion and activation are (directly and exclu-
sively) induced by genetic alteration, there
is no partial effect and, furthermore, no
side-effects due to inhibition or alteration
of other molecules. These transgenic mice
can be used in disease models in order to
evaluate the in vivo potential of a certain
therapy before investing a lot of time and
money in the development of small-molec-
ular-weight inhibitors or activators. In
summary, transgenic mice serve as disease
models for sophisticated in vivo target vali-
dation in the desired physiological context.
Executive Summary LXXVII
For example, they allow us to induce a ge-
netic alteration after disease development
in order to test whether inactivation of a
certain molecule (in the specific physiolog-
ical setting) could result in tumor remis-
sion or not.
After highlighting the advantages of
mice, however, it should be noted that,
although most genes are highly conserved
between man and mouse (as the respective
genome projects revealed), there are still
many differences in the biology of man
and mice. These differences can result in
a different susceptibility to certain diseases.
For example, while in human skin about
five mutations are required for transforma-
tion into tumor cells, only two to three are
necessary in mice. Therefore each model
should be carefully tested to ascertain
whether the situation in the mouse suffi-
ciently mimics the situation in humans
and to what degree it might be necessary
to humanize the mouse by exchanging
specific mouse genes against human ones.
It should also be mentioned that, although
the generation of mice carrying point muta-
tions or deletions of specific genes is quite
well established, it is still a long and labor-
ious procedure, which cannot necessarily
be applied on the required large scale. This
is the reason why other model systems are
also being further developed and applied
in parallel. One of the model systems in
use is Dictyostelium discoideum, this was first
described in the late 19th century and since
then there has been a lively debate about the
evolutionary descent of the Dictyostelids.
Molecular phylogenicity studies placed the
Dictyostelid cellular slime molds at the root
of the Crown group of eukaryotes, within
the clade of lobose amoebae. Thus, the
Dictyostelids are very distantly related to
the eukaryotic protozoans and also have lit-
tle in common with plants. Instead, D. dis-
coideum is among the closest living relatives
to animals and fungi, and therefore is well
suited for studies of fundamental biological
phenomena that play important roles in hu-
man health and disease, e.g., cytokinesis is
critical in cell proliferation, and is therefore
an integral part of immune response, tissue
maintenance, and cancer. Cell motility is an
essential early event in metastasis of tumor
cells and in angiogenesis by endothelial
cells. Chemotaxis and signal transduction
by chemoattractant receptors play a key role
in inflammation, arthritis, asthma, lympho-
cyte trafficking and also in axon guidance.
Phagocytosis is a critical process involved
in immune surveillance and antigen pre-
sentation. Cell-type determination, cell sort-
ing and pattern formation are basic features
of embryogenesis, and alteration of these
events can lead to neoplasms. Many such
phenomena are obviously easier to analyze
in unicellular D. discoideum than in com-
plex models such as mice, plus it was
shown that embryotoxicity assay could re-
produce teratogenic activities of valproic
acid (VPA) analogs previously characterized
in animal models. This assay could in prin-
ciple be used to make predictions about the
potential embryotoxicity of drugs not yet
tested in animals and to suggest a common
molecular mechanism of action for some
biopharmaceuticals in D. discoideum and
in humans! Consequently, D. discoideum
has been chosen by the NIH as a nonmam-
malian model organism for biopharmaceu-
tical research.
I was always amazed by the lectures from
Theodor Dingermann who worked at Yale
University before he became full Professor
for Pharmaceutical Biology at the Goethe-
University in Frankfurt/Main. He chairs
the section Pharmaceutical Biology of
the German Pharmacopoeia Committee,
has an impressive track record, and has
published extensively on the exciting bug
D. discoideum in Nature Biotechnology and
Executive Summary LXXVIII
other peer-reviewed journals. Together with
Thomas Winckler, who previously worked
at the Institut Pasteur in Paris, and Ilse
Zndorf, who was visiting scientist at Uni-
versity of Kentucky, Lexington, in their ex-
cellent review they describe the state-of-
the-art of biopharmaceutical and biomedical
research, starting with tools available for
gene discovery and analysis, and then sum-
marizing experiments aimed at optimizing
D. discoideum even as an expression host
for the production of biopharmaceuticals.
In addition, they highlight some selected as-
pects of biomedical research: (a) D. discoi-
deum as a model for the study of host
pathogen interactions in infectious diseases
such as Legionnaires disease and Pseudo-
moniasis, (b) screening future biopharma-
ceuticals for potential embryotoxicity in hu-
mans using recombinant D. discoideum
strains carrying reporter genes, and (c) de-
velopment of new concepts for the improve-
ment of gene transfer vectors in human
gene therapy by studying mobile genetic
elements found in the D. discoideum ge-
nome.
Altogether, the goal is to combine re-
sults from systems biology experiments
(e.g., signal transduction) with findings
from disease models (e.g., mice) and em-
bryotoxicity assays using model systems
(e.g., D. discoideum strains carrying report-
er genes) to draw the right conclusion and
to achieve functional validation of the bio-
pharmaceutical target as early as possible in
the development process.
Revolution by Evolution Rational Design for
Desire and the Scientific Art of Optimization
In the next section we will see two impres-
sive examples of how biopharmaceuticals
can be optimized, e.g., with enhanced or
altered specific functions, by rational de-
sign. As we will see, rational design, mo-
lecular irrational design and directed evo-
lution can work in a synergistic way to
accelerate the development of modern bio-
pharmaceuticals. This synergism is high-
lighted with the examples of enhancing
PCR performance of a B-type DNA poly-
merase from Thermococcus aggregans
through molecular rational design and in-
creasing the activity of FIXa.
As discussed earlier FIXa is a key player
in the activation cascade of blood coagula-
tion. Subtle structural characteristics result
in an almost latent protease, and distin-
guish this enzyme from closely related co-
agulation factors like IIa, VIIa and Xa.
Thereby, FIXa can serve its dual capacity
to both boost and throttle blood coagula-
tion. Again, I have the pleasure to intro-
duce another friend from the Max-Planck-
Institute: Hans Brandstetter. After working
at Harvard Medical School and MIT, Hans
now serves as CSO at Proteros Biostruc-
tures, the company which I founded to-
gether with Robert Huber during my PhD
thesis at the Max-Planck-Institute. At that
time we won the first McKinsey Business
Plan contest good reputation and money
which helped to make the first steps with
the start-up company, still called Struc-
ture LifeSciences at these early days. Two
hearts were beating in my chest: on the
one hand, for me this was an excellent en-
trepreneurial experience; on the other
hand, the coaches from McKinsey made
me excited about business consulting
(which I then actually did). Interestingly,
our business concept to adopt biopharma-
ceutical drugs to the desired functionality
(being it specificity, selectivity or activity)
by means of rational structure-based de-
sign, was the right choice. I am happy that
Hans presents a really impressive example
of this approach (in this specific case link-
ing enzymatic and structural properties),
Executive Summary LXXIX
which was also published in Nature and
Proceedings of the National Academy of
Sciences of the USA: a FIXa triple mutant
featuring 7000-fold enhanced catalytic ac-
tivity!
In his chapter he reviews the substrate
preference of FIXa and homologous en-
zymes, and relates them to structural ele-
ments critical for substrate recognition.
His review illustrates how nature has opti-
mized FIXa as a strictly regulated enzyme
with multiple control mechanisms. The
evolutionary optimization reflects the ex-
treme danger of any misfiring of this en-
zyme due to its strategic role at the intersec-
tion between extrinsic and intrinsic coagula-
tion pathways, as well as between initiation
and amplification of coagulation.
Another example is presented by Harald
Sobek, who was working on postdoctoral
studies at Gesellschaft fr Biotechnologi-
sche Forschung before joining Roche, to-
gether with his colleague Zhixin Shao,
who received a PhD from Peking University,
and was working on postdoctoral studies at
CalTech before joining Roche. As also pub-
lished in Nature Biotechnology and Science,
they describe how rational design, molecu-
lar irrational design and directed evolution
work in a synergistic way to accelerate the
development of modern biopharmaceuticals
such as diagnostic products.
The primary goal of diagnostic testing is
the detection and quantification of disease-
specific analytes ranging from simple spe-
cies, such as ions, through complex bio-
molecules, such as drugs, hormones and
proteins, to complex analytes, such as cells
and viruses. As the most analytes occur at
low concentrations in complex biological
matrices, such as blood, plasma, sweat,
urine, feces or tissue biopsy, a high analyt-
ical sensitivity and specificity is required
in diagnostic testing. Many of the diagnos-
tic test methods mimic the way in which
biological molecules are recognized in or-
ganisms by specific molecular interactions.
For example, antibodies are used for the
detection of an antigen (e.g., a virus) in
immunological tests. If the analyte of in-
terest can be used as a substrate for a spe-
cific enzyme, an enzymatic assay can be
applied to determine the concentration of
this analyte. Therefore, proteinaceous bio-
molecules are extremely useful for the ap-
plication in diagnostic tests and further
development and optimization of these
proteins is continuously ongoing to match
new diagnostic challenges. The method
applied to improve the crucial properties
of these molecules ultimately determines
whether an enzyme or antibody can be
successfully used for innovative diagnostic
processes, lower manufacturing costs and
robust assay applications. With more and
more three-dimensional protein structures
available in databases, and with rapid de-
velopment of powerful molecular model-
ing tools, rational design is going to be
more efficient and broadly applicable.
Complex multiparameter optimization
problems are quite common to new and
innovative diagnostic applications. For
these applications, generating best-fit
biomolecules is particularly challenging,
mainly because of the rapidity with which
desired biomolecules must be created and
diagnostic processes using the newly cre-
ated molecules developed. For some diag-
nostic applications, rational engineering of
some intensively characterized proteins for
novel functions has been achieved. For en-
gineering most other diagnostic proteins
which have not been well characterized, a
primary irrational approach to introduce
random mutations into the whole or part
of the gene sequence and then screen or
select for expressed variants with the de-
sired properties has become more widely
used to enhance or to alter specific func-
Executive Summary LXXX
tions. Synergism of these methods is high-
lighted in the chapter with the examples
of enhancing PCR performance of a B-type
DNA polymerase from Thermococcus aggre-
gans through molecular rational design
and evolving the highly active calf intesti-
nal alkaline phosphatase (cIAP) in E. coli.
Part IV: Production of Biopharmaceuticals
The Industrys Workhorses
Mammalian Expression Systems
As we have now learned about methods to
accelerate the development of biopharma-
ceuticals by (ultra) high-throughput, to re-
duce their attrition rate by early target vali-
dation and to improve them for their spe-
cific application by rational design, we will
now focus on the production of biophar-
maceuticals. Manufacturing of biopharma-
ceuticals is crucial, because they represent
the fastest-growing sector within the phar-
maceutical industry. Biopharmaceuticals
have been successfully expressed in E. coli
for 25 years and Humulin (recombinant
human insulin developed by Genentech in
collaboration with Eli Lilly) was the first to
receive marketing authorization in the US
in 1982. This was the true beginning of
the biopharmaceutical industry, and since
then nearly 150 biopharmaceuticals have
gained approval for general human use in
the EU and/or US. More than 250 million
people worldwide have been treated to date
with biopharmaceuticals and the vast ma-
jority were protein-based either recombi-
nant proteins or monoclonal/engineered
antibodies. As stated before, the most fre-
quently used expression system was E. coli
for several years and nearly half of all pro-
tein biopharmaceuticals approved to date
were produced with this established host.
There is a plethora of literature available
about E. coli expression systems its ad-
vantages and limitations so that we pres-
ent only one example in this book and fo-
cus on recent developments with other
protein production systems. Several bio-
pharmaceuticals are produced using engi-
neered Saccharomyces cerevisiae and, as we
will see, this includes various insulin-
based products manufactured by Novo,
recombinant HBsAg produced by Smith-
Kline Beecham as well as a recombinant
form of the anticoagulant hirudin. The
majority of approved biopharmaceuticals
are however expressed in animal cell lines,
mainly Chinese hamster ovary (CHO), but
also baby hamster kidney (BHK) cells.
Although expression in animal cell lines is
technically more complex and expensive
when compared to E. coli-based systems,
eukaryotic cell lines, unlike prokaryotic
ones, are capable of carrying out posttrans-
lational modifications such as glycosylation
which is essential for some biopharma-
ceuticals (biological activity, stability, circu-
lating half-life, immunogenicity). In such
instances, expression in a eukaryotic sys-
tem becomes desirable, if not necessary.
While expression of biopharmaceuticals in
lower eukaryotes such as S. cerevisiae is
possible, glycosylation patterns more simi-
lar to native human proteins are obtained
if expressed in an animal cell line.
Florian M. Wurm, Professor of Biotech-
nology at the famous Swiss Federal Insti-
tute of Technology, is one of the pioneers
who has had experience with animal cell
lines from the very beginning, when he
was still working with Genentech. He pre-
sents a comprehensive overview on the im-
pressive development of cultivated mam-
malian cells. In 1986 (already 4 years after
approval of E. coli-derived insulin Humu-
lin), the first protein biopharmaceutical
made in such cells obtained market ap-
proval and made the use of CHO cells in
Executive Summary LXXXI
large-scale bioreactors known to a wider
public. These cells are now the dominat-
ing host system for recombinant protein
production as more than 60% of all new
target proteins in the clinical pipelines of
pharmaceutical and biotechnology compa-
nies are being produced in hamster-de-
rived cells. Professor Wurm covers aspects
of gene transfer, cell line development and
process development for mammalian pro-
tein expression systems using CHO cells
as the main example, but makes reference
to other mammalian cells that are used for
large-scale production of biopharmaceuti-
cals. Most importantly, the scientific and
technological insights that resulted in the
rapid and surprising yield improvements
from such processes, bringing the volu-
metric productivity of mammalian cell cul-
ture processes into the gram per liter
range (a level of productivity equal to that
of microbial systems!), are discussed. The
know-how and technology behind large-
scale processes for mammalian cells have
evolved over 20 years and have, as he re-
cently also published in Nature Biotechnol-
ogy, resulted in more than 100-fold im-
provement in volumetric productivity. So-
phisticated processes in extended batch
cultures of up to 12000 L were developed
in the meantime. The chapter concludes
with discussing the regulatory framework
for the use of mammalian host systems,
since the perceived risks of transmission
of adventitious agents [e.g., viruses, patho-
gens, bovine spongiform encephalopathy
(BSE), etc.] to patients resulted in strin-
gent rules to which all manufacturers
must adhere.
As stated before, eukaryotic cells in gen-
eral are capable of carrying out posttransla-
tional modifications such as glycosylation.
Although this is essential for some bio-
pharmaceuticals to ensure biological activ-
ity, stability and increase circulating half-
life, glycosylation patterns provided by ani-
mal cells differ from human glycosylation.
When applied to the patient, this can lead
to an immunogenic reaction which can
only be avoided by using human cells for
expression. Such a human cell line was
jointly developed by Crucell in The Nether-
lands and DSM Biologics Inc. in Canada.
PER.C6
was generated by immortaliza-

tion of primary human retina cells with
E1 sequences of human adenovirus sero-
type 5. As previously published in Human
Gene Therapy, the cell line was initially de-
veloped for the safe production of pharma-
ceutical-grade recombinant human adeno-
viral vectors, but more recently also for the
production of therapeutic proteins. Bram
Bout, who gained his PhD at the Aca-
demic Medical Center in Amsterdam, is
Vice President at Crucell and shares his
PER.C6 experience with us. His colleague
Chris Yallop was at University College
London before he moved to Novo Nordisk
and later on to Crucell. Another colleague,
Dirk-Jan Opstelten, was at the Karolinska
Institute before joining Crucell as Director
of Protein R & D, being responsible for
developing PER.C6 as a platform for pro-
duction of mAbs and therapeutic proteins.
In their contribution, they focus on one
group of therapeutic proteins, the mAbs,
which have shown particularly rapid
growth in recent years, increasing from
approximately 1% of therapeutic protein
sales in 1995 to 14% in 2001. There are
currently nearly 30 approved antibodies on
the market and many more in the late
stages of clinical development. Due to the
high doses required for many antibody
therapies, high product yields are particu-
larly desirable. Therefore, it was the aim
of Crucell and DSM Biologics to establish
the human PER.C6 cell line as a platform
for the production of biopharmaceuticals
with particular emphasis on mAbs. The
Executive Summary LXXXII
approach taken has been to develop an in-
tegrated production platform, that com-
bines the rapid generation of high-yielding
production cell lines with high-yielding
generic production (batch, fed-batch and
perfusion) and purification processes as
well as metabolically characterized host
cell lines. Data generated from the meta-
bolic characterization of PER.C6 cell lines
were used to design generic, high-yielding
batch, fed-batch and perfusion production
processes, matched to the metabolic re-
quirements of the cells. Adhering to this
unique approach, it becomes possible to
evaluate cell lines as early as possible in
the desired production process, so that
lead clones are selected which match and
will perform optimally in the desired pro-
duction process. The excellent work de-
scribed here gives an overview of clone
generation, fed-batch and perfusion pro-
cess development, as well as detailing the
history of the PER.C6 cell line and how it
has been characterized for gaining ap-
proval from regulatory authorities.
Another approach towards human glyco-
sylation is followed by the utilization of a
humanized mouse cell line, a human/
mouse heterohybridoma presented by Dr.
Volker Sandig from ProBioGen. I have
known Volker for a couple of years, since
when I invented a real-time PCR test kit
for the detection of mycoplasma contami-
nation in pharmaceutical products. Using
an internal standard we developed this
method for rapid in-process (IPC) control
for production of biopharmaceuticals and
ProBioGen was one of the partners partici-
pating in the validation of the system as
they wanted to use it for rapid quality
analysis of their designer cell lines. As we
have learned from previous examples, the
development of mammalian super-produ-
cer cells from CHO (or also from the
mouse myeloma cell line, NS0) starter cell
lines is an unpredictable and time-con-
suming effort, requiring the identification
of rare clones which combine integration
of the expression unit into a highly active
genomic locus with superior folding, pro-
cessing and secretion capabilities. Fine
tuning the selection and vector, which in-
cludes new cellular promoters, allows us
to reproducibly generate productive clone
pools of CHO cells suitable for immediate
production of test material and improves
identification of superior clones. Alterna-
tively, the fast and reliable generation of
clones is achieved by site-specific cassette
exchange based on, for example, the Gate-
way system as described earlier by John
Chesnut from Invitrogen. Volker Sandig
and colleagues from ProBioGen have ex-
panded this strategy by using the strong
IgH locus of their G-line (human/mouse
heterohybridoma). Replacement of the en-
dogenous human IgM heavy chain gene
provides the environment for efficient
transcription, secretion and a mostly hu-
man glycosylation pattern for Ig fusion
proteins.
As a new platform alternative to CHO
and NS0, which supports the production
of fully human proteins, they describe the
evaluation of human designer cell lines of
various tissues created directly from pri-
mary cells. The hybridoma was shown to
secrete the antibody in a stable manner at
a rate of 45 pg (cell day)
1
over a period
of 2 years. This is in striking contrast to
the majority of heterohybridomas, which
tend to quickly lose human chromosomes,
resulting in unstable immunoglobulin ex-
pression. Moreover, expression was pre-
served when the cell line was cultivated in
high-density fermentation and remained
stable in five independent fermenter runs
which had a mean duration of 66 days.
Expression did not decrease below
30 pg (cell day)
1
when the medium was
Executive Summary LXXXIII
exchanged for a protein-free medium in a
continuous fermentation run.
Based on the presence of the set of hu-
man chromosomes in the heterohybrido-
ma, a glycosylation pattern different from
that of mouse myelomas such as NS0 was
expected. Indeed, the single N-linked
oligosaccharide chain located in the Fc re-
gion was sialylated at 37% (a rate close to
average sialylation on antibodies in human
blood) and sialic acids were mainly the N-
acetylneuraminic acid typical of human
cells. Only 2% were represented by N-gly-
colylneuraminic acid, the immunogenic
form dominating in mouse myeloma cells.
a(1,3)Gal structures, not made in human
cells and recognized by pre-existing anti-
bodies, were only found in 1.3% of the gly-
cans. This suggests that the G-line indeed
executes glycosylation in a way that better
resembles the pattern of human compared
to mouse cell lines. Based on the selection
systems in the targeting vectors, the G-line
allows the simple introduction of second-
ary target genes via recombinase-mediated
cassette exchange. Multiple glycoproteins
have been introduced into the IgH locus.
For example, for a
1
-antitrypsin introduced
into the IgH locus, expression levels
reached 9 pg (cell day)
1
. In general, ex-
pression levels of secondary transgenes
were comparable to those achieved in
CHO cells and the G-line seems particular-
ly well suited for Fc-fusion proteins which
are more difficult to produce in other sys-
tems. Several of the clones secreted more
than 75 pg (cell day)
1
, at least 6-fold more
than the best CHO producers isolated for
this protein so far. Titers of up to 0.5 g L
1
,
accumulated over 17 days of stationary cul-
ture in T flasks, could be obtained.
As previously described, a high-yielding
biopharmaceutical protein manufacturing
process is the result of using a number of
approaches that affect the cell line per se,
the cell culture process, product recovery
and purification activities. The optimiza-
tion of cell culture processes by, for exam-
ple, improving media and by developing
advanced feeding strategies that support
high space-time yields of viable biomass
has substantially increased the product
concentrations achieved in the bioreactor.
One such strategy was developed by Lon-
zas CSO, Professor John Birch, and collea-
gues. John received his PhD from London
University, where he was giving lectures
before he went into industry working for
Searle, Celltech and now Lonza. His focus
is on the production of therapeutic pro-
teins, particularly from mammalian cells
and as Professor at University College
London he has also published extensively
in this field. The chapter reviews an ex-
pression technology based on the use of
the glutamine synthetase (GS) gene and
the integration of this technology into the
development of high-yielding, large-scale
manufacturing processes for biopharma-
ceuticals at Lonza. Professor Birch pro-
vides examples of improvements in both
the creation of cell lines and in cell culture
optimization.
The enzyme GS catalyses the formation
of glutamine from glutamic acid and am-
monia, driven by hydrolysis of ATP. Gluta-
mine has multiple roles in cell metabo-
lism, particularly as an energy source, pro-
tein constituent, and as a nitrogen donor
in purine and pyrimidine synthesis. Cell
lines that do not produce GS have an ab-
solute requirement for glutamine and do
not grow in glutamine-free culture media;
hence, this provides the basis for using
the enzyme as a selectable marker in gene
expression vectors. The utility of GS as a
selectable marker is increased by the avail-
ability of an efficient inhibitor of GS,
methionine sulfoximine (MSX), which can
be used to improve the stringency of selec-
Executive Summary LXXXIV
tion, to select for gene amplification and
to inhibit enzyme activity in those cell
lines which produce endogenous GS. Most
myeloma and hybridoma cells have an ab-
solute requirement for glutamine. In con-
trast, many other cell types such as BHK-
21, L-cells and the widely used CHO do
not require glutamine, provided glutamic
acid is present in the culture medium. In
these cases GS can still be used as a se-
lectable marker, but it is necessary to use
a specific inhibitor of GS, such as MSX, to
inhibit the endogenous enzyme. In prac-
tice, the most commonly used cell lines
are NS0 and CHO, and whilst GS-NS0 has
been used most commonly for antibody
production, GS-CHO has been used to ex-
press a large range of other protein types.
Altogether, antibodies, enzymes, interleu-
kins and membrane-bound proteins were
successfully produced using the GS ex-
pression system.
The chapter outlines how the collegues
at Lonza successfully increased productiv-
ity and, because one of the concerns there-
by is to maintain product quality character-
istics, they have monitored the product
quality of the GS-NS0 cell line throughout
the entire optimization process. Through
successive rounds of optimization involv-
ing changes in the composition of the
feeds, culture pH and extending culture
duration, the product concentration from
the GS-NS0 process was increased from
0.37 to 1.4 g L
1
and there were no major
changes observed in the oligosaccharide
profiles during this optimization process.
From these impressive examples we
have seen that at least for recombinant
antibodies it is possible to regularly
achieve yields in excess of 1 g L
1
in com-
pletely chemically defined media. It is
hoped (and personally I am convinced)
that yields of at least 10 g L
1
will be
achieved for modern biopharmaceuticals
in the foreseeable future in mammalian
expression systems.
Vivat, Crescat, Floreat
A Ripe and Blooming Market for Transgenic
Animals and Plants
The next section deals with transgenic ex-
pression systems, either plants or animals,
also known as molecular pharming. I
am working in this area myself and I am
always amazed about the opportunities
that lie in these alternative expression
hosts others obviously share my enthu-
siasm. Last year I had the pleasure to pro-
vide a chapter on plant expression systems
for a biotech book my contribution was
promptly selected by the Board (consisting
of 10 Nobel Prize laureates!) of the out-
standing reference Encyclopedia of Molecu-
lar Cell Biology and Molecular Medicine and
has just been published there. One reason
why alternative expression systems are so
exciting and important is because it be-
came obvious that production capacities
for biopharmaceuticals with conventional
bioreactors would be a bottleneck and that
worldwide fermentation capacities are lim-
ited. Plant expression systems are an excit-
ing solution to these capacity crunches
and the next section describes different
plant expression systems, their advantages
and limitations, and some of the innova-
tions and trends likely to influence the fu-
ture of plant-based biopharmaceuticals.
The section concludes with another no-
velty the first biopharmaceutical from a
transgenic animal: approval of ATryn
[an-
tithrombin (AT) III] from goat milk.
A comprehensive overview on plant- as
well as animal-based production of bio-
pharmaceuticals is given by my friend
Julio Baez, who worked for several years
with Monsanto as VP of Research and was
leading a team to produce the first inject-
Executive Summary LXXXV
able mAb from transgenic corn used in an
FDA-approved clinical trial. Julio provides
a comprehensive overview on the technical
advances during the past 20 years that
have enabled the genetic transformation
and regeneration of transgenic plants and
animals for the tissue-specific accumula-
tion of recombinant human proteins.
These transgenic systems are able to also
produce biopharmaceuticals requiring
complex multisubunit assembly, such as
vaccines and secretory antibodies. They are
also used for proteins that cannot be effi-
ciently synthesized by currently commer-
cialized microbial or mammalian cell cul-
ture systems. Julio, who is also lecturing
at Stanford University, focuses his article
on biopharmaceuticals derived from trans-
genic animals and plants that are currently
commercialized or that have human clini-
cal experience. Manufacturing biopharma-
ceuticals in transgenic animals and plants
grown based on conventional agronomic
and farming practices also offers the op-
portunity to produce practically unlimited
supplies of life-saving products at low cost.
In addition to providing enabling technol-
ogy at significant lower cost and with ad-
vantages in product availability, production
of biopharmaceuticals using selected trans-
genic systems, such as milk, offers the
highest accumulation level of heterologous
proteins ever obtained from any recom-
binant production systems. Transgenic
plants offer the possibility to produce bio-
pharmaceuticals free of potential animal-
derived contaminants and pathogens such
as prions in a matrix that can be used for
oral delivery without additional purifica-
tion and not requiring refrigeration. Seeds,
for example, provide a stable matrix for
handling and storing biopharmaceuticals
for years after harvest, decoupling down-
stream processing from biosynthesis. In
summary, transgenic systems can deliver
all kinds of innovative biopharmaceuticals
for the treatment of cancer, infectious dis-
eases, inflammation, organ rejection, skin
conditions, genetic deficiencies and respi-
ratory ailments, which will be affordable
and accessible to broad segments of the
population and developing regions of the
world that currently do not have access to
them.
I am delighted that Harry Meade from
GTC Biotherapeutics and his colleagues
Yann Echelard and Carol A. Ziomek
agreed to provide their excellent contribu-
tion on the first biopharmaceutical from
transgenic animals: ATryn (AT III from
goat milk). AT concentrates derived from
pooled human plasma have been used for
the management of hereditary and ac-
quired AT deficiencies since the early
1980s. The development of a recombinant
version of AT would alleviate supply and
safety concerns associated with the use of
the plasma-derived biopharmaceutical.
However, the complex structure of the AT
molecule and the large doses usually re-
quired in supplementation treatments
have precluded the use of traditional bacte-
rial and cell culture bioreactors for com-
mercial production. GTC Biotherapeutics
has applied their transgenic animal expres-
sion system to the production of recombi-
nant human AT (rhAT), trade name ATryn.
This approach provides the opportunity to
produce recombinant forms of proteins
that are difficult to express in conventional
production methods. The progress made
with this system over the years was pub-
lished by the authors a couple of times in
Nature Biotechnology: a herd of transgenic
dairy goats expressing high levels of rhAT
in milk was generated, characterized, bred
and expanded, providing a homogeneous,
well-defined and abundant supply of rhAT.
Their review describes the clinical develop-
ment of ATryn (including eight clinical
Executive Summary LXXXVI
studies) and the production of this modern
biopharmaceutical in transgenic goats.
GTC has submitted and discussed in a
meeting, its responses to the consolidated
list of questions generated by the European
Medicines Agency (EMEA) as part of the re-
view of a Marketing Authorization Applica-
tion (MAA) for ATryn. The MAA covers
the use of ATryn in the prophylactic treat-
ment of patients with hereditary AT defi-
ciency during high-risk situations such as
surgery and childbirth. GTC expects the
EMEA to respond with additional questions,
or provide an opinion on the MAA, soon.
We believe we had very constructive and
valuable meetings with the agency recently
and we are pleased that the EMEA has
granted us this extension to complete our
response to all outstanding issues, noted
Geoffrey F. Cox, GTCs Chairman and
CEO. This provides us with the opportu-
nity to bring our MAA to a successful con-
clusion and we will be continuing to work
diligently towards this goal. Subject to ap-
proval of the MAA, GTC is planning for a
European market launch of ATryn in 2005,
representing the first recombinant thera-
peutic protein produced using transgenic
technology to be approved by regulatory
authorities anywhere in the world. Thus, I
hope that by the time when Modern Biophar-
maceuticals is available on the book shelves,
ATryn will be approved and available on the
pharmacy shelves!
Now we shift gear slightly and switch
from transgenic animals to transgenic
plants. ICON Genetics in Halle, which I
consider as one of the leading experts in
plant expression, use tobacco for the ex-
pression of biopharmaceutical proteins.
Professor Yuri Gleba from the Ukrainian
Academy of Sciences is one of the early
pioneers of biopharmaceutical production
in plants, with over 30 years of experience,
and also founded the International Insti-
tute of Cell Biology, Kiev, Ukraine, in
1988, where he still serves as its Director.
With more than 200 research papers (Na-
ture, Nature Biotechnology, Science and Pro-
ceedings of the National Academy of Sciences
of the USA), several books, book chapters
and over 20 patents, Yuri has earned the
respect of the international scientific com-
munity as is evidenced by his election to
the World Academy of Arts and Science
(Rome) or receiving the USSR State Prize
(former Stalin Prize). He joined American
Cyanamid Company, Princeton, in 1992
and served as Director of the Crop Engi-
neering Department, before he co-founded
ICON Genetics in 1999, where he still
serves as CEO. His colleague Victor Klim-
yuk has over 20 years of research and
management experience in the fields of
plant molecular biology, plant genetics and
biotechnology, including time spent at the
Russian Academy of Sciences and Hungar-
ian Academy of Science. Victor has pub-
lished numerous research and review pa-
pers, as well as over 20 patents in the field
of plant biotechnology. He joined ICON in
1999 and is currently serving as CSO.
It was a pleasure for me to not only
share my thoughts with these very experi-
enced collaborators in a joint project on
molecular pharming, but also to co-
author this chapter. Here, we review the
progress and challenges in the area of pro-
duction of recombinant proteins, in partic-
ular biopharmaceuticals, in plants. Differ-
ent expression platforms are summarized,
including those based on the use of trans-
genic, transplastomic or transfected plants
as production hosts. The quality and yield
of recombinant proteins produced in and
purified from plants, as well as progress
in clinical trials with plant-made biophar-
maceuticals, are described.
Whereas, initially, the emphasis in mo-
lecular pharming was on unlimited scala-
Executive Summary LXXXVII
bility and the low cost of plant-based pro-
duction, yield and biosafety issues were
not necessarily properly addressed. How-
ever, the last two parameters are crucial
for determining the economics and, conse-
quently, the chances for commercial suc-
cess of each specific plant-based system.
Hence, the advantages, limitations and
biological safety aspects of plant-based pro-
tein production are also discussed.
A fresh boost was actually given to
plant-based molecular pharming in re-
cent years, as the biopharmaceutical indus-
try is trying to eliminate manufacturing
processes that rely on production in ani-
mal cells due to the possible contamina-
tion of these products by human patho-
gens such as BSE or (variant) Creutzfeldt-
Jacob disease (CJD, vCJD) as described in
the previous chapters. In summary, we
discuss different expression systems that
are being developed. We consider the po-
tential of each system by taking into ac-
count the impact of several parameters on
economics and regulatory acceptability of
the system: productivity (absolute and rela-
tive yield), biological safety (in particular
transgene containment), scalability, versa-
tility (ability to accommodate diverse pro-
teins and to express a recombinant protein
identical to the natural one), speed of re-
search, development and commercial scal-
ability provided by each of these systems.
The technique of cloning and growing en-
tire plants, and subsequently performing
scale-up in the field, is also shown on the
supplementary CD-ROM. One system is
described in detail, which is ICON Genet-
ics straightforward viral system: based on
realistic yields of 100 tons of plant leaf bio-
mass per hectare of a greenhouse per year,
a 1 ha facility should be capable of produc-
ing 280400 kg of recombinant protein a
year! This means that for the vast majority
of biopharmaceutical proteins, industrial-
scale production could be done entirely in
a partially or fully contained greenhouse
facility.
Ongoing public fears from the food in-
dustry and the public, particularly in Eu-
rope (Franken Food), could have spillover
effects on plant-derived biopharmaceuti-
cals. Mistakes and misunderstandings
have already cost the genetically enhanced
grain industry hundreds of millions of
Euros, and as I have stated earlier, in Ger-
many, for example, open-field studies are
literally impossible, especially since the
Bundestag decided in November 2004 to
implement an even more restrictive law
for genetic engineering.
One solution to this problem could be
another interesting technology: the moss
bioreactor from greenovation in Freiburg.
This system shares the advantage of utiliz-
ing nonedible plants (nonfood and non-
feed) and being able to be kept in a fer-
menter to avoid any segregation risk. An-
other obvious advantage is secretion of the
protein into the medium so that no grind-
ing or extraction is required. This is very
important in light of downstream process-
ing, because protein purification is often
as expensive as the biomanufacturing and
should never be underestimated in the to-
tal cost of goods sold (COGS) equation.
Gilbert Gorr, CSO of greenovation and
co-inventor, together with Sabrina Wagner,
co-founder and CEO, successfully ex-
pressed active human VEGF (vascular en-
dothelial growth factor) a homodimer
linked via a disulfide-bridge in moss.
They discuss additional unique properties
of this moss bioreactor: mosses are culti-
vated as haploid, photoautotrophically ac-
tive and fully differentiated gametophytic
tissue performed as suspension cultures.
In addition, moss is the only known plant
system which shows a high frequency of
homologous recombination which allows
Executive Summary LXXXVIII
for gene knockouts, opening the possibility
of genetic engineering of the glycosylation
pathway and this is exactly what they de-
scribe: human-like glycosylation of bio-
pharmaceuticals expressed in the glyco-en-
gineered moss Phycomitrella patens.
Production of human proteins in moss
was first shown by the expression of
rhVEGF. The rhVEGF was successfully tar-
geted to the secretory pathway, resulting in
efficient secretion of the recombinant pro-
tein into the medium and it was shown
that the moss-derived rhVEGF was biologi-
cally active. One important criterion for
successful expression of a therapeutic pro-
tein from a recombinant cell is to obtain a
transgenic plant that maintains stability of
production and, in addition, stability at the
molecular level. Several transgenic moss
strains aged 2 and 7 years were therefore
examined concerning expression of the tar-
get protein rhVEGF and neomycin phos-
photransferase as an antibiotic resistance
marker. Protein levels of rhVEGF were
measured by ELISA and found to be un-
changed after 7 years. Furthermore, 100%
of the transgenic plant material showed re-
sistance to the antibiotic G418 even after
several years of cultivation without selec-
tion pressure. Protein analysis data at the
molecular level revealed that the VEGF is
fully active. This makes the moss bioreac-
tor an ideal production system for biophar-
maceuticals, even under strict regulatory
requirements.
In the next chapter, a semiclosed sys-
tem is described by Ning Huang and Dai-
chang Yang from Ventria Bioscience, Sa-
cramento, CA. As published in Proceedings
of the National Academy of Sciences of the
USA, they have developed ExpressTec
TM
to
produce biopharmaceuticals cost-effectively
and in large quantities in seeds. Ning de-
scribes the success of ExpressTec, because
it utilizes the latest developments in plant
molecular biology with the use of strong,
endosperm-specific promoters; signal pep-
tides targeting the subcellular compart-
ments to prevent proteolytic degradation
of the recombinant protein; optimized co-
dons to maximize translational efficiency;
and transcriptional activators that increase
target gene transcription and control of
the expression of competitive molecules.
Several recombinant proteins could be pro-
duced using the ExpressTec system with
levels of 0.11% of brown rice weight or
2560% of soluble protein. The data pre-
sented show that both the transgenes and
their expression are stable over 5 years
and 10 generations. The physical and bio-
chemical properties of the recombinant
proteins are the same as of the native pro-
teins. Scale-up processing has shown that
recombinant proteins are easily extracted
from cereal grains and Ventrias economic-
al analysis has placed the cost of biophar-
maceuticals produced by ExpressTec at
about US$ 6/g.
So far, we have talked about whole plants,
or parts thereof, as expression systems. The
next chapter, provided by my colleagues
from the Fraunhofer-Institute for Molecular
Biology and University of York, therefore fo-
cuses on cultured plant cells. Stefan Schill-
berg, Rainer Fischer and Richard Twyman
are the real experts for this topic: they have
published their work in several papers in
Nature Biotechnology and Proceedings of the
National Academy of Sciences of the USA, as
well as the recent Wiley book Molecular
Pharming: Plant-made Pharmaceuticals and
Technical Proteins. They describe why pro-
duction systems that utilize whole plants
lack several of the intrinsic benefits of cul-
tured cells, including precise control over
growth conditions, batch-to-batch product
consistency, high level of containment and
ability to produce biopharmaceuticals in
compliance with current GMP (cGMP).
Executive Summary LXXXIX
Plant cell cultures combine the merits of
plant-based systems with those of microbial
and animal cell cultures, particularly in
terms of downstream processing. In their
chapter they discuss the benefits of plant
cell cultures compared to other systems,
the technological requirements for produc-
ing biopharmaceutical proteins in plant
cells and the unique aspects of downstream
processing which are applied to this expres-
sion platform.
All transgenic plant production technol-
ogies presented in this section, although a
relatively new arena in modern plant
science, have immense potential to change
the shape of agriculture and offer new op-
portunities in the field of modern agrobio-
technology. As we have seen, the use of
transgenic plants in expressing biopharma-
ceuticals is feasible; however, several envi-
ronmental conditions such as salinity,
drought and extreme temperatures are ma-
jor limiting factors for plant growth and
crop productivity. High-temperature stress
is highly unfavorable in optimal growth of
plants, but nearly 25% of total arable land
is affected by heat and drought stress. The
annual mean air temperature of 23% of
the Earths land surface is above 408C al-
ready and with the increasing concentra-
tion of greenhouse gases, the Earths sur-
face temperature is expected to increase by
up to a further 58C by 2100. This rise in
ambient temperature would obviously
warm the climate in most parts of the
world. In addition, it is estimated that
more than one-third of all of the irrigated
land in the world is presently affected by
salinity, and this is exclusive of the regions
classified as arid and desert lands already
(which comprise 25% of the total land of
our planet anyway). The problem of
drought stress is even more severe and
economically damaging. Drought and sa-
linity are predicted to cause serious salini-
zation of more than 50% of all arable land
by the year 2050!
Having said that, the problem is that
conventional plant-breeding tools have
been of only limited help so far in alleviat-
ing these abiotic stress problem. Recent
microarray studies have been employed for
examination of expression profiles of the
whole genomes of some plants in re-
sponse to saline and drought stress. The
use of microarray techniques have signifi-
cantly accelerated efforts in assigning the
functional role of genes involved in plant
responses to saline and drought stresses,
and have shed light on the possible in-
volvement of regulatory pathways in stress
tolerance. This information is used for
planning new strategies for the production
of abiotic stress-tolerant transgenic plants.
Shimon Gepstein from Technion Institute,
Haifa, Israel, together with his colleagues
Anil Grover from the University of Delhi,
New Delhi, India, and Eduardo Blumwald
from the University of California, Davis,
present their recent breakthroughs, which
they have also published in Nature, Nature
Biotechnology and Science. Professor Gep-
stein, who was at Stanford University and
currently serves as Dean at Technion, fo-
cuses on plant senescence and plant re-
sponses to abiotic stress. Professor Anil
Grover, who was a Fellow of the National
Academy of Sciences and the Rockefeller
Foundation, was recently awarded the Na-
tional Bioscience award for his contribu-
tions to understanding the roles of heat-
shock proteins in the stress response. Pro-
fessor Eduardo Blumwald contributed tre-
mendously to the understanding and
engineering of salt tolerance in plants, is
on the Editorial Board of Trends in Plant
Sciences, and was organizer of the Gordon
Conference on Salt and Drought Stress in
Plants. Of late, transgenic plants have
been raised that in fact show increased
Executive Summary XC
resistance to abiotic stresses like heat,
drought and salinity, e.g., transgenic Arabi-
dopsis plants expressing heatshock protein
(Hsp) could tolerate temperatures as high
as 508C, whereas while wild-type plants
were already killed above 458C.
Recently, tomatoes with increased yields
under drought conditions could be ob-
tained by introducing yield-promoting
genomic regions from the drought-tolerant
green-fruited wild species. The yield of the
hybrids was more than 50% higher than
that of a control market leader variety un-
der dry field conditions that received only
10% of the irrigation water. Salt tolerance
can be attained by limiting Na
+
accumula-
tion in plant cells and indeed compart-
mentalization of Na
+
ions into vacuoles
provides an efficient mechanism to avert
the toxic effects of Na
+
in the cytosol. The
transport of Na
+
into the vacuoles is
mediated by a Na
+
/H
+
antiporter and its
overexpression resulted in transgenic
plants that were able to grow in high salt
concentrations. Transgenic tomato plants,
for example, grown in the presence of
200 mM NaCl were able to grow, flower
and set fruit. Although the leaves accumu-
lated high sodium concentrations, the to-
mato fruits displayed very low amounts of
sodium. Similar results could be obtained
with transgenic canola: leaves grown in
the presence of 200 mM NaCl accumu-
lated sodium to up to 6% (!) of their dry
weight, but the seed yields and oil quality
were not affected, demonstrating the po-
tential use of this technology for agricul-
tural use in saline soils.
Here, we have summarized the stress re-
sponse molecular mechanisms and their
biotechnological applications. Based on the
various stress response mechanisms and
the identification of the corresponding
stress-induced genes, genetically engi-
neered plants have been produced and
some of them display significant improved
abiotic stress tolerance for salt, heat and
drought. Thus, in the future we will be
able to produce even high-value traits,
such as biopharmaceuticals, in areas on
our planet which are today not farmable
and cannot be used for any agriculture at
all. Combining the advantages of advanced
plant expression systems (which we dis-
cussed in the previous chapters) with im-
proved abiotic stress tolerance (such as for
salt, heat and drought) needs to be the
next stage of development. As plant-
derived biopharmaceuticals demonstrate
widespread, tangible benefits to the popu-
lation, and as the plant expression indus-
try develops a longer safety track record,
public acceptance of the technology is
likely to improve continuously. Plants are
by far the most abundant and cost-effective
renewable resource uniquely adapted to
complex biochemical synthesis. The in-
creasing cost of energy and chemical raw
materials, combined with the environmen-
tal concerns associated with conventional
pharmaceutical manufacturing, will make
plants even more compatible in the future.
With the words of Max Planck (1858
1947) How far advanced Mans scientific
knowledge may be, when confronted with
Natures immeasurable richness and capacity
for constant renewal, he will be like a marvel-
ing child and must always be prepared for
new surprises, we will definitely discover
more fascinating features of plant expres-
sion systems. But there is no need to wait:
combining the advantages of some tech-
nologies that we have in hand by now
could already lead to the ultimate plant ex-
pression system. This is what we should
focus on. And this is really my vision, be-
cause then, at the dawn of this new mil-
lennium, this would for the first time yield
large-enough amounts of biopharmaceuti-
cals to treat everybody on our planet!
Executive Summary XCI
Alea Non Iacta Est Improving Established
Expression Systems
Following the same goal, in the next sec-
tion we discuss the continuous efforts and
recent results in improving established ex-
pression systems. We start with a contribu-
tion from Martin Fussenegger, Professor at
the Swiss Federal Institute of Technology
(ETH Zrich), and his colleague Wilfried
Weber. Martin Fussenegger did his PhD
together with Werner Arber at the Univer-
sity of Basel, worked at the Max-Planck-In-
stitute before he was awarded by the Swiss
National Science Foundation and became
Professor of Molecular Biotechnology. Wil-
fried Weber has a Masters in Biotechnol-
ogy from the European School of Biotech-
nology Strasbourg, did his diploma at No-
vartis and is co-founder, together with Mar-
tin Fussenegger, of the biotech startup
company Cistronics Cell Technology. Both
have extensive experience on the use of
baculovirus-based production of biophar-
maceuticals using insect cell cultures. The
baculovirus expression vector system
(BEVS) developed for heterologous protein
production in insect cell cultures almost
three decades ago is one of todays pre-
ferred pilot-production technology due to
the BEVSs superior protein titers and un-
matched gene-to-protein process speed,
which still surpasses currently available
mammalian cell-based production pro-
cesses. It represents a well-established
technology for straightforward pilot-scale
production of desired heterologous pro-
teins, advances in the generic process, in-
cluding (a) optimized expression vectors,
(b) development of chemically defined cul-
ture media, (c) elaboration of the best nu-
tritional and kinetic parameters as well as
(d) the design of novel cultivation hard-
ware, and boosted the overall product
yield, while reducing process time and
costs. With product yield at its near maxi-
mum, the BEVS community is currently
focusing on the improvement of product
quality by humanizing insect cell glycosy-
lation patterns. This comprehensive over-
view of the baculovirus-based system, with
special emphasis on the development of
an integrated process design, illustrates
key process parameters by a case study
covering the production of a mammalian
kinase. The unique combination of transi-
ent expression implementation, high-yield
protein production capacity and the pro-
spect of human glycoprofiles in insect cul-
tures indicates a bright future for BEVS
technology in the production of human
biopharmaceuticals.
In the next contribution we learn about
hands-on experience and recent improve-
ments with different production systems
for biopharmaceuticals at Bayer Health-
Care. As previously also published in Na-
ture by Heiner Apeler, Head of Expression,
an E. coli host/vector system was originally
developed for the efficient production of
an interleukin-4 variant, but afterwards it
was optimized for the expression of other
proteins and even Fab fragments. Process
development and optimization of the yeast
secretory Saccharomyces cerevisiae for ex-
pression of a protease inhibitor will also
be presented. The focus, however, is on
the use of a recently developed mamma-
lian HKB11 (hybrid clone of human kid-
ney and B cells) expression system for re-
combinant human glycoprotein biophar-
maceuticals. HKB11 is a favorable cell host
for the production of human proteins, be-
cause it delivers biopharmaceuticals that
are structurally identical to the natural
product. The host/vector system supports
the production of gram quantities of pro-
teins in a large-scale transient transfection
format as well as the development of
stable cell lines. These systems together
Executive Summary XCII
with the baculovirus insect system are
used routinely within Bayer HealthCare
Pharma Biotechnology for the production
of biopharmaceutical proteins for research
purposes, for proof-of-concept studies and
also for therapeutic applications.
Another case study with hands-on ex-
perience using the yeast S. cerevisiae is pre-
sented from Novo Nordisk, the leading in-
sulin producer. Ivan Diers, who studied at
the Technical University of Denmark be-
fore joining Novo Nordisk in 1967, has a
long track record with expression of pro-
teins in S. cerevisiae. His colleague Asser
Sloth Andersen, who studied at Kings Col-
lege, University of London before joining
Novo, also has many years experience with
host cell engineering. Both present how S.
cerevisiae is used as the work horse to
manufacture insulin on an industrial
scale. Insulin is a naturally occurring pep-
tide hormone produced by the b-cells in
the Langerhans islets of the pancreas in
response to hyperglycemia. This is nicely
shown on the supplementary CD-ROM.
Insulin facilitates entry of glucose into tar-
get tissues such as muscle, adipose tissue
and liver by binding to and activating spe-
cific membrane receptors on these cells.
The WHO estimates that some 170 mil-
lion people suffer from diabetes, a figure
that is likely to double by 2030. Although
only a minority of these sufferers actually
require daily insulin injections, the current
world market for insulin is valued at in ex-
cess of US$ 4.5 billion, a figure that is
likely to reach US$ 8 billion before the
end of the decade. For Novo Nordisk, the
insulin business is continuously growing
and the company expects an overall
growth of up to 14% in 2005, forecasting a
general expansion in the world insulin
market.
Diabetes mellitus is a group of meta-
bolic diseases characterized by high blood
sugar (glucose) levels, which result from
defects in insulin secretion, action or both.
In Type 1 diabetes this may be due to b-
cell destruction, and in Type 2 diabetes to
a combination of b-cell failure and resis-
tance of target tissues to insulin action (in-
sulin resistance). The latter disease can in
its early stages be helped by a low-calorie,
nonsugar diet and/or treatment with oral
antidiabetic drugs, while the later stages
and Type 1 diabetes require insulin treat-
ment. In the first 60 years after the discov-
ery of insulin by Frederick G. Banting
(18911941; Nobel Prize in 1923) and
Charles Best (18991978) in 1921 and the
successful treatment of diabetics, only in-
sulin extracted from bovine or porcine
pancreases was available to treat Type 1
diabetics. Unfortunately, with the rapid in-
crease in the incidence of diabetes, it is no
longer possible to satisfy the pharmaceuti-
cal requirement (estimated to be 1520
metric tones per year in 2005) from ani-
mal sources. Furthermore, the insulins ex-
tracted from animals are slightly different
from human insulin, which might cause
the formation of insulin-binding antibod-
ies and allergic reactions. Porcine insulin,
which only deviates by a single amino acid
in position B30 (the last amino acid of the
B chain) from human insulin, can be con-
verted to human insulin in a transpeptida-
tion reaction, in which the alanine is re-
placed with a threonine. The biosynthesis
of insulin and its conversion from the por-
cine to the human version are shown in
an animation on the supplementary CD-
ROM.
The developments in molecular biology
and biotechnology opened up for new pos-
sibilities, among these the biosynthesis of
human insulin. Insulin is composed of
two disulfide-linked peptide chains re-
ferred to as the A chain and B chain, and
the first recombinant approach used E. coli
Executive Summary XCIII
as host for the expression as fusion pro-
teins. In a later approach in E. coli proin-
sulin (B chainconnecting peptideA
chain) was expressed also as a fusion pro-
tein. In both of these systems the fusion
proteins were isolated as inclusion bodies
and several chemical steps were needed
for dissolution, cleavage, folding and for-
mation of disulfide bridges. Later a single-
chain insulin precursors with a mini C-
peptide could successfully be produced
(also containing the correct disulfide
bridges) and secreted in the yeast S. cerevi-
siae. Eventually, other mini C-peptide insu-
lin precursors of human insulin, with
minimal postfermentation chemistry, and
purification could be achieved with an S.
cerevisiae expression system. The demand
for a more optimal treatment of the pa-
tient has called for the design and develop-
ment of new fast- and slow-acting insulin
analogs, and has required alterations of
the yeast process. As nicely shown on the
supplementary CD-ROM, a number of
fast- and slow-acting derivatives of insulin
have been developed over the years. One
such example is Levemir
from Novo Nor-

disk, an unusual long-acting insulin prod-
uct that has just recently gained marketing
approval and will also be discussed. The
major structural alteration characteristic of
this insulin analog is the attachment of a
C
14
fatty acid via the side-chain of lysine
residue 29 of the insulin B chain. This
promotes binding of the insulin analog to
albumin, both at the site of injection and
in the plasma, which in turn leads to a
constant and prolonged release of free in-
sulin into the blood, giving it a duration of
action of up to 24 h.
After learning about optimizing an es-
tablished expression host like S. cerevisiae,
we will now enter a completely different
route. Luke Anthony Miles from St. Vin-
cents Institute of Medical Research, Vic-
toria, Australia describes his experience
with cell-free protein synthesis systems
from E. coli and wheat embryo. The wide-
spread use of cell-free systems in biomedi-
cal research laboratories reflects their use-
fulness in producing functional proteins.
However, cell-free methods have typically
yielded only nanogram to microgram
quantities of proteins, which has limited
their utility to functional studies. Cell-free
systems derived from many cell types have
been described in the scientific literature.
For a small number of these cell types,
significant advances made in recent years
have seen the development of robust, cost-
effective and highly efficient cell-free ex-
pression systems suitable for the prepara-
tion of proteins in milligram quantities. In
his chapter, Luke describes the advantages
of cell-free protein synthesis methods, with
particular emphasis on applications to
structural biology. E. coli and wheat em-
bryo systems are the best-characterized
prokaryotic and eukaryotic high-efficiency
systems, respectively, and are therefore the
focus. The chapter is devoted to a discus-
sion of recent advances in cell-free synthe-
sis methods that have facilitated the pro-
duction of proteins in high yield that are
soluble, intact and functional. Recent ad-
vances in cell-free expression systems that
are amenable to automation and high-
throughput screening and therefore well
suited for accelerating the development of
biopharmaceuticals are also discussed.
When Success Raises its Ugly Head Out-
sourcing to Uncork the Capacity Bottleneck
Now we have gained insight into a variety
of different expression systems, their ad-
vantages and limitations, and why certain
companies use certain production systems
for their specific manufacturing require-
ments. However, not all pharmaceutical
Executive Summary XCIV
companies manufacture their biopharma-
ceuticals themselves, so it is compulsory to
conclude with a common trend in the in-
dustry: outsourcing. While the commerciali-
zation of biopharmaceuticals is gathering
momentum, the sector is facing a world-
wide shortfall of available biomanufactur-
ing capacity (although one solution to es-
cape from capacity crunches could be al-
ternative expression systems as discussed
earlier) that is becoming a critical strategic
limitation, especially for companies with-
out established market access. However,
there are also several other reasons to out-
source the production process to contract
manufacturers and these will be thor-
oughly discussed by J. Carsten Hempel.
Carsten and myself have known each other
for several years now, when we both were
at the Society for Biotechnology Research
(GBF), where I did my first Diploma the-
sis in Biotechnology. Already during that
time, and since then, Carsten worked on
the development of manufacturing pro-
cesses and outsourcing strategies for
biopharmaceuticals. He is now head of
Biotechnology at Chemgineering, and ex-
plains why expertise and knowledge along
the whole value chain from high-through-
put screening, lead identification and pro-
cess development to clinical development,
manufacturing and marketing are essen-
tial for the successful development of a
biopharmaceutical, but also why this can-
not be available within every pharmaceuti-
cal company. Potential alternative to in-
house expertise are discussed, e.g., part-
nerships between smaller companies spe-
cialized in certain fields such as contract
research organizations (CRO), drug devel-
opment or contract manufacturing.
Outsourcing obviously reflects a wide
range of benefits such as avoidance of cap-
ital expenditure, increased cash flow, mini-
mized fixed assets by avoiding investments
into manufacturing facilities, cost contain-
ment and flexibility, but also reduction of
time-to-market and acceleration of the en-
tire development process by creating ac-
cess to additional expertise. Focusing on
outsourcing of manufacturing processes,
the need for production is the major driver
for such a decision. Key aspects of the out-
sourcing strategy (including calculations of
development cost and burn rates versus
revenue potential) and management ef-
forts of certain outsourcing steps and the
respective processes will be discussed in-
depth as well as how to select the right
contract manufacturing organization
(CMO) out of a continuously growing list
of major global players.
Part V: Biopharmaceuticals used
for Diagnostics and Imaging
From Hunter to Craftsman Engineering
Antibodies with Natures Universal Toolbox
We have discussed the use of biopharma-
ceuticals, their different mechanisms and
modes of action in general, some specific
applications in detail, and different tech-
nologies relating to how they are manufac-
tured. The next section focuses on the
mainstay products of the biotechnology in-
dustry: antibodies their development and
different areas of application. Only about
30 years ago Khler and Milstein set the
stage for one of the key technologies revo-
lutionizing human life: the invention of
mAbs provided the basis for new tools in
biochemical research and applications
spread throughout all medical areas. To-
gether with Niels K. Jerne, George D. Kh-
ler and Csar Milstein were later awarded
the 1984 Nobel Prize in Medicine For
theories concerning the specificity in de-
velopment and control of the immune sys-
Executive Summary XCV
tem and the discovery of the principle for
production of mAbs. The 1980s were char-
acterized by a hype that antibodies as ma-
gic bullets would provide a major break-
through in oncology. Drug targeting, immu-
notoxins and radioimmunotherapy were
key words that lead to the foundation of re-
search programs, expert groups and even-
tually also companies. Since then various
drawbacks have hit the sector and antibod-
ies have survived in niches, as research
agents, diagnostic tools and for very specific
indications. Now with the maturation of the
whole biotech industry, mAbs show a signif-
icant renaissance and appear stronger than
ever. mAbs now represent the fastest-grow-
ing pharmaceutical market segment, with
a potential to reach worldwide sales of
US$ 20 billion by 2010. Since 1982 some
20 mAb-based biopharmaceuticals have
been approved for the treatment of chronic
and life-threatening disease, and there are
hundreds of second-generation products
under clinical investigation. As discussed
previously, this development is further ac-
celerated by validated disease targets that
are becoming accessible through the hu-
man genome sequence and many innova-
tive research avenues. The most prominent
feature of all antibodies, which makes them
the most important class of biopharmaceu-
ticals, is the molecular recognition of their
specific target: binding of antigen and anti-
body (epitopeparatope interaction) is a re-
sult of multiple noncovalent interactions.
Their sum leads to a considerable binding
energy, so that antigen recognition shows
a high level of specificity with thermody-
namic affinity constants even in the nano-
molar range!
My friend Uwe Gottschalk, who was
working for Bayer for several years before
he joint Sartorius, together with Kirsten
Mundt from Amgen, have prepared an ex-
cellent overview on 30 years of mAbs. It
was a must to get a contribution from Am-
gen, since this is THE biotech company
(besides Genentech). Amgen is the most
profitable biotechnology company, and ery-
thropoietin (EPO
) permitted it to become
one of the second-level pharmaceutical
houses and the first in terms of sales of
biotechnology companies. Thus, in 2003,
the sales figures for the Amgen-derived
EPO-based drugs alone amounted to over
US$ 9 billion when the various trade
names were combined. Similarly, Amgens
CSF (colony stimulating factor)-based
drugs amounted to over US$ 2.5 billion
for the same time period.
Uwe and Kirsten start with the initial ex-
periments, process optimization like hu-
manization with transgenic mice that
have functionally replaced the mouse anti-
body genes with the human equivalents
and discuss different antibody formats and
their application. The 30-year history of
mAbs is described as a roller coaster ride
to success: from hype to depression and
back to hype is probably the shortest sum-
mary of what happened. After the market-
ing authorization of the first therapeutic
antibody OKT3 (Ortho Biotech) in 1986 for
treatment of acute transplant rejection,
premature hopes and unrealistic expecta-
tions were raised, and mAbs had a diffi-
cult time to live up to their original prom-
ise. For the following period, antibodies
were of only limited use (e.g., reagents for
analytical tests) and it took until 1995
when Centocors ReoPro (Abciximab), a
chimeric mAb fragment, gained approval
for the prevention of thrombotic side-ef-
fects in patients undergoing coronary ar-
tery angioplasty. Excitingly, both authors
share with us their own experience with
the development of marketed antibodies,
e.g., Enbrel
(Etanercept, Amgen), an
antibody-based biopharmaceutical that tar-
gets and inhibits tumor necrosis factor
Executive Summary XCVI
(TNF)-a a highly effective approach to
the treatment of chronic inflammatory ill-
nesses. Originally approved for rheuma-
toid arthritis (approximately 5 million peo-
ple globally suffer from rheumatoid arthri-
tis), Enbrel has now gained approval for
additional indications, including psoriatic
arthritis, juvenile rheumatoid arthritis and
ankylosing spondylitis.
As we have learned, antibodies are now
the mainstay of biopharmaceuticals and by
the end of 2003, 17 marketed therapeutic
antibodies generated over US$ 5 billion in
combined annual sales, with market
growth at 30%. Ten years earlier, this class
of biopharmaceutical drugs was almost
written off, based on disappointments ex-
perienced with the first generation of mur-
ine mAbs. Two experts who contributed
most to the revival and recent success
story of antibodies are Andreas Plckthun
from the University of Zrich, Switzerland
and Simon Moroney, CEO of THE anti-
body company MorphoSys. Before becom-
ing Full Professor at the University of Zr-
ich, Andreas Plckthun worked at Harvard
University and the Max-Planck-Institute,
where I met him. He was finalist in the
World Technology Awards and recipient of
the Swiss Technology Award 2005. Profes-
sor Plckthun is co-founder of MorphoSys
together with Simon Moroney, who had
previously enjoyed a distinguished career
at the Universities of Cambridge and Ox-
ford, ETH Zrich, and Harvard Medical
School. In 2002, Dr. Moroney was awarded
the German Cross of the Order of Merit
for his services to the biotechnology indus-
try. Their excellent chapter looks at how
new technologies have provided solutions
to problems that hampered early efforts to
develop effective antibody therapeutics and
transformed the market for antibody
drugs. As they have extensively published
in Nature, Nature Biotechnology, Science and
Proceedings of the National Academy of
Sciences of the USA, this includes the gen-
eration of fully human antibodies, affinity
maturation and the selection of antibodies
to bind to particular epitopes on disease-
relevant targets. The article also highlights
what distinguishes a therapeutic from a
simple binding molecule: different modes
of actions of antibodies in different molec-
ular and cellular settings are compared. Fi-
nally, some of the available formats of the
antibody and their effect on molecular and
pharmacological properties are discussed.
Starting with antibodies generated in an-
imals, historically, the first generation of hy-
brid antibodies (part mouse/part human)
comprised the entire murine variable do-
mains of the original mAb, with the re-
mainder of the IgG (constant C
L
domain,
usually a, plus C
H
1, hinge, C
H
2 and C
H
3
domains) coming from a human antibody.
Thus, in these so-called chimeric antibodies,
four out of 12 domains in the IgG remain of
murine origin (two V
L
and two V
H
). In total,
approximately two-thirds of the sequence is
of human origin, the remaining one-third
being murine. The next improvement was
humanization: the grafting of the comple-
mentary determining regions (CDRs) of a
mouse antibody onto a human framework.
For this purpose, a human framework is
chosen from the database of human genes
for V
H
, another for V
L
(j or k), and the mur-
ine and human sequences are aligned. De-
spite the outstanding achievements made
with the techniques of chimerization and hu-
manization, a means of routinely accessing
fully human antibodies has always been a
goal for developers of therapeutic antibod-
ies. Historically, the first method for mak-
ing human IgGs was the immortalization
of human B cells with Epstein-Barr virus.
Since this method is rather inefficient, it
has not been widely used. Recently, how-
ever, a new method was introduced that dra-
Executive Summary XCVII
matically increased the efficiency of trans-
formation. This offers the opportunity to
immortalize memory B cells from patients
after an infection (and potentially even from
cancer patients), and thus complements
cloning of such antibodies from patients
and recovery of antibodies by display tech-
nologies. Today, the most widely used tech-
nologies for making fully human antibodies
are either library-based methods or trans-
genic mouse approaches. The fact that over
30 antibodies based on these technologies
are currently in clinical trials indicates
how well established they have become.
The technology platform enabling this suc-
cess is MorphSyss fully synthetic human
combinatorial antibody library (HuCAL
)
based on modular consensus frameworks
and CDRs randomized with trinucleotides.
The HuCAL GOLD
library also incor-

porates unique restriction sites bracketing
the CDRs, a feature made possible by the
use of chemical synthesis for construction
of the encoding genes. The ends of the
CDR cassettes match the restriction sites
bracketing their positions in the HuCAL li-
brary this results in a fully modular sys-
tem. The benefit of such a system is that
antibody optimization can be rapidly and
systematically carried out by replacing
CDRs in turn to create new sublibraries
based on one or more hits from a first
screening. Multiple examples have shown
that this is a reliable means of generating
antibodies with predefined properties,
while still retaining 100% humanness in
the sequence. Such systematic optimiza-
tion of antibodies builds on their inherent
affinity and specificity to create substances
with drug-like characteristics, including
predefined crossreactivity patterns as well
as the ability to activate, deactivate and/or
block certain biological processes.
However, antibodies are not exclusively
helpful and protect mankind from diseases,
they can also cause diseases: in the next
contribution, Constanze Breithaupt from
Robert Hubers group at the Max-Planck-In-
stitute for Biochemistry, describes the role
of autoantibodies in autoimmune diseases.
More than 80 clinically distinct autoim-
mune diseases have been identified includ-
ing systemic disorders like rheumatoid ar-
thritis and systemic lupus erythematosus,
or organ-specific disorders, like multiple
sclerosis, Type I diabetes mellitus and auto-
immune thyroid diseases. Although many
autoimmune diseases are individually rare,
they collectively affect an estimated 58%
of the US population (and presumably a
similar percentage of the population else-
where in the industrialized world). Affect-
ing women disproportionally, autoimmune
diseases are among the 10 leading causes
of death for young and middle-aged wom-
en. Constanze presents the molecular as-
pects of the recently determined structure
of the multiple sclerosis autoantigen myelin
oligodendrocyte glycoprotein (MOG) com-
plexed with the Fab of the pathogenic auto-
antibody 8-18C5, and its implications for di-
agnosis and understanding of autoimmu-
nity. Since the production of antibodies
directed at self-proteins is a hallmark of
many autoimmune diseases, the detection
of specific autoantibody responses is used
increasingly to aid diagnosis of various
autoimmune diseases. Molecular character-
ization of autoantibodyantigen interaction
sites may help to identify subsets of patients
with certain clinical features or prognostic
outcomes and serve as kind of personal-
ized medicine. Moreover, it can facilitate
the development of immunoassays that
use recombinant or synthetic antigens as
substrates for autoantibody detection. Many
investigators are now making use of the
growing number of known three-dimen-
sional structures of autoantigens to guide
mapping and mutagenesis studies. How-
Executive Summary XCVIII
ever, detailed information about strictly con-
formational epitopes can only be obtained
by crystallographic studies, which are so
far confined to the structure of IgG4 Fc
complexed with the Fab of an IgM rheuma-
toid factor and the recently determined
structure of the above mentioned multiple
sclerosis autoantigen MOG complexed with
the Fab of the pathogenic autoantibody 8-
18C5. As published in Proceedings of the Na-
tional Academy of Sciences of the USA, the
MOG-(8-18C5) crystal structure identified
a highly discontinuous epitope centered
about MOG residues 101108. These resi-
dues encompass a strained tight turn that
is kept in its conformation by the protein
environment explaining the failure to de-
tect this antigenic region by conventional
peptide mapping. This phenomenon, the
three-dimensional structure of MOG-(8-
18C5) and complex formation are shown
as fascinating video animations on the sup-
plementary CD-ROM.
Despite in the case of autoimmune dis-
eases, antibodies can be considered as a
helpful toolbox for medicine and their use
as biopharmaceuticals to treat some of the
most serious diseases affecting mankind to-
day. And this arises directly from impres-
sive technological developments that have
been made in this field over the last 20
years. I think that the improvements in
antibody technology have been one of the
most significant achievements in the field
of modern biotechnology and the develop-
ments described in the previous chapters
promise that this class of modern biophar-
maceutical will play an even more impor-
tant role in the clinicians armamentarium
for the foreseeable future. Molecular engi-
neering holds the promise that the remain-
ing problems, many of them due to incom-
plete molecular understanding of the most
important diseases, will be addressable in
the future. Eventually, this will further lead
to the ultimate biopharmaceutical, which,
for example, targets the desired payload
to the tumor or enables a precise image of
its shape and dimension, thus fully realiz-
ing Paul Ehrlichs vision of the proverbial
magic bullet.
Find, Fight and Follow Target-specific
Troika from Mother Natures Pharmacopoeia
mAbs have been rapidly introduced into a
number of applications within and outside
the medical field. Analytical in vitro meth-
ods such as enzyme-linked immunosorbent
assay (ELISA), radioimmunoassay (RIA),
various blotting techniques, flow cytometry,
immunofluorescence, confocal imaging and
immunohistochemistry are dependent on
the use of antibodies. However, most impor-
tantly, they have become the ultimate vehi-
cles for enabling target-specific in vivo imag-
ing techniques. Although full-sized IgG
molecules are the naturally occurring for-
mat of antigen-binding molecules, mostly
fragments and/or derivatives are developed
and employed for the various diagnostic
and target-specific therapeutic applications,
including Fab fragments, single-chain anti-
body fragments (scFv) and even single-do-
main antibodies. It has been shown that
the much smaller protein size allows faster
tissue penetration, which is obviously a
great advantage for targeting solid tumors
and for local applications as well. The Fv
fragment represents the smallest antibody
domain that retains an acceptable affinity,
but is usually very unstable and aggregates
as a result of only weak, noncovalent forces.
Mostly, it can be stabilized with a linker
polypeptide yielding the scFv and allows ex-
pression from a single gene in a number of
hosts. Interestingly, this approach also facil-
itates additional genetic engineering, such
as fusion to effector domains (e.g., TNF-a)
leading to a biopharmaceutical with intrin-
Executive Summary XCIX
sic target-specific cytotoxicity (e.g., against
cancer cells). Through variation in the linker
size, multimeric variants of the scFv such as
diabodies and tribodies, but also other linear
antibody fragments, can be generated that
possess a higher avidity than monovalent
versions. A variation thereof is the genera-
tion of bispecific antibodies. In certain ap-
plications they are designed to bind to a can-
cer-specific surface molecule with one bind-
ing domain, while the other binding do-
main recruits cytotoxic T cells to the
pathogenic cell, inducing T-cell-dependent
cytotoxicity and apoptosis.
Before we discuss the target-specific ther-
apeutic approaches in detail, we will first fo-
cus on methods of how to make diseases
visible. Classically, medical imaging pro-
vides structural information of a patients
body and is used mainly for diagnostic pur-
poses. Joke Orsel, who received her PhD
from the Biozentrum of the University of
Basel, Switzerland and was a postdoctorate
fellow at Stanford University School of
Medicine, is now testing and developing
new materials for molecular imaging at Phi-
lips. Together with her colleague Tobias
Schaeffter, a Lecturer on Biomedical Imag-
ing at the Technical University Hamburg-
Harburg, they present recent advances in
the development of contrast agents that
can highlight molecules or molecular struc-
tures, and allow researchers and physicians
to obtain ever more detailed information on
diseases. This field of molecular imaging
promises to dramatically improve the future
of healthcare, shifting the emphasis toward
much earlier diagnosis and treatment. The
combination of molecular imaging with
the advent of devices for the imaging of
small animals renders it increasingly inter-
esting for use in drug discovery and drug
development. Their chapter gives an intro-
duction to the different imaging techniques
available, provides examples of contrast
agents and applications for molecular imag-
ing, and then focuses on the potential and
implications of the technique for drug dis-
covery and development of modern biophar-
maceuticals.
The next contribution from Professor Eric
Aboagye describes current trends and op-
portunities of positron emission tomog-
raphy (PET). Eric worked at Johns Hopkins
University before he became Head of Mo-
lecular Therapy at Imperial College London
and now shares his first-hand experience
with us on molecular imaging with PET,
which is evolving as a unique noninvasive
method for studying tumor and normal tis-
sue biochemistry, physiology and pharma-
cology. In oncology, a range of drugs can
be radiolabeled for pharmacokinetic studies
including microdosing of human subjects
prior to phase I trials. Gene delivery can be
assessed by incorporating a reporter gene
that is detectable by PET within the vector
of interest. As Eric published earlier in
The Lancet, he also reviews progress made
in the PET imaging field in the design of
pharmacodynamic markers including as-
says of cell surface receptor status, angio-
genesis, apoptosis, proliferation, glucose
metabolism and hypoxia. Such probes are
potentially useful in patient management
and for drug development. PET is particu-
larly attractive for the development of can-
cer-targeted therapies where assessment of
plasma drug levels or assays of the target pro-
tein in peripheral blood cells are less specif-
ic than direct assessment of the tumor or tu-
mor material. The recent development of
dedicated small animal scanners has helped
bridge in vitro science with in vivo clinical
studies to efficiently develop modern bio-
pharmaceuticals, and this is also the topic
of the next chapter.
It is pleasure for me to introduce the next
author who presents some of his impressive
results on ligand-based specific targeting
Executive Summary C
approaches: Dario Neri. Dario is Professor
at the famous Swiss Federal Institute of
Technology (ETH Zrich) and also founder
and CSO of Philogen, Milano, Italy. We
have known each other for a couple of years
and I am happy to say that we do not only
share scientific interests, but also some
friends in the biotech community. Together
with his colleague from ETH Zrich, Mi-
chela Silacci, Dario describes the targeted
delivery of molecules to sites of disease in
vivo, and shares with us the promises to
open new avenues for the imaging of
pathologies and for the development of
more selective therapeutic agents.
Chemotherapy (i.e., the administration of
chemical compounds in order to confer a
therapeutic benefit to the patient) is often
limited by the doses of drug which can be
reached, without observing limiting toxici-
ties. For example, in oncology, many thera-
peutic strategies rely on the expectation that
anticancer drugs will preferentially kill rap-
idly dividing tumor cells, rather than nor-
mal cells. Since a large proportion of the tu-
mor cells has to be killed in order to obtain
and maintain a complete remission, large
doses of drugs are typically used, with sig-
nificant toxicity towards proliferating non-
malignant cells. In principle, several strate-
gies could be considered in order to develop
better, more selective therapeutic agents. In
many cases, research is driven by the hope
of identifying macromolecular targets
which are not essential in normal physiol-
ogy, but whose inhibition may revert the
pathological condition that one intends to
fight. While such prerequisites may be
met in certain therapeutic areas (e.g., the
use of antibiotics inhibiting microbial pro-
tein targets which do not have a counterpart
in the host), the discovery of selective tar-
gets remains a formidable challenge for
many relevant pathologies. The selective de-
livery of bioactive compounds to a site of
disease (the earlier mentioned magic bul-
lets first envisioned by Paul Ehrlich) ap-
pears to be a general strategy for the devel-
opment of better, more selective therapeutic
agents. In most cases, the selective accumu-
lation of drugs at the site of disease will
spare normal tissues and will hence in-
crease the therapeutic index of the drug
(i.e., the relative activity towards the dis-
eased tissue, compared to normal organs).
Dario reviews progress made (which he
published recently in Nature Biotechnology)
in the identification of pathology-associated
antigens and in the development of binding
molecules (antibodies, peptides and small
organic molecules), and furthermore pre-
sents his view on molecular strategies for
the conversion of binding molecules into
novel imaging or therapeutic biopharma-
ceutical agents.
Another very good friend of mine, An-
dreas Briel from Schering AG, presents, to-
gether with his colleagues Michael Rein-
hardt, Matias Murer and Peter Hauff, the
principle of Find, Fight and Follow an in-
novative theranostic approach just pub-
lished partly in Journal of Neurology and
Radiology. Andreas studied chemistry with
the main focus on physical-chemistry of
polymers and completed his PhD at the
Max-Planck-Institute of Colloids and Inter-
faces. He is an expert in nanotechnology,
target-specific in vivo diagnostics and drug
delivery systems, and acts as chairman of
the Association Colloids and Interfaces
Berlin/Brandenburg, and also Lecturer for
Novel Technologies and Innovation at
the University of Applied Sciences in Berlin.
He is also involved in several initiatives to
consult the German government (e.g.,
BMBF-NanoForLive) and the European
Commission (e.g., European Nanotechnol-
ogy Platform on Nanomedicine) to define
a common future vision on the field of na-
no(bio)medicine. One major challenge fac-
Executive Summary CI
ing the pharmaceutical industry today is to
develop contrast-enhancing agents for mo-
lecular imaging. Classic contrast agents pri-
marily document the anatomy. They are
only suitable to a limited degree for patho-
physiological examination using differential
diagnostic techniques, i.e., characterizing
the development of a disease. As previously
seen, molecular imaging selectively tracks
down molecules and cell structures to be
able to establish proof of diseases at a very
early stage and then to make decisions
on a highly individual treatment. The next
straightforward vision of medical imaging
quite clearly lies in the aforementioned con-
cept of Find, Fight and Follow. In radio-
pharmaceuticals we are already pursuing
the approach of a troika consisting of early
diagnosis, therapy and monitoring the ther-
apy success.
For ultrasound imaging, in general, there
are no limits with regard to examination
time and frequency, and even investigations
of embryos in utero are now clinical routine
an impressive video of three-dimensional
scanning of an embryo is shown on the sup-
plementary CD-ROM. Utilizing the nano-
technological concepts of colloid and inter-
face science, imaging even on a molecular
level can be achieved via diagnostic ultra-
sound using tiny gas-filled polymer particles
coupled to target-specific ligands. Addition-
ally, nanosized polymeric drug carriers for
targeting and controlled release have been
extensively studied in the past. Here, a na-
noparticle or capsule acts like a container
for a pharmacologically active agent. Passive
and active targeting can be attained by care-
fully chosen size and surface modification
of the carrier. As Andreas and colleagues de-
scribe, drug release can be controlled via de-
sorption of surface-bound drugs, diffusion
through the particle matrix or the capsule
wall, or matrix erosion. Moreover, smart
release can be achieved by using smart poly-
mers (pH or temperature sensitive) or, more
interestingly, by applying an external stress
to the drug carrier. If the drug carrier is ap-
propriately designed, release can be induced
by diagnostic ultrasound.
Building this bridge between therapy
and diagnosis opens the field of the afore-
mentioned theranostics. With the Find,
Fight and Follow strategy, the tissue of in-
terest can first be imaged via target-specific
ultrasound contrast particles. In a second
step, the same particles, now filled with a
pharmacologically active agent, can be se-
lectively guided to the diseased tissue.
Here, the bubble is bursted by applying ul-
trasound, thereby releasing the biopharma-
ceuticals used for a specific therapy. Finally,
monitoring of treatment effects is possible
by sequential imaging.
This approach demonstrates the success
of a resolute implementation of nanobio-
technological concepts in medical applica-
tions. Polymer nanoparticle and microcap-
sule formation, the control of colloidal
structure, surface modification, antibody
coupling strategies, and the resulting in vi-
tro properties eventually lead to an ultra-
sound theranostic. The chapter presents
in vivo results with special emphasis on
antibody-based targeting and gene delivery.
Investigations with different drugs and
targeting sites demonstrate that this
approach can already serve as a biophar-
maceutical platform technology.
Gaining Insight Sense the Urgency for
Early Diagnostics
Obviously, and as mentioned in previous
chapters, the earliest possible diagnosis of
a disease is imperative for efficient therapy.
This is true for virtually any type of disease,
including infectious diseases, Parkinsons
disease as well as other dementias and is
best documented for tumor diseases, where
Executive Summary CII
the current compilation of internationally
acknowledged markers in human blood
contains not more than a handful of pro-
teins: prostate-specific antigen (PSA), pros-
tatic acid phosphatase (PAP), carcinoem-
bryonic antigen (CEA), a-fetoprotein
(AFP), human chorionic gonadotropin
(HCG), lactate dehydrogenase (LDH), neu-
ron-specific enolase (NSE), and CA 19-9,
CA 15-3, CA 27-29 and CA 125. Most of
them were first described a long time ago
(e.g., HCG already in 1927) and are mainly
used for prognostic purposes. Application
in tumor diagnosis is very limited, espe-
cially for the early cancer stages where the
success rates are disappointingly low. De-
spite huge research efforts, only a few tar-
geted therapeutics, such as the earlier men-
tioned Herceptin (trastuzumab) directed
against the HER-2 receptor of breast cancer
cells, are in use and, in general, nonspecific
cytotoxic agents are administered with lim-
ited success.
It is very obvious that only an earlier diag-
nosis could improve the situation and, re-
gardless of the type of disease, there is this
urgency for early diagnosis because that ear-
lier any disease is diagnosed, the more effi-
cient a therapy can be. This simple fact is
true for virtually any type of disease, but
best documented for tumor diseases.
Over the last few years, comparative pro-
tein profiling by SELDI TOF-MS (surface
enhanced laser desorption/ionization time-
of-flight mass spectrometry) has become
acknowledged as a promising way for the
large-scale detection of specific and predic-
tive protein patterns reflecting certain
stages of cancer, neurological disorders
and infectious diseases. One such example
is the ProteinChip
system from Cipher-

genBiosystems. This system has proven to
be a valuable tool for the discovery and val-
idation of newly detected protein biomark-
er patterns. According to Andreas Wiesner,
CSO from Ciphergen and Assistant Pro-
fessor at the Free University, it is used by
more than 1000 customers worldwide, in-
cluding academic and governmental insti-
tutes as well as pharmaceutical companies
(e.g., Abbott, Bayer, BASF, Eli Lilly, GSK).
The SELDI process with its unique combi-
nation of chromatographic principles and
mass spectrometric detection meets the
challenges of the new proteomic era by en-
abling comparative protein profiling with
Expression Difference Mapping
TM
analysis
of several hundreds of samples per day on
a single technology platform, with soft-
ware support for the construction of multi-
marker predictive models. The Interaction
Discovery Mapping platform is introduced
as the next methodical step for antibody-
based assays and investigations into pro-
tein binding partners of importance in di-
agnosis and development. It should be
mentioned that all assays developed thus
far are still of research grade and not yet
released for commercial use, but current,
multi-institutional studies are ongoing to
evaluate the clinical robustness of this
SELDI-based multimarker detection with
the necessary scientific diligence before
making it available for the public. These
studies will help establish diagnostic as-
says, which in turn should foster the de-
velopment of modern biopharmaceuticals.
The next chapter also deals with the
early detection of diseases; however, the
difference is that it is not specimens like
blood, tissue, urine or feces that are used,
but human breath. This sounds really in-
triguing early detection of lung cancer
by metabolic profiling of human breath
with ion mobility spectrometers. Jrg Ingo
Baumbach from the Institute of Analytical
Sciences and Lutz Freitag from Lung Hos-
pital Hemer, Dortmund, Germany answer
the question whether this is a dream or
reality. Dr. Baumbach, who received his
Executive Summary CIII
PhD in Physics, was working at the Acade-
my of Sciences for several years before he
was appointed Director of the Metabolo-
mics Department. He explains how volatile
metabolites, which occur in human ex-
haled air, can be directly correlated to dif-
ferent kinds of diseases. Some metabolites
are biomarkers: acetone is related to dia-
betes, nitric acid to asthma and ammonia
to hepatitis, others arise from bacteria. In
their chapter, an ion mobility spectrometer
(IMS) coupled to a multicapillary-column
(MCC) as a preseparation unit is used to
identify and quantify volatile metabolites
occurring in human breath down to the
nanogram or picogram per liter range of
analytes. The spectra obtained from pa-
tients suffering from chronic obstructive
pulmonary disease (COPD) and pneumo-
nia are discussed in detail. Furthermore,
IMS chromatograms of metabolites of Ser-
ratia marcescens, Enterobacter aerogenes and
E. coli are compared. In addition, the effect
of drug delivery on a patient showing angi-
na lateralis is presented as an example to
show the potential of a method developed
in the field of detection of pathways, effec-
tive dosage and decision of effective time
intervals to deliver biopharmaceuticals.
The delivery of biopharmaceuticals will
also be the topic of the next section,
which will present different drug delivery
approaches, achievements as well as, lim-
itations and future trends.
Part VI: Advanced Application Routes for
Biopharmaceuticals
Getting Inside Quest for the Best and
How to Improve Delivery
The successful administration of biomole-
cules to the body has proven to be one of
the most challenging aspects in the develop-
ment of biopharmaceuticals, although the
functions of the different absorption bar-
riers in the body are well known. This is de-
picted by the biopharmaceuticals in the
market, which are mainly applied by paren-
teral administration, e.g., subcutaneous or
intravenous injection. Currently, depot sys-
tems for the delivery of peptide hormone
analogs, peptide or antibody solutions and
vaccines are the only marketed formula-
tions. However, the increasing number of
new biological entities (NBEs) which are
currently entering the development phase
is going to increase the need for advanced
drug delivery systems allowing biophar-
maceuticals to develop their whole potential
for the treatment of various diseases. Drug
delivery systems, ranging from modified re-
lease tablets to sustained release implants
and transdermal patches, are widely used
for small molecules. Such systems may be
a starting point for advanced drug delivery
systems modified with regard to the rather
multilayered requirements of modern bio-
pharmaceuticals. The next chapter therefore
deals with advanced drug delivery systems
and is provided by my colleagues Gesine
Hildebrand and Stephan Harnisch from
Schering AG. Gesine did her PhD on liquid
crystals for transdermal application and was
Head of Drug Delivery Systems at Schering
AG, and Stephan has worked with her for
several years. They give an overview of the
diverse drug delivery strategies pursued to
overcome or circumvent absorption bar-
riers, and how the delivery systems are
going to affect the pharmacokinetics, stabil-
ity, degradation, efficacy and efficiency of
the respective incorporated biopharmaceuti-
cal. In general, the ideal delivery system for
all routes of administration should release
its contents only at a desired region of ab-
sorption, where the drug delivery system at-
taches by specific interaction with surface
determinants specific for that region. The
Executive Summary CIV
delivery system should travel independently
of the transitory constraints that are typical
for the route of administration. Such a de-
livery system is unavailable so far, but it
would benefit biopharmaceuticals as well
as other poorly absorbed drugs. The use of
novel injection devices in parenteral admin-
istration of proteins has nowadays become
more common for daily protein administra-
tion (e.g., for the administration of insulin
and recombinant human growth hormone).
Over the next few years it will continue to be
the quickest and least expensive new deliv-
ery system to commercialize. Depot delivery
systems already provide opportunities to
improve patient compliance through fewer
injections. These systems might also in-
crease the viability of local delivery for bio-
pharmaceuticals that are not well tolerated
systemically or that are degraded too fast
to result in therapeutic drug levels after sys-
temic administration. Of the noninvasive
routes, pulmonary delivery has the greatest
promise because of the higher protein bio-
availability compared to transdermal or oral
delivery. Phase III clinical trials of pulmo-
nary insulin delivery are ongoing and the
results will show the capability of that meth-
od. The next chapter therefore summarizes
the current status and provide a case study
of oral insulin. Later on we will also learn
about the orphan drug status of inhaled
Aviptadil
in pulmonary arterial hyperten-

sion and chronic thromboembolic pulmo-
nary hypertension both being life-threat-
ening conditions.
Pathfinder New Ways for Peptides,
Proteins and Co
As described earlier, the current advocacy
of intensive insulin therapy regimens, in-
volving multiple daily subcutaneous injec-
tions, places a heavy burden of compliance
on patients and has prompted interest in
developing alternative, less invasive routes
of delivery. There have been various efforts
to develop such alternative methods for ad-
ministering insulin. Pankaj Modi, Vice
President of R & D at Generex in Toronto,
presents recent results of clinical studies
with Oralin
, an oral spray insulin. Pan-

kaj, who is editor of Expert Opinion in
Drug Delivery, has developed a special treat-
ment for diabetes using the novel Rapid-
Mist
TM
Diabetes Management System,
which is based on a proprietary formula-
tion technology, allowing a liquid aeroso-
lized pharmaceutical formulation to be de-
livered accurately into the mouth of the
patients via a spray. This system intro-
duces a high-velocity fine-particle aerosol
into the patients mouth, thereby inducing
a markedly increased deposition of the
preparation over the thin mucosa mem-
brane a deposition that is much larger
than that observed with conventional tech-
nology. Thus, the fast-moving, fine-particle
aerosol is able to traverse this thin mem-
brane. Once they have penetrated through
these superficial layers, insulin molecules
rapidly get absorbed into the blood stream,
with the timely adjusted aid of absorption
enhancers, and therefore appear in the pe-
ripheral circulation within 10 min of its
application. Several studies conducted in
subjects with Type 1 and 2 diabetes dem-
onstrated very clearly that Oralin has faster
absorption and metabolic control compar-
able to subcutaneously injected insulin.
This novel, pain-free, oral insulin formula-
tion has a series of positive attributes: rap-
id absorption, simple (user-friendly) ad-
ministration technique, precise dosing
control (comparable to injection within
one unit) and bolus delivery. Altogether
these patient compliance features make it
a unique modern biopharmaceutical.
Another pulmonary insulin is currently
being co-developed by Pfizer, Aventis and
Executive Summary CV
Nektar Therapeutics, Huntsville, AL: Exu-
bera
has completed phase III clinical

trials, although additional safety studies
are currently being undertaken. Formed
through the fusion of the leading compa-
nies Inhale Therapeutic Systems (San Car-
los, CA), Shearwater Corporation (Hunts-
ville, AL) and Bradford Particle Design
(Bradford, UK), Nektar Therapeutics offers
a suite of leading drug delivery technologies
that encompasses molecule engineering,
advanced drug delivery solutions for oral,
injectable and pulmonary administration,
and particle engineering comprising both
pulmonary particle technology and super
critical fluid technology (SCF). One focus
area of Nektar Therapeutics is the advanced
conjugation with poly(ethylene) glycol (PE-
Gylation) of biopharmaceuticals, described
by Michael Bentley, Vice President of Re-
search, and colleagues in their chapter. Mi-
chael Bentley was NIH postdoctoral fellow
at Berkeley University, a Fulbright Senior
Research Scholar, and a Japan Science and
Technology Agency Fellow before he be-
came Professor at the University of Maine.
Dr. Mary Bossard was NIH postdoctoral fel-
low at Berkeley University, before she
moved on to SmithKline and then to Nek-
tar, where she currently serves as Director
of Biopharmaceuticals. Dr. Tacey X. Viegas,
Senior Director of R & D, and Dr. Kevin W.
Burton, Head of Product Development with
focus on PEGylation, complete the team of
experts from Nektar. PEGylation has come
into its own as a powerful approach for en-
hancing the properties of biopharmaceuti-
cals. There are six marketed PEG biophar-
maceutical products utilizing this technol-
ogy and many more currently in clinical
trials. Benefits which can be achieved
through application of PEGylation include
extended circulation lifetime, improved bio-
distribution, decreased immunogenicity, in-
creased solubility, decreased proteolytic deg-
radation and greater stability of the drug
product on storage. In their chapter they
focus on applications which have led to
marketed products, contrast early, first-gen-
eration approaches to PEGylation with cur-
rent second-generation technology, and dis-
cuss improvements in properties of the
products as well as clinical benefits which
result from application of current reagents
and methods. Finally, formulation proper-
ties of PEG drug products compared to
those of the native biopharmaceutical prod-
ucts will be discussed in depth.
Although these are successful examples
of PEGylation, the oral bioavailability of bio-
pharmaceuticals is generally very poor, since
they are poorly absorbed and easily de-
graded by proteolytic enzymes in the gastro-
intestinal tract. Therefore, for systemic de-
livery of peptide and protein drugs, parente-
ral administration is currently required in
order to achieve their therapeutic activities.
An animation comparing parenteral and
oral administration of drugs is provided
on the supplementary CD-ROM. However,
parenteral administration routes are poorly
accepted by patients and may cause an aller-
gic reaction. Thus, alternative routes in ad-
dition to oral administration, i.e., nasal, buc-
cal, rectal, vaginal, conjunctival, and trans-
dermal routes, are being investigated for
peptide and protein delivery. Of all the
routes mentioned, the oral route is the most
common and convenient for the administra-
tion of biopharmaceutical drugs. However,
the intestinal absorption of peptide and pro-
tein biopharmaceuticals is known to be very
poor due to their extensive degradation by
various peptidase and digestive enzymes,
as well as poor membrane permeability
characteristics. Therefore, various strategies
have been examined to improve the intest-
inal absorption of biopharmaceuticals.
Professor Akira Yamamoto from Kyoto
Pharmaceutical University in Japan is an
Executive Summary CVI
expert in this field. He has an impressive
track record on the delivery of peptides
and proteins, was a postdoctorate fellow at
the University of Southern California and
has been the recipient of several awards,
including from the Pharmaceutical Society.
In his chapter, he introduces strategies for
improving the intestinal absorption of pro-
tein biopharmaceutical drugs. He starts
with the effects of various absorption en-
hancers and protease inhibitors on the in-
testinal absorption of peptide and protein
drugs, and then introduces the effect of
chemical modification (acylation) on the
intestinal absorption of peptide and pro-
tein drugs including insulin and tetragas-
trin. Professor Yamamoto also describes
the colon-specific delivery of insulin using
chitosan capsules and demonstrates that
these capsules proved effective for improv-
ing the intestinal absorption of insulin.
In addition to protease inhibitors and
chemical modifications, another aid to
enhance delivery of biopharmaceuticals,
especially for vaccines, are adjuvants. Ed-
ward Jenner (17491823) first showed that
infection with the cowpox virus could pro-
tect against human smallpox more than
200 years ago. Amazingly, at that time
nothing was known about the pathogens
causing diseases and it was only later that
Robert Koch (18431910) discovered that
infections are caused by microorganisms.
These and other discoveries enabled the
production of vaccines, and some diseases
like smallpox have subsequently been era-
dicated. Other infections such as diphthe-
ria or polio, although not yet eradicated,
have become very rare in the Western
world because of early childhood vaccina-
tion programs. Despite the tremendous
success of vaccination programs, many in-
fectious diseases like hepatitis C or tuber-
culosis are still widespread among the hu-
man population. It is estimated that 170
million people worldwide are infected with
hepatitis C (plus a further 2 billion with
hepatitis B) and that one-third of the world
population is carrying pathogens causing
tuberculosis. Despite many efforts, no effi-
cient treatment of these diseases is avail-
able to date and the necessity for the de-
velopment of new vaccines is obvious.
Professor Alexander von Gabain, CEO of
Intercell AG, Vienna, and colleagues Mi-
chael Buschle and Karen Lingnau are pre-
senting their efforts in developing novel ad-
juvants based on cationic drug delivery sys-
tems for a therapeutic vaccine against hepa-
titis C. Karen Lingnau is Head of the
Department of Pharmacology/Toxicology
at Intercell and Michael Buschle, who
worked at the Royal Free Hospital School
of Medicine, London, UK, and at Boehrin-
ger Ingelheim before, is serving as Chief
Technical Officer. Alexander von Gabain is
co-founder of Intercell AG and Professor
of Microbiology at the University of Vienna.
Earlier he was at Stanford University and
Karolinska Institute where he served as Pro-
fessor of medical biotechnology and later as
foreign adjunct professor. The colleagues
from Intercell make use of specific protein
or peptide subunits of the pathogens for
vaccination, which contributes to the design
of effective and safe vaccines, and lowers
the costs of production. However, peptides
are generally not very immunogenic on
their own and require the aforementioned
adjuvants to induce an adequate immune
response. Professor Gabain and colleagues
have therefore developed two novel adju-
vants, IC30 and IC31, that strongly enhance
the immune response. IC30, a cationic poly-
amino acid (poly-l-arginine, pR), was identi-
fied as an adjuvant that transferred in a
highly efficient manner peptides to anti-
gen-presenting cells (APCs) in an investiga-
tion to use tumor antigens as therapeutic
vaccines in mice. This enhanced uptake of
Executive Summary CVII
peptides subsequently led to a strongly im-
proved peptide-specific T cell response and
a reduction in tumor growth. The length
of the pR molecule and the negative charge
of the peptides influence the uptake of the
peptides. A therapeutic vaccine against hep-
atitis C has been developed subsequently
using this IC30, formulated with five syn-
thetic peptides. Results from clinical trials
both in healthy volunteers and chronically
infected patients are discussed in their
chapter. This success with IC30 has
prompted the search for further adjuvants
with even better characteristics. Cationic
antimicrobial peptides (CAMPs) are used
by the immune system as a defense mecha-
nism against infections by microbes.
Hence, they have been used as antibiotic
therapeutics, but it was not known that
CAMPs could also function as adjuvants.
The adjuvant effect was first shown for an
artificial CAMP, KLKL
5
KLK, when co-in-
jected with the ovalbumin-derived peptide
OVA. Furthermore, it was shown that, like
IC30, KLKL
5
KLK enhances the association
of antigen to APCs and induces the forma-
tion of an antigen depot at the site of injec-
tion. The adjuvant properties of KLKL
5
KLK
could be further enhanced by combination
with a novel immunostimulatory deoxynu-
cleotide containing repeats of deoxy-ino-
sine/deoxy-cytosine. This novel adjuvant,
IC31, has the unique capacity of being able
to stimulate Tand/or B lymphocytes in vivo.
Professor Gabain and colleagues discuss
the results of further preclinical models in
which IC30 and IC31 have been tested in
existing and novel vaccines. The latest ex-
periments to elucidate the mechanisms of
actions of these two adjuvants are pre-
sented.
Via Mala The Stony Road of DNA
Delivery: Back-pack, Feed-back, and Pay-back
So far, we have focused more on the deliv-
ery of peptides and proteins, and we will
now switch to the delivery of DNA-based
biopharmaceuticals this is not an easy
task, because the enzymes that degrade
DNA are ubiquitous and are basically lurk-
ing everywhere.
Professor Robert Langer and colleagues
from MIT share their experience in DNA
protection and delivery, and in the design
of materials specially developed for this pur-
pose in biology and medicine. Professor
Langer has written more than 800 articles
(several of them in Nature, Nature Biotech-
nology, Nature Medicine and Science) and is
the only active member of all three US Na-
tional Academies the National Academy of
Sciences, the Institute of Medicine and the
National Academy of Engineering. Bob
has also over 500 patents, which have been
licensed to over 100 pharmaceutical and bio-
technology companies, and has served as
Chairman of the Science Board, the FDAs
highest advisory board. In addition, he
was ranked one of the 18 top people in
science in America and has received several
honorary doctorates, including ones from
ETH Zrich and Technion Israel. His co-
authors also have an impressive track-re-
cord. Professor Chun Wang was an NIH
postdoctoral fellow at MITwith publications
in Nature and Nature Materials, and his col-
league Herman Eisen from the Center for
Cancer Research at MIT, has also published
extensively in Nature, Cell and Proceedings of
the National Academy of Sciences of the USA.
Dr. Jorge Heller was Director of the Con-
trolled Release Department at the Stanford
Research Institute, and is founder and for-
mer Editor-in-Chief of the Journal of Con-
trolled Release as well as Past President of
the Controlled Release Society, with more
Executive Summary CVIII
than 200 papers and 50 patents. In their
excellent and comprehensive contribution
they describe how biodegradable poly(ortho
ester) microspheres (POEs) were specifically
designed to deliver plasmid DNAvaccines to
antigen-presenting cells. POEs degrade by a
hydrolysis reaction to nontoxic products
and, most importantly, do not generate an
acidic environment that could adversely af-
fect plasmid DNA bioactivity. Two types of
POEs were prepared one type, referred
to as POE 1, lacks tertiary amine groups
in the polymer backbone, while the other
type, referred to as POE 2, has tertiary
amines in the polymer backbone. In vitro ex-
periments have shown that at pH 7.4 both
polymers release plasmid DNA at a slow,
steady rate, but when the pH is abruptly
changed to 5.0 (to simulate the environment
within the phagosomes) both polymers rap-
idly released DNA. The rapid release of plas-
mid DNA at pH 5 is due to the known pH-
dependent rate of POE hydrolysis. While
POE 1 rapidly released DNA as soon as
the pH was lowered, POE 2 released plas-
mid DNA only after a 24-h induction period.
Both polymers were found to suppress the
growth of tumor cells bearing a model anti-
gen, but POE 2 was significantly more effec-
tive than POE 1. The greater effectiveness of
POE 2 is due to the delay in plasmid DNA
release that prevents release before the
APCs become activated and reach the lymph
nodes. The delay in plasmid DNA release is
most likely due to an electrostatic interac-
tion between the negative charges on the
plasmid DNA and the positive charges on
POE 2 created when the tertiary amines be-
come protonated at the low pH.
Another approach to in vivo gene DNA
delivery is presented by Mitsuru Hashida,
VP of the Academy of Pharmaceutical
Sciences and Technology, and Dean and
Professor of Graduate School of Pharmaceu-
tical Sciences, Kyoto University, together
with his colleagues Fumiyoshi Yamashita
and Shigeru Kawakami. These three collea-
gues share with us their tremendous knowl-
edge and achievements with cationic lipo-
some-mediated in vivo gene transfer, and
also discuss the many fold goals, including:
(a) controlled transfection efficacy, (b) con-
trolled cell specificity of transfected cells
and (c) controlled duration of transgene ex-
pression after intravenous administration.
After systemic administration of naked plas-
mid DNA or oligonucleotides, these gene
therapies are not effective because of their
susceptibility to degradation by nucleases
and/or low membrane permeability. In the
early 1990s it was demonstrated that sus-
tained and efficient gene transfection could
be achieved following the local administra-
tion of naked pDNA and, so far, several
methods involving the local administration,
i.e., intramuscular and intratumoral, of
naked pDNA have been studied for applica-
tion in gene therapy. Moreover, electropora-
tion, the application of a controlled electric
field to facilitate cell permeabilization, has
been shown to enhance the transfection ac-
tivity of administered pDNA. However, com-
pared with local applications of naked
pDNA, systemic application by, for example,
liposomal in vivo delivery enables the trans-
fection of a large number of cells through-
out the entire tissue. Various transfection
characteristics are important when highly
efficient gene therapy for the treatment of
a variety of refractory diseases is required.
Non-viral vectors should be preferred to cir-
cumvent some of the problems occurring
with viral vectors, such as endogenous virus
recombination, oncogenic effects and unex-
pected side-effects. These nonviral vectors
can be divided into two general groups: cat-
ionic liposomes and polymers. Among var-
ious types of nonviral vectors, cationic lipo-
some-mediated gene transfection is one of
the most promising approaches due to the
Executive Summary CIX
high transfection efficiency this is espe-
cially true in the lung after intravenous ad-
ministration. Moreover, recent advances in
gene delivery technologies now enable us
to deliver the pDNA into liver, heart (as de-
scribed earlier for the adenovirus-based
gene therapy), macrophages and cancer
cells via uptake by cell-specific receptors.
In their review, the colleagues from Kyoto
University focus on the progress of research
of cationic liposome-mediated in vivo gene
transfer.
The next contribution is also on gene
delivery and also comes from Japan: Pro-
fessor Hideyoshi Harashima from Hokkai-
do University, who graduated at Tokyo
University, presents together with his col-
leagues Kentaro Kogure, Hidetaka Akita
and Hiroyuki Kamiya their latest results
on a very innovative delivery system. They
describe a novel nonviral gene delivery sys-
tem: a Multifunctional Envelope-type Nano
Device (MEND). In their chapter they
present the concept of programmed
packaging, which they also recently pub-
lished in the Journal of Gene Medicine and
Molecular Therapy. Dramatic advancements
have occurred and tumor targeting with
long-term circulating liposomes that con-
tain antitumor agents (passive targeting)
has been successfully demonstrated in
clinical trials. Active targeting with ligands
specific to cell surface receptors has also
been developed. Consequently, the next
generation of drug delivery must be pro-
grammed packaging, in which both the
intracellular trafficking and the disposition
of DNA for gene therapy are controlled at
the same time.
However, for efficient gene delivery into
the nucleus of target cells, the nonviral vec-
tors must overcome several barriers, such as
the plasma membrane, the endosomal
membrane and the nuclear membrane.
Thus, to overcome the barriers, the nonviral
gene delivery system must be equipped
with various functional devices such as li-
gands for specific receptors, pH-sensitive
fusogenic peptides for endosomal escape
and a nuclear localization signal (NLS) for
enhanced nuclear delivery. It is impossible
to integrate all these functional devices into
a single system by simple mixing, and to
have each function exerted at the appropri-
ate time and correct place. Therefore, pro-
grammed packaging is a new packaging
concept which consists of three compo-
nents: (1) programming, i.e., a program to
overcome all barriers, (2) design, i.e., the de-
velopment of functional devices and their
three-dimensional assignment, and (3) as-
sembly, i.e., the use of nanotechnology to
assemble all devices into a nano-size struc-
ture. Professor Hideyoshi Harashima and
colleagues have therefore recently proposed
their novel nonviral gene delivery system
MEND to realize this type of programmed
packaging. As they explain, the ideal
MEND consists of a condensed DNA core
and a lipid envelope structure equipped
with the various functional devices. The
compacted core has some advantages, such
as protection of DNA from DNase, size con-
trol and an improvement in packaging effi-
ciency. Furthermore, as published in Drug
Delivery, separate structures (and not a dis-
ordered mixture) of the DNA core and lipid
envelope are necessary to control the topol-
ogy of this functional device.
Getting Beyond
Rocket Science vs. Science Fiction
The aforementioned nanotechnology is
also topic of the next chapter from Profes-
sor Oliver Kayser from the University of
Groningen. I know Oliver from the EAPB
and also because I recently contributed a
book chapter on plant-based expression of
biopharmaceuticals for his textbook Phar-
Executive Summary CX
maceutical Biotechnology. Olivers main re-
search fields are the development of drug
delivery systems, pharmaceutical biotech-
nology and nanotechnology. As he de-
scribes in his chapter, nanotechnology is
the key technology of the 21st century.
The possibility to exploit the structures
and processes of high-molecular-weight
biomolecules like proteins, nucleic acids
and synthetic polymers (silicon, poly-
methylacrylate) for novel functional mate-
rials, biosensors, biomicroelectrochemical
systems and smart drug delivery systems
has created the rapidly growing field of
bionanotechnology. He reviews the current
state-of-the-art and availability of nano-
technologies in medical and pharmaceuti-
cal sciences. Drug and gene delivery, tissue
engineering, biosensors, and safety aspects
as examples of the main fields are dis-
cussed to show the potential applications,
but also the limitations of the discussed
techniques. The main focus, however, is
on the fabrication, miniaturization and
pharmaceutical use of smart drug delivery
systems like microneedles, BioMEMS (mi-
croelectromechanical systems) and bio-
sensing microchips in which release from
a particular reservoir is initiated by apply-
ing an electric potential. Filling of a micro-
CHIP (implant) with a biopharmaceutical
drug under GMP, the disintegration of the
golden foil and the subsequent release of
the pharmaceutical is nicely shown in the
presentation on the supplementary CD-
ROM. Safety aspects and biohazards of na-
nosystems are also discussed in his chap-
ter.
Part VII: From Transcription to Prescription
Dosis Facit Venenum The Therapeutic
Window between Systemic Toxicity and Lack
of Efficacy
As we have seen how to manufacture, for-
mulate and deliver biopharmaceuticals, we
will now discuss analytical characterization
and approval. The analytical characteriza-
tion of a biopharmaceutical is quite com-
plex and a very time-consuming task,
which goes along with the development of
the biopharmaceutical through the differ-
ent stages of maturity. The required analyt-
ics to thoroughly characterize a biophar-
maceutical are described by one of my
colleagues at Schering AG in Berlin. Dr.
Michael Hildebrand is Head of Global
Pharmaceutical Development at Schering
AG, Berlin and Associate Professor for
Pharmaceutical Chemistry at the Friedrich
Schiller University in Jena, Germany. He
is also acknowledged Expert Pharmacolo-
gist by the German Society of Pharmacol-
ogy and Toxicology, and shares with us his
extensive knowledge on established tools
to describe and verify the quality of bio-
pharmaceutical drugs.
Apart from pure quality control topics,
reliable analytical methods are important
to assess the pharmacokinetic profile of
drugs and their metabolites, and often a
correlation of pharmacokinetic and phar-
macodynamic parameters helps to predict
therapeutic effects. In the case of biophar-
maceuticals, a new set of analytical meth-
ods had to be developed due to the differ-
ent nature of these products. Potency as-
says were rarely used in conventional
small-molecule analysis, but have gained
special attention for biopharmaceuticals. A
series of assays was established which are
useful both in quality control and in in
vivo analytics. A lot of attention is paid to
Executive Summary CXI
quality characteristics, especially in the
case of modifications to the manufacturing
process. The broad range of different types
of new products, covering, for example,
antibodies, recombinant proteins, and cell
and gene therapy principles, offers many
new challenges to the analytical scientists.
New test systems often require highly spe-
cific equipment and know-how, and are far
from the routine methods as established
for small molecules. The chapter gives a
comprehensive overview of specific analyti-
cal aspects for biopharmaceuticals and also
refers to quality standards required by reg-
ulatory authorities.
In the next contribution we learn more
about another class of biologicals. Pamela
Williams and her colleagues Dijana Vinko-
vc, Sheena Whyte, Jose Cosme, and Harren
Jhoti, from Astex Technology Ltd, Cam-
bridge, UK present recent advances in the
structure and function elucidation of cyto-
chrome P450 metabolizing enzymes. In
the past decade the use of protein structures
in drug discovery and development has sig-
nificantly advanced the design of drug mol-
ecules. However, achieving potency towards
a target protein is just one step in the pro-
cess of compound optimization that leads
to a marketed drug. Cytochrome P450s play
a major role in phase I metabolism of many
drugs, with a nonideal P450 profile often ne-
cessitating significant redesign of otherwise
promising compounds. Knowledge of how
compounds bind to the cytochromes there-
fore presents the possibility of our being
able to successfully modify a compound, re-
taining potency towards the target protein
whilst adjusting its metabolic profile.
Although the cytochromes themselves are
per se not biopharmaceuticals in the narrow
sense, they are biomolecules that closely in-
teract with pharmaceuticals and hence are an
integral part of drug development. Still, one
could also imagine developing biopharma-
ceuticals out of certain cytochromes. In
their chapter, Pamela and colleagues de-
scribe findings which they have recently
published in Nature and Science, and review
what is known currently about the struc-
tures of the drug metabolizing cytochrome
P450 family and the models of ligand bind-
ing to the respective cytochromes.
As we have learned in previous chapters,
the rate of marketing approvals for novel
drugs has declined dramatically in recent
years and those currently submitted for ap-
proval reflect the prenomination selection
criteria which prevailed in the 1990s. The
paradigm in that period was heavily skewed
towards strong binding affinities in recep-
tor-based assays as a primary selection crite-
rion. The development of more sophisti-
cated libraries and improvements in the
capacity of high-throughput screens have
generated many new compounds for
consideration. The demand is now for tools
to enrich very early in preclinical develop-
ment for those compounds whose profiles
include desirable solubility, bioavailability
and efficacy traits, whilst avoiding metabolic
and toxicological liabilities. Preferably such
tools will provide decision-strength data
(supportive of go/no go choices) at a rate
commensurate with the output of high-
throughput approaches. Therefore, emer-
gent technologies need to be speedy, accu-
rate, easily integrated into existing test pro-
grammes and must provide unequivocal re-
sults. The most useful assays will be predic-
tive of tests to be carried out in animal
models or even drug behavior in humans.
For metabolism, there are already many ap-
proaches available to assess the likely behav-
ior of a compound upon exposure to man.
As explained by Pamela Williams earlier,
one important class of drug-modifying reac-
tions is catalyzed by liver cytochrome P450s
which account for about 90% of the known
metabolism of clinical drugs. Assays cap-
Executive Summary CXII
able of indicating which compounds are
likely targets for given cytochrome P450s
now exist. Computer-based modeling tech-
niques are steadily gaining approval within
the pharmaceutical industry. These so-called
in silico systems enable unlimited attempts
to model the interaction of a candidate drug
with a drug-modifying enzyme (DME). This
process does not require actual chemicals
and can provide valuable insights into
drug/enzyme interactions, allowing itera-
tive drug design and optimization. In vitro
assays to determine the action of human cy-
tochromes and other DMEs have been de-
veloped in the last 10 years. In more recent
times, their use in preclinical testing has in-
creased rapidly in response to the pressures
of prioritizing drug candidates and the ethi-
cal drive to reduce animal experimentation.
As drug metabolism is better understood at
the molecular level, it is proving possible to
design assays to look at particular aspects of
this complex process. There are still limita-
tions which reflect the discrete nature of the
tests with respect to each other, whereas the
in vivo reactions they attempt to mimic oc-
cur in a complex environment, with individ-
ual drugs and their metabolites being acted
upon by a plethora of systems designed to
rid the body of xenobiotics.
As discussed by Mike Murray from BTG,
London, clinical toxicology, on the other
hand, presents a completely different chal-
lenge, in that toxic effects are seen at a con-
siderable distance from the supposed mo-
ment of action of the suspect drug. In addi-
tion, toxicological endpoints are extremely
varied where the mode of toxic action is con-
cerned and particular assays are required to
test for particular toxic outcomes. There
needs to be as many specific toxicology as-
says as there are known toxic pathways
clearly, there is no easy way to develop tests
ab initio for unprecedented toxic events.
Thus, much of the prediction of mecha-
nisms of toxicology is experience-based.
Genotoxicology is arguably an exception, be-
cause damage to DNA can be detected in a
number of relatively simple test systems
(e.g., the Ames test, in which mutagenicity
of a substance is simply measured by corre-
lating its ability to induce the conversion
from a his
to a his
+
mutant of Salmonella ty-
phimurium) and can provide the investigator
with unequivocal results. ADME/Tox predic-
tion has two separate, but coincident aims:
some assays, used in early preclinical devel-
opment, are aimed at gaining insights into
the likely performance of a compound in a
later-stage preclinical assay. Ultimately,
however, assays are aimed at gaining a
strong indication of how a compound will
react once it is in a human metabolic envi-
ronment. Mike describes the technology
platforms in Mettox
TM
, a BTG business fo-
cused on preclinical assays with the capabil-
ity to predict phase I metabolism mediated
by liver cytochrome P450s. In addition, he
introduces GreenScreen, another system
in which positive results are predictive of
genotoxicity in mandated regulatory tests
by detecting potential genotoxic carcino-
gens. This recently launched yeast (i.e., eu-
karyotic) microplate screening assay detects
the DNA damage-induced transcription of
the RAD54 gene using a GFP (green fluo-
rescent protein) reporter. The assay provides
some of the eukaryotic targets missed by the
bacterial screen as well as good detection of
the DNA-breaking agents (clastogens) not
readily detected in prokaryotic tests.
Now, as we have discussed in detail the
required analytical methods to thoroughly
characterize a biopharmaceutical, includ-
ing the importance of, for example, drug
drug interactions, we will now look into
the requirements from a regulatory author-
ity perspective to finally approve such a
new biological entity.
Executive Summary CXIII
Happy End Claim to Fame and Approval
The last, but by far not the least, step be-
fore launching a new modern biopharma-
ceutical to market is the approval by a reg-
ulatory authority. In the US the regulatory
authority is the FDA and in Europe it is
the EMEA. Since the requirements as well
as the procedures for approval can vary
quite significantly between these two
authorities, we will present both.
Kurt Brorson and colleagues from FDA
describe the regulatory aspects of approv-
ing a biopharmaceutical in the US. Dr.
Brorson, who obtained his PhD from the
California Institute of Technology (Cal-
Tech), did postdoctoral studies at the NIH
and then joined the FDAs division of
mAbs in 1992. Dr. Patrick Swann, who re-
ceived his PhD from Purdue University
and also did postdoctoral studies at NIH
before joining FDA, currently serves as ex-
pert biologist for mAbs. This group of ex-
perts describes why they anticipate that
INDs, Biologics License Applications
(BLAs) and NDAs for novel targets, prod-
ucts and indications will continue to be
submitted to the FDA fueled by sequenc-
ing of the human genome. While it is es-
sential for companies to follow existing
regulations and to obtain various guid-
ance, the unique nature of many biotech
products calls for FDA regulators to apply
a flexible, case-by-case, science-based
approach when evaluating safety, product
quality, clinical development and market-
ing authorization. Contacting the appropri-
ate review office at FDA before submitting
an IND, BLA or NDA is key in avoiding
misunderstandings and/or misperceptions
regarding regulatory expectations for par-
ticular products and applications. Effecti-
ve communication between the FDA and
the submission sponsor is a crucial ele-
ment of the pathway from drug discovery
to the clinic. While the regulatory pathway
is complex, an early understanding of
the regulatory process and careful product
and preclinical characterization enhances
the chances of success. In their review,
Kurt and colleagues discuss biopharma-
ceutical development, manufacturing and
preclinical testing from an FDA CMC or
product reviewers perspective. They dis-
cuss issues identified by Agency personnel
that in the past have adversely impacted
product development and success. Finally,
they describe recent initiatives within FDA
to streamline and facilitate the product de-
velopment pathway and recent activities
concerning follow-on biopharmaceuticals
(biogenerics). For this purpose, in Septem-
ber of 2004, FDA sponsored a public work-
shop on scientific and technical considera-
tions related to the development of follow-
on protein biopharmaceutical products
and how they are perceived in the US an
update on this is also provided.
The EU counterpart is described by my
friend Axel F. Wenzel. Axel, a PhD in Biol-
ogy, is the managing director of p.ss.t (Phar-
ma Scientific Services Team), a service pro-
vider and consulting company for drug and
medical device development. Axel is a Lec-
turer at the University of Witten and mem-
ber of the Board of Directors of TOPRA
(The Organization for Professionals in Reg-
ulatory Affairs) and is Editor-in-Chief of its
journal, the Regulatory Rapporteur. He has
more than 20 years of experience in phar-
maceutical R & D, and worked for several
pharmaceutical companies such as Sandoz
and MSD Sharp & Dohme. His colleague,
Carina E. A. Sonnega, is a biotechnology
consultant in regulatory and quality affairs
with a Doctorate in Molecular Biology. Her
professional experience started at Chiron
and from 1994 to 1996 she was also a mem-
ber of the Laboratory for Medicines and
Medical Devices, National Institute of Pub-
Executive Summary CXIV
lic Health and Environment, Department of
Biotechnology. Both have extensive knowl-
edge and expertise in this field, and describe
the regulatory environment for approval of
biopharmaceuticals in the EU.
In the EU, the development of medical
products for human and veterinary use is
governed by a variety of laws, legislations,
directives and guidelines, some of them
have very specifically developed for bio-
pharmaceuticals. The market value (at ex-
factory prices) of the total EU pharmaceu-
tical market is just over Euro 62000 mil-
lion (i.e., approximately 30% of the world
market); its retail value now exceeds Euro
90000 million. In 1997, the pharmaceuti-
cal industry employed nearly 500000 peo-
ple within the EU, including 71000 in
R&D. In addition to a substantial R&D-
based sector, the pharmaceutical industry
in Europe also has active sectors dealing
in generic (i.e., patent-expired) and OTC
medicines.
On the biotechnology side, Europe has
made a particularly poor start compared
with the progress in the US, as was noted
in a 1994 Communication of the European
Commissions Directorate DGIII. Figures
compiled in 1995 on the invention and
marketing of biotechnology-derived new
active substances put the US share at
76%, Japans at 14% and Europes at 10%.
Data based on a total of 770 biotechnol-
ogy-derived medicines (including 206 ge-
netically engineered ones) under develop-
ment at the end of 1995 indicated that
25% of the biopharmaceutical develop-
ment work is currently located in Europe
(63% in the US and 7% in Japan): in gene
therapy specifically, 22% of the develop-
ment work is located in Europe (70% in
the US and 1% in Japan). It is remarkable
that the percentage of medicinal products
launched since the early 1980s is steadily
increasing up to approximately 20% in the
late 1990s. As described by Axel and Cari-
na, biopharmaceuticals are already repre-
sented in many medical indications me-
tabolic diseases, growth disorders (e.g.,
growth hormones) and cancer being the
most important ones.
Part VIII: From Bench to Bedside
The Aftermath
Thinking Big and Deal Making for Growth
Global Changes in the Healthcare Sector
Twenty-first century biopharmaceutical
medicine offers an unprecedented number
of pharmaceutical and other treatment op-
tions for more diseases and conditions
than ever before. However, the underlying
advances have fueled an equally unprece-
dented growth in healthcare costs, leading
to widespread concerns about the funding
of healthcare systems. As explained by my
colleagues from McKinsey, in both the US
and Europe, the two largest markets in
terms of expenditures, healthcare spend-
ing has been rising for decades with only
a few intermittent slowdowns, and began
to spike upwards again even more steeply
in the early 2000s. The upwards trend is
expected to continue, given the aging of
both the US and EU populations, which
will continue to propel the demand for
healthcare, particularly drugs. As pointed
out by Alexander Moscho, Markus A.
Schfer and Kristin Yarema, the real stick-
ing point of course is sluggish economic
growth, particularly in Europe, as all three
experts state uni sono. Dr. Moscho, who
holds a Degree in Biotechnology, was
working at Stanford University before
founding a Swiss biotech company and
then joining McKinsey. Dr. Schfer re-
ceived his PhD from the Max-Planck-Insti-
tute for Molecular Genetics, before he
Executive Summary CXV
joined McKinsey where he is now serving
as member of the European strategy prac-
tice. Their US counterpart, Dr. Yarema, re-
ceived her academic degrees from Stanford
University, and currently focuses on bio-
pharmaceutical R & D strategy develop-
ment and healthcare policy setting. The
three explain why if increases in gross
domestic products were sufficiently vigor-
ous society could theoretically absorb
ever-higher spending on healthcare with-
out cutting back in other areas. However,
as they also explain, the growing weight of
healthcare budgets versus other public
spending has provoked a wave of new
healthcare cost-containment reforms, par-
ticularly in Europe. The reforms designed
to curb drug costs are particularly severe,
and have permanently raised the bar for
pharmaceutical innovation and cost perfor-
mance. In the two decades since Eli Lilly
began distributing Genentech-licensed hu-
man insulin in 1982, the status of biophar-
maceutical drugs as a new industry and
source of specialty treatments largely
shielded this segment from the mounting
pressures on price and patient access faced
by traditional chemistry-based pharmaceu-
ticals. Now, however, biopharmaceuticals
are coming of age in a harsh industrial
landscape, marked by tighter limits on
prices and patient access, tougher tests of
product efficacy and cost-efficiency, in-
creasing scrutiny by policy makers and the
public and as we will see below espe-
cially due to foreseeable revenue pressure
from biogenerics!
As the colleagues from McKinsey de-
scribed in Nature Biotechnology, every com-
pany will need to work through these is-
sues in its own way, but there are a few
key questions that all biologic-focused
companies need to examine. What is the
impact of current regulatory/cost-contain-
ment trends on biopharmaceuticals? As
the industry matures, what assets can bio-
pharmaceuticals players draw on or devel-
op to keep on generating profitable
growth? What action should a company
take to defend and strengthen its competi-
tiveness? In answering these questions,
they begin with describing the general cli-
mate and the most prominent features of
the new landscape, i.e., the forces shaping
the pharmaceutical industry as a whole.
Subsequently, they take a closer look at the
biopharmaceutical segment and the ele-
ments that will help biopharmaceutical
players succeed in the increasingly chal-
lenging business and regulatory environ-
ment. Finally, they discuss three critical
areas beyond operational excellence (which
will be a must) in which most biophar-
maceutical companies will need to adapt
their strategies and business models: keep-
ing the biogeneric threat at bay, adjusting
portfolios to reflect the changing value of
innovation, and remaking the marketing
organization to better address the needs of
multiple markets and decision makers.
News and Views
Quo Vadis, Biopharmaceuticals?
One very interesting business model ad-
justed to the aforementioned global
changes in the healthcare system and sub-
sequently combining the required adjust-
ments is presented by Dorian Bevec, CSO
and co-founder of mondoBIOTECH Group.
Dorian, who has extensively published in
Science and Proceedings of the National Aca-
demy of Sciences of the USA, has 10 years
of experience with biopharmaceutical devel-
opment, gained with the Sandoz/Novartis
Research Institute in Vienna. We know
each other from a PharmaManagement
workshop, which I initiated to bring to-
gether people from different areas of the
pharmaceutical business to exchange ex-
Executive Summary CXVI
periences and ideas, and to develop success-
ful strategies for pharma companies. To-
gether with Fabio Cavalli, CEO and co-foun-
der of mondoBIOTECH, they describe their
successful business concept, focusing on
development of innovative therapeutics
and diagnostics in severe and rare lung dis-
eases, also offering the patient the opportu-
nity to meet the biotech world. Headquar-
tered in Lugano, Switzerland, they specia-
lize in redirecting approved biopharmaceu-
tical drugs and clinical stage compounds
into new medical indications. As an integral
part of mondoBIOTECHs philosophy, they
organize seminars and workshops on the
own campus where science meets science,
business meets business and, finally,
science meets business. In this way usually
unattended patients profit quicker from ac-
cumulated academic know-how. Founded in
2000, mondoBIOTECH focuses on three
different drug platforms to address unmet
medical needs in rare diseases of the lungs
by licensing projects and creating strategic
business alliances with pharmaceutical or
biotech partners. Thus, the company grows
with the strategic partner in a global phar-
maceutical business. Of particular note,
mondoBIOTECH entered a strategic alli-
ance with Bachem AG for development of
Aviptadil
(the human vasoactive intestinal

peptide) for late clinical development
phases in pulmonary arterial hypertension
and lung sarcoidosis, and received orphan
drug status designation from the European
Commission for Aviptadil in pulmonary ar-
terial hypertension and chronic throm-
boembolic pulmonary hypertension both
are life-threatening Aviptadil-deficiency
conditions. A smart decision was to deliver
Aviptadil locally via inhalation to the lungs,
as this is the target organ for disease inter-
vention. Clinical data obtained from pa-
tients suffering from pulmonary hyperten-
sion and treated for 3 months with inhaled
Aviptadil demonstrates safety and efficacy
for this modern biopharmaceutical. mondo-
BIOTECH is developing inhaled Aviptadil
as a new treatment option for pulmonary ar-
terial hypertension and, as mentioned
above, have just received orphan drug status
designation by EMEA. When reaching regu-
latory approval, this orphan drug status
guarantees mondoBIOTECH 10-year mar-
ket exclusivity for the drug on the European
market. Due to its biological and medicinal
features, Aviptadil promises great upside
potential as a therapeutic option also in
other different lung diseases. Its predomi-
nant localization in the lungs, and the vast
body of pharmacological and clinical experi-
ence makes it an attractive candidate for al-
ternative treatment for acute and chronic
pulmonary disorders, e.g., lung sarcoidosis,
and beyond.
After this enjoyable example of how to
successfully sneak into a niche, we will
now talk about a real threat for the entire
pharmaceutical business: generics. As stat-
ed before, biogenerics are currently one of
the greatest fears for innovative biophar-
maceutical companies and there has never
been a better time to be a generic manu-
facturer: the generic market is growing fas-
ter than the pharmaceutical industry as a
whole. With nearly US$ 45 billion of
drugs expected to go off-patent over the
next 4 years, the US and countries in Eu-
rope are moving towards increased generic
usage due to the rising cost of healthcare.
In the US, the generic drug utilization rate
has moved up steadily since 1984 after the
Hatch-Waxman Act was passed. The Act
proposed an increased usage of generic
drugs. Around a dozen generic companies
now exceed US$ 1 billion in annual sales,
with more set to join them over the next
few years, e.g., the huge markets of Brazil,
Russia, India and China (the BRIC econo-
mies) have promised much and delivered
Executive Summary CXVII
little yet. Ongoing political and economic
issues have held them back from realizing
their true potential in the global market,
and common characteristics to be found
among the four markets include low levels
of health expenditure, wide unmet clini-
cal/social needs and variable health provi-
sion; however, this is all changing. If one
considers the long term according to a
Goldman Sachs global economics paper
in less than 40 years time the BRIC econo-
mies (and their huge markets) together
could be larger than the G6 in US dollar
terms; China alone could overtake Ger-
many in the next 4 years, Japan by 2015
and the US by 2039.
Indian biotechnology companies claim
that the industry could reach a value of
US$ 5 billion by 2010, and the worldwide in-
terest in Indian biotech is currently growing
in the wake of new data protection laws and
the relaxation of clinical trial rules. A re-
cently published study indicates that the bio-
tech industrys revenues rose by an annual
39% during the 2003/04 financial year, with
biopharmaceuticals accounting for the lions
share of the sectors income. Rapid growth
in stem cell research and clinical studies
in general has been observed in recent years,
and recent modernization of the regulatory
regime governing the sector is set to expand
investment further. Following the amend-
ment of Indias so-called Schedule Y in early
2005, full clinical trials through phases IIII
are now legal in India. Provisions allowing
the export of human tissue samples are also
testimony to the improved environment for
foreign biotech companies, which are also
set to benefit from low local operating costs.
Obviously the real opportunities (and
threats at the same time) lie in the future,
where steady growth in all BRIC markets
will erode the commercial differences with
the established markets of North America,
Japan and Europe.
According to FDA News, the CEO of the
UK-based drug major GlaxoSmithKline,
Jean-Pierre Garnier, indicated that the com-
pany will call on the UKs government to
make new intellectual property laws a con-
dition and prerequisite of trade with the
BRIC countries. For example, he highlights
the fact that although Indias new patents
framework would be a positive move, the
authorities failure to resource patent pro-
tection efficiently could pose a threat to fu-
ture investment and hold off investment
plans until India can provide adequate intel-
lectual property protection. However, with
thousands of filings likely to be submitted
by local producers, it is feared that bottle-
necks in patent approvals could effectively
render the reforms meaningless.
In 2004, more than 20 leading drugs
lost patent protection, implying substantial
losses for companies like AstraZeneca and
the leading French producer Sanofi-Aven-
tis. Industry sources have claimed that by
2007, some 30% of Frances current reim-
bursement list will carry generic alterna-
tives following patent expiry. However, sev-
eral leading multinationals are set to pros-
per from the trend, with the generics busi-
nesses of Germanys Merck KGaA and
Switzerlands Novartis all reporting strong
sales in 2004. Indeed, Novartis who have
just bought over Hexal in an US$ 8.3 bil-
lion deal estimated that the market share
of generics in France is set to double, with
the sector accounting for 10% of total mar-
ket value by 2009.
Although at the time of writing there
are no clear-cut decisions from the regula-
tory authorities on how to treat biogenerics
and how to show bioequivalence, it is sure
that this will become a blooming market
and that the pharmaceutical companies
are preparing. Again, Novartis, for exam-
ple, has revived the name Sandoz (one of
the companies that previously formed this
Executive Summary CXVIII
pharma giant) as brand for their new bio-
generics business.
The terms bioequivalence, biosimilar and
biogeneric are interchangeable, but bioequi-
valence is often associated with biogenerics.
Specifically, bioequivalence describes the
mode of action of a substance and compares
that action to the innovators product in
terms of its outcome in patient response.
G-CSF, for example, generically known as
Filgrastim, is produced by recombinant
DNA technology. Amgen was the first com-
pany to receive approval and markets its G-
CSF, which stimulates the production of
white blood cells (specifically neutrophils
which protect the body from infection) un-
der the trade name Neupogen
. By increas-
ing neutrophils, the risk of infection de-
creases in conditions such as cancer, bone
marrow transplant, prechemotherapy blood
cell collection and severe chronic neutrope-
nia. G-CSF is a blockbuster with annual
sales of US$ 1.5 billion and EPO, also from
Amgen, has annual sales several times that.
One very experienced person with a long
industry track record, who was involved in
both of these two major product launches,
is James Harris from Dragon Pharmaceuti-
cals. As described by James, these blockbus-
ters and other biopharmaceuticals are of
course patent protected; however, as he also
explains, it is anticipated that the first pa-
tents are due to expire on or about 2006.
In countries where the patents are not en-
forceable (e.g., BRIC countries), many com-
petitors have developed their own therapies
already and eagerly await the expiry of pa-
tents in the US, Europe and Japan. Prior
to the patent expirations, competitors will
work on showing bioequivalence to the reg-
ulatory authority, and determine if and
when to target a particular market for pene-
tration with their biogeneric. One thing is
obvious: this will definitely have a tremen-
dously negative impact on future opportu-
nities of pharmaceutical and biotech com-
panies to develop new, innovative, modern
biopharmaceuticals.
Light at the End of the Tunnel
or Back to the Roots?
Since biopharmaceuticals are more com-
plex and consequently more expensive to
develop, we conclude with two chapters fo-
cusing on small-molecule drugs. The first
contribution will explain methods how to
develop a small-molecule compound out
of a biopharmaceutical drug. The second
contribution is an impressive example on
how to modify and genetically engineer
the biosynthetic pathways of microorgan-
isms to make them produce a desired
small-molecule compound.
My long-time colleague and friend, Pro-
fessor Paul Wrede from Charit, Berlin, de-
scribes some examples of how to apply
bioinformatic means to design new bioac-
tive small molecules derived from the avail-
able knowledge of the biopharmaceutical
counterpart. For example, the Pep2Lead
strategy first integrates all available infor-
mation of the natural peptide ligand to
identify key interaction points and subse-
quently switch from peptide backbones to
nonpeptidic scaffolds. In other words, Paul
describes methods of how to develop from
a potent peptide to a small-molecule drug
with high efficacy, and he presents several
possibilities and computer algorithms to
make drug discovery more dependent on ra-
tional approaches (whereby the underlying
principles are always based on advanced
pattern recognition approaches). Paul, who
received his PhD from the Max-Planck-In-
stitute of Molecular Genetics, Berlin, did
postdoctoral studies with Alexander Rich
at MIT, Cambridge, MA. He is founder
and CEO of CallistoGen, a biotechnology
company focusing on virtual screening of
Executive Summary CXIX
drugs, developing prediction algorithms for
pharmacokinetic profiling and designing
the first inhibitors for BACE (b-amyloid-
converting enzyme) against Alzheimers
disease. In his interesting chapter he de-
scribes three different examples of how to
use bioinformatic means to design new
bioactive small molecules starting from a
biopharmaceutical. The first example is
the identification of new thrombin inhibi-
tors starting from a peptide. The second ex-
ample considers the search for small mole-
cules starting from a peptide inhibitor: the
identification of antagonists of the neuroki-
nine receptor, a transmembrane protein re-
ceptor regulating brain functions related to
depression and anxiety. The third example
is the search for a BACE inhibitor for the
treatment of Alzheimers disease.
However, these illustrated successful ex-
amples are more the exception than the
rule and the techniques described are still
in their infancy. Therefore, although very
powerful and quickly evolving, the entire
field of bioinformatics will not be capable
of replacing biopharmaceutical develop-
ment in the near future.
A smart way to employ natures versatility
and millions of years development experi-
ence to produce desired small-molecule
compounds is presented by Chaitan Khosla
from Stanford University. Chaitan, who is
founder of Kosan Biosciences, received his
PhD at the California Institute of Technol-
ogy, and has extensively published in Nature
and Science. After completing postdoctoral
studies at the John Innes Centre in the
UK, he joined Stanford in 1992, where he
is now Professor of Chemistry, Chemical
Engineering and Biochemistry. His collea-
gue, Martha Lovato Tse, received her BA
in Chemistry from Rice University and com-
pleted her PhD at the Scripps Research In-
stitute. She then worked with Professor
Khosla at Stanford University and is now
working for Genentech THE biopharma-
ceutical company. Chaitan and Martha focus
on polyketides, which are a family of com-
plex natural products synthesized from a se-
ries of small carbon precursors. The mem-
bers of the polyketide family exhibit consid-
erable structural diversity and complexity,
and although the actual cellular roles of
these compounds in the native producing
organisms remain unclear, many of them
have found use as important pharmaceuti-
cal and agricultural agents. Polyketides are
synthesized through the action of polyketide
synthases (PKSs), multienzyme complexes
that consist of numerous catalytic domains.
Importantly, these megasynthases are mod-
ular in design, which provides the opportu-
nity for the production of novel polyketides.
By splicing together modules from different
PKSs, such that a hybrid PKS incorporates
structural elements from different polyke-
tides, completely new hybrid molecules
can be created. Ultimately, PKSs are capable
of producing molecules with a complexity
that is unattainable through reasonable syn-
thetic routes and, thus, the large-scale pro-
duction of polyketides for commercial uses
depends on biosynthetic routes. As exten-
sively published (also on applying this
approach for producing taxol) in Nature,
Science and Proceedings of the National Aca-
demy of Sciences of the USA, in order to
achieve reasonable levels of polyketide pro-
duction or realize the potential of creating
novel polyketides, a high-resolution me-
chanistic understanding of PKSs is neces-
sary. Additionally, appropriate production
technology must be developed, including
high-volume production processes and flex-
ible, high-producing heterologous hosts. In
their chapter, they discuss polyketide pro-
duction as a goal of biotechnology, starting
with the chemistry and microbiology of
polyketide biosynthesis, and the current un-
derstanding of PKS mechanisms. Their ex-
Executive Summary CXX
cellent chapter closes with a discussion of
the current efforts towards novel polyketide
production and the development of scalable
production processes for a class of mole-
cules which may at one point of time re-
place biopharmaceuticals?
This is obviously more a rhetorical ques-
tion rather than a real one, because we have
learned from the exciting examples pre-
sented in Modern Biopharmaceuticals that
no other class of molecules covers such a
broad spectrum of diverse applications as
pharmaceutical drugs! However, to guaran-
tee future excitement about an increasing
number of emerging biotechnologies (and
hence newly developed biopharmaceuti-
cals), it might be worthwhile for some offi-
cial decision makers to rethink their conclu-
sions and present alternative solutions.
But for now, I hope that you will like
this first four volume compilation of cut-
ting edge biotechnologies, written by the
most knowledgeable experts from acade-
mia and industry enjoy reading Modern
Biopharmaceuticals.
Executive Summary CXXI
Eric O. Aboagye
Molecular Therapy & PET Oncology
Research Group
Imperial College London
Faculty of Medicine
Hammersmith Hospital
Du Cane Road
London W12 0NN
United Kingdom
Peter Ahnert
Center for Biotechnology and Biomedicine
University of Leipzig
Institute for Clinical Immunology and
Transfusion Medicine
Johannisallee 30
04103 Leipzig
Germany
Hidetaka Akita
Graduate School of Pharmaceutical
Sciences
Hokkaido University
Kita 12, Nishi 6
Kita-ku, Sapporo City
Hokkaido 060-0812
Japan
Asser Sloth Andersen
Novo Nordisk A/S
Novo Alle
2880 Bagsvaerd
Denmark
Heiner Apeler
PH-OP-BT
Bayer HealthCare AG Pharma
Friedrich-Ebert-Strasse 217
42096 Wuppertal
Germany
Julio Baez
FibroGen Inc.
225 Gateway Boulevard
South San Francisco, CA 94080
USA
Christoph Bagowski
Department of Integrative Zoology
University of Leiden
Wassenaarseweg 64
2333 AL Leiden
The Netherlands
Jan Barciszewski
Institute of Bioorganic Chemistry
Polish Academy of Sciences
Noskowskiego 1214
61-704 Poznan
Poland
CXXIII
List of Contributors
ISBN: 3-527-31184-X
Jrg Ingo Baumbach
Department of Metabolomics
ISAS Institute for Analytical Sciences
Bunsen-Kirchhoff-Strasse 11
44139 Dortmund
Germany
Michael D. Bentley
Nektar Therapeutics
490 Discovery Drive
Huntsville, AL 35806
USA
Dorian Bevec
mondoBIOTECH Group
Via Pasque 23
6925 Gentilino
Switzerland
John R. Birch
Lonza Biologics plc
228 Bath Road
Slough
Berkshire SL1 4DX
United Kingdom
Eduardo Blumwald
Department of Pomology
University of California
1035 Wickson Hall
One Shields Avenue
Davis, CA 95616-8683
USA
Queta Boese
Dharmacon, Inc.,
2650 Crescent Dr., Suite no. 100
Lafayette, CO 80026
USA
Mary J. Bossard
Nektar Therapeutics
490 Discovery Drive
USA
Abraham Bout
Crucell Holland N.V.
Archimedesweg 4
2333 CN Leiden
The Netherlands
Cord Brakebusch
Department for Molecular Medicine
Max-Planck-Institute of Biochemistry
Am Klopferspitz 18 a
82152 Martinsried
Germany
Hans Brandstetter
Department of Natural Sciences/
Structural Biology
University of Salzburg
Billrothstrae 11
5020 Salzburg
Austria
Structure Research
Am Klopferspitz 18a
82152 Martinsried
Germany
Andreas Briel
Research Laboratories
Schering AG
Mllerstrasse 178
13342 Berlin
Germany
Kurt Brorson
Office of Biotechnology Products
Center for Drug Evaluation and Research
Food and Drug Administration
29 Lincoln Drive
Bethesda, MD 20892
USA
List of Contributors CXXIV
Janice Brown
Offices of Biotechnology Products
New Drug Chemistry and Drug
Evaluation VI
29 Lincoln Drive
Bethesda, MD 20892
USA
Elisabeth Brundke
ProBioGen AG
Goethestrasse 54
13089 Berlin
Germany
Kevin W. Burton
Nektar Therapeutics
490 Discovery Drive
USA
Michael Buschle
Intercell AG
Campus Vienna Biocenter 6
1030 Vienna
Austria
Oren Caspi
Sohnis Family Research Laboratory for the
Regeneration of Functional Myocardium
and the Rappaport Family Institute
for Research in the Medical Sciences
The Bruce Rappaport Faculty of Medicine
Technion-Israel Institute of Technology
Technion City
Haifa 32000
Israel
Frank Castillo
Department of Gene Therapy
and Corporate Center for Biologics
Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Fabio Cavalli
mondoBIOTECH group
Via Pasque 23
6925 Gentilino
Switzerland
Jonathan D. Chesnut
Corporate Research Lab.
Invitrogen Corp.
1600 Faraday Avenue
Carlsbad, CA 92008
USA
Thomas R. Cech
and Biochemistry
Howard Hughes Medical Institute
University of Colorado
Boulder, CO 80309
USA
Wayne M. Coco
DIREVO Biotech AG
Nattermannallee 1
50829 Kln
Germany
Jose Cosme
Astex Technology Ltd.
436 Cambridge Science Park
Milton Road
Cambridge CB4 0QA
United Kingdom
List of Contributors CXXV
Johanne Cote
DSM Biologics Company Inc.
6000, Royalmount Avenue
Montral, QC H4P 2T1
Canada
Gordon M. Cragg
Natural Products Branch
Developmental Therapeutics Program
Division of Cancer Treatment
and Diagnosis
National Cancer Institute
1003 West 7th Street, Suite 206
Frederick, MD 21702
USA
John Crowley
DSM Biologics
Poststraat 1
6130 AA Sittard
The Netherlands
Benjamin Dekel
Department Immunology
Weizmann Institute of Science
PO Box 26
Rehovot 76100
Israel
Ivan Diers
Novo Nordisk A/S
Novo Alle
2880 Bagsvrd
Denmark
Theodor Dingermann
Institute of Pharmaceutical Biology
Goethe-University of Frankfurt (Biocenter)
Marie-Curie-Strasse 9
60439 Frankfurt am Main
Germany
Friedrich Dorner
Baxter AG
Industriestrasse 67
1220 Vienna
Austria
Yann Echelard
Department of Research and Development
GTC Biotherapeutics, Inc.
5 the Mountain Road
Framingham, MA 01701
USA
Manfred Eigen
DIREVO Biotech AG
Nattermannallee 1
50829 Kln
Germany
Herman N. Eisen
Department of Chemical Engineering
Massachusetts Institute of Technology
Building E25-342
Cambridge, MA 02139
USA
Volker Erdmann
Institute for Chemistry and Biochemistry
Free University of Berlin
Takustrasse 3
14195 Berlin
Germany
Farah Fawaz
Department of Gene Therapy and
Corporate Center for Biologics
Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
List of Contributors CXXVI
Matthias Filter
Institute for Molecular Biology
and Bioinformatics
Free University of Berlin
Arnimallee 22
14195 Berlin
Germany
Rainer Fischer
Fraunhofer-Institute for Molecular Biology
and Applied Ecology (IME)
Worringerweg 1
52074 Aachen
Germany
Lutz Freitag
Lung Hospital
Theo-Funccius-Strasse 1
58675 Hemer
Germany
Rainer Friedrich
Max-PLanck-Institute for Biochemistry
Am Klopferspitz 18
82152 Martinsried/Planegg
Germany
Martin Fussenegger
Biotechnology and Bioengineering Group
Institute for Chemical and Bio-Engineer-
ing (ICB)
Swiss Federal Institute of Technology
Zurich
ETH Hnggerberg
Wolfgang-Pauli-Strasse 10
8093 Zurich
Switzerland
Alexander von Gabain
Intercell AG
1030 Vienna
Austria
Rodney Gagne
DSM Biologics Company Inc.
6000, Royalmount Avenue
Montreal, QC H4P 2T1
Canada
Joerg Geistlinger
Array-On GmbH
Am Schwabeplan 1 b
06466 Gatersleben
Germany
Lior Gepstein
Technion City
Haifa 32000
Israel
Shimon Gepstein
Department of Biology
and the Rappaport Family Institute for
Research in the Medical Sciences
Technion City
Haifa 32000
Israel
Yuri Gleba
Icon Genetics AG
Maximilianstrasse 38/40
80539 Munich
Germany
List of Contributors CXXVII
Gilbert Gorr
greenovation Biotechnologie GmbH
Boetzingerstrasse 29b
79111 Freiburg
Germany
Uwe Gottschalk
Purification Technologies
Sartorius AG Biotechnology
Weender Landstrasse 94108
37075 Gttingen
Germany
Hermann Graf
Special Therapeutics
Schering AG
Mllerstrasse 178
13342 Berlin
Germany
Anil Grover
Department of Plant Molecular Biology
University of Delhi South Campus
New Delhi-110021
India
Thorsten S. Gutjahr
Pharmaceuticals Division
F. Hoffmann-La Roche Ltd
Grenzacherstrasse 124
4070 Basel
Switzerland
Hideyoshi Harashima
Sciences
Hokkaido University
Kita 12, Nishi 6
Hokkaido 060-0812
Japan
Stephan Harnisch
Pharmaceutical Development/Parenterals
Schering AG
Mllerstrasse 170
13342 Berlin
Germany
James Harris, III
Sales and Marketing
Dragon Pharmaceuticals Inc.
1055 West Hastings Street
Vancouver, BC V6E 2E9
Canada
Mitsuru Hashida
Department of Drug Delivery Research
Sciences
Kyoto University
Sakyo-ku, Kyoto 606-8501
Japan
Peter Hauff
Schering AG
Mllerstrasse 178
13342 Berlin
Germany
Kirsten Hegmans-Brouwer
Archimedesweg 4
2333 CN Leiden
The Netherlands
Jorge Heller
AP Pharma
P.O. Box 3519
Ashland, OR 97520
USA
List of Contributors CXXVIII
J. Carsten Hempel
Biotechnology
Chemgineering AG
Habichthorst 36
22459 Hamburg
Germany
Philipp N. Hess
Lavendeltuin 11
2317 NB Leiden
The Netherlands
Gesine E. Hildebrand
Drug Delivery Systems
Schering AG
Mllerstrasse 170
13342 Berlin
Germany
Michael Hildebrand
Corporate Pharmaceutical Development
Schering AG
Mllerstrasse 178
13342 Berlin
Germany
Ning Huang
Ventria Bioscience
4110 North Freeway
Sacramento, CA 95834
USA
Shiew-Mei Huang
Office of Clinical Pharmacology
and Biopharmaceutics
HFD-850, CDER Office I, Rm 4546
White Oak
10903 New Hampshire Avenue
Silver Spring, MD 20993
USA
Woo Suk Hwang
Department of Theriogenology
and Biotechnology
College of Veterinary Medicine
Seoul National University
San 56-1 Shillim-Dong, Kwanak-Gu
Seoul 151-742
Korea
Harren Jhoti
Milton Road
Cambridge CB4 0QA
United Kingdom
Ingo Jordan
ProBioGen AG
Goethestrasse 54
13089 Berlin
Germany
Rolf Kalhammer
Medical Enzymes AG
Anna-Louisa-Karsch-Strasse 7
10178 Berlin
Germany
Hiroyuki Kamiya
Sciences
Hokkaido University, Kita 12, Nishi 6
Hokkaido 060-0812
Japan
Sung Keun Kang
and Biotechnology
Seoul 151-742
Korea
List of Contributors CXXIX
Joachim-Friedrich Kapp
Schering AG
Mllerstrasse 178
13342 Berlin
Germany
Shigeru Kawakami
Sciences
Kyoto University
Japan
Oliver Kayser
Pharmaceutical Biology
University of Groningen
Antonius Deusinglaan 1
9713 AV Groningen
The Netherlands
Izhak Kehat
Technion City
Haifa 32000
Israel
Ulrich Kettling
DIREVO Biotech AG
Nattermannallee 1
50829 Kln
Germany
Chaitan Khosla
Departments of Chemistry,
Chemical Engineering
and Biochemistry
Stanford University
Stanford, CA 94305-5025
USA
Anastasia Khvorova
Dharmacon, Inc.
2650 Crescent Drive, Suite #100
Lafayette, CO 80026
USA
Christoph Klade
Intercell AG
1030 Vienna
Austria
Victor Klimyuk
Icon Genetics
Weinbergweg 22
06120 Halle (Saale)
Germany
Jrg Knblein
Microbiological Chemistry
Schering AG
Mllerstrasse 178
13342 Berlin
Germany
Kentaro Kogure
Hokkaido University
Laboratory for Molecular Design
of Pharmaceutics
Sciences
Kita 12, Nishi 6, Sapporo City
Hokkaido 060-0812
Japan
List of Contributors CXXX
Andre Koltermann
DIREVO Biotech AG
Nattermannallee 1
50829 Kln
Germany
Jacob Kung
Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Fija Lagerwerf
Archimedesweg 4
2333 CN Leiden
The Netherlands
Robert Langer
Department of Chemical Engineering
Massachusetts Institute of Technology
Building E25-342
Cambridge, MA 02139
USA
Byeong Chun Lee
and Biotechnology
Seoul 151-742
Korea
Elisabeth Lehmberg
Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Lawrence J. Lesko
Office of Clinical Pharmacology
and Biopharmaceutics
29 Lincoln Drive
Bethesda, MD 20892
USA
Heiko E. von der Leyen
Avontec GmbH
Fraunhoferstrasse 15a
82152 Martinsried
Germany
Karen Lingnau
Intercell AG
1030 Vienna
Austria
Francisco Luque Vzquez
Department of Experimental Biology
University of Jan
Campus de las Lagunillas s/n
23071 Jan
Spain
David O. Mainwaring
Lonza Biologics plc
228 Bath Road
Slough
Berkshire SL1 4DX
United Kingdom
Abeel A. Mangi
Division of Cardiac Surgery
Columbia Presbyterian Medical Center
Milstein Hospital
Building 7GN-435
177 Fort Washington Avenue
New York, NY 10035
USA
List of Contributors CXXXI
Bruce Mann
Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Sylvestre Marillonnet
Icon Genetics
Weinbergweg 22
06120 Halle (Saale)
Germany
William S. Marshall
Dharmacon, Inc.
2650 Crescent Drive, Suite #100
Lafayette, CO 80026
USA
Jose Coco Martin
DSM Biologics Company B.V.
Poststraat 1
6130 AA Sittard
The Netherlands
Matias Murer
Department of Neurology
Clinical Research Group for Multiple
Sclerosis
Julius Maximilians-University
of Wrzburg
Josef-Schneider-Strasse 11
97080 Wrzburg
Germany
Michael McCaman
Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Harry M. Meade
Department of Research
and Development
5 the Mountain Road
USA
Luke Anthony Miles
St. Vincents Institute
of Medical Research
9 Princes Street
Fitzroy, Victoria 3065
Australia
Pankaj Modi
Research and Development
Generex Biotechnology Corp.
33 Harbour Square, Suite 202
Toronto, ON M5J 2G2
Canada
Shin Yong Moon
Deptartment of Theriogenology
and Biotechnology
Seoul 151-742
Korea
Luis Moroder
Bioorganic Chemistry
Am Klopferspitz 18a
82152 Martinsried
Germany
Simon Moroney
MorphoSys AG
Lena-Christ-Strasse 48
82152 Martinsried/Planegg
Germany
List of Contributors CXXXII
Kevin V. Morris
Division of Molecular Medicine
Beckman Research Institute of the
City of Hope National Medical Center
Duarte, CA 91010
USA
Alexander Moscho
McKinsey & Company
Prinzregentenstrasse 22
80538 Mnchen
Germany
Kirsten Mundt
AMGEN Switzerland AG
Alpenquai 30
PO Box 2046
6002 Luzern
Switzerland
Michael Murray
Business Development
Amura Therapeutics Ltd.
Incenta House
Horizon Park
Barton Road
Cambridge CB3 7AJ
United Kingdom
Dario Neri
and Applied Biosciences
Institute of Pharmaceutical Sciences
Zurich (ETH)
8093 Zrich
Switzerland
David J. Newman
Natural Products Branch
1003 W 7th Street, Suite no. 206
Frederick, MD 21702
USA
Yasushi Ogawa
Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Barry R. OKeefe
Molecular Targets Development Program
Center for Cancer Research
Frederick, MD 21702
USA
Nico Oosterhuis
DSM Biologics Company B.V.
Zuiderweg 72-2
Groningen 9744 AP
The Netherlands
Dirk-Jan Opstelten
Archimedesweg 4
2333 CN Leiden
The Netherlands
Joke G. Orsel
Orsel Philips Research
Prof. Holstlaan 4
5656 AA Eindhoven
The Netherlands
Ricardo Oya
STI (Research Central Services)
University of Jan
Campus de las Lagunillas, s/n
23071 Jan
Spain
List of Contributors CXXXIII
Carol E. A. Pea
CuraGen Corporation
555 Long Wharf Drive
New Haven, CT 06511
USA
Andreas Plckthun
Biochemistry Department
University Zurich
Winterthurer Strasse 190
8057 Zrich
Switzerland
Erno Pungor, Jr.
and Corporate Center
for Biologics Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Andrew J. Racher
Lonza Biologics plc
228 Bath Road
Slough
Berkshire SL1 4DX
United Kingdom
Markus Rarbach
Operations
DIREVO Biotech AG
Nattermannallee 1
50829 Kln
Germany
Raymond M. Reilly
Departments of Pharmaceutical Sciences
and Medical Imaging
Leslie Dan Faculty of Pharmacy
University of Toronto
19 Russell Street
Toronto, ON M5S 2S2
Canada
Carsten Reinhardt
Pharmaceuticals Division
F. Hoffmann-La Roche Ltd.
Grenzacherstrasse 124
4070 Basel
Switzerland
Michael Reinhardt
Schering AG
Mllerstrasse 178
13342 Berlin
Germany
Yair Reisner
Department of Immunology
Weizmann Institute of Science
PO Box 26
Rehovot 76100
Israel
Elke Reissig
Schering AG
Mllerstrasse 178
13342 Berlin
Germany
Norbert Riedel
Baxter International
One Baxter Parkway
Deerfield, IL 60015
USA
Thomas Rose
ProBioGen AG
Goethestrasse 54
13089 Berlin
Germany
List of Contributors CXXXIV
John Rossi
Division of Molecular Biology
Beckman Research Institute of the
City of Hope National Medical Center
Duarte, CA 91010
USA
Bonnie E. Gould Rothberg
Division of Chronic Disease Epidemiology
Yale University School of Public Health
60 College Street
P.O. Box 208034
New Haven, CT 06520-8034
USA
Jonathan M. Rothberg
CuraGen Corporation
555 Long Wharf Drive
New Haven, CT 06511
USA
Gabor M. Rubanyi
Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Vera Ruzsnyi
G.A.S. Gesellschaft fr analytische
Sensorsysteme mbH
Joseph-von-Fraunhofer-Strasse 13
44227 Dortmund
Germany
Volker Sandig
ProBioGen AG
Goethestrasse 54
13086 Berlin
Germany
Markus Schfer
McKinsey & Company
Kurfrstendamm 185
10707 Berlin
Germany
Stefan Schillberg
Fraunhofer Institute for Molecular Biology
and Applied Ecology (IME)
Worringerweg 1
52074 Aachen
Germany
Manuel Schmidt
Mologen AG
Fabeckstrasse 30
14195 Berlin
Germany
Rainer Schmuck
Department of Enzymology and Genetics
Roche Diagnostics GmbH
Roche Centralized Diagnostics
Nonnenwald 2
82377 Penzberg
Germany
Natarajan Sethuraman
Research Foundation
University of South Carolina
10 College Hill
Hanover, NH 03755
USA
Zhixin Shao
Department of Enzymology and Genetics
Roche Centralized Diagnostics
Nonnenwald 2
82377 Penzberg
Germany
List of Contributors CXXXV
Marjorie A. Shapiro
Evaluation VI
29 Lincoln Dr.
Bethesda, MD 20892
USA
Katrin Sichler
Quality Management Centralizied
Diagnostics
Sandhofer Strasse 116
68305 Mannheim
Germany
Michela Silacci
Institute of Pharmaceutical Sciences
Department of Chemistry and Applied
Biosciences
Zurich (ETH)
8093 Zrich
Switzerland
Harald Sobek
Global Biotechnology Operations
Roche Applied Science
Nonnenwald 2
82377 Penzberg
Germany
Carina E. A. Sonnega
Casa Oberti, U Pianu
20225 Muro, Corsica
France
Patrick G. Swann
Evaluation VI
29 Lincoln Drive
Bethesda, MD 20892
USA
Marciej Szymanski
Institute of Bioorganic Chemistry
Polish Academy of Sciences
Noskowskiego 1214
61-704 Poznan
Poland
Mei Tan
and Corporate Center
for Biologics Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Martha Lovato Tse
Genentech
One DNA Way
South San Francisco, CA 94080-4990
USA
Richard M. Twyman
Department of Biology
University of York
Heslington, York YO10 5DD
United Kingdom
Wolfgang Vautz
ISAS Institute for Analytical Science
Bunsen-Kirchhoff-Strasse 14
44139 Dortmund
Germany
List of Contributors CXXXVI
Tacey X. Viegas
Nektar Therapeutics
490 Discovery Drive
USA
Dijana Matak Vinkovic
Milton Road
Cambridge CB4 0QA
United Kingdom
Barbara Volz
Mologen AG
Fabeckstrasse 30
14195 Berlin
Germany
Andreas H. Wagner
Avontec GmbH
Fraunhoferstrasse 15a
82152 Martinsried
Germany
Sabrina Wagner
greenovation Biotechnologie GmbH
Boetzingerstrasse 29b
79111 Freiburg
Germany
Gary Walsh
Industrial Biochemistry Program
University of Limerick
Castleroy, Limerick City
Ireland
Chun Wang
Department of Biomedical Engineering
University of Minnesota
7-105 BSBE, 312 Church Street S.E.
Minneapolis, MN 55455
USA
Wilfried Weber
Biotechnology and Bioengineering Group
Institute for Chemical and Bio-
Engineering (ICB)
Zurich
ETH Hnggerberg
Wolfgang-Pauli-Strasse 10, HCI F105
8093 Zrich
Switzerland
Axel F. Wenzel
PSST Pharma Scientific Services Team
Kreillerstrasse 65
81673 Munich
Germany
Erik Whiteley
Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Sheena Whyte
Milton Road
Cambridge CB4 0QA
United Kingdom
Andreas Wiesner
Ciphergen Biosystems
Hannah-Vogt-Strasse 1
37085 Gttingen
Germany
List of Contributors CXXXVII
Barbara Wilcox
Evaluation VI
29 Lincoln Drive
Bethesda, MD 20892
USA
Pamela A. Williams
Milton Road
Cambridge CB4 0QA
United Kingdom
Thomas Winckler
60439 Frankfurt/Main
Germany
Burghardt Wittig
Mologen AG
Fabeckstrasse 30
14195 Berlin
Germany
Paul Wrede
Institute of Molecular Biology
and BioInformatics
Charit Campus Benjamin Franklin
Arnimallee 22
14195 Berlin
Germany
Florian M. Wurm
Lausanne (EPFL)
1015 Lausanne
Switzerland
Chris Yallop
Archimedesweg 4
2333 CN Leiden
The Netherlands
Akira Yamamoto
Department of Biopharmaceutics
Kyoto Pharmaceutical University
Misasagi Yamashina-ku, Kyoto, 607-8414
Japan
Fumiyoshi Yamashita
Sciences
Kyoto University
Japan
Daichang Yang
Ventria Bioscience
4110 North Freeway
Sacramento, CA 95834
USA
Kristin Yarema
McKinsey & Company
600 Campus Drive
Florham Park, NJ 07932-1046
USA
Carol A. Ziomek
Department of Development
Suite 410, 175 Crossing Blvd.
USA
Ilse Zndorf
60439 Frankfurt/Main
Germany
List of Contributors CXXXVIII
Abstract
Biopharmaceuticals represent the fastest
growing and, in many ways, the most ex-
citing sector within the pharmaceutical in-
dustry. Within this Introduction we first
consider what category of product falls
within the description of a biopharmaceu-
tical. An overall global snapshot of the cur-
rent status of the biopharmaceutical sector
is then presented, followed by an overview
of upstream and downstream processing
operations typical of protein-based biophar-
maceuticals. General trends in product ap-
provals are next overviewed and this is fol-
lowed by a summary of the main actual
biopharmaceutical products that have
gained approval to date (within the EU
and/or US). These are considered by prod-
uct type, the most significant of which are
blood-related products, hormones, cyto-
kines, vaccines and monoclonal antibodies.
Biopharmaceuticals that have gained ap-
proval for veterinary application are then
considered, and the Introduction con-
cludes by considering some of the innova-
tions and trends likely to influence the
shape of the biopharmaceutical sector in
the future.
Abbreviations
AIDS aquired immunodeficiency
syndrome
BHK baby hamster kidney
BHV bovine herpes virus
BMP bone morphogenic protein
CHO chinese hamster ovary
CSF colony-stimulating factor
dsRNA double-stranded RNA
EL eurifel
EPO erythropoietin
EU Europian Union
FSH follicle-stimulating hormone
G-CSF granulocyte colony-stimulating
factor
GH growth hormone
GM-CSF granulocyte macrophage colony-
stimulating factor
HAMA human anti-mouse antibodies
HBsAg hepatitis B surface antigen
1
ISBN: 3-527-31184-X
Current Status of Biopharmaceuticals: Approved Products
and Trends in Approvals
Gary Walsh
Introduction
HER2 herceptin
HIV human immunodeficiency virus
IB inclusion body
IFN interferon
IL interleukin
LH luteinizing hormone
mAb monoclonal antibody
MS multiple sclerosis
NPV nuclear polyhedrosis virus
PhRMA Pharmaceutical Research and
Manufacturers of America
PDGF platelet-derived growth factor
PEG polyethylene glycol
r recombinant
rh recombinant human
RNAi RNA interference
siRNA small interfering RNA
TNF tumor necrosis factor
tPA tissue plasminogen activator
1
What are Biopharmaceuticals?
What exactly is a biopharmaceutical? The
term has now become an accepted one in
the pharmaceutical vocabulary, but it can
mean different things to different people.
A clear, concise definition is absent from
pharmaceutical dictionaries, books on the
subject, or even in the home pages of reg-
ulatory agencies or relevant industry orga-
nizations. The term biopharmaceutical
appears to have originated in the 1980s,
when a general consensus evolved that it
represented a class of therapeutic product
produced by modern biotechnological tech-
niques. These incorporated protein-based
products produced by genetic engineering
or, in the case of monoclonal antibodies
(mAbs), produced by hybridoma technol-
ogy (see also Part IV, Chapter 16 and Part
V, Chapters 1 and 2). During the1990s the
concept of nucleic acid-based drugs for
use in gene therapy and antisense technol-
ogy came to the fore (see also Part I,
Chapters 68 and Part VI, Chapter 6).
Such products are also considered to be
biopharmaceuticals. On that basis biophar-
maceuticals may be defined or at least
described as proteins or nucleic acid-
based pharmaceuticals, used for therapeu-
tic or in vivo diagnostic purposes (see also
Part III, Chapter 7 and Part V, Chapters
47), and produced by means other than
direct extraction from a non-engineered
biological source. By defining the method
of manufacture in negative terminology,
proteins obtained by direct extraction from
native sources are excluded. The descrip-
tion also encompasses nucleic acid-based
products, be they produced by biotechnolo-
gical means or by direct chemical synthesis
as is the case for most antisense-based
products (see also Part II, Chapters 7 and
8 and Part III, Chapter 3). Also, small inter-
fering RNAs (siRNAs) and decoy oligonu-
cleotides are of course considered biophar-
maceuticals, regardless of how they were
produced (see also Part I, Chapters 9 and
10). However, even that definition is becom-
ing somewhat restrictive as, for example,
cell-based products become more promi-
nent (see also Part I, Chapters 1115).
2
A Global Snapshot
It is now 22 years since approval of hu-
mulin (recombinant human insulin), pro-
duced in Escherichia coli and developed by
Genentech in collaboration with Eli Lilly
[1]. Lilly received marketing authorization
in the US for the product in 1982. This
marked the true beginning of the biophar-
maceutical industry. Currently some 142
biopharmaceuticals have gained approval
for general human use in the EU and/or
US (see also Part II, Chapter 4, Part VII,
Chapter 4 and Part VIII, Chapter 1). The
major companies marketing one or more
approved biopharmaceutical products in
these regions are listed in Tables 19, as
presented later. Additional relevant com-
pany and product information is generally
available via the company web pages, the
details of which are also provided in Tables
19. Approximately one in four of all gen-
uinely new drugs currently coming on to
the market is a biopharmaceutical and the
biopharmaceutical sector is estimated to
be worth in excess of $ 30 billion, approxi-
mately double its global value in 1999 [2].
Some 250 million people worldwide
have been treated to date with biopharma-
ceuticals. The vast majority are protein
based either recombinant proteins or
monoclonal/engineered antibodies [3]. A
small number of cell-based products con-
tinue to gain marketing approval and one
antisense-based product (Vitravene, ISIS
Pharmaceuticals) has also been approved
for general medical use (see also Part III,
Chapter 3). Thus far only a single gene-
therapy product has gained approval any-
where. The product, trade name Gendi-
cine, is a human adenovirus engineered to
contain the human p53 tumor suppressor
gene. It was approved in October 2003 in
China, and is indicated for the treatment
of head and neck squamous cell carcino-
ma [4].
The major categories of product indica-
tions are as one might expect; mirroring
major killers in the first world, including
various forms of cancer and heart attacks.
The single most lucrative product is that
of erythropoietin (EPO). Combined sales
of the recombinant EPO products Procrit
(Ortho biotech) and Epogen (Amgen)
have reportedly surpassed the $ 6.5 billion
mark. The biopharmaceutical sector has
matured rapidly over the last decade and
is set to continue to grow into the foresee-
able future and follow-on biopharmaceu-
ticals are also coming onto the scene (see
also Part VIII, Chapter 3).
3
Upstream and Downstream Processing
Protein-based biopharmaceuticals are in-
variably produced by an initial cell culture/
microbial fermentation step (upstream
processing), followed by product recovery,
purification and formulation into final
product format (downstream processing)
[5, 6]. In the region of 40% of all protein
biopharmaceuticals approved to date are
produced by recombinant means in E. coli.
E. coli displays several advantages as a pro-
duction system. Its molecular genetics are
well characterized. It is easy to grow, and
grows rapidly and on relatively inexpensive
media. Furthermore, high product expres-
sion levels are generally achieved. Many of
the earlier approved E. coli-based products
accumulate intracellularly in the form of
inclusion bodies. This complicated subse-
quent downstream processing as it neces-
sitated inclusion body recovery, solubiliza-
tion and renaturation of the product. How-
ever, some E. coli product expression sys-
tems now used promote export of the
desired protein into the periplasmic space
in fully folded format, from where it can
be conveniently recovered without the ne-
cessity for cellular disruption (see also Part
IV, Chapters 7 and 12).
Several products are produced using en-
gineered Saccharomyces cerevisiae. These in-
clude various insulin-based products man-
ufactured by Novo [7] (see also Part IV,
Chapter 13), recombinant hepatitis B sur-
face antigen (rHBsAg) produced by
SmithKline Beecham as well as a recombi-
nant form of the anticoagulant hirudin [8,
9]. The majority of approved biopharma-
ceuticals are, however, expressed in animal
cell lines, mainly Chinese hamster ovary
(CHO) (see also Part IV, Chapters 1 and
4), but also baby hamster kidney (BHK)
cells [10] (see also Part IV, Chapter 12).
Although expression in animal cell lines
is more technically complex and expensive
when compared to E. coli-based systems,
eukaryotic cell lines, unlike prokaryotic
ones, are capable of carrying out post-
translational modifications such as glycosy-
lation (see also Part IV, Chapters 2 and 7).
Many key biopharmaceuticals are naturally
glycosylated. Examples include EPO, many
(although not all) interferons (IFNs), blood
factor VIII (see also Part II, Chapter 3),
and gonadotrophins such as follicle-stimu-
lating hormone (FSH) and luteinizing hor-
mone (LH). In some instances unglycosy-
lated versions of a naturally glycosylated
protein retain the therapeutic properties of
the native protein and several such prod-
ucts produced in E. coli have gained regu-
latory approval. A prominent example is
that of Filgrastim a recombinant gran-
ulocyte colony-stimulating factor (G-CSF)
produced in E. coli and which displays a
biological activity similar to the native gly-
cosylated protein [11] (see also Part VIII,
Chapter 3). Additional examples include
Betaferon (Schering, Berlin) and Neu-
mega, nonglycosylated versions of IFN-b
and interleukin (IL)-11, respectively, both
of which are produced in E. coli [12, 13].
The glycocomponent of many glycopro-
teins, however, may be necessary for/im-
pact upon the biological activity of a pro-
tein, or may influence protein stability or
its circulating half-life [14]. In such in-
stances, expression in a eukaryotic system
becomes desirable, if not necessary. While
expression in lower eukaryotes such as S.
cerevisiae is possible, glycosylation patterns
more similar to a native human protein
are obtained if the protein is expressed in
an animal cell line. Although glycosylation
represents the most common post-transla-
tional modification characteristic of such
modified biopharmaceuticals, some other
forms of post-translational modification
can also occur and be relevant to the thera-
peutic/biological activity of the protein. A
prominent example is that of the anticoag-
ulant activated protein C (trade name Xi-
gris), which harbors several c-carboxylated
glutamic acid residues and one b-hydroxy-
lated aspartic acid residue [15]. Both forms
of post-translational modification are nec-
essary to underpin full functional anticoag-
ulant activity.
Although the research literature con-
tains numerous examples of high-level re-
combinant protein production using insect
cell lines, this approach has not been used
thus far to produce any commercial bio-
pharmaceutical for human use. Insect-
based systems are, however, employed in
the manufacture of several protein-based
veterinary biopharmaceuticals, as de-
scribed later. Insect cell line culture is
usually straightforward and inexpensive,
and cell growth is rapid (see also Part IV,
Chapter 14). Many insect cell lines are
sensitive to infection by baculovirus. Upon
infection, up to 50% of all cell protein pro-
duced is that of the viral protein polyhe-
drin. A common recombinant production
strategy used therefore entails introducing
the gene coding for the protein of interest
into an engineered baculovirus, under the
influence of the polyhedron promoter [16].
Downstream processing for virtually all
protein biopharmaceuticals follows a fairly
predictable sequence of events (outlined in
Fig. 1) [17]. Following initial product recov-
ery and concentration, multiple chromato-
graphic steps are undertaken (usually be-
tween three and six individual fraction-
ation steps). While gel filtration and ion
exchange are particularly common, down-
Fig. 1 Generalized overview of the down-
stream processing procedures applied to the
production of therapeutic proteins. Note that
additional/alternative chromatographic steps
are undertaken for different proteins and that
viral inactivation steps are often also included.
Final product sterilization is usually under-
taken by filtration. (Reproduced from [17],
by kind permission of the publisher.)
stream processing of several products also
entail the use of the more bioselective
technique of affinity chromatography. Ex-
amples include the incorporation of an im-
munoaffinity step in the purification of re-
combinant factor VIII (see also Part II,
Chapter 3) and the use of Protein A affin-
ity columns in the purification of some
antibody-based products. Several other
downstream processing procedures employ
at least one pseudoaffinity step, e.g., the
purification of the veterinary product Vi-
bragen x (discussed later), which involves
the use of both dye affinity and immobi-
lized metal affinity chromatography. Prep-
arative high-performance liquid chromato-
graphy systems have also been included in
the downstream processing of some prod-
ucts, including some recombinant insulins
and Leukine, a recombinant granulocyte
macrophage colony-stimulating factor
(GM-CSF) sold by Schering.
Inclusion of a viral inactivation step is
generally also characteristic of downstream
processing procedures, especially if the
product is derived from an animal cell line
[18]. Chromatographic steps are them-
selves usually quite effective in removing
viruses from product streams (e.g., gel fil-
tration will separate most viruses quite ef-
fectively from much smaller therapeutic
proteins), and, of course, the ability of the
downstream processing procedure to re-
move typical animal cell viruses from the
product will have been tested and validated
during the purification design stage (see
also Part I, Chapter 6 and Part VII, Chap-
ter 1). Amongst the various safety net
viral removal approaches often included in
downstream processing procedures are
multiple repeat filtration through a 0.1 lm
filter, heat/UV treatment or treatment with
chemical inactivation agents such as b-pro-
piolactone (used for some veterinary prod-
ucts at least).
Final product may be formulated in liq-
uid or freeze-dried form, and virtually all
biopharmaceutical products are sterilized
by filtration followed by aseptic processing.
The most commonly employed excipients
include human serum albumin, polysor-
bate 20 or 80, mannitol, sucrose or mal-
tose, amino acids (usually glycine, arginine
or histidine) and a buffer (often citrate,
acetate or phosphate based).
4
Trends in Approvals
4.1
Protein Engineered Products
The bulk of first-generation (early ap-
proved) biopharmaceuticals were unaltered
mAbs or simple replacement proteins
such insulin, blood factor VIII and IFNs.
An increasing number of modern biophar-
maceuticals, however, have been engi-
neered in order to tailor their therapeutic
properties. The most common form of
protein engineering involves the alteration
of amino acid sequence in order to achieve
one or more of the following goals:
Alteration of biological half-life of the
protein.
Reduction or elimination of issues of
product immunogenicity.
Generation of either fast- or slow-acting
product (e.g., variants of insulin).
Generation of novel, hybrid protein ther-
apeutics.
Second-generation tissue plasminogen ac-
tivator (tPA) products represent the most
prominent example of a biopharmaceutical
engineered in order to alter biological half-
life [19]. Unmodified tPA, although an ef-
fective thrombolytic agent, displays a half-
life of some 3 min after i.v. administration.
From a practical standpoint this necessi-
tated product administration by continu-
ous infusion over a 90-min period. Do-
main-deleted engineered variants (trade
names Ecokinase and Retavase), however,
display half-lives in the region of 15
20 min, facilitating product administration
by a single i.v. injection.
First-generation mAbs approved for
medical use were invariably unmodified
murine monoclonals produced by classical
hybridoma technology [20]. These suffered
from a number of clinical disadvantages,
not least the fact that they were highly im-
munogenic when administered to man.
Administration elicited the production of
human anti-mouse antibodies (the HAMA
response) that limited efficacy, particularly
upon repeat administration. Murine mono-
clonals also displayed relatively short half-
lives (typically 3040 h) when administered
to man and they proved to be poor triggers
of human immune effector functions,
such as the activation of complement. The
majority of antibody-based biopharmaceu-
ticals gaining marketing approval in recent
years are engineered in order to reduce or
effectively eliminate such problems [21,
22]. Chimeric antibodies are murinehu-
man hybrid antibodies produced by spli-
cing the gene sequences coding for the
mouse-derived antibody variable regions
(which contain the antigen binding site) to
nucleotide sequences coding for the con-
stant regions of a human antibody (see
also Part IV, Chapter 16 and Part V, Chap-
ters 1 and 2). Such chimeric products,
when compared to first generation murine
mAbs, display significantly extended half-
lives (of up to 250 h), are capable of acti-
vating human immune effector functions
and are significantly less immunogenic.
Humanized antibodies are more exten-
sively engineered, effectively produced by
grafting (at the DNA level) the actual mur-
ine antibody antigen binding regions (the
complementarity-determining regions) into
a human antibody sequence. Such prod-
ucts display half-lives essentially identical
to fully native antibodies and are signifi-
cantly less immunogenic in man, even
when compared to chimeric antibodies.
Engineered insulin analogs represent
the most prominent group of second-gen-
eration biopharmaceuticals modified in or-
der to generate either short- or long-acting
forms of the native therapeutic protein
(see also Part IV, Chapter 13 and Part VI,
Chapter 4). Both long- and short-acting in-
sulin, and the engineering principles un-
derpinning their generation are considered
subsequently in this Introduction. A num-
ber of novel hybrid proteins have also been
generated by protein engineering and have
gained medical approval for various condi-
tions. Examples include Enbrel [a tumor
necrosis factor (TNF) receptor fragment
linked to an antibody fragment, indicated
in the treatment of rheumatoid arthritis]
and Ontak (an IL-2diphtheria toxin fu-
sion protein used to treat cutaneous T cell
lymphoma).
4.2
Engineering via Post-translational
Modification
Although the majority of engineered bio-
pharmaceuticals have been altered specifi-
cally in terms of their amino acid se-
quence, several products have now come
on stream that are engineered post-synthe-
sis. The changes introduced normally en-
tail the covalent attachment of a chemical
group to the proteins polypeptide back-
bone or the alteration of a specific pre-ex-
isting post-translational modification, i.e.,
the glycocomponent of glycoproteins (see
also Part IV, Chapters 2 and 7).
4 Trends in Approvals 7
Thus far, engineering by attachment of
a chemical group has centered around
PEGylation and, to a lesser extent, attach-
ment of fatty acid groups. PEGylation
entails the covalent attachment of one or
more molecules of polyethylene glycol
(PEG) to the polypeptide backbone [23].
PEGylation is technically straightforward
to achieve and chemically activated PEG
molecules for conjugation can be gener-
ated in-house or purchased commercially.
PEGylation generally increases the plasma
half-life of a protein, by decreasing the rate
of systemic clearance. This decreases the
frequency of administration required, with
consequent economic savings and im-
proved patient experience, often along with
reduced treatment side-effects.
Native IFNs have relatively short plasma
half-lives, typically of the order of 4 h.
PEGylation can increase this value up to
24 h (see also Part VI, Chapter 2). Intron
A, for example, is a recombinant human
IFN-a2b produced by Schering Plough. It
is approved for the treatment of various
cancers including leukemia, as well as for
some viral infections such as hepatitis B
and C. Generally, administration schedules
entail product injection 3 times a week.
PEGylated Intron A, however, need only
be administered once weekly to achieve
the same effect [24].
Levemir is a recently approved insulin
analog (Table 3) whose principal engineer-
ing feature related to the covalent attach-
ment of a fatty acid side-chain, as de-
scribed later.
Thus far, at least two approved therapeu-
tic proteins are engineered by modification
of their glycocomponent. Nespo (Aranesp
in the US) is a recombinant human EPO
molecule expressed in a CHO cell line.
Native EPO harbors three N-linked carbo-
hydrate side-chains, whereas the engi-
neered recombinant product displays five
such carbohydrate side-chains. The in-
creased carbohydrate content extends the
products serum half-life significantly,
again facilitating once weekly administra-
tion [25].
Cerezyme is the trade name given to
recombinant human glucocerebrosidase, a
lysosomal enzyme central to the metabo-
lism of glucocerebrosides (glycolipids found
naturally in the body). Lack of glucocerebro-
sidase activity triggers Gauchers disease, a
genetic condition characterized by accumu-
lation of glucocerebrosides, particularly in
tissue-based macrophages. An obvious ther-
apeutic strategy in treating Gauchers dis-
ease would be the direct administration of
the missing enzyme. However, injected glu-
cocerebrosidase is quickly removed from
the bloodstream by the liver. Cerezyme is
produced in an engineered CHO cell line.
However, downstream processing includes
an enzyme-based processing step using an
exoglucosidase enzyme. The exoglucosidase
removes the sialic acid sugar caps of the oli-
gosaccharide side-chains. This exposes side-
chain mannose residues, which in turn pro-
motes macrophage-specific enzyme uptake,
mediated by mannose-specific receptors
present on the macrophage cell surface. In
this way the sugar engineering promotes
targeted delivery of the biopharmaceutical
to the cell type most affected [26] (see also
Part VI, Chapters 5 and 3, and Part VI,
Chapter 1).
5
Declining Number of Approvals
A marked decrease in the number of new
biopharmaceuticals gaining marketing ap-
proval has become evident over the last 2
3 years, both in Europe and the US (see
also Part II, Chapter 4, Part VII, Chapter 4
and Part VIII, Chapter 1). Fig. 2 presents
numbers of biopharmaceuticals approved
within the EU per year since the introduc-
tion of the centralized European applica-
tions procedure in 1995 (see also Part VII,
Chapter 5). The first product approved un-
der that new centralized evaluation system
was Gonal F, Seronos follicle-stimulating
hormone product. It was one of three bio-
tech products approved that year. Approval
numbers peaked in 2000, when 18 new
biopharmaceutical products gained mar-
keting approval. In 2001 the European fig-
ure was 12. Although 14 distinct products
were approved in 2002, eight of these were
various formulations of the same active in-
gredient (recombinant insulin produced by
Novo in an engineered strain of S. cerevi-
siae) (see also Part IV, Chapter 13). There-
fore, in reality only seven genuinely differ-
ent biopharmaceuticals were approved that
year. The downward trend continued in
2003, with the approval of just four bio-
pharmaceuticals, although eight products
were approved within the EU in 2004. The
reason this decline in number of
approvals over the last few years is not
immediately apparent, but the large num-
bers currently under clinical evaluation
likely renders this decline a short-term
phenomenon. New approaches for the ac-
celerated development of biopharmaceuti-
cals (see also Part III, Chapter 2 and Part
III, Chapter 1) will most likely also
support a healthy increasing trend in
approval.
6
Products Approved for Human Use
Here, an overview of biopharmaceutical
products thus far approved (within the EU
and US at least) is presented. The prod-
ucts have been grouped into nine catego-
ries: recombinant blood factors, recombi-
nant thrombolytics, recombinant insulins,
additional recombinant hormones, recom-
binant hematopoietic growth factors, re-
combinant IFNs and ILs, recombinant vac-
cines, monoclonal and engineered anti-
bodies, and additional biopharmaceuticals
(e.g., cell therapy, gene therapy, siRNA).
6.1
Recombinant Blood Factors
A total of seven recombinant blood factors
have gained marketing approval, mainly
throughout the 1990s (Table 1). All aim to
treat either hemophilia A or B and all are
Fig. 2 Biopharmaceutical numbers approved for human use per year since
the introduction of the centralized approvals system in 1995.
produced in engineered animal cell lines
in order to facilitate product glycosylation.
Blood factor VIII-based products are in-
dicated for the treatment and prophylaxis
of patients with hemophilia A (see also
Part II, Chapters 13). This is a genetic
disease characterized by the total lack or
presence only at low levels of blood clot-
ting factor VIII. Lack of adequate levels of
this clotting factor results in prolonged
bleeding episodes, occurring sponta-
neously or after trauma/surgery.
Recombinant blood factor products have
proven to be as effective as the plasma-de-
rived product, without suffering the disad-
vantage of the potential risk of transmis-
sion of human blood-borne pathogens (see
also Part II, Chapter 3).
Blood factor IX-based products are indi-
cated for the control and prevention of
bleeding episodes in patients with hemo-
philia B. Hemophilia B again is a heredi-
tary disorder caused by a deficiency in cir-
culating levels of coagulation factor IX, re-
sulting in impaired blood clotting ability
(see also Part III, Chapter 6).
NovoSeven is an unusual product in that
it is (a recombinant form of) human coag-
ulation factor VII (FVII). The product is
converted in an autocatalytic fashion into
the active two-chain form (FVIIa) during
its chromatographic purification. NovoSe-
Table 1 Recombinant blood factors approved to date
Product Company Therapeutic
indication
Approved
Bioclate (rhFactor VIII produced
in CHO cells)
Centeon hemophilia A 1993 (US)
Benefix (rhFactor IX produced
in CHO cells)
Genetics Institute hemophilia B 1997 (US, EU)
Kogenate (rhFactor VIII produced
in BHK cells; also sold as Helixate
by Centeon via a license agreement)
Bayer hemophilia A 1993 (US),
2000 (EU)
Helixate NexGen (octocog a; rhFactor
VIII produced in BHK cells
Bayer hemophilia A 2000 (EU)
NovoSeven (rhFactor VIIa produced
in BHK cells)
Novo-Nordisk some forms
of hemophilia
1995 (EU);
1999 (US)
Recombinate (rhFactor VIII produced
in an animal cell line)
Baxter Healthcare/
Genetics Institute
hemophilia A 1992 (US)
Advate (octocog-a, rhFactor VIII
produced in CHO cell line; the product
is similar to Recombinate except it is
expressed in culture media free from
animal-derived proteins and formulated
without plasma-derived human
albumin)
Baxter hemophilia A 2004 (EU)
ReFacto (Moroctocog-a, i.e., B-domain-
deleted rhFactor VIII produced in
CHO cells)
Genetics Institute hemophilia A 1999 (EU),
2000 (US)
ven is employed to stimulate the coagula-
tion process in hemophilic patients with
inhibitors to factor VIII and IX. It achieves
its therapeutic effect by inducing the acti-
vation of factor X to factor Xa, which con-
verts prothrombin into thrombin (see also
Part II, Chapters 1 and Part III, Chapters
6). This, in turn, triggers the final clotting
step, where fibrinogen is converted into fi-
brin to form the hemostatic plug. The
whole clot-triggering process is therefore
achieved by bypassing the action of factor
VIII and IX. This activation occurs only in
the presence of tissue factor (a membrane
protein not present in plasma), calcium
and phospholipids, so that coagulation is
stimulated only when an injury has oc-
curred to a vessel, with resulting local he-
mostasis. This complex process is nicely
shown in a video animation on the supple-
mentary CD-ROM.
6.2
Recombinant Thrombolytics
Six tPA-based thrombolytic products have
gained approval thus far (Table 2). Native
tPA is a 527-amino-acid glycosylated serine
protease synthesized predominantly in
vascular endothelial cells, from where it
enters the bloodstream. It is the major
activator of the natural thrombolytic pro-
cess and therefore has obvious application
in the accelerated removal of blood clots
that form under inappropriate conditions.
A recombinant form of native human tPA
was first marketed by Genentech in 1987.
Most subsequent tPA-based products are
engineered in some way, in order to ex-
tend their plasma half-lives, as previously
mentioned.
Table 2 Recombinant tPA-based products thus far approved
indication
Approved
Activase (Alteplase, rhtPA produced
in CHO cells)
Genentech acute myocardial
infarction
1987 (US)
Ecokinase (Reteplase, rtPA; differs from
human tPA in that three of its five
domains have been deleted; produced
in E. coli)
Galenus Mannheim acute myocardial
infarction
1996 (EU)
Retavase (Reteplase, rtPA; see Ecokinase) Boehringer
Mannheim/
Centocor
acute myocardial
infarction
1996 (US)
Rapilysin (Reteplase, rtPA;
see Ecokinase)
Boehringer
Mannheim
acute myocardial
infarction
1996 (EU)
Tenecteplase (also marketed as Metalyse;
TNK-tPA, modified rtPA produced
in CHO cells)
Boehringer Ingel-
heim
myocardial
infarction
2001 (EU)
TNKase (Tenecteplase; modified rtPA
produced in CHO cells; see Tenecteplase
entry above)
Genentech myocardial
infarction
2000 (US)
6.3
Recombinant Insulins
In many ways insulin remains the proto-
typic biopharmaceutical. Used to treat dia-
betes mellitus, early commercial prepara-
tions were extracted directly from the pan-
creatic tissue of slaughterhouse animals.
The WHO estimates that some 170 mil-
lion people suffer from diabetes, a figure
that is likely to double by 2030. Although
only a minority of these sufferers actually
require daily insulin injection, the current
world market for insulin is valued at in ex-
cess of $ 4.5 billion, a figure that is likely
to reach $ 8 billion before the end of the
decade (see also Part IV, Chapter 13 and
Part VI, Chapter 4). Commercially feasible
methods of enzymatically converting por-
cine insulin into a product identical to hu-
man insulin were developed in the 1970s,
but recombinant DNA technology has had
the greatest impact upon this sector. The
biosynthesis of insulin in the human body
and the procedure to enzymatically convert
porcine insulin is nicely shown on the
supplementary CD-ROM.
Initially produced in 1978, recombinant
human insulin (trade name Humulin) was
the first biopharmaceutical to gain ap-
proval in any world region (approved in
1982). Since then a number of additional
recombinant insulin products have come
on the market (Table 3). The modern insu-
lin industry is dominated by Lilly, Novo
and, to a lesser extent, Aventis, and these
companies manufacture and market a
range of both first-generation and engi-
neered (second-generation) insulin prod-
ucts (Table 3). Engineered second-genera-
tion insulin analogs display an amino acid
sequence altered in order to generate either
fast-acing or slow-acting product. Un-
modified human insulin molecules, when
stored at typical commercial therapeutic
dose concentrations (around 10
3
M), exist
primarily in oligomeric form, as zinc-con-
taining hexamers. Each hexamer consists
of three identical dimers, exhibiting strong
inter-subunit interactions. Three dimers
are coordinated to central zinc ions. Upon
s.c. administration, hexamers must first dis-
associate into monomeric form before entry
into the bloodstream. As a result, injected
insulin has a slower onset (and a longer
duration) of action when compared to en-
dogenous insulin secretion. A practical con-
sequence is that such insulins must be
administered to the diabetic 30 min or so
before meal times and the planned meal
time should not subsequently be altered.
In addition to such traditional short-acting
insulins, insulin may be formulated in order
to actually retard the rate of insulin entry
into the bloodstream from the injection site.
Such long-acting insulins are usually ad-
ministered (in combination with short-act-
ing insulins) in order to mimic low baseline
endogenous insulin levels.
Insulin lispro (sold under the trade
names Humalog and Liprolog) exemplifies
engineered short-acting insulin products.
This product displays an amino acid se-
quence identical to native human insulin,
with the exception that the natural pro-
linelysine sequence characteristics of po-
sitions 28 and 29 of the insulin B chain
have been reversed. The sequence inver-
sion leads to local conformational changes,
eliminating hydrophobic interactions criti-
cal to dimer stabilization. As a result, de-
oligomerization occurs rapidly upon injec-
tion and the product can be administered
at meal times rather than 30 min before.
The different forms and formulations of
insulin are shown on the supplementary
CD-ROM.
Levemir is the trade name given to an
unusual long-acting insulin product that
has just recently gained marketing ap-
Table 3 Recombinant insulins/insulin analogs thus far approved
indication
Approved
Humulin (rhInsulin produced in E. coli) Eli Lilly diabetes mellitus 1982 (US)
Novolin (rhInsulin produced in
S. cerevisiae)
Novo Nordisk diabetes mellitus 1991 (US)
Humalog (Insulin lispro, an insulin
analog produced in E. coli)
Eli Lilly diabetes mellitus 1996
(US and EU)
Insuman (rhInsulin produced in E. coli) Hoechst diabetes mellitus 1997 (EU)
Liprolog (Bio Lysprol, short-acting
insulin analog produced in E. coli)
Eli Lilly diabetes mellitus 1997 (EU)
NovoRapid (Insulin Aspart, short-acting
rhInsulin analog produced in
S. cerevisiae)
Novo Nordisk diabetes mellitus 1999 (EU)
Novomix 30 [contains insulin Aspart,
short-acting rhInsulin analog produced
in S. cerevisiae (see NovoRapid) as one
ingredient]
Novolog (Insulin Aspart, short-acting
S. cerevisiae; see also NovoRapid)
Novolog mix 70/30 (contains insulin
Aspart, short-acting rhInsulin analog
produced in S. cerevisiae as one
ingredient; see also Novomix 30)
Actrapid/Velosulin/Monotard/Insula-
tard/Protaphane/Mixtard/Actraphane/
Ultratard (all contain rhInsulin pro-
duced in S. cerevisiae formulated as
short/intermediate/long-acting product)
Lantus (Insulin glargine, long-acting
rhInsulin analog produced in E. coli)
Aventis diabetes mellitus 2000
(US and EU)
Optisulin (Insulin glargine, long-acting
rhInsulin analog produced in E. coli,
see Lantus)
Aventis diabetes mellitus 2000 (EU)
Levemir (Insulin detemir, long-acting
S. cerevisiae)
Apidra (Insulin Glulisine, rapid-acting
insulin analog produced in E. coli)
Aventis diabetes mellitus 2004 (US)
proval (Table 3). The major structural al-
teration characteristic of this insulin ana-
log is the attachment of a C14 fatty acid
via the side-chain of lysine residue num-
ber 29 of the insulin B chain. This pro-
motes binding of the insulin analog to al-
bumin, both at the site of injection and in
the plasma. In turn, this promotes a con-
stant and prolonged release of free insulin
into the blood, giving it a duration of ac-
tion of up to 24 h.
6.4
Additional Recombinant Hormones
Nineteen additional recombinant hor-
mones have been approved thus far. These
include a number of recombinant versions
of human growth hormone (hGH), various
gonadotropins, glucagon, parathyroid hor-
mone and calcitonin (Table 4). Amongst
these Somavert is notable in that it is a
recombinant PEGylated analog of hGH. It
Table 4 Additional recombinant hormones approved for general medical use
indication
Approved
Protropin (rhGH, differs from human
hormone only by containing an addi-
tional N-terminal methionine residue;
produced in E. coli )
Genentech hGH deficiency
in children
1985 (US)
Glucagen (rhGlucagon produced in S.
cerevisiae)
Novo Nordisk hypoglycemia 1998 (US)
Thyrogen (Thyrotrophin-a, rhTSH pro-
duced in CHO cells)
Genzyme detection/treatment
of thyroid cancer
1998 (US),
2000 (EU)
Humatrope (rhGH produced in E. coli) Eli Lilly hGH deficiency
in children
1987 (US)
Nutropin (rhGH produced in E. coli) Genentech hGH deficiency
in children
1994 (US)
Nutropin AQ (rhGH produced in E. coli) Schwartz Pharma growth failure,
Turners syndrome
2001 (EU)
BioTropin (rhGH produced in E. coli) Biotechnology
General
hGH deficiency
in children
1995 (US)
Genotropin (rhGH produced in E. coli) Pharmacia &
Upjohn
hGH deficiency
in children
1995 (US)
Saizen (rhGH produced in an engi-
neered mammalian cell line)
Serono hGH deficiency
in children
1996 (US)
Serostim (rhGH produced in an
engineered mammalian cell line)
Serono Laboratories treatment of AIDS-
associated catabo-
lism/wasting
1996 (US)
Norditropin (rhGH produced in E. coli) Novo Nordisk treatment of growth
failure in children
due to inadequate
GH secretion
1995 (US)
has been engineered so as to introduce
nine mutations into the hGH amino acid
sequence. It binds the hGH cell surface re-
ceptor, but fails to trigger an intracellular
response. As such it functions in an antag-
onistic fashion, reducing the effects of en-
dogenous hGH, underlining its use for
the treatment of acromegaly. The molecule
is PEGylated so as to increase its serum
half-life (see also Part VI, Chapter 2).
6.5
Recombinant Hematopoietic Growth Factors
Recombinant hematopoietic growth factors
consist of several EPOs and colony-stimu-
lating factors (Table 5). EPO represents the
single most lucrative biopharmaceutical of
all and is indicated for the treatment of
anemia associated with various medical
conditions (see also Part VIII, Chapter 3).
As previously described, Nespo (Aranesp
in the US) is an engineered EPO analog
displaying an extended serum half-life.
Leukine (owned by Schering) is a recombi-
nant human GM-CSF, a 127-amino-acid
glycosylated hematopoietic growth factor
produced in an engineered strain of S. cer-
evisiae. It was initially approved for use fol-
lowing induction of chemotherapy in adult
patients with acute myelogenous leukemia
(or acute nonlymphocytic leukemia) in or-
der to shorten time to neutrophil recovery
and reduce the incidence of severe infec-
tion. Neupogen (filgrastim) is a recombi-
nant human G-CSF (see also Part VIII,
Table 4 (continued)
indication
Approved
Gonal F (rhFSH produced in CHO
cells)
Serono anovulation and
superovulation
1995 (EU),
1997 (US)
Puregon (rhFSH produced in CHO
cells)
Organon anovulation and
superovulation
1996 (EU)
Follistim (Follitropin-b, rhFSH produced
in CHO cells)
Organon some forms of
infertility
1997 (US)
Luveris (lutropin-a; rhLH produced
in CHO cells)
Ares-Serono some forms of
infertility
2000 (EU)
Ovitrelle (also termed Ovidrelle;
rhCG produced in CHO cells)
Serono used in selected as-
sisted reproductive
techniques
2001 (EU),
2000 (US)
Forcaltonin (rSalmon calcitonin
produced in E. coli)
Unigene Pagets disease 1999 (EU)
Forteo (Forsteo in EU; teriparatide;
recombinant shortened form of human
parathyroid hormone, produced in
E. coli).
Eli Lilly treatment of osteo-
porosis in selected
postmenopausal
women
2002 (US),
2003 (EU)
Somavert [pegvisomant; recombinant
engineered hGH analog (antagonist),
produced in E. coli]
Pharmacia
Enterprises
treatment of
selected patients
suffering from
acromegaly
2002 (EU)
Chapter 3) produced by recombinant
means in E. coli. It regulates the produc-
tion of neutrophils and is indicated for the
treatment of neutropenia associated with
various medical conditions. Neulasta is a
PEGylated form filgrastim, also used to
treat neutropenia. It exhibits an extended
duration of action, due to the PEG-
mediated reduction in product renal clear-
ance rate.
6.6
Recombinant IFNs and ILs
Quite a number of IFN-based products
have gained marketing approval over the
last decade and a half (Table 6). The a
IFNs have found application mainly in the
treatment of certain cancer types and viral
diseases. The trend in this case is toward
the development of PEGylated forms of
these products. As described earlier, the
extended plasma half-life associated with
such PEGylated forms renders possible
their administration as single as apposed
to thrice weekly injection. IFN-b prepara-
tions (Betaferon, Schering) have found
application in the treatment of multiple
sclerosis (MS), a chronic disabling disease
of the central nervous system. The major-
ity of MS patients develop significant dis-
abilities, either gradually or due to relaps-
ing/remitting symptoms, which involves a
worsening of the disease followed by a
temporary recovery (see also Part V, Chap-
Table 5 Recombinant hemopoietic growth factors approved for general medical use
indication
Approved
Epogen (rhEPO produced
in a mammalian cell line)
Amgen treatment
of anemias
1989 (US)
Procrit (rhEPO produced
in a mammalian cell line)
Ortho Biotech treatment
of anemias
1990 (US)
Neorecormon (rhEPO produced
in CHO cells)
Boehringer
Mannheim
treatment
of anemias
1997 (EU)
Aranesp (Darbepoetin-; long-acting
rEPO analog produced in CHO cells)
Amgen treatment
of anemia
2001
(EU and US)
Nespo (Darbepoetin-; see also Aranesp;
long-acting rEPO analog produced
in CHO cells)
Dompe Biotec treatment
of anemia
2001 (EU)
Leukine (rGM-CSF, differs from the
native human protein by 1 amino acid,
Leu23; produced in S. cerevisiae)
Immunex autologous bone
marrow
transplantation
1991 (US)
Neupogen (Filgrastim, rG-CSF, differs
from human protein by containing an
additional N-terminal methionine;
Amgen chemotherapy-in-
duced neutropenia
1991 (US)
Neulasta (Pegfilgrastim, recombinant
PEGylated filgrastim see Neupogen;
also marketed in the EU as Neupopeg)
Amgen neutropenia 2002
(US and EU)
Table 6 Recombinant IFNs and ILs approved for general medical use
indication
Approved
Intron A (rIFN-a2b produced in E. coli) Schering Plough cancer, genital
warts, hepatitis
1986 (US),
2000 (EU)
PegIntron A (PEGylated rIFN-a2b
Schering Plough chronic hepatitis C 2000 (EU),
2001 (US)
Viraferon (rIFN-a2b produced in E. coli) Schering Plough chronic hepatitis B
and C
2000 (EU)
ViraferonPeg (PEGylated rIFN-a2b
Schering Plough chronic hepatitis C 2000 (EU)
Roferon A (rhIFN-a2a, produced
in E. coli)
Hoffmann-La Roche hairy cell leukemia 1986 (US)
Actimmune (rhIFN-c1b produced
in E. coli)
Genentech chronic granulo-
matous disease
1990 (US)
Betaferon (rIFN-b1b, differs from
human protein in that Cys17 is replaced
by Ser; produced in E. coli)
Schering MS 1995 (EU)
Betaseron (rIFN-b1b, differs from
human protein in that Cys17 is replaced
by Ser; produced in E. coli)
Berlex Laboratories
and Chiron
relapsing/remitting
MS
1993 (US)
Avonex (rhIFN-b1a, produced
in CHO cells)
Biogen relapsing MS 1997 (EU),
1996 (US)
Infergen (rIFN-a, synthetic type I IFN
Amgen (US) and
Yamanouchi Europe
(EU)
chronic hepatitis C 1997 (US),
1999 (EU)
Rebif (rh IFN-b1a produced
in CHO cells)
Ares-Serono relapsing/remitting
MS
1998 (EU),
2002 (US)
Rebetron (combination of ribavirin
and rhIFN-a2b produced in E. coli)
Schering Plough chronic hepatitis C 1999 (US)
Alfatronol (rhIFN-a2b produced
in E. coli)
Schering Plough hepatitis B and C,
and various cancers
2000 (EU)
Virtron (rhIFN-a2b produced in E. coli) Schering Plough hepatitis B and C 2000 (EU)
Pegasys (Peginterferon-a2a produced
in E. coli)
Hoffmann-La Roche hepatitis C 2002
(EU and US)
Proleukin (rIL-2, differs from human
molecule in that it is devoid of an N-
terminal alanine and Cys-125 has been
replaced by a Ser; produced in E. coli)
Chiron renal cell carcinoma 1992 (US)
Neumega (rIL-11, lacks N-terminal
proline of native human molecule;
Genetics Institute prevention of che-
motherapy-induced
thrombocytopenia
1997 (US)
Kineret (anakinra; rIL-1 receptor
antagonist produced in E. coli)
Amgen rheumatoid arthritis 2001 (US)
ter 3). The underlining mechanism is not
completely understood, but it is known
that it is mediated by specific cell receptors
and involves the modulation of the im-
mune response.
In addition to IL-2 and -11, an IL-1 re-
ceptor antagonist (trade name Kineret) has
also gained general marketing approval in
the US for the treatment of rheumatoid ar-
thritis (Table 6). Kineret binds IL-1a and
IL-1b cell surface receptors, but without in-
ducing a biological response. The product
therefore blocks IL-1 biological activity,
which is a critical mediator of the inflam-
mation and joint damage characteristic of
this condition.
6.7
Vaccines
A number of recombinant vaccines have
also gained marketing approval for general
medical use. By far the most prominent ex-
ample is that of rHBsAg, which is used to
vaccinate against hepatitis B. While the re-
combinant antigen can be used on its
Table 7 Recombinant vaccines approved for general medical use
Product Company Therapeutic indication Approved
Recombivax (rHBsAg produced
in S. cerevisiae)
Merck hepatitis B prevention 1986 (US)
Comvax (combination vaccine,
containing rHBsAg produced
in S. cerevisiae as one component)
Merck vaccination of infants
against hemophilus
influenzae type B and
hepatitis B
1996 (US)
Engerix B (rHBsAg produced
in S. cerevisiae)
SmithKline
Beecham
vaccination against
hepatitis B
1998 (US)
Tritanrix-HB (combination vaccine,
SmithKline
Beecham
vaccination against
hepatitis B, diphtheria,
tetanus and pertussis
1996 (EU)
Lymerix (rOspA, a lipoprotein found
on the surface of Borrelia burgdorferi,
the major causative agent of Lymes
disease; produced in E. coli)
SmithKline
Beecham
Lyme disease vaccine 1998 (US)
Infanrix-Hep B (combination vaccine,
containing rHBsAg produced in
S. cerevisiae as one component)
SmithKline
Beecham
immunization against
diphtheria, tetanus,
pertussis and hepatitis B
1997 (EU)
InfanrixHexa (combination vaccine,
SmithKline
Beecham
pertussis, polio,
hemophilus influenzae B
and hepatitis B
2000 (EU)
Infanrix-Penta (combination vaccine,
SmithKline
Beecham
pertussis, polio and
hepatitis B
2000 (EU)
own, it more generally represents one com-
ponent of multicomponent vaccine prepara-
tions (Table 7). The emphasis upon hepati-
tis B no doubt reflects the global signifi-
cance of this condition. Two billion people
are infected worldwide, with 350 million in-
dividuals suffering from lifelong chronic in-
fections. In excess of 1 million sufferers die
each year from liver cancer and/or cirrhosis
triggered by the condition.
Table 7 (continued)
Ambirix (combination vaccine,
Glaxo SmithKline immunization against
hepatitis A and B
2002 (EU)
Twinrix (adult and pediatric formsin
EU; combination vaccine containing
rHBsAg produced in S. cerevisiae as
one component)
SmithKline
Beecham (EU) and
Glaxo SmithKline
(US)
hepatitis A and B
1996
(EU, adult),
1997 (EU,
pediatric),
2001 (US)
Primavax (combination vaccine,
Pasteur Merieux
MSD
diphtheria, tetanus and
hepatitis B
1998 (EU)
Pediarix (combination vaccine,
Glaxo SmithKline immunization of chil
dren against various
conditions, including
hepatitis B
2002 (US)
Procomvax (combination vaccine,
containing rHBsAg as one
component)
Pasteur Merieux
MSD
hemophilus influenzae
type B and hepatitis B
1999 (EU)
Hexavac (combination vaccine,
Aventis Pasteur immunization against
diphtheria, tetanus, per-
tussis, hepatitis B, polio
and hemophilus
influenzae type B
2000 (EU)
Triacelluvax [combination vaccine
containing recombinant (modified)
pertussis toxin]
Chiron immunization against
diphtheria, tetanus and
pertussis
1999 (EU)
Hepacare [rS, pre-S and pre-S2
HBsAgs, produced in a mammalian
(murine) cell line]
Medeva immunization against
hepatitis B
2000 (EU)
HBVAXPRO (rHBsAg produced
in S. cerevisiae)
Aventis immunization of children
and adolescents against
hepatitis B
2001 (EU)
Dukoral (Vibrio cholerae and
recombinant cholera toxin B subunit)
SBL Vaccine active immunization
against disease caused by
V. cholerae subgroup O1
2004 (EU)
6.8
Monoclonal and Engineered Antibodies
Antibody-based products represent the sin-
gle largest category of biopharmaceutical
approved for general medical use (Table 8)
(see also Part IV, Chapter 16 and Part V,
Chapters 1 and 2). As previously outlined,
first-generation products were invariably
murine monoclonals produced by classical
hybridoma technology. The development
of engineered products, i.e., chimeric and
humanized antibodies, overcame many of
the therapeutic difficulties associated with
such early preparations and the majority
of recently approved products are engi-
neered in this way. The majority of anti-
body-based products are used to either de-
tect or treat various forms of cancer. The
antibodies are raised against specific tu-
mor-associated antigens found on the sur-
face of specific cancer types and therefore
will bind specifically to those cell types
upon administration. Conjugation of radio-
isotopes, toxins or other chemotherapeutic
agents should allow selective delivery of
these agents directly to the tumor surface,
courtesy of antibody-binding specificity
(see also Part II, Chapter 5 and Part V,
Chapter 6). In practice, some difficulties
can arise with this approach, e.g., due to
antibody cross-reactivity with nontrans-
formed cells or the presence of the tumor-
associated antigen, even in low numbers,
on the surface of unrelated, healthy cells.
Herceptin is the trade name given to
one such antibody-based cancer therapy
product (Table 8). It is a humanized mAb
with binding specificity for the human epi-
dermal growth factor receptor 2 (HER2),
which is overexpressed on the surface of
2530% of metastatic breast cancers.
HER2 overexpression induces abnormal
proliferation of cells (see also Part I, Chap-
ter 5). The mAb binds specifically to the
tumor cells overexpressing HER2, thus in-
hibiting their proliferation and inducing
antibody-directed cell-mediated cytotoxicity.
This results in reducing metastasis while
not affecting normal cells, therefore limit-
ing side-effects.
Detection (as opposed to treatment) of
tumors is facilitated by the conjugation of
a c-emitting radioactive tag to an appropri-
ate antibody. The radioactivity congregated
at the tumor site can penetrate outward
from the body, facilitating its detection by
equipment such as a planar c-camera (see
also Part V, Chapters 4, 5 and 7). Examples
of such products approved include Leukos-
can and Prostascint (Table 8).
Several antibody-based products are indi-
cated for non-cancer applications. Zena-
pax, for example, is used for the preven-
tion of acute kidney transplant rejection.
The product is a humanized mAb that
specifically binds the a-chain (also known
as CD25 or Tac) of the IL-2 receptor. This
receptor is expressed on the surface of ac-
tivated lymphocytes. It acts as an antago-
nist of the receptor, thus blocking the
binding of IL-2 that in turn prevents the
stimulation of lymphocytes mediating or-
gan rejection.
6.9
Additional Biopharmaceuticals
A number of additional products that do
not fall into any of the categories dis-
cussed thus far have also gained market-
ing approval. Amongst these are several
enzymes, including glucocerebrosidase
(used for the treatment of Gauchers dis-
ease, discussed previously), DNase (used
to treat cystic fibrosis), a-galactosidase
(used to treat Fabry disease) and enzymes
for amino acid depletion (see also Part II,
Chapter 6). Recombinant a-galactosidase is
produced in mammalian cell lines. The
Table 8 Monoclonal/engineered antibody-based products approved for general medical use
CEA-scan [Arcitumomab; murine
mAb fragment (Fab), directed against
human carcinoembryonic antigen]
Immunomedics detection of recurrent/
metastatic colorectal
cancer
1996
(US and EU)
MyoScint (Imiciromab-Pentetate;
murine mAb fragment directed
against human cardiac myosin)
Centocor myocardial infarction
imaging agent
1996 (US)
OncoScint CR/OV (Satumomab
Pendetide; murine mAb directed
against TAG-72, a high-molecular-
weight tumor-associated glycoprotein)
Cytogen detection/staging/
follow-up of colorectal
and ovarian cancers
1992 (US)
Orthoclone OKT3 (Muromomab CD3;
murine mAb directed against the T
lymphocyte surface antigen CD3)
Ortho Biotech reversal of acute kidney
transplant rejection
1986 (US)
ProstaScint (Capromab Pentetate;
murine mAb directed against the
tumor surface antigen PSMA)
Cytogen detection/staging/
follow-up of prostate
adenocarcinoma
1996 (US)
ReoPro (Abciximab; Fab fragments
derived from a chimeric mAb, direc-
ted against the platelet surface recep-
tor GPII
b
/III
a
)
Centocor prevention of blood clots 1994 (US)
Rituxan (Rituximab; chimeric mAb
directed against CD20 antigen found
on the surface of B lymphocytes)
Genentech/IDEC
Pharmaceuticals
non-Hodgkins
lymphoma
1997 (US)
Verluma [Nofetumomab; murine
mAb fragments (Fab) directed against
carcinoma associated antigen]
Boehringer-Ingel-
heim/NeoRx
detection of small cell
lung cancer
1996 (US)
Zenapax (Daclizumab; humanized
mAb directed against the a chain of
the IL-2 receptor)
Hoffmann-La Roche prevention of acute kid-
ney transplant rejection
1997 (US),
1999 (EU)
Simulect (Basiliximab; chimeric mAb
directed against the a chain of the
IL-2 receptor)
Novartis prophylaxis of acute
organ rejection in
allogeneic renal
transplantation
1998
(EU and US)
Remicade (Infliximab, chimeric mAb
directed against TNF-a)
Centocor treatment of Crohns
disease
1998 (US),
1999 (EU)
Synagis (Palivizumab; humanized
mAb directed against an epitope on
the surface of respiratory syncytial
virus)
MedImmune (US)
and Abbott (EU)
prophylaxis of lower
respiratory tract disease
caused by respiratory
syncytial virus in
pediatric patients
1998 (US),
1999 (EU)
Table 8 (continued)
Herceptin (Trastuzumab; humanized
antibody directed against HER2)
Genentech (US) and
Roche Registration
(EU)
treatment of metastatic
breast cancer if tumor
overexpresses HER2
protein
1998 (US),
2000 (EU)
Indimacis 125 (Igovomab; murine
mAb fragment (Fab
2
) directed against
the tumor-associated antigen CA 125)
CIS Bio diagnosis of ovarian
adenocarcinoma
1996 (EU)
Tecnemab KI [murine mAb
fragments (Fab/Fab
2
mix) directed
against high-molecular-weight
melanoma-associated antigen]
Sorin diagnosis of cutaneous
melanoma lesions
1996 (EU)
LeukoScan [Sulesomab; murine mAb
fragment (Fab) directed against NCA
90, a surface granulocyte nonspecific
cross-reacting antigen]
Immunomedics diagnostic imaging for
infection/inflammation
in bone of patients with
osteomyelitis
1997 (EU)
Humaspect (Votumumab; human
mAb directed against cytokeratin
tumor-associated antigen)
Organon Teknika detection of carcinoma
of the colon or rectum
1998 (EU)
Mabthera (Rituximab; chimeric mAb
directed against CD20 surface antigen
of B lymphocytes)
Hoffmann-La Roche
(see also Rituxan)
non-Hodgkins
lymphoma
1998 (EU)
Mabcampath (EU) or Campath (US)
(Alemtuzumab; humanized mAb
directed against CD52 surface antigen
of B lymphocytes)
Millennium and
ILEX (EU), and Ber-
lex, ILEX Oncology
and Millennium
Pharmaceuticals
(US)
chronic lymphocytic
leukemia
2001
(EU and US)
Mylotarg (Gemtuzumab zogamicin;
humanized antibody-toxic antibiotic
conjugate targeted against CD33
antigen found on leukemic blast
cells)
Wyeth Ayerst acute myeloid leukemia 2000 (US)
Zevalin (Ibritumomab Tiuxetan;
murine mAb, produced in a CHO
cell line, targeted against the CD20
antigen)
IDEC pharmaceuti-
cals (US) and Scher-
ing (EU)
non-Hodgkins
lymphoma
2002 (US),
2004 (EU)
Humira [EU and US; also sold as
Trudexa in EU; Adalimumab;
recombinant (anti-TNF) human mAb
created using phage display
technology]
Cambridge Anti-
body Technologies
and Abbott (US),
and Abbott (EU)
rheumatoid arthritis 2002 (US),
2003 (EU)
429-amino-acid glycoprotein spontaneously
dimerizes, yielding the 100-kDa biologi-
cally active enzyme. Fabry disease is a rare
genetic condition characterized by a defi-
ciency of the lysosomal enzyme a-galacto-
sidase A. As a result, sufferers exhibit an
inability to break down certain glycolipids,
particularly the glycosphingolipid ceramide
trihexoside or globotriaosylceramide (GL-
3). Glycolipid accumulates in the walls of
vascular cells, particularly in the kidney,
heart and nervous system.
Regranex is an interesting product in
that it is administered not by direct in-
jection as is the case for most biopharma-
ceuticals. The active product ingredient
is a recombinant human platelet-derived
growth factor (PDGF). Active PDGF is a
homodimer. Each 109-amino-acid, glycosy-
lated polypeptide is aligned in an antipar-
allel fashion relative to the other, yielding
the 24.5-kDa mature molecule. Regranex
is produced by recombinant DNA technol-
ogy in S. cerevisiae and the product is pre-
sented in a gel formulation containing
0.1% active ingredient for external topical
use. Regranex is indicated for the treat-
ment of chronic diabetic ulcers that do not
Table 9 Additional biopharmaceuticals approved for general medical use
indication
Approved
Beromun (rhTNF-a produced in E. coli) Boehringer-
Ingelheim
adjunct to surgery
for subsequent tu-
mor removal, to pre-
vent or delay
amputation
1999 (EU)
Revasc (anticoagulant; recombinant
hirudin produced in S. cerevisiae)
Ciba Novartis
Europharm
prevention of venous
thrombosis
1997 (EU)
Refludan (anticoagulant; recombinant
hirudin produced in S. cerevisiae)
Hoechst Marion
Roussel (US) and
Behringwerke (EU)
anticoagulation ther-
apy for heparin-asso-
ciated thrombocyto-
penia
1998 (US);
1997 (EU)
Cerezyme (rb-glucocerebrosidase
produced in CHO cells; differs from
native human enzyme by 1 amino acid,
Arg495 is substituted with a His, also has
modified oligosaccharide component)
Genzyme treatment of
Gauchers disease
1994 (US);
1997 (EU)
Pulmozyme (Dornase-a, rDNase
produced in CHO cells)
Genentech cystic fibrosis 1993 (US)
Fabrazyme (rha-Galactosidase produced
in CHO cells)
Genzyme Fabry disease
(a-galactosidase A
deficiency)
2001 (EU)
Replagal (rha-Galactosidase produced
in a continuous human cell line)
TKT Europe Fabry disease
(a-galactosidase A
deficiency)
2001 (EU)
heal with normal wound care practice. It
is usually administered daily for up to a
maximum of 20 weeks.
Bone morphogenic proteins (BMPs) rep-
resent another interesting class of biophar-
maceutical (Table 9). As their name sug-
gests, these proteins can promote the de-
position and growth of new bone, and are
often administered by implantation as part
of a medical device. InductOs, for exam-
ple, consists of a recombinant human
BMP-2 that promotes the differentiation of
mesenchymal cells into bone cells (see
also Part I, Chapter 13). The biologically
active form is a glycosylated heterodimer,
consisting of 114- and 131-amino-acid
Table 9 (continued)
indication
Approved
Fasturtec (Elitex in US; rasburicase;
recombinant urate oxidase produces
in S. cerevisiae)
Sanofi-Synthelabo hyperuricaemia 2001 (EU),
2002 (US)
Aldurazyme (Laronidase; rh-a-l-
iduronidase produced in an engineered
CHO cell line)
Genzyme long-term enzyme
replacement therapy
in patients suffering
from mucopolysac-
charidosis
2003 (EU)
Regranex (rhPDGF produced
in S. cerevisiae)
Ortho-McNeil
Pharmaceuticals
(US) and Janssen-
Cilag (EU)
lower extremity
diabetic neuropathic
ulcers
1997 (US),
1999 (EU)
Vitravene (Fomivirsen; an antisense
oligonucleotide)
ISIS Pharmaceuti-
cals
treatment of cytome-
galovirus retinitis in
AIDS patients
1998 (US)
Ontak (rIL-2-diphtheria toxin fusion
protein which targets cells displaying
a surface IL-2 receptor)
Seragen/Ligand
Pharmaceuticals
cutaneous T cell
lymphoma
1999 (US)
Enbrel (rTNF receptorIgG fragment
fusion protein produced in CHO cells)
Immunex (US) and
Wyeth Europa (EU)
rheumatoid arthritis 1998 (US),
2000 (EU)
Osteogenic protein 1 (rhOsteogenic
protein-1: BMP-7 produced in CHO cells)
Howmedica (EU)
and Stryker (US)
treatment of non-
union of tibia
2001
(EU and
US)
Infuse (rhBMP2 produced in CHO cells) Medtronic Sofamor
Danek
promotes fusion of
vertebrae in lower
spine
2002 (US)
Inductos (dibotermin-a; rBone morpho-
genic protein-2 produced in CHO cells)
Genetics Institute treatment of acute
tibia fractures
2002 (EU)
Xigris [drotrecogin-a; rh activated protein
C produced in a mammalian (human)
cell line]
Eli Lilly severe sepsis 2001 (US),
2002 (EU)
polypeptide subunits. It is produced in a
CHO cell line. InductOs is used in skele-
tally mature patients for the treatment of
acute tibia fractures in adjunct to standard
care using fracture reduction and intrame-
dullary nail fixation. It is applied during
surgical procedure at the site of fracture.
The use of a bovine collagen sponge en-
sures retention of the active substance at
the site of the fracture for the time re-
quired for healing, with the matrix com-
pletely dissolving over time.
7
Products Approved for Veterinary Use
While the majority of pharmaceuticals pro-
duced by modern biotechnological means
are destined for human use, several veter-
inary biopharmaceuticals have also gained
approval (Table 10). One of the earliest
such examples is bovine somatotropin (re-
combinant bovine GH), used to boost the
milk yields of dairy cattle. The majority of
veterinary biopharmaceuticals, however,
are engineered vaccines. The importance
of effective vaccination to prevent rapid
spread of disease through high-density ani-
mal populations characteristic of modern
agricultural practice is obvious and most
vaccines are destined for use in agricultu-
rally important species. Porcillis Porcoli,
for example, is a multisubunit vaccine con-
taining a combination of recombinant E.
coli-derived adhesin proteins. These pro-
teins are essential for colonization of the
gut by pathogenic E. coli. Immunization of
sows effectively provides passive immunity
to progeny via colostrum for the first few
days of life, when piglets are particularly
susceptible to E. coli infections. Two
additional veterinary biopharmaceuticals,
which are particularly interesting, are
7 Products Approved for Veterinary Use 25
Table 10 Recombinant veterinary medicinal products approved in EU via the centralized application process
Porcilis Porcoli (combination vaccine
containing recombinant E. coli
adhesins)
Intervet active immunization of
sows
1996
Fevaxyn Pentofel (combination
vaccine containing recombinant
feline leukemia viral antigen as one
component)
Fort Dodge
Laboratories
immunization of cats
against various feline
pathogens
1997
Neocolipor (vaccine containing four
inactivated E. coli strains; two wild-
type strains expressing E. coli
adhesins F6 and F41, and two
recombinant strains, engineered to
express F4 and F5 adhesins)
Merial reduction of neonatal
enterotoxicosis of young
piglets caused by E. coli
strains expressing F4,
F5, F6 or F41 adhesins
1998
Porcilis AR-T DF (combination
vaccine containing a modified toxin
from Pasteurella multocida expressed
in E. coli)
Intervet reduction in clinical
signs of progressive
atrophic rhinitis in
piglets: oral
administration
2000
Ibraxion and Vibragen x, as discussed
below.
The advent of genetic engineering has
facilitated the development of engineered
vaccines capable of allowing subsequent
immunological differentiation between in-
fected and vaccinated animals. Using this
form of vaccination allows veterinary in-
spectors to tell if a seropositive animal has
simply been vaccinated (and is noninfec-
tious) or if it has been infected with the
wild-type pathogen (and is likely to be in-
fectious, thereby requiring treatment/isola-
tion).
Table 10 (continued)
Porcilis pesti (vaccine containing
recombinant classical swine fever virus
E
2
subunit antigen produced in an
insect cell baculovirus expression
system)
Intervet immunization of pigs
against classical swine fe-
ver
2000
Ibraxion (vaccine consisting of an
inactivated, BHV type 1 engineered
by removal of the viral glycoprotein
gE gene)
Merial active immunization of
cattle against infectious
bovine rhinotracheitis
2000
Bayovac CSF E2 (vaccine consisting
of recombinant classical swine fever
virus E2 subunit antigen produced
using a baculovirus vector system)
Bayer immunization of pigs
against classical swine
fever virus
2001
Eurifel FELV (vaccine consisting of
an engineered canarypox virus into
which the gag, env and a partial pol
gene of feline leukemia virus have
been inserted)
Merial Immunization of cats
against feline leukemia
virus
2000
Vibragen x (rFeline IFN-x) Virbac reduce mortality/clinical
signs of canine
parvovirusis
2001
Eurifel RCPFEVL (multicomponent
vaccine containing as one component
an engineered canarypox virus into
which the gag, env and a partial pol
gene of feline leukemia virus have
been inserted (see Eurifel FELV
above)
cats against viral
pathogens, including
feline leukemia virus
2002
Gallivac HVT IBD (live multic-
omponent vaccine containing as one
component an engineered herpes virus
of turkeys housing a gene coding for
the protective VP2 antigen of the
infectious bursal disease virus)
chickens against,
amongst others, the viral
causative agent of
infectious bursal disease
2002
Ibraxion is an example of such an engi-
neered vaccine. It is an engineered bovine
herpes virus (BHV) from which one struc-
tural gene (the gE gene) has been deleted.
BHV induces infectious bovine rhinotra-
cheitis, a condition characterized by losses
in animal production and abortions. Ibrax-
ion induces immunological protection in
cattle, but the serum of Ibraxion-vacci-
nated animals is devoid of anti-gE antibod-
ies, whereas infected animals will have
high titers of such antibodies (Fig. 3).
Vibragen x is IFN-x, a novel type 1
IFN. Like other type 1 IFNs, it displays
antiviral activity which is the basis of its
use in treating parvoviral infections, espe-
cially in young dogs for whom such an
infection can be fatal. Vibragen x is also
somewhat unusual in that it is manufac-
tured using insect-based biosynthesis oc-
curring in whole silkworms (Fig. 4). The
process entails the use of the silkworm nu-
clear polyhedrosis virus (NPV), engineered
to carry cDNA for feline IFN-x. Initial vi-
ral amplification is first undertaken to pro-
duce sufficient quantities of virus to seed
the process. Amplification is undertaken
by viral incubation with an insect cell line
(originated from Bombix mori) grown in
conventional culture flasks. The product is
then manufactured by rearing several
thousand silkworms on heat-treated, syn-
thetic chow in sterile cabinets. After 24
48 h each silkworm is inoculated with
NPV using an automatic microdispenser.
Five days later the silkworms are mechani-
cally incised and the acid-stable IFN is ex-
tracted from the body parts. Downstream
processing is more conventional, and em-
ploys both dye and metal affinity chroma-
tography to achieve product purification.
8
Likely Future Directions
The main aim of this Introduction is to
provide the reader with a snapshot of the
profile of biopharmaceutical products ap-
proved to date. How the sector will devel-
op over the next decade or two will likely
echo, at least in part, many of the in-
novations discussed within the remaining
chapters of this book. A summary over-
view of some such likely innovations and
directions is presented below.
Fig. 3 Diagrammatic representation of the altera-
tion made to the engineered BHV in order to pro-
duce the product Ibraxion. By means of genetic
engineering, the structural gene gE is deleted from
the genome. BHV induces infectious bovine rhino-
tracheitis, a condition characterized by losses in
animal production and abortions. Ibraxion induces
immunological protection in cattle, but the serum
of Ibraxion-vaccinated animals is devoid of anti-gE
antibodies, whereas infected animals will have
high titers of such antibodies.
8.1
What is in the Pipeline?
Globally, in excess of 500 candidate bio-
pharmaceuticals are undergoing clinical
evaluation. The Pharmaceutical Research
and Manufacturers of America (PhRMA),
which represents the US drug industry, es-
timates that some 371 biotech medicines
are undergoing trials in the US [27]. Of
these, around half (178) aim to treat can-
cer (see also Part II, Chapter 4), and other
notable target indications include infec-
tious diseases (47 products) (see also Part
VI, Chapter 3), autoimmune disorders (26
products) (see also Part V, Chapter 3), neu-
rological disorders (22 products) (see also
Part I, Chapter 14) and AIDS/HIV related
conditions (21 products) (see also Part II,
Chapters 7 and 8). The single largest cate-
Fig. 4 Overview of the manufacture of the veterin-
ary medicinal product Vibragen x. The process
entails inoculation of whole silkworms (grown on
synthetic food in pre-sterile cabinets) with an en-
gineered silkworm NPV housing the feline IFN-x
gene. Product extraction, purification and formula-
tion ensues.
gory is vaccines, of which there are 98 in
development (see also Part I, Chapter 7).
Fifty-three of these vaccines aim to treat or
prevent cancers, whereas an additional 29
aim to treat various infectious diseases, in-
cluding hepatitis and HIV. The second
largest product category is that of mono-
clonal/engineered antibodies (see also Part
IV, Chapter 16 and Part V, Chapters 1 and
2). Of the 75 such products in develop-
ment, 39 (52%) target cancers and 10 aim
to treat various autoimmune conditions,
most notably rheumatoid arthritis. The
number of IL- and antibody-based prod-
ucts in trials has increased modestly over
the past 23 years. The past few years has
also witnessed a significant decrease in the
number of growth factors and gene-thera-
py-based products undergoing clinical eval-
uation by PhRMA-associated companies, at
least (see also Part VI, Chapter 6). The lat-
ter reflects the continued difficulties asso-
ciated with making nucleic acid-based
products a therapeutic reality [28, 29] (see
also Part I, Chapters 69).
8.2
Alternative Production Systems
for Biopharmaceuticals
Essentially all recombinant therapeutic
proteins approved thus far are expressed
either in E. coli, S. cerevisiae (see also Part
IV, Chapters 12 and 13), or in an engi-
neered animal cell line (see also Part IV,
Chapters 1 and 4), hybridoma (mouse/hu-
man) cells (see also Part IV, Chapter 2) or
even human cells (see also Part IV, Chap-
ter 3).
Research continues into the develop-
ment of alternative production systems
and of particular note is the use of trans-
genic animals or plants (see also Part IV,
Chapter 5). A number of recombinant
therapeutic proteins (including a
1
-antitryp-
sin, a-glucosidase and antithrombin III)
have been successfully produced in trans-
genic animals, mainly in the milk of mice
(proof-of-concept stage) or goats (putative
production-scale systems) (see also Part IV,
Chapter 11). While this approach has prov-
en to be technically possible, a range of
problems has thus far delayed/prevented
approval of any product produced in this
manner, although ATryn
(antithrombin
III) will most likely be approved at the
time when this book will be published.
Difficulties have included modest/variable
production levels, regulatory issues and
cost. The major companies besides GTC
Biotherapeutics (US) sponsoring this
technology are PPL therapeutics (Scotland)
and Pharming (The Netherlands). Work
also continues on the development of
transgenic plant-based systems for the pro-
duction of, for example, oral vaccines or
other therapeutic proteins. Again, regula-
tory and cost issues are complicating fac-
tors, as are issues such as the significant
difference between glycosylation patterns
characteristic of plant versus animal cell-
based production systems. A comprehen-
sive overview is given by Knblein in two
excellent reviews [30, 31]. Due to the im-
portance of such emerging systems, this
book has dedicated a complete section to
alternative expressions systems for bio-
pharmaceuticals especially plant-based
expression systems (see also Part IV, Chap-
ter 6, Part IV, Chapters 7 and 8, Part IV,
Chapter 9 and Part IV, Chapter 10). The
section concludes with the engineering of
plant expression systems for abiotic stress
tolerance. By making plants tolerant for
high salt concentrations, heat and drought,
this might in the near future lead to grow-
ing plants in areas which today cannot be
used for agriculture at all.
8.3
Alternative Delivery Methods
for Biopharmaceuticals
Thus far, therapeutic proteins are invari-
ably administered parenterally. Drug ad-
ministration by nonparenteral means is
generally less invasive, requires less tech-
nical training and is normally associated
with improved patient compliance (see
also Part VI, Chapter 1, and Part VI, Chap-
ters 5 and 3). Some progress has also been
recorded relating to the development of
nonparenteral delivery routes for biophar-
maceuticals. Most prominent in this re-
gard is pulmonary delivery [32, 33]. Macro-
molecules are absorbed from the lung sur-
prisingly well, likely due to the lungs
large surface area, thin diffusional layer
and the presence of proteolytic inhibitors.
Nebulizer technology allows product deliv-
ery into the deep lung and drugs adsorbed
in this way avoid first-bypass metabolism
(see also Part VI, Chapter 4). Exubera is
the name given to an insulin product ad-
ministered by pulmonary means (see also
Part IV, Chapter 13). Developed by Pfizer,
Aventis and Nektar (see also Part VI,
Chapter 2), this product has completed
phase III clinical trials, although addi-
tional safety studies are currently being
undertaken.
8.4
The Advent of Generic Biopharmaceuticals?
Patent protection for many early biophar-
maceuticals, such as recombinant insulin,
EPO, hGH and IFN-a, is now nearing or
at an end (see also Part VIII, Chapter 3).
Most of these are blockbuster products
each commands annual sales well in ex-
cess of $ 1 billion and hence these repre-
sent attractive targets for the fledgling bio-
pharmaceutical generics industry [34] (see
also Part II, Chapter 4 and Part VIII,
Chapter 1). Companies producing generic
biopharmaceuticals include Sicor (Irvine,
CA), Ivax (Miami, FL), Dragon (Vancouver,
Canada), Genemedix (Suffolk, UK) and
BioGenerix (Mannheim, Germany). Sicor
already markets hGH and a-IFN-a in East-
ern Europe, whereas Genemedix markets
a recombinant colony-stimulating factor in
China and is soon to manufacture EPO
(see also Part VIII, Chapter 3). Major gen-
erics companies such as Teva, Sandoz and
Merck will also likely develop/consider de-
veloping biopharmaceutical portfolios.
However, the regulatory framework re-
quired to underpin generic biopharmaceu-
tical approvals within Europe and North
America is not yet finalized, although it
appears to be at a more advanced stage in
Europe as compared to the US (see also
Part VII, Chapter 4). The concept of simi-
lar biological medicinal products is one
now enshrined in the EU regulatory fra-
mework. Within the US the regulatory
terminology includes phrases such as fol-
low-on biologics and well-characterized
protein. Although generics will probably
be reviewed by regulators on a case-by-case
basis, substantial in vitro work as well as
some clinical data will almost certainly be
required to show comparability/product
equivalence.
8.5
Genomics and Proteomics
Most pharmaceutical companies have re-
search programs in genomics/proteomics.
The omics revolution was initially hailed
as a revolution in drug discovery. While
these modern technologies may well help
identify a host of putative new biopharma-
ceuticals (see also Part I, Chapters 4 and
5), they almost certainly will have a far
more significant impact upon identifying
new drug targets as well as disease diag-
nostic markers (see also Part I, Chapters 2
and 3 and Part V, Chapter 8).
8.6
Gene Therapy
There have been well in excess of 400
gene-therapy-based trials undertaken to
date. The vast majority have reported a dis-
appointing lack of efficacy. Safety concerns
have also been raised and these concerns
were heightened in 1999 when one partici-
pant in a gene-therapy trial died. Largely
prompted by this event, the US National
Institute of Health requested detailed in-
formation from a large number of trials
and uncovered several hundred reports of
serious trial-associated adverse events in
the process. In addition, allegations en-
sued that other gene-therapy trial deaths
went unreported/misreported, particularly
in trials using retroviral-based delivery vec-
tors. As a result, more stringent reporting
requirements were introduced.
Gene therapy did receive a much-needed
boost in 2002 when French scientists re-
ported that they had apparently corrected
severe combined immunodeficiency in a
number of children using a retroviral-
mediated protocol [35]. Severe combined
immunodeficiency is a genetic condition
caused by a deficiency of the enzyme
adenosine deaminase, which triggers se-
vere B and T lymphocyte dysfunction.
However, celebrations were halted when a
number of trial participants developed un-
controlled lymphoproliferation, a condition
similar to leukemia. This halted at least
temporarily several gene-therapy trials in
various regions of the world [36].
Up until then retroviruses had been
used as the delivery vectors in over 75% of
all gene-therapy clinical trials, mainly be-
cause their molecular biology was well un-
derstood, the efficiency of gene transfer to
sensitive cells was extremely high, subse-
quent gene expression is usually high and
high level stocks of replication-deficient
retroviral particles can be produced. De-
spite such undoubted advantages, retro-
viruses also display certain disadvantages
in the context of gene delivery, including
their ability to infect only actively replicat-
ing cells and the fact that their proviral
DNA integrates randomly into the host
chromosome [37].
The emphasis is now shifting somewhat
away from retroviruses and towards alter-
native viral as well as nonviral vectors.
Adenoviruses are receiving attention due
to their stability, easy manufacture, ability
to infect nondividing cells and their ability
to promote high-level gene expression (see
also Part I, Chapters 6 and 7). However,
this category of vector is not without its
own difficulties, as adenoviruses tend to
be highly immunogenic in humans, dis-
play broad cell specificity and the duration
of resultant gene expression can be transi-
ent. The most prominent nonviral vector
type remains liposome based [38] (see also
Part VI, Chapters 7 and 8).
While genetic diseases constitute an ob-
vious target for gene therapy, cancer re-
mains the major indication (see also Part
I, Chapter 1). A wide range of strategies
continue to be pursued in this regard, in-
cluding selected delivery of toxins/tumor-
suppressor genes/suicide genes into tumor
cells, modifying tumor cells to increase
their immunogenicity or modifying lym-
phocytes in order to enhance their antitu-
mor activity [39].
8.7
Antisense and RNA Interference (RNAi)
Antisense technology is based upon the
manufacture of short single-stranded
stretches of nucleic acids (DNA or RNA
based) or chemically modified versions
thereof (see also Part I, Chapter 8). The
nucleotide sequence specificity of these
antisense molecules allows them to bind
to specific gene or (more commonly)
mRNA sequences, thereby preventing
gene expression by blocking either tran-
scription or translation [40]. The therapeu-
tic rationale underlining this approach
stems from the fact that many diseases are
triggered or are exacerbated by inappropri-
ate expression/overexpression of specific
genes. Antisense, in principle, provides a
mechanism by which this can be blocked.
While the underlining concept is straight-
forward, like gene therapy, it is proving
more difficult to apply in practice. Major
difficulties have arisen in relation to prod-
uct nuclease sensitivity, product targeting,
delivery and cellular uptake.
Vitravene (fomivirsen sodium, ISIS
Pharmaceuticals; see Table 9) remains the
only antisense-based biopharmaceutical ap-
proved for general medical use (see also
Part III, Chapter 3). The product is a 21-
base phosphorothioate nucleotide that dis-
plays a base sequence complementary to
certain human cytomegaloviral mRNA
transcripts. Its administration inhibits viral
replication through an antisense mecha-
nism. Approved in the US in 1998 and in
the EU in 1999, the product is indicated
for the treatment of cytomegalovirus reti-
nitis by intraocular injection in AIDS pa-
tients. It was withdrawn from the EU mar-
ket in May 2002 for commercial reasons.
RNAi represents an alternative and
more recently pursued mechanism of
downregulating gene expression ([41] and
references therein) (see also Part II, Chap-
ter 8 and Part I, Chapter 1 and 10). The
RNAi pathway was first discovered in
plants, but it is now known to function in
most if not all eukaryotes. RNAi repre-
sents the sequence-specific post-transla-
tional inhibition of gene expression, in-
duced ultimately by double-stranded RNA
(dsRNA). Be it produced naturally or
synthesized in vitro and introduced into a
cell by researchers, the (sequence-specific)
dsRNA is then cleaved into short (20- to
25-nt) fragments. The RNA strands therein
are separated one is degraded and the
other binds to a cellular protein complex.
This strand will bind to target (comple-
mentary) mRNA, which is then cleaved by
an endonuclease within the complex. Be-
cause of its ability to downregulate gene
expression, RNAi technology has obvious
therapeutic potential, and initial therapeu-
tic targets of RNAi include viral infection,
neurological diseases and cancer therapy.
The synthesis of dsRNA displaying the de-
sired nucleotide sequence is straightfor-
ward. However, as in the case of additional
nucleic acid-based therapeutic approaches,
major technical hurdles remain to be over-
come before RNAi becomes a therapeutic
reality.
8.8
Stem Cell-based Therapies
The therapeutic application of stem cells
has long been a dream of medical
sciences, but recent discoveries and techni-
cal advances have brought this dream
much closer to being a reality. Stem cells
are usually defined as undifferentiated
cells capable of self-renewal, which can dif-
ferentiate into more than one specialized
cell type. Pluripotent stem cells are cap-
able of essentially differentiating into any
cell type, whereas multipotent stem cells,
often found (be it in low numbers) within
specific organs, give rise to lineage-re-
stricted, tissue-specific cell types (see also
Part I, Chapter 13). Human embryonic
stem cells, harvested from the inner mass
of the blastocyst, are the most convenient
source of pluripotential cells. A quantum
leap was taken in 2004 with the generation
of an unlimited source of human embryo-
nic stem cells by Woo Suk Hwang from
Seoul University. Hwang et al. were able
to obtain pluripotent embryonic stem cells
from somatic cell nuclear transfer of re-
programmed human adult cells (see also
Part I, Chapter 11). When cultured in the
presence of various specific growth factors
these pluripotential cells have be induced
to differentiate into various mature cell
types, including liver, hematopoietic cells,
neurons, pancreatic, skeletal and endothe-
lial cells. Therefore stem cells harbor the
potential to form replacements for dam-
aged/diseased body cells, tissue or even
entire organs (see also Part I, Chapters 12
and 15). This technology could eventually
give rise to cell-based therapies for various
neurological diseases, kidney, heart, lung
or other organ failure, diabetes, cardiac
damage, etc. The scope and limitations
currently underpinning stem cell technol-
ogy are well beyond the scope of Introduc-
tion, but will be discussed in the respec-
tive chapters in this book. The interested
reader is also referred elsewhere [4245]
for sources of additional information.
9
Concluding Remarks
Overall, the biopharmaceutical sector is
one that is now maturing rapidly. The fact
that biopharmaceuticals now generate in
excess of $ 30 billion revenue annually
from a zero starting point just over 20
years ago illustrates the medical and, in-
deed, commercial importance of these
drugs. Even more excitingly, continued ad-
vances in both pure and applied medical
research will fuel continued growth within
the biopharmaceutical sector for many
years to come.
Many of the concepts described in here
are explored in more detail in subsequent
chapters. It was a real pleasure for me to
write this Introduction and I am im-
pressed, because this book brings together
world-class contributions from world-class
scientists, drawn from both industry and
academia. This compilation is one of the
most comprehensive books published to
date in this area and it is a must-read for
everybody working in this field.
References
1 Chance, R. E., Frank, B. H. 1993. Research,
development, production and safety of biosyn-
thetic human insulin. Diabetes Care 16, 133
142.
2 Robinson, K. 2002. An industry comes of age.
Biopharm Int 15, 2024.
3 Walsh, G. 2003. Biopharmaceutical bench-
marks 2003. Nature Biotechnol 21, 865870.
4 Anonymous. 2004. The genesis of gendicine.
Biopharm Int 17, 4250.
5 Walsh, G. 2003. Biopharmaceuticals, Biochemis-
try and Biotechnology. Wiley, Chichester.
6 Crommelin, D., Sindelar, R. 2002. Pharmaceu-
tical Biotechnology, 2nd edn. Tailor & Francis,
London.
7 Kjeldsen, T. 2000. Yeast secretary expression
of insulin precursors. Appl Microbiol Biotechnol
54, 277286.
8 Woodrow, G. 1997. New Generation Vaccines.
Decker, New York.
9 http://www.refludan.com.
10 Hu, W., Peshwa, M. 1993. Mammalian cells
for pharmaceutical manufacturing. Am Soc
Microbiol News 59, 6568.
11 Foote, M., Boone, T. 1999. Biopharmaceutical
drug development: a case history. In Biophar-
maceuticals, An Industrial Perspective. Walsh G.,
Murphy, B. (eds.). Kluwer, Dordrecht, pp. 109
123.
12 http://www.betaseron.com.
13 Kaye, J. A. 1998. FDA licensure of NEUMEGA
to prevent severe chemotherapy-induced
thrombocytopenia. Stem Cells 16, 207223.
References 33
14 Kobata, A. 1992. Structure and function of the
sugar side-chains of glycoproteins. Eur J Bio-
chem 209, 483501.
15 Lyseng-Williamson, K. A., Perry, C. M. 2002.
Drotrecogin alfa (activated). Drugs 62, 617632.
16 Pfeifter, T. 1998. Expression of heterologous
proteins in stable insect cell culture. Curr
Opin Biotechnol 9, 518521.
17 Walsh, G. 2002. Proteins Biochemistry and Bio-
technology. Wiley, Chichester.
18 Aranha, H. 2001. Viral clearance strategies for
biopharmaceutical safety: part 1, general con-
siderations. Biopharm Int 14, 2835.
19 Waller, M., Kohnert, U. 1999. Reteplase, a re-
combinant plasminogen activator. In Biophar-
maceuticals, An Industrial Perspective. Walsh,
G., Murphy, B. (eds.). Kluwer, Dordrecht,
pp. 185217.
20 Kohler, G., Milstein, C. 1975. Continuous cul-
ture of fused cells secreting antibody of de-
fined specificity. Nature 256, 495497.
21 Kontermann, R. 2001. Antibody Engineering.
Springer, Berlin.
22 Breedveld, F. 2000. Therapeutic monoclonal
antibodies. Lancet 355, 735740.
23 Katre, N. 1993. The conjugation of proteins
with polyethylene glycol and other polymers
altering properties of proteins to enhance
their therapeutic potential. Adv Drug Deliv Rev
10, 91114.
24 Patel, K., McHutchison, J. 2001. Peginterferon
alpha-2b: a new approach to improving re-
sponse in hepatitis C patients. Expert Opin
Pharmacother 2, 13071315.
25 http://www.aranesp.com.
26 Hoppe, H. 2000. Cerezyme recombinant
protein treatment for Gauchers disease. J Bio-
technol 76, 259261.
27 Anonymous. 2003. New biotechnology medi-
cines in development. Available at
http://www.pharma.org.
28 Pfeifer, A., Verma, I. 2001. Gene therapy:
promises and problems. Annu Rev Genom
Hum Genet 2, 177211.
29 Lebedeva, I., Stein, C. 2001. Antisense oligo-
nucleotides: promise and reality. Annu Rev
Pharmacol Toxicol 41, 403419.
30 Knblein, J. 2004. Biopharmaceuticals ex-
pressed in plants a new era in the new
Millennium. In Applications in Pharmaceutical
Biotechnology. Mller, R., Kayser, O. (eds.).
Wiley-VCH, Weinheim, pp. 3556
31 Knblein, J. 2005. Plant-based expression of
biopharmaceuticals. In Encyclopedia of Molecu-
lar Cell Biology and Molecular Medicine, 2nd
edn. Meyers, R. A. (ed). Wiley, New York, vol.
10, pp. 489510.
32 Owens, D. R., Zinman, B., Bolli, G. 2003.
Alternative routes of insulin delivery. Diabetic
Med 20, 886898.
33 Patton, J. 1996. Mechanisms of macromole-
cule absorption by the lungs. Adv Drug Deliv
Rev 19, 336.
34 Griffiths, S. 2004. Betting on biogenerics. Nat
Rev Drug Disc 3, 197198.
35 Cavazzana-Calvo, M., Hacein-Bey, S., de Saint
Basile, G., Gross, F., Yvon, E., Nusbaum, P.,
Selz, F., Hue, C., Certain, S., Casanova, J. L.,
Bousso, P., Deist, F. L., Fischer, A. 2000. Gene
therapy of human severe combined immuno-
deficiency (SCID)-X1 disease. Science 288,
669672.
36 Gunzburg, W. H. 2003. Retroviral gene thera-
py where now? Trends Mol Med 9, 277278.
37 Blankenstein, T. 1999. Gene Therapy: Principles
and Applications. Birkhauser, Basel.
38 Schatzlein, A. 2001. Non-viral vectors in can-
cer gene therapy: principles and progress. An-
ticancer Drugs 12, 275304.
39 Shinohara, E. T., Lu, B., Hallahan, D. E. 2004.
The use of gene therapy in cancer research
and treatment. Technol Cancer Res Treat 3,
479490.
40 Crooke, S. (ed.). 2001. Antisense Drug Technol-
ogy. Decker, New York.
41 Jana, S., Chakraborty, C., Nandi, S., Deb, J. K.
2004. RNA interference: potential therapeutic
targets. Appl Microbiol Biotechnol 65, 649657.
42 Montanya, E. 2004. Islet- and stem-cell-based
tissue engineering in diabetes. Curr Opin Bio-
technol 15, 435440.
43 Sadiq, T. S., Gerber, D. A. 2004. Stem cells in
modern medicine: reality or myth? J Surg Res
122, 280291.
44 Mayhall, E. A., Paffett-Lugassy, N., Zon, L. I.
2004. The clinical potential of stem cells. Curr
Opin Cell Biol 16, 713720.
45 Gerecht-Nir, S., Itskovitz-Eldor, J. 2004. The
promise of human embryonic stem cells. Best
Pract Res Clin Obstet Gynecol 18, 843852.
Part I
Biopharmaceuticals Used in Molecular Medicine
ISBN: 3-527-31184-X
Abstract
In their 1985 Cell paper, Greider and
Blackburn announced the discovery of an
enzyme that extended the DNA at chromo-
some telomeres in the ciliate, Tetrahymena.
Since then, there has been an explosion of
knowledge about both the RNA and pro-
tein subunits of this unusual ribonucleo-
protein enzyme in organisms ranging
from the ciliates to yeast to humans. The
regulation of telomerase is now under-
stood to take place both at the level of syn-
thesis of the enzyme and via the state of
its substrate, the telomere itself. The roles
of telomerase in both cellular immortality
and cancer are vibrant areas of current re-
search.
Here we show that e.g. telomerases are an
attractive target for the development of an
anti-cancer drug. Other innovative ap-
proaches for the development of biopharma-
ceuticals against cancer (e.g. Herceptin) will
be presented in subsequent chapters (see also
Part I, Chapter 5; Part II, Chapters 5 and
6).
1.1
Introduction
It is unusual for an enzyme to be a topic
of widespread conversation. While millions
may marvel at the lower cholesterol levels
theyve achieved by taking a statin, few of
them know that the drug inhibits their
HMG-CoA reductase enzyme. Telomerase,
on the other hand, is connected to notions
of our mortality and longevity and has
been popularized by articles and books in-
cluding Merchants of Immortality [1].
The purpose of this review is to celebrate
the Greider and Blackburn paper that
started it all and then to highlight some of
the major advances that followed. My re-
view will be highly selective, covering only
about 1% of the hundreds of scientific pa-
pers written each year that deal with telo-
merase, and I refer the reader to other
recent reviews for a more comprehensive
treatment [2, 3].
37
ISBN: 3-527-31184-X
1
Beginning to Understand the End of the Chromosome
Thomas R. Cech
From Genome to Clinic Correlation Between Genes,
Diseases and Biopharmaceuticals
1.2
Telomere Terminal Transferase
Carol Greider, a graduate student in Liz
Blackburns group at the University of Ca-
lifornia, Berkeley, had chosen an ambitous
PhD thesis project: identify the molecular
entity responsible for replicating chromo-
some ends. There was a basis for thinking
that such an activity would exist: when lin-
ear DNA molecules from ciliated protozoa
were propagated in yeast, their ends were
extended not by the ciliate telomeric DNA
sequence (repeats of TTGGGG in Tetrahy-
mena or T
4
G
4
in hypotrichous ciliates), but
instead by the more heterogeneous G
13
T
yeast telomeric sequence [4, 5, 6]. One rea-
sonable interpretation was that the ciliate
telomeres served as seeds for addition of
a DNA sequence that was specified by a
nucleotide addition activity intrinsic to
yeast. Greiders project was to purify the
corresponding Tetrahymena enzyme.
The identification and characterization
of this new enzymatic activity was the sub-
ject of Greider and Blackburn [7]. The ac-
tivity added TTGGGG repeats, one nucleo-
tide at a time, to the ends of GT-rich prim-
ers that represented either the Tetrahymena
or the yeast telomeric sequence. This pa-
per marked the first appearance in the lit-
erature of the six nucleotide ladder of ex-
tension products that would appear in a
hundred subsequent papers a hallmark
of telomerase activity not just in Tetrahy-
mena, but in human extracts as well. The
authors made the reasonable proposal that
the activity might be related to known ter-
minal transferases, such as the enzyme
that adds CCA to the 3' ends of transfer
RNAs. The real nature of the enzyme
turned out to be much more novel.
1.3
Telomerase Contains an Essential RNA
In the following years, Greider and Black-
burn found that the enzyme (now called
telomerase) was even much more interest-
ing than one might have thought. It con-
tained an essential RNA component, a por-
tion of which served as a template to spe-
cify the sequence that was added to the
chromosome end [8], Fig. 1.1. The rules
were those of Watson-Crick base-pairing;
Cs and As in the RNA template specified
Gs and Ts, respectively, in the sequence
Fig. 1.1 Telomerase in action.
The RNA subunit (purple) has
the template sequence of Tetrahy-
mena telomerase. Proteins in-
clude the catalytic subunit, TERT
(yellow), and additional less-con-
served proteins (orange).
Reprinted with permission from
Cech [9]. Illustration by
K. Sutliff. Copyright 1994 AAAS.
that was laid down at chromosome ends.
Indeed, the formal proof of this model
was provided in an elegant paper from
Blackburns lab, in which site-specific mu-
tagenesis of nucleotides in the RNA tem-
plate lead to the deposition of complemen-
tary nucleotides in Tetrahymena telomeres
[10].
Subsequently, Greiders group at Cold
Spring Harbor Laboratory collaborated
with scientists at Geron Corporation to
identify and sequence the telomerase
RNAs from mouse and human the ribo-
nucleoprotein nature of telomerase was
general [11]. She also worked with Ron
DePinho to construct a mouse knockout
for the RNA. The homozygous null mice
were remarkable in that they were viable
for six generations [12]. The explanation is
that the mice start with long telomeres
(10160 kb); the failure to replicate those
telomeres leads to gradual loss of terminal
DNA sequences, but it takes many cell di-
visions before the shortest telomeres reach
a critical length. Embryonic fibroblasts cul-
tured from the fourth generation of these
mice onward showed aneuploidy and chro-
mosome end-to-end fusions, indicative of
failure to cap chromosome ends.
1.4
Finally, the Protein:
Telomerase Reverse Transcriptase
Greider and Blackburn [7] provided indi-
rect evidence that telomerase contained at
least one essential protein component,
which presumably provided the catalytic
center for nucleotide addition. The protein
proved elusive, however, and it took ten
years for two different approaches to con-
verge on its identification. Joachim
Lingner purified active telomerase from
Euplotes, a ciliated protozoan with an extra-
ordinary number of telomeres per macro-
nucleus (ca. 10
8
) and a correspondingly
high dose of telomerase [13]. It contained
the telomerase RNA and two proteins,
p123 and p43. At the same time, a genetic
screen in yeast in Vicki Lundblads lab
identified three genes whose deletion re-
sulted in an EST (Ever Shorter Telomeres)
phenotype [14]. It turned out that the best
sequence match to Euplotes p123 was
Lundblads Est2p.
Better yet, both p123 and Est2p con-
tained the amino acid sequence hallmarks
of reverse transcriptases. It had long been
thought that telomerase resembled a re-
verse transcriptase (RT), in that it synthe-
sized DNA using an RNA template. It now
appeared that it was directly related to
other RTs in terms of its protein structure
and evolution. The critical evidence came
from a collaboration between the Lundblad
and Cech groups, in which mutations of
amino acids implicated in RT activity were
shown to eliminate telomerase activity
both in vivo and in vitro [15].
As often happens in these days of the
human genome project, the human ver-
sion (hTERT) was found shortly thereafter
[16, 17]. It was expressed in a variety of
transformed cells but not detectable in pri-
mary cultures of human somatic cells, al-
ready giving a simple answer as to why tel-
omerase activity was deficient in somatic
cells. Thus, over a short time span, we
went from having no telomerase protein
to a whole family of TERTs (Telomerase
Reverse Transcriptases).
TERT is now known to have several
functions. The reverse transcriptase mo-
tifs, present in the C-terminal half of the
protein, provide the active site for catalysis.
Several conserved amino acid sequence
motifs in the N-terminal half rivet the
RNA component to the protein, assuring
maintenance of a stable RNP while allow-
1.4 Finally, the Protein: Telomerase Reverse Transcriptase 39
ing the template to move through the ac-
tive site [18], Fig. 1.2. Movement is essen-
tial, as a single active site (the triangle in
Fig. 1.2) must accommodate addition of
multiple nucleotides, after which transloca-
tion of the template relative to the DNA
product is necessary for multiple rounds
of addition. Finally, the very N terminus of
yeast TERT recruits another telomerase
subunit, Est3p, to the complex [19], and
additional protein-protein interactions may
remain to be discovered.
1.5
Current Picture of Telomerase
We currently view telomerase as composed
of an RNA molecule with a well-defined
secondary structure, best characterized in
ciliates and vertebrates [20]; the conserved
TERT catalytic subunit; and a number of
additional protein subunits, only some of
which are conserved phylogenetically.
Three classes of proteins in addition to
TERT are now known to be associated with
various telomerases. Est1p was first found
in yeast [21]; it is essential for activity in
vivo, but seems entirely dispensable for en-
zyme activity per se as judged by in vitro
assays. Est1p interacts directly with the
yeast telomere DNA end binding protein,
Cdc13p [22]; this interaction appears to re-
cruit telomerase to the chromosome end
[22] and somehow activate telomerase that
is already associated with the telomere
[23]. A human Est1 ortholog, EST1A, is as-
sociated with most or all active telomerase
in human cell extracts and is involved,
either directly or indirectly, in chromo-
some end-capping and telomere elongation
[24, 25]. Another yeast subunit, Est3p, is
similarly important for activity in vivo but
not in vitro, with its specific function un-
known.
The two subunits of the Ku heterodimer
comprise the second class of telomerase
proteins. Ku is responsible for nonhomolo-
gous end-joining of broken chromosomes,
and initially it appeared odd or perhaps
even dangerous that it would be telomere
associated; after all, telomeres protect chro-
mosome ends from fusion events that re-
sult in genomic instability. A solution to
this conundrum was recently provided by
Stellwagen et al. [26], who showed that Ku
binds directly to telomerase RNA and pro-
motes the de novo addition of telomeres to
broken chromosome ends, thereby helping
Fig. 1.2 Telomerase reaction cycle. The RNA sub-
unit is held into the complex by interactions with
the N-terminal domain of TERT. This leaves the
template (CA-containing sequence) free to move
through the RT domain so that a single active site
(triangle) can catalyze nucleotide addition at mul-
tiple positions.
heal DNA damage by capping the broken
end with telomeric DNA.
Finally, a large variety of proteins con-
tribute to the assembly and maturation of
the telomerase RNP, and these vary much
more in evolution than TERT and the
other proteins listed above. Budding yeast
telomerase RNA is an RNA polymerase II
transcript, and its intracellular transport
and assembly are mediated by the same
Sm proteins found in the small nuclear
RNPs involved in RNA splicing [27]. Cur-
rent evidence suggests that the RNA may
be made in the nucleus, exported to the
cytoplasm to pick up protein components,
and then reimported into the nucleus
where it functions [28, 29]. Human telo-
merase RNA, also a pol II transcript, has a
snoRNP (small nucleolar RNP) domain,
appears to be matured in the nucleolus,
and binds dyskerin and other snoRNP pro-
teins [30, 31]. Defects in the RNA or the
dyskerin protein that interrupt this ma-
turation can lead to a human disease, dys-
keratosis congenita [30, 32]. Ciliate telo-
merase RNA is transcribed instead by pol
III and, at least in Euplotes, is bound by a
telomerase-specific La-motif protein p43
that may shepherd its maturation or con-
fer nuclear localization [33].
Returning to the RNA component, it is
now seen to provide much more than the
template. One class of additional functions
is to provide specific binding sites for
many of the proteins listed above. All telo-
merases exist as stable RNPs, and the
RNA sequences responsible for TERT
binding have been identified in several
organisms [3436]. In yeast, the Ku pro-
tein binds to one RNA secondary structure
element [26] and Est1p to a separate
bulged stem [37] (Fig. 1.3). RNA binding
sites for the yeast Sm proteins and the hu-
man snoRNP proteins have similarly been
defined. In the second class of function,
the RNA may be acting directly to promote
a specific feature of catalysis. Clear exam-
ples are the base-paired RNA elements
that form template boundaries to termi-
nate each cycle of reverse transcription in
yeast and human telomerase [38, 39].
One key question about telomerase RNA
concerns how the template portion is iden-
tified by TERT. After all, the RNA subunits
1.5 Current Picture of Telomerase 41
Fig. 1.3 Telomerase RNA
functions. The S. cerevisiae
telomerase RNA (1.2 kb) func-
tions in part to bring proteins
into the complex. Curved
lines represent regions where
the RNA secondary structure
has not yet been reported.
contain hundreds or even a thousand nu-
cleotides, depending on the organism, and
the template is not the only single-
stranded region in the RNA that could in
principle make a few base pairs with a
DNA primer and be reverse transcribed.
This problem has been most successfully
tackled in Tetrahymena, where a short
template-recognition sequence element
directs the use of 5' adjacent nucleotides
as the template for DNA synthesis [40].
Whether this template-recognition element
directly binds to TERT or interacts with
another portion of the RNA remains a
question for future research.
The yeast and human telomerases have
been observed as dimers containing two
functionally interacting RNA molecules
[41]. Thus, telomerase and retroviruses re-
semble each other, not just in their reverse
transcriptase proteins, but also in their
packaging of their RNA template as a
dimer. Because recombinant Tetrahymena
telomerase is active as a monomer (one
RNA+one TERT), dimerization is not al-
ways required for core enzymatic activity
[42].
1.6
Regulation of Telomerase
In the most general sense, telomere length
either is maintained at a steady-state distri-
bution or undergoes progressive shorten-
ing or lengthening depending on at least
two considerations: the level of the telo-
merase RNP and the state of the telomere
itself.
Human telomerase is regulated during
development by the first of these factors
telomerase expression is dramatically re-
duced in many somatic cells during em-
bryonic development, and therefore chro-
mosome ends shrink with successive cell
divisions [43]. In these cells, the limiting
component is hTERT, and the transcrip-
tional repression of the hTERT gene leads
to a loss of telomerase activity. An uniden-
tified repressor encoded on chromosome 3
controls the state of hTERT chromatin,
leading to transcriptional silencing [44].
Specifically, three tumor suppressor path-
ways have been identified as negative reg-
ulators of hTERT transcription: Mad1, a re-
pressor of c-Myc; TGF-b, acting through
SIP1; and Menin, binding directly to the
hTERT promoter [45]. Human cells that re-
tain readily detectable telomerase activity
include some proliferating epithelial cells,
lymphocytes, and testis. Stem cells have
weak telomerase activity (reviewed by Col-
lins and Mitchell, 2002). Even in somatic
cells where the level of telomerase activity
is undetectable by standard assays, one
needs to be aware of the limit of detection.
Recently, immunopurification has been
used to reveal that there is in fact some ex-
pression of hTERT and telomerase activity
in cycling human fibroblasts, and that this
low level of activity has biological conse-
quences [46].
Some cells that lack telomerase activity,
on the other hand, still have a high level
of hTERT transcription. In these cases,
regulation at the level of alternative splic-
ing leads to skipping of exons that encode
reverse transcriptase function, so any
translation product would not give an
active enzyme [47].
In addition to the developmental regula-
tion mentioned above, telomere length
regulation in all organisms from yeast to
human involves the accessibility of the
telomere to telomerase. This appears to oc-
cur at four different levels, as follows:
(1) Double-stranded telomeric DNA
binding proteins such as Rap1p in bud-
ding yeast are involved in telomere length
regulation [48, 49]. Data support the pro-
tein counting model, in which Rap1p
binds Rif1p and Rif2p to nucleate the for-
mation of a folded chromatin structure at
the telomere, thereby preventing access by
telomerase. As the telomere shortens due
to incomplete replication, the number of
protein binding sites decreases and the
chromatin opens up to restore access to
telomerase. The human telomeric dsDNA
binding proteins, TRF1 and TRF2 [50],
may act by a similar mechanism, TRF1 re-
cruits TIN2 [51] and TRF2 recruits hRAP1
[52] through protein-protein interactions.
Similarly, in fission yeast, the dsDNA
binding protein Taz1 recruits Rap1 and
Rif1 [53]. Surprisingly, some aspects of
telomere chromatin structure and func-
tion, including the binding of Taz1, are
maintained in circular chromosomes that
have no telomeric repeats [54].
(2) Long telomeres, including those in
human cells, can form a t loop structure
in which the entire telomeric DNA forms
a large circle; the 3' single-stranded DNA
tail invades the double-stranded telomeric
DNA to form a D loop [55]. The t loop pre-
sumably provides chromosome end protec-
tion and also renders the DNA terminus
inaccessible to telomerase. Double-
stranded DNA binding proteins such as
TRF1 and 2 might exert their effects on
telomere length regulation in part by mod-
ulating t loop formation.
(3) Proteins that bind the 3' single-
stranded DNA tail are involved in regula-
tion of telomerase since the 3' end cannot
simultaneously bind the protein and the
alignment region of the telomerase RNA.
The role of yeast Cdc13p in recruiting telo-
merase was described above. The Protec-
tion of Telomeres (POT1) protein appears
to be the analog of Cdc13 in fission yeast,
plants, mice, and humans [56]. A recent X-
ray structure shows the ssDNA compacted
and sequestered within the protein
(Fig. 1.4) [57], consistend with human
POT1 acting as a repressor of telomerase
[58]. Under other conditions, human
POT1 can stimulate telomere elongation
[59], perhaps in analogy to yeast Cdc13p.
Telomeric single-stranded DNA binding
proteins may also control the action of nu-
cleases that generate the G strand over-
hangs at chromosome ends [60].
(4) Single-stranded telomeric DNA tails
also become resistant to telomerase exten-
sion when they fold into quadruplex
1.6 Regulation of Telomerase 43
Fig. 1.4 Pot1 (protection of telomeres) protein
from S. pombe binds single-stranded telomeric
DNA (GGTTAC). Two views of the X-ray crystal
structure of the complex are shown (Lei et al.,
2003). The six bases are bound as three stacked
pairs; arrows indicate the 5'-most pair (GG).
structures, a proclivity of guanine-rich se-
quences. The extent to which this occurs
in vivo is unknown. However, small mole-
cules that bind to quadruplex structures
can push the equilibrium toward this
folded form, providing a credible approach
to telomerase inhibition [61].
1.7
Cellular Immortality
Telomeres shorten during serial passage of
human fibroblasts in vitro [62]. Early pro-
posals that telomere length determines the
number of cell divisions a cell can under-
go the Hayflick Limit were based on
such correlation between telomere length
and proliferative potential. The availability
of hTERT allowed a direct test of this pro-
posal. When hTERT was transfected into
fibroblasts or retinal epithelial cells, they
had a greatly extended lifespan, apparently
limitless, while control cells transfected
with empty vector underwent senescence
at the Hayflick Limit as expected [63]. This
apparent immortalization different from
oncogenic transformation in that the
hTERT-transfected cells did not develop
chromosome abnormalities, were unable
to grow on soft agar, and were not tumori-
genic. T lymphocytes also achieved dra-
matic extension of their replicative lifespan
upon ectopic expression of hTERT [64, 65].
Mammary epithelial cells and keratino-
cytes, on the other hand, required inactiva-
tion of the Rb/p16 tumor suppressor path-
way in addition to activation of hTERT in
order to achieve extended lifespan [66].
This simple picture repression of hu-
man telomerase initiates telomere shorten-
ing, telomere length then serves as a yard-
stick of proliferative potential now
appears incomplete. For example, overex-
pression of TRF2 in primary human fibro-
blasts uncouples telomere shortening from
senescence [67]. Moreover, dividing pri-
mary human fibroblasts, which show pro-
gressive telomere shortening, nevertheless
have recently been shown to have low lev-
els of hTERT expression and telomerase
activity. Disruption of this activity by ecto-
pic expression of a catalytically inactive
mutant of hTERT (DN-hTERT) or by RNA
interference (RNAi) leads to premature se-
nescence [46]. The phenomenon of RNAi
and its potential as biopharmaceutical will be
discussed in details by John Rossi and also
Anastasia Khvorova (see also Part I, Chapter
10; Part II, Chapter 8).
1.8
Cancer
Human cancers are invariably associated
with activation of some mechanism to
maintain telomere length: approximately
8590% show reactivation of telomerase,
while the remainder maintain telomeres
by ALT (alternative lengthening of telo-
meres), which occurs by exchange of se-
quences between telomeres [68]. Hahn et
al. [69, 70] have shown that one pathway
to transformation of cultured human cells
involves three steps: activation of prolifera-
tion, e.g., induced by expression of a mu-
tant ras oncogene and the SV40 small t
antigen; inactivation of tumor suppressors
p53 and Rb; and activation of telomerase
by expression of hTERT. This differs from
the situation with rodent cells, which can
be transformed by the first two events
alone. Lin and Elledge (2003) achieved
transformation of human cells by a
slightly different pathway, inactivating the
hTERT repressor, Menin, instead of ex-
pressing hTERT ectopically. Thus, telomer-
ase activation may not be just a marker for
neoplastic growth in humans, but a causal
event. This makes telomerase an attractive
target for pharmaceutical development of
anti-cancer chemotherapeutics.
As an initial proof of principle of the
usefulness of telomerase inhibition, two
groups have shown that ectopic expression
of DN (dominant-negative) mutants of
hTERT in transformed human cells lead to
growth inhibition and apoptosis [71, 72]. A
very different result was obtained with a
synthetic non-nucleoside drug candidate
that inhibits telomerase. The drug induces
senescence rather than apoptosis in hu-
man cancer cells, and the inhibition of
both cell growth and telomerase are fully
reversible upon removal of the drug [73].
Although the difference in results could
be due to the different cell lines used, it is
also possible that the way in which telo-
merase is inhibited is the decisive factor.
DN-TERT may inhibit by titrating telomer-
ase or telomere components from the
chromosome end, thereby perturbing cap-
ping, while the small molecule inhibitor
may simply decrease the catalytic activity
of the telomerase enzyme without perturb-
ing the amount or location of telomerase
or its protein-protein contacts. For this rea-
son, it will be interesting to test whether
small molecule inhibitors of catalytic activ-
ity are less toxic for normal cells than DN-
TERT or RNAi, which may be more likely
to perturb chromosome capping. The great
potential of siRNAs as biopharmaceuticals
also for other indications will be discussed in
depth elsewhere in Modern Biopharmaceuti-
cals.
Acknowledgments
I thank Bod Weinberg, Bill Hahn, and Car-
ol Greider for helpful comments and Ming
Lei and David Zappulla for preparation of
illustrations.
Reprint with courtesy by Cell, Vol. 116,
273279, January 23, 2004, Copyright
2004 by Cell Press
References
1 Hall, S. S. (2003) Merchants of Immortality
(New York: Houghton Miffin Company).
2 Blackburn, E. H. (2001). Switching and signal-
ing at the telomere. Cell 106, 661673.
3 Collins, K., and Mitchell, J. R. (2002). Telo-
merase in the human organism. Oncogene
21, 564579.
4 Szostak, J. W., and Blackburn, E. H. (1982).
Cloning yeast telomeres on linear DNA plas-
mid vectors. Cell 29, 245255.
5 Pluta, A. F., Cani, G. M., Spear, B. B., and Za-
kian, V. A. (1984). Elaboration of telomeres in
yeast: recognition and modification of termini
from Oxytricha macronuclear DNA. Proc.
Natl. Acad. Sci. USA 81, 14751479.
6 Shampay, J., Szostak, J. W., and Blackburn,
E. H. (1984). DNA sequences of telomeres
maintained in yeast. Nature 310, 154157.
7 Greider, C. W., and Blackburn, E. H. (1985).
Identification of a specific telomere terminal
transferase activity in Tetrahymena extracts.
Cell 43, 405413.
8 Greider, C. W., and Blackburn, E. H. (1989). A
telomeric sequence in the RNA of Tetrahyme-
na telomerase required for telomere repeat
synthesis. Nature 337, 331337.
9 Cech, T. R. (1994). Chromosome end games.
Science 266, 387388.
10 Yu, G. L., Bradley, J. D., Attardi, L. D., and
Blackburn, E. H. (1990). In vivo alteration of
telomere sequences and senescence caused by
mutated Tetrahymena telomerase RNAs. Na-
ture 344, 126132.
11 Feng, J., Funk, W. D., Wang, S., Weinrich,
S. L., Avilion, A. A., Chiu, C., Adams, R. R.,
Chang, E., Allsopp, R. C., Yu, J., et al. (1995).
The RNA component of human telomerase.
Science 269, 12361241.
12 Blasco, M. A., Lee, H. W., Hande, M. P., Sam-
per, E., Lansdorp, P. M., DePinho, R. A., and
Greider, C. W. (1997). Telomere shortening
and tumor formation by mouse cells lacking
telomerase RNA. Cell 91, 2534.
13 Lingner, J., and Cech, T. R. (1996). Purification
of telomerase from Euplotes aediculatus: Re-
References 45
quirement of a primer 3' overhang. Proc. Natl.
Acad. Sci. USA 93, 1071210717.
14 Lendvay, T. S., Morris, D. K., Sah, J., Balasu-
bramanian, B., and Lundblad, V. (1996). Se-
nescence mutants of Saccharomyces cerevisiae
with a defect in telomere replication identify
three additional EST genes. Genetics 144,
13991412.
15 Lingner, J., Hughes, T. R., Shevchenko, A.,
Mann, M., Lundblad, V., and Cech, T. R.
(1997). Reverse transcriptase motifs in the cat-
alytic subunit of telomerase. Science 276,
561567.
16 Meyerson, M., Counter, C. M., Eaton, E. N., El-
lisen, L. W., Steiner, P., Caddle, S. D., Ziaugra,
L., Beijersbergen, R. L., Davidoff, M. J., Liu,
Q., et al. (1997). hEST2, the putative human
telomerase catalytic subunit gene, is up-regu-
lated in tumor cells and during immortaliza-
tion. Cell 90, 785795.
17 Nakamura, T. M., Morin, G. B., Chapman,
K. B., Weinrich, S. L., Andrews, W. H., Lingner,
J., Harley, C. B., and Cech, T. R. (1997). Telo-
merase catalytic subunit homologs from fis-
sion yeast and human. Science 277, 955959.
18 Bryan, T. M., Goodrich, K. J., and Cech, T. R.
(2000). Telomerase RNA bound by protein
motifs specific to telomerase reverse transcrip-
tase. Mol. Cell 6, 493499.
19 Friedman, K. L., Heit, J. J., Long, D. L., and
Cech, T. R. (2003). N-terminal domain of yeast
telomerase reverse transcriptase: recruitment
of Est3p to the telomerase complex. Mol. Biol.
Cell 14, 113.
20 Chen, J.-L., Blasco, M. A., and Greider, C. W.
(2000). Secondary structure of vertebrate telo-
merase RNA. Cell 100, 503514.
21 Lundblad, V., and Szostak, J. W. (1989). A mu-
tant with a defect in telomere elongation leads
to senescence in yeast. Cell 57, 633643.
22 Pennock, E., Buckley, K., and Lundblad, V.
(2001). Cdc13 delivers separate complexes to
the telomere for end protection and replica-
tion. Cell 104, 387396.
23 Taggart, A. K. P., Teng, S.-C., and Zakian, V. A.
(2002). Est1p as a cell cycle-regulated activator
of telomere-bound telomerase. Science 297,
10231026.
24 Reichenbach, P., Hoss, M., Azzalin, C. M.,
Nabholz, M., Bucher, P., and Lingner, J. (2003).
A human homolog of yeast est1 associates with
telomerase and uncaps chromosome ends
when overexpressed. Curr. Biol. 13, 568574.
25 Snow, B. E., Erdmann, N., Cruickshank, J.,
Goldman, H., Gill, R. M., Robinson, M. O.,
and Harrington, L. (2003). Functional conser-
vation of the telomerase protein est1p in hu-
mans. Curr. Biol. 13, 698704.
26 Stellwagen, A. E., Haimberger, Z. W., Veath,
J. R., and Gottschling, D. E. (2003). Ku inter-
acts with telomerase RNA to promote telo-
mere addition at native and broken chromo-
some ends. Genes Dev. 17, 111.
27 Seto, A. G., Zaug, A. J., Sobel, S. G., Wolin,
S. L., and Cech, T. R. (1999). Saccharomyces cer-
evisiae telomerase in an Sm small nuclear ri-
bonucleoprotein particle. Nature 401, 177180.
28 Ferrezuelo, F., Steiner, B., Aldea, M., and
Futcher, B. (2002). Biogenesis of yeast telo-
merase depends on the importin Mtr10. Mol.
Cell. Biol. 22, 60466055.
29 Teixeira, M. T., Frstermann, K., Gasser, S. M.,
and Lingner, J. (2002). Intracellular trafficking
of yeast telomerase components. EMBO
Rep. 3, 652659.
30 Mitchell, J. R., and Collins, K. (2000) Human
telomerase activation requires two indepen-
dent interactions between telomerase RNA
and telomerase reverse transcriptase in vivo
and in vitro. Mol. Cell 6, 361371.
31 Pogacic, V., Dragon, F., and Filipowicz, W.
(2000). Human H/ACA small nucleolar RNPs
and telomerase share evolutionarily conserved
proteins NHP2 and NOP10. Mol. Cell. Biol.
20, 90289040.
32 Vulliamy, T., Marrone, A., Goldman, F., Dear-
love, A., Bessler, M., Mason, P. J., and Dokal,
J. (2001). The RNA component of telomerase
is mutated in autosomal dominant dyskerato-
sis congenita. Nature 413, 432435.
33 Aigner, S., Postberg, J., Lipps, H. J., and Cech,
T. R. (2003). The Euplotes La motif protein p43
has properties of a telomerase-specific sub-
unit. Biochemistry 42, 57365747.
34 Mitchell, J. R., Wood, E., and Collins, K.
(1999). A telomerase component is defective
in the human disease dyskeratosis congenita.
Nature 402, 551555.
35 Chen, J.-L., Opperman, K. K., and Greider,
C. W. (2002). A critical stem-loop structure in
the CR4CR5 domain of mammalian telomer-
ase RNA. Nucleic Acids Res. 30, 592597.
36 Livengood, A. J., Zaug, A. J., and Cech, T. R.
(2002). Essential regions of Saccharomyces cere-
visiae telomerase RNA: separate elements for
Est1p and Est2p interaction. Mol. Cell. Biol.
22, 23662374.
37 Seto, A. G., Livengood, A. J., Tzfati, Y., Black-
burn, E. H., and Cech, T. R. (2002). A bulged
stem tethers Est1p to telomerase RNA in bud-
ding yeast. Genes Dev. 16, 28002812.
38 Tzfati, Y., Fulton, T. B., Roy, J., and Blackburn,
E. H. (2000). Template boundary in a yeast tel-
omerase specified by RNA structure. Science
288, 863867.
39 Chen, J.-L., and Greider, C. W. (2003). Tem-
plate boundary definition in mammalian telo-
merase. Genes Dev. 27, 27472752.
40 Miller, M. C., and Collins, K. (2002). Telo-
merase recognizes its template by using an
adjacent RNA motif. Proc. Natl. Acad. Sci.
USA 99, 65856590.
41 Ly, H., Xu, L., Rivera, M.A., Parslow, T. G.,
and Blackburn, E. H. (2003). A role for a novel
trans-pseudoknot RNA-RNA-interaction in
the functional dimerization of human telo-
merase. Genes Dev. 17, 10781083.
42 Bryan, T. M., Goodrich, K. J., and Cech, T. R.
(2003). Tetrahymena telomerase is active as a
monomer. Mol. Biol. Cell 14, 47944804.
43 Wright, W. E., Piatyszek, M. A., Rainey, W. E.,
Byrd, W., and Shay, J. W. (1996). Telomerase
activity in human germline and embryonic tis-
sues and cells. Dev. Genet. 18, 173179.
44 Szutorisz, H., Lingner, J., Cuthbert, A. P.,
Trott, D. A., Newbold, R. F., and Nabholz, M.
(2003). A chromosome 3-encoded repressor of
the human telomerase reverse transcriptase
(hTERT) gene controls the state of hTERT
chromatin. Cancer Res. 63, 689695.
45 Lin, S.-Y., and Elledge, S. J. (2003). Multiple
tumor suppressor pathways negatively regu-
late telomerase. Cell 113, 881889.
46 Masutomi, K., Yu, E. Y., Khurts, S., Ben-Por-
ath, I., Currier, J. L., Metz, G. B., Brooks,
M. W., Kaneko, S., Murakami, S., DeCaprio,
J. A., et al. (2003). Telomerase maintains telo-
mere structure in normal human cells. Cell
114, 241253.
47 Ulaner, G. A., Hu, J., Vu, T. H., Giudice, L. C.,
and Hoffman, A. R. (1998). Telomerase activity
in human development is regulated by hTERT
transcription and by alternate splicing. Cancer
Res. 58, 41684172.
48 Marcand, S., Gilson, E., and Shore, D. (1997).
A protein-counting mechanism for telomere
length regulation in yeast. Science 275, 986
990.
49 Ray, A., and Runge, K. W. (1999). The yeast
telomere length counting machinery is sensi-
tive to sequences at the telomere-nontelomere
junction. Mol. Cell. Biol. 19, 3145.
50 Smogorzewska, A., van Steensel, B., Bianchi,
A., Oelmann, S., Schaefer, M. R., Schnapp, G.,
and de Lange, T. (2000). Control of human tel-
omere length by TRF1 and TRF2. Mol. Cell.
Biol. 20, 16591668.
51 Kim, S. H., Kaminker, P., and Campisi, J.
(1999). TIN2, a new regulator of telomere
length in human cells. Nat. Genet. 23, 405
412.
52 Li, B., Oestreich, S., and de Lange, T. (2000).
Identification of human Rap1: implications
for telomere evolution. Cell 101, 471483.
53 Kanoh, J., and Ishikawa, F. (2001). spRap1
and spRif1, recruited to telomeres by Taz1,
are essential for telomere function in fission
yeast. Curr. Biol. 11, 16241630.
54 Sadaie, M., Naito, T., and Ishikawa, F. (2003).
Stable inheritance of telomere chromatin
structure and function in the absence of telo-
meric repeats. Genes Dev. 17, 22712282.
55 Griffith, J. D., Comeau, L., Rosenfield, S.,
Stansel, R. M., Bianchi, A., Moss, H., and de
Lange, T. (1999). Mammalian telomeres end
in a large duplex loop. Cell 97, 503514.
56 Baumann, P., and Cech, T. R. (2001). Pot1, the
putative telomere end-binding protein in fis-
sion yeast and humans. Science 292, 1171
1175.
57 Lei, M., Podell, E., Baumann, P., and Cech,
T. R. (2003). DNA self-recognition in the struc-
ture of Pot1 bound to telomeric single-
stranded DNA. Nature 426, 198203.
58 Loayza, D., and de Lange, T. (2003). POT1 as
a terminal transducer of TRF1 telomere
length control. Nature 424, 10131018.
59 Colgin, L. M., Baran, K., Baumann, P., Cech,
T. R., and Reddel, R. R. (2003). Human POT1
is a positive regulator of telomere length.
Curr. Biol. 13, 942946.
60 Jacob, N. K., Kirk, K. E., and Price, C. M.
(2003). Generation of telomeric G strand over-
hangs involves both G and C strand cleavage.
Mol. Cell 11, 10211032.
61 Kim, M. Y., Vankayalapati, H., Shin-Ya, K.,
Wierzba, K., and Hurley, L. H. (2002). Telo-
mestatin, a potent telomerase inhibitor that
interacts quite specifically with the human
telomeric intramolecular G-quadruplex. J. Am.
Chem. Soc. 124, 20982099.
References 47
62 Harley, C. B., Futcher, A. B., and Greider, C. W.
(1990). Telomeres shorten during ageing of
human fibroblasts. Nature 345, 458460.
63 Bodnar, A. G., Quellette M., Frolkis, M., Holt,
S. E., Chiu, C. P., Morin, G. B., Harley, C. B.,
Shay, J. W., Lichtsteiner, S., and Wright, W. E.
(1988). Extension of life-span by introduction
of telomerase into normal human cells.
Science 279, 349352.
64 Hooijberg, E., Ruizendaal, J. J., Sniders, P. J.,
Kueter, E. W., Walboomers, J. M., and Spits, H.
(2000). Immortalization of human CD8+T
cell clones by ectopic expression of telomerase
reverse transcriptase. J. Immunol. 165, 4239
4245.
65 Rufer, N., Migliaccio, M., Antonchuk, J.,
Humphries, R. K., Roosnek, E., and Lansdorp,
P. M. (2001). Transfer of the human telomer-
ase reverse transcriptase (TERT) gene into T
lymphocytes results in extension of replicative
potential. Blood 98, 597603.
66 Kiyono, T., Foster, S. A., Koop, J. I., McDougall,
J. K., Galloway, D. A., and Klingelhutz, A. J.
(1998). Both Rb/p16INK4a inactivation and
telomerase activity are required to immortalize
human epithelial cells. Nature 396, 8488.
67 Karlseder, J., Smogorzewska, A., and de
Lange, T. (2002). Senescence induced by al-
tered telomere state, not telomere loss.
Science 295, 24462449.
68 Dunham, M. A., Neumann, A. A., Fasching,
C. L., and Reddel, R. R. (2000). Telomere main-
tenance by recombination in human cells.
Nat. Genet. 26, 447450.
69 Hahn, W. C., Counter, C. M., Lundberg, A. S.,
Beijersbergen, R. L., Brooks, M. W., and Wein-
berg, R. A. (1999a). Creation of human tu-
mour cells with defined genetic elements. Na-
ture 400, 464468.
70 Hahn, W. C., Dessain, S. K., Brooks, M. W.,
King, J. E., Elenbaas, B., Sabatini, D. M., DeCa-
prio, J. A., and Weinberg, R. A. (2002). Enu-
meration of the simian virus 40 early region
elements necessary for human cell transfor-
mation. Mol. Cell. Biol. 22, 21112123.
71 Hahn, W. C., Stewart, S. A., Brooks, M. W.,
York, S. G., Eaton, E. N., Kurachi, A., Beijers-
bergen, R. L., Knoll, J. H. M., Meyerson, M.,
and Weinberg, R. A. (1999b). Inhibition of tel-
omerase limits the growth of human cancer
cells. Nat. Med. 5, 11641170.
72 Zhang, X., Mar, V., Zhou, W., Harrington, L.,
and Robinson, M. O. (1999). Telomerer short-
ening and apoptosis in telomere-inhibited hu-
man tumor cells. Genes Dev. 13, 23882399.
73 Damm, K., Hemmann, U., Garin-Chesa, P.,
Hauel, N., Kauffmann, I., Priepke, H., Nies-
troj, C., Daiber, C., Enenkel, B., Guilliard, B.,
et al. (2001). A highly selective telomerase in-
hibitor limiting human cancer cell prolifera-
tion. EMBO J. 20, 69586968.
Abstract
As part of the FDAs strategic action plan,
the Agency is developing standards to ef-
fectively handle emerging technologies,
especially in the areas of pharmacoge-
nomics, in order to provide efficient and
rapid translation of new scientific develop-
ments and breakthroughs into safe and ef-
fective medical products. A guidance for
industry on genomic data submission has
been recently published. This is intended
to encourage voluntary genomic data sub-
mission and to facilitate the agencys use
of genomic data during regulatory deci-
sion-making. This chapter will focus on
the current state of knowledge with regard
to DNA-based differences in pharmacody-
namics and pharmacokinetics of medica-
tions. It will provide an update on how the
pharmacogenetics/pharmacogenomics in-
formation is being applied in IND (Investi-
gational New Drug) and NDA (New Drug
Application) submissions, and provide ex-
amples of when voluntary submissions
may be appropriate. Critical issues in the
regulatory review and labeling implications
of pharmacogenomic data and various
challenges in the effective translation of
pharmacogenomic information into phar-
maceuticals will be also discussed.
Abbreviations
ADH alcohol dehydrogenase
ADR adverse drug reactions
ALDH acetaldehyde dehydrogenase
AUC area under the curve
BIO Biotechnology Industry
Organization
BLA biologics license application
CDER Center for Drug Evaluation
Research
COMT catechol-O-methyl transferase
DIA Drug Information Association
DPD dihydropyrimidine
dehydrogenase
EM extensive metabolizer
FDA Food and Drug Administration
GST gold sodium thiomalate
HMT histamine metabolism tissue
IND investigational new drug
NAT N'-nitrosoanatabine
NDA new drug application
NME new molecular entity
PD pharmacodynamics
PET positron emission tomography
PG pharmacogenetics
PK pharmacokinetic
PM poor metabolizer
PWG Pharmacogenomics Working
Group
SNP single nucleotide polymorphism
ST ster
49
2
The Role of Pharmacogenetics/Pharmacogenomics
in Drug Development and Regulatory Review: Current Status
ISBN: 3-527-31184-X
TPMT thiopurine methyl transferase
UGT 1A1 uridine disphosphate glucuro-
nosyl transferase 1A1
VGDS voluntary genomic data
submission
2.1
Introduction
The mission of the U.S. Food and Drug Ad-
ministration (FDA) is to advance the public
health by helping to speed innovations that
make medicines and foods more effective,
safer, and more affordable [1]. Drug research
and development, although reasonably suc-
cessful, has been hampered by high cost [2],
high investigational new drug (IND) failure
rate [3], and multiple new drug application
(NDA) review cycles [4]. The number of ap-
plications for new molecular entities sub-
mitted to the Agency has declined steadily
[5] (see also Part VIII, Chapter 1). As part
of the FDAs strategic plan [6], the Agency
is developing standards to handle emerging
technologies, especially in the area of phar-
macogenomics, in order to provide efficient
and rapid translation of new scientific devel-
opments and breakthroughs into safe and
effective medical products. A recent docu-
ment by the Agency [5] stressed that The
product development problems we are see-
ing today can be addressed, in part, through
an aggressive, collaborative effort to create a
new generation of performance standards
and predictive tools. The new tools will
match and move forward new scientific in-
novations and will build on knowledge de-
livered by recent advances in science, such
as bioinformatics, genomics, imaging tech-
nologies, and materials science (see also
Part I, Chapters 3 and 4; Part V, Chapters
4 and 8; and Part VI, Chapter 6). There
are various initiatives within the Center
for Drug Evaluation and Research (CDER)
to address issues in the area of pharmacoge-
nomics. A guidance for industry on geno-
mic data submission has been published
[7, 85]. The guidance was intended to en-
courage voluntary genomic data submission
by sponsors using pharmacogenomics in ex-
ploratory research during drug develop-
ment. A workshop was held in November
2003 to discuss issues related to pharmaco-
genomic data submissions, and the proceed-
ings have been published [812]. In addition
to the guidance on pharmacogenomic data
submissions, the FDA is developing a new
guidance for drug/test combinations when
a DNA-based test is used prior to prescrib-
ing a drug. Another workshop [13] was held
in July 2004 to identify issues in the develop-
ment of these combination products and a
concept paper was published [86]. With
the increasing knowledge and available
tools in pharmacogenomics, the FDA will
continue to encourage genomic-based re-
search, where scientifically appropriate,
and the translation of the resultant scientific
data to clinical practice [14, 15].
This chapter will focus on the current
state of knowledge with regard to DNA-
based differences in pharmacokinetics and
pharmacodynamics of medications. It will
provide an update on how the pharmaco-
genomic information is being applied and
reviewed in IND and NDA submissions
and provide examples of when voluntary
submissions may be appropriate. Critical
issues in the regulatory review and label-
ing implications of pharmacogenomic data
will be also discussed.
2.2
Variability in Drug Response
Many factors can affect a patients response
to a drug. These include intrinsic factors
such as age, gender, race/ethnicity, genetics,
2 The Role of Pharmacogenetics/Pharmacogenomics in Drug Development and Regulatory Review: Current Status 50
disease states, organ dysfunctions, and
other physiological changes, including preg-
nancy, lactation, and extrinsic factors such
as smoking, diet (food, juice, dietary supple-
ments), and concomitant medications [16].
A recent review of post-approval dosage
changes between 1980 and 1999 indicates
that, of the evaluable drug products
(n=354), 21% had dosage changes, and
79% of these changes were safety-related
dose reduction [17]. Many of these changes
were based on new information that was
obtained after the initial approval. These
changes included dosing recommenda-
tions for specific populations, such as pa-
tients at various stages of renal or hepatic
impairment, patients taking specific conco-
mitant medications, or patients who are
pregnant. This study pointed out the im-
portance of having accurate dosage recom-
mendation for individuals with various in-
trinsic or extrinsic factors prior to market-
ing to avoid risks of adverse drug reactions
(ADR). ADRs, accounting for 5% of hospi-
tal admissions, were also experienced by
10% of hospitalized patients, have led to
700000 injuries/deaths per year, and were
estimated to be the fourth or the sixth
leading cause of death in the United States
for hospitalized patients [18]. Serious
ADRs have contributed to market withdra-
wals. Table 2.1 lists drugs withdrawn from
the US market in the past 7 years due to
safety reasons [19].
In order to optimize drug therapy and
reduce adverse events, it is critical that in-
formation on how various intrinsic and ex-
trinsic factors may affect drug treatment
be available for the healthcare providers
and patients. It is recommended that
when a drug is being developed, the varia-
bility in drug response and the factors con-
tributing to it be investigated, and this in-
formation be included in the labeling. As
part of the good review practices during
the regulatory review of clinical pharma-
cology and biopharmaceutics data in an
IND or NDA submission, key pharmacoki-
netic (PK) and pharmacodynamic (PD) pa-
rameters and their variability in various
population groups are reviewed in an inte-
grated approach. Detailed data are in-
cluded in the important clinical pharma-
cology findings section. And key results
summarized in the executive summary
section [20]. For example, changes in PK
parameters, such as AUC (area under the
plasma concentrationtime curve) or C
max
(maximum plasma concentration), due to
various extrinsic and intrinsic factors may
be displayed in graphic or table forms.
The changes in PK in various population
groups of two recently approved drugs, ro-
suvastatin and atomoxetine [21], are de-
picted in Figs. 2.1 and 2.2. These PK
changes, coupled with information on the
exposureresponse relationship derived
from the doseresponse, PK/PD, and/or
efficacy/safety database (Fig. 2.3) and other
considerations (e.g., labeling of other
drugs in the same pharmacologic class),
form the basis of labeling recommenda-
tions for patients with these extrinsic or
intrinsic factors. Tables 2.2 and 2.3 show
the corresponding labeling recommenda-
tions for these two approved drugs in spe-
cific patient groups [21]. Atomoxetine is
metabolized by CYP2D6, a polymorphic
enzyme. It is a credit to the sponsor that
in addition to PK data, the efficacy and
safety data in patients identified as exten-
sive metabolizers (EM) of CYP2D6 have
been compared with those identified as
poor metabolizers (PM) of CYP2D6, and
the results were stated in the label. Many
of the studies evaluating the effect of var-
ious intrinsic and extrinsic factors on PK
of atomoxetine were conducted in EMs of
CYP2D6. With the exclusion of PM sub-
jects, the evaluation of changes in PK in
2.2 Variability in Drug Response 51
patients with hepatic impairment or in pa-
tients taking CYP2D6 inhibitors, will not
be confounded by the patients intrinsic
CYP2D6 enzyme status.
2.3
Drug-metabolizing Enzymes
and Transporters
A recent analysis of 18 ADR studies con-
ducted between 1995 and 2000 showed that
59% of drugs causing ADRs are metabolized
by polymorphic enzymes, while only 722%
of other randomly selected drugs are sub-
strates for polymorphic enzymes [22]. These
results suggest that doses based on an indi-
viduals genotype may reduce ADRs. Impor-
tant human metabolizing enzymes are
listed Fig. 2.4 [23]. An updated list of CYP
enzymes and literature references for in-vi-
tro or in-vivo activities for various alleles is
available on-line [24]. In addition to poly-
morphism in metabolizing enzymes, there
are polymorphisms in transporters, recep-
tors and other therapeutic targets. The ex-
tent for which the metabolizing enzyme
genotypes affect pharmacokinetics and/or
pharmacodynamics and clinical responses
has been the subject of various recent re-
Table 2.1 Drugs withdrawn from the US market between 1997 and 2001 [19].
Year Drug name
a)
Use Risk
Withdrawn Approval
1997 1973 Fenfluramine
(Pondimin)
Obesity Heart valve abnormality
1997 1996 Dexfenfluramine
(Redux)
Obesity Heart valve abnormality
1998 1997 Mibefradil (Posicor) High blood pressure/
Chronic stable angina
Drugdrug interactions
Torsades de Pointes
1998 1997 Bromfenac (Duract) NSAID Acute liver failure
1998 1985 Terfenadine
(Seldane/Seldane-D)
Antihistamine Torsades de Pointes
1999 1988 Astemizole
(Hismanal)
Antihistamine Torsades de Pointes
1999 1997 Grepafloxacin (Raxar) Antibiotics Torsades de Pointes
2000 2000 Alosetron
b)
(Lotronex)
Irritable bowel
syndrome in women
Ischemic colitis; compli-
cations of constipation
2000 1993 Cisapride
(Propulsid)
Heartburn Torsades de Pointes
2000 1997 Troglitazone
(Rezulin)
Diabetes Acute liver failure
2001 1997 Cerivastatin
(Baycol)
Cholesterol lowering Rhabdomyolysis
Drug-drug interactions
2001 1999 Rapacuronium bromide
(Raplon)
Anesthesia Bronchospasm
a) Tradenames are in parentheses.
b) Reintroduced to the market in 2002 with use restricted to pa-
tients severely affected with irritable bowel syndrome.
2.3 Drug-metabolizing Enzymes and Transporters 53
Fig. 2.1 Fold-changes in the area under the plasma
concentration-time curve (AUC) of rosuvastatin in
the presence of intrinsic/extrinsic factors, as com-
pared to the control group, Group 1 (without the
specific intrinsic or extrinsic factor). Groups 2 and 3
under hepatic: subjects with hepatic impairment
as defined by Child-Pugh A and B, respectively.
Groups 2, 3, and 4 under renal: subjects with renal
impairment as defined by creatinine clearance val-
ues of 5080, 3050 and <30 mL min 1.73 m2,
respectively. Group 2 under race: Japanese sub-
jects residing in Japan and Chinese subjects resid-
ing in Singapore when compared with Caucasians
residing in North America and Europe (Group 1).
Group 2 under Cyclosporine (C), Gemfibrozil
(G), and Itraconazole (I): subjects taking rosu-
vastatin concomitant with C, G, or I, respectively.
The control group (Group 1): subjects not taking
C, G, or I, respectively. These AUC fold-change
values were extracted from PDR labeling of Cres-
tor* (AstraZeneca) [21].
Fig. 2.2 Fold-changes in the area under the plas-
ma concentration-time curve (AUC) of atomoxe-
tine in the presence of intrinsic/extrinsic factors,
as compared to the control group, Group 1 (with-
out the specific intrinsic or extrinsic factor).
Groups 2 and 3 under hepatic: subjects with he-
patic impairment as defined by Child-Pugh B and
C, respectively. Group 2 under renal: subjects
with end-stage renal disease. Group 2 under Pe-
diatric: adolescents and children older than 6
years old. Group 2 under Gender: female sub-
jects as compared to male subjects (control:
Group 1). Group 2 under (Fluoxetine): subjects
taking atomoxetine concomitant with fluoxetine as
compared to the control group (Group 1): sub-
jects not taking fluoxetine concomitantly. Group 2
under Genotype: subjects with a poor metaboli-
zer genotype (PM) as compared to subjects with
at least one wild allele, EM (control, Group 1).
Note that the studies were carried out in EM only
when EM was denoted. These AUC fold-change
values were extracted from the PDR labeling of
Strattera (Lilly) [21].
views [2527]. Several enzymes that are con-
sidered known valid or probable valid
biomarkers based on the criteria described
in the draft guidance on pharmacogenomic
data submission [7, 85] are listed in Table
2.4. These valid biomarkers are defined as
being measured in an analytical test system
with well-established performance charac-
teristics and for which there is evidence
about the physiologic, toxicologic, pharma-
cologic, or clinical significance of the results
[7]. Table 2.4 also summarizes some of the
published correlation data between the gen-
otypes and outcome measures (e.g., clinical
efficacy, ADR, doses, PK and PD) for some
model drugs.
Table 2.5 lists enzymes and transporters
that have not reached the known valid or
probable valid biomarker status, and are
considered exploratory biomarkers. For
some genes (e.g., CYP3A4), the correlation
between certain genotypes and enzyme or
transporter activities was observed in vitro
only. For others (e.g., ABCB1), contradic-
tory data have been published for different
drugs and the correlation between SNP
genotype or haplotype and the phenotype
(PK parameters, other response measures)
will need to be further defined.
Additional exploratory biomarkers not
related to metabolism or transport of
drugs are listed in Table 2.6. Although the
cases listed in Tables 2.42.6 are mostly
from monogenic studies, many drugs dis-
play polygenic traits. The interplay of
genotypes of the enzymes, transporters
and receptors, among other factors (such
as concomitant medications and disease
states), can affect the risk/benefit ratio for
individual patients [2830], and need to be
considered when evaluating varied results
from many genotyping studies with small
number of subjects.
2.4
Applications of Pharmacogenetics
and Pharmacogenomics in Drug Develop-
ment and Regulatory Review
A recent internal, informal survey of the
IND and NDA submissions received in
CDER indicated that, of the 70 submis-
sions with pharmacogenomic data received
between 1992 and 2001, many evaluated
the status of drug-metabolizing enzymes
with CYP2D6 on the top of the list. The
distribution of submissions evaluating var-
ious polymorphic enzymes is depicted in
Fig. 2.5 [62]. Many of the submissions re-
ceived between 1992 and 1999 used phe-
notyping (e.g., urinary metabolic ratios of
dextromethorphan and dextrorphan) to es-
timate CYP2D6 activities. Most of the later
submissions (received between 2000 and
2001) used genotyping.
The goals of these studies include the
following:
To evaluate PK differences in subjects
with different genotypes or phenotypes
(such as the dextromethorphan urinary
Fig. 2.3 Determination of therapeutic range based
on exposure (dose, AUC, C
max
, etc.) and response
data [82].
metabolic ratio); many are evaluated in
Phase 1/2 clinical pharmacology studies
using a small panel of subjects.
To use genotype as one of the covariates
in population PK or PD analysis of clini-
cal trial data.
To explain (post-hoc) outliers in PK or
PD observed in clinical studies.
To stratify patients based on their geno-
types in the clinical evaluation of effec-
tiveness (prospective stratification or ret-
rospective analysis).
To determine if an ADR is related to cer-
tain genotypes (retrospective analysis).
More recent IND submissions showed that
many sponsors are banking blood (geno-
mic DNA) samples for future evaluation of
the role of multiple enzymes, transporters
and/or receptors, when future study re-
2.4 Applications of Pharmacogenetics and Pharmacogenomics in Drug Development and Regulatory Review 55
Table 2.2 Rosuvastatin (CRESTOR
) label recommendations in
patients defined by various intrinsic and extrinsic factors [21].
Extrinsic or
intrinsic factors
Rosuvastatin
AUC
fold-change
Rosuvastatin
C
max
fold-change
Rosuvastatin
Labeling
Approved dosing: 540 mg once daily
Hepatic
(Child-Pugh B)
1.2 2 Child-Pugh: no recommendation for change
Contraindicated in patients with active liver
disease or with unexplained persistent eleva-
tions of serum transaminases
Renal 3
a)
Mild to moderate renal insufficiency; no
modification of dosage
Severe renal impairment (CL
cr
<30 mL min 1.73 m
2
) not on hemodialysis, be
started at 5 mg once daily; not to exceed
10 mg once daily
Race 2
b)
These increases should be considered when
making rosuvastatin dosing decisions for
patients of Japanese and Chinese ancestry
Cyclosporine 7 11 Patients taking cyclosporine
limited to 5 mg once daily
Gemfibrozil 2 2 Combination with gemfibrozil
limited to 10 mg once daily
Itraconazole 1.31.4 No recommendation for change
Note: the AUC fold-change was calculated by dividing AUC of ro-
suvastatin with specific extrinsic/intrinsic factor by AUC of rosu-
vastatin of the control group (without the specific extrinsic/in-
trinsic factor).
a) Plasma concentrations increased to 3-fold
(CL
cr
<30 mL min 1.73 m
2
as compared to
CL
cr
>80 mL min 1.73 m
2
). Cl
cr
: creatinine clearance.
b) Japanese subjects residing in Japan and in Chinese subjects
residing in Singapore when compared with Caucasians resid-
ing in North America and Europe.
sults may indicate the evaluation to be ap-
propriate [62, 63].
2.5
Determination of Different Genotype Groups
based on Known Valid and Probable Valid
Biomarkers
Literature data provide evidence that those
enzymes listed in Table 2.4, CYP2D6,
CYP2C9, CYP2C19, TPMT, and UGT1A1,
are known valid or probable valid me-
tabolizing enzyme biomarkers. Based on a
recent FDA guidance [7], data related to
genotypes of these enzymes will need to
be submitted for review in NDA, with var-
ious reporting format (full report, abbre-
viated report or synopsis) depending on
the purpose of the genomic evaluation and
the validity of the genomic biomarker [7].
The type of genomic data (e.g., which al-
leles, what genotypes) needed to be evalu-
ated is one of the critical issues in drug
development and regulatory review, and
was the subject of a recent discussion at a
Table 2.3 Atomoxetine (STRATTERA
) label recommendations in
patients defined by various intrinsic and extrinsic factors [21].
Extrinsic or
intrinsic factors
Atomoxetine
AUC
fold-change
Atomoxetine
C
max
fold-change
Atomoxetine
labeling
Approved dosing: 0.5 mg kg
1
initially up to
1.2 mg kg
1
(no more than 1.4 mg kg
1
per day
or 100 mg, whichever is less)
Hepatic
a)
(Child-Pugh C)
4 Reduced to 25% of the normal dose
Hepatic
a)
(Child-Pugh B)
2 Reduced to 50% of the normal dose
Renal
a)
1 No recommended dose change
Pediatric
(> 6 years old)
similar No recommended dose change
Gender (female) 1 No recommended dose change
Co-administra-
tion with fluoxe-
tine, paroxetine,
quinidine*
68 34 Dosage adjustment of STRATTERA in EMs may
be necessary when coadministered with CYP2D6
inhibitors (e.g., paroxetine, fluoxetine, and quini-
dine)
In-vitro studies suggest that co-administration of
cytochrome P450 inhibitors to PMs will not in-
crease the plasma concentrations of atomoxetine
CYP2D6
genotype
10 5 Approximately 7% of a Caucasian population are
PMs. Laboratory tests are available to identify
CYP2D6 PMs. The blood levels in PMs are similar
to those attained by taking strong inhibitors of
CYP2D6. The higher blood levels in PMs lead to a
higher rate of some adverse effects of STRATTERA
a) Studies conducted in EM of CYP2D6; renal: subjects with
end-stage renal disease; pediatric: adolescents and children
under 6 years of age.
FDA/PhRMA/Johns Hopkins University
educational workshop [64]. Because of
race/ethnic differences in the distribution
of various alleles with no or reduced en-
zyme activities for various metabolizing
enzymes [2527], it is important to consid-
2.5 Determination of Different Genotype Groups based on Known Valid and Probable Valid Biomarkers 57
Fig. 2.4 Contribution of genetic polymorphisms to drug metabolism [23].
Table 2.4 DNA-based biomarkers of enzyme activities considered as valid biomarkers.
Enzyme Model drugs Outcome measures Study results Reference
CYP2C9 Warfarin Maintenance dose
Time to reach stable dosing
Patients with *2 and *3
maintained with lower doses
and took longer time to reach
stable dosing
3133
CYP2C19 Proton pump
inhibitors
Plasma levels
Gastric pH
Gastroesophageal reflux
disease cure rate
Higher in PM (20 mg)
Higher dose (40 mg)
showed no difference
34
66
CYP2D6 Codeine Morphine formation
Analgesic effects
Higher in EM 35
Atomoxetine Pharmacokinetic measure PM higher AUC (10-fold) 21
UGT1A1 Irinotecan Grade 3/4 neutropenia UGT1A1 7/7 and 6/7 more
frequent than 6/6
36
Pharmacokinetic parameters
(AUC ratio of SN38G/SN38)
UGT1A1*28 and *6 with
reduced ratios
37, 38
TPMT 6-MP Dose-limiting hematopoietic
toxicity
More in TPMT deficiency
or heterozygosity
3941
UGT 1A1: uridine diphosphate glucuronosyl transferase 1A1;
TPMT: thiopurine methyl transferase; SN-38: an active metabo-
lite of ironotecan: SN-38G: a glucuronide metabolite of SN-38.
er the allelic distribution in different race/
ethnic groups when evaluating dosere-
sponse. For example, in conducting clini-
cal studies of CYP2D6 substrates, evaluat-
ing *3, 4, 5, 6, 8 (and possibly *41) may
capture a high percentage of Caucasians
with low or no CYP2D6 enzyme activities
[65]. It is important to measure, in addi-
tion, *10 (and possibly *21) in Asians and
*17 (and possibly *29) in African Ameri-
cans to ascertain that genotypes corre-
sponding to medium or low activity have
been assessed across populations that will
receive the drug [6568]. A recent study on
desipramine suggested that additional gen-
otyping (and molecular haplotyping) of al-
leles with intermediate metabolizing activ-
ities (IM) may be necessary to fully charac-
terize CYP2D6 genotypephenotype rela-
tionships [69]. It is also critical to evaluate
the presence of multiple copies of *2 in
order to understand the doseresponse of
CYP2D6 substrates in Caucasians and
African Americans [65]. For CYP2C19,
measuring only *2 and *3 may capture 84,
>99 and 90% of the main variant
Table 2.5 Genes encoding metabolizing enzymes/transporters currently considered as exploratory biomarkers.
Enzyme/
Transporter
Model drugs Outcome
measures
Study results Reference
CYP3A4 Testosterone In-vitro meta-
bolism rate
*17 lower activity while *18 higher
activity
42
CYP3A5 Tacrolimus
Cyclosporine
Pharmacokinetic
parameters
*3 (non-expressor) associated with
higher trough plasma concentrations
43, 44
CYP2B6 Efavirenz Pharmacokinetic
parameters
*6 homozygous associated with
higher plasma concentrations
45
CYP2C8 Repaglinide Pharmacokinetic
parameters
*3 associated with lower plasma
concentrations
46
CYP2A6 Nicotine Pharmacokinetic
parameters
*7, *10 associated with higher
nicotine and lower cotinine plasma
concentrations
47
ABCB1
(MDR1)
Digoxin Pharmacokinetic
parameters
TT homozygous C3435T associated
with higher plasma concentrations
48
Fexofenadine Pharmacokinetic
parameters
TT homozygous C3435 associated
with lower plasma concentrations
49
Nelfinavir
Efavirenz
Pharmacokinetic
parameters and
Immune recovery
TT homozygous C3435 associated
with lower plasma concentrations,
and greater rise in CD4 responses
50
Antiepileptic
drugs
Clinical responses CC homozygous C3435 associated
with drug-resistant epilepsy
51
ABCA1 Atorvastatin
Simvastatin,
Pravastatin
LDL-cholesterol
lowering
Higher adjusted mean change
in certain HAP markers
52
OATP-C Pravastatin Pharmacokinetic
parameters
*15 lower clearance 53
ABCB1: ATP-binding cassette family (ABC) B1; multi-drug resis-
tance (MDR1): a human gene that encodes P-glycoprotein; MRP:
multi-drug resistance protein;
OATP-C: organic anion transporting peptide-C.
CYP2C19 genotypes in Caucasian, Asian
and African American populations, respec-
tively [70, 71]. The addition of *4, 5, and 6
will assure that 92% of the variant alleles
in the Caucasian population has been cap-
tured [70]. For CYP2C9, the assessment of
alleles *4, 5 and *6, in addition to *2 and
*3, the major variant alleles in Caucasians,
may be necessary to capture CYP2C9 vari-
ant genotypes in various populations [72].
For UGT1A1, while *28 in Caucasians ap-
peared to be correlated with adverse events
of irinotecan (e.g., diarrhea or neutrope-
nia) [7375] and may be appropriate to as-
sess in Caucasians, it may be critical to
evaluate additional alleles in other popula-
tion groups (e.g., *6 in Japanese or *60 in
African Americans) [7375].
2.5 Determination of Different Genotype Groups based on Known Valid and Probable Valid Biomarkers 59
Table 2.6 Other exploratory pharmacogenomic biomarkers.
Pharmacogenomic
markers
Model drugs, drug
classes or diseases
or trials
Outcome measures Study results Refer-
ence
QT genes Beta blockers Cardiac events Higher rates in LQT2
and LQT3 genotypes
54
KvLQT1 (KCNQ1),
HERG (KCNH2),
SCN5A
Acquired long-QT
syndrome (patients
with drug-associated
torsades de pointes)
Mutations in
these genes
5/92 patients has muta-
tions
55
b
2
-adrenergic
receptor
Albuterol Respiratory flow Homozygous for
arginine at codon 16
appears to be asso-
ciated with lower
response
56
HMG-CoA Pravastatin LDL-cholesterol
lowering
Different genotypes
of HMG-CoA asso
ciated with different
responses
57
IL-10 Adult lung
Transplant patients
Acute persistent
rejection
Increased IL-10
production genotype
has lower rejection rate
58
Multiple genes Abacavir Hypersensitivity
cases
HLA-B57 was present in
39 (46%) of 84 patients
with hypersensitivity
versus four (4%) of 113
controls (without hyper-
sensitivity)
59
EGFR Gefitinib Response in non-
small cell lung
cancer
8/9 responders
(vs 0/7 in non-respon
ders) with mutations
60
FCGR3A Infliximab Response in Crohns
disease
Higher clinical re-
sponse in V/V vs. V/F
or F/F
61
IL-10 interleukin-10; EGFR: epidermal growth factor receptor;
FCGR3A: the gene coding for FccRIIIa.
2.6
Drug Interactions
While pharmacogenetics of metabolizing
enzymes can affect the patients response
to treatment, concomitant drug or dietary
supplement administration is another im-
portant factor in altered drug response. Re-
cent studies have shown that the extent of
drug interaction may be impacted by geno-
types. Some examples are listed in Table
2.7. This type of information has begun to
appear in the product labeling. For exam-
ple, Table 2.3 shows the dosing recom-
mendation of atomoxetine. In contrast to
the warning for EMs of CYP2D6 that
Dosage adjustment of STRATTERA in
EMs may be necessary when coadminis-
tered with CYP2D6 inhibitors, e.g., paroxe-
tine, fluoxetine, and quinidine, no similar
warnings for PM of CYP2D6 are in the la-
beling. The labeling indicates that in vitro
studies suggest that co-administration of
cytochrome P450 inhibitors to PMs will
not increase the plasma concentrations of
atomoxetine. [21].
2.7
Voluntary versus Required Submissions
Whether certain type of pharmacogenomic
data need to be submitted to the Agency
as required by regulation for review is dis-
cussed in a FDA guidance [7, 85] and pre-
sented at a FDA/DIA/PWG/PhRMA/BIO
workshop [812]. The following cases high-
light scenarios in drug development, and
illustrate the basis for submitting pharma-
cogenomic information to the FDA as vol-
untary or required data submissions.
Fig. 2.5 Distribution of pharmacogenetic/pharma-
cogenomic studies evaluating the impact of differ-
ent genotypes of CYP2D6, CYP2C19, CYP2C9,
CYP3A, CYP1A2, and other metabolizing enzymes,
transporters, receptors in 70 IND- and NDA-sub-
mitted reports between 1992 and 2001 [62].
2.7.1
Scenario 1
During an IND stage, a sponsor conducts
single- and multiple-dose pharmacokinetic
studies of a new molecular entity (NME)
in healthy volunteers enrolled to represent
the major racial demographic groups. The
NME is metabolized primarily by
CYP2C19 to inactive metabolites. The
sponsor assesses the CYP2C19 genotypes
in the volunteers to determine the clear-
ance phenotype in order to determine if
drug dosing needs to be individualized
based on the genotype groups.
Type of submission: Full report (IND).
Rationale: The sponsor is using the test
results to support scientific arguments
pertaining to the selection of drug dos-
ing (see Fig. 2.6).
2.7.2
Scenario 2
A sponsor conducts a Phase III clinical
trial of a NME in patients with the target
indication. The NME is metabolized pri-
marily by CYP2D6 to an active metabolite
equipotent to the parent molecule. The
sponsor genotypes a randomly selected
subset of the patients for their CYP2D6 al-
leles in order to explore the association be-
tween genotype, drug dosing and clinical
outcome. The results show minor differ-
ences in clinical outcomes among the gen-
otypes. The information is included in the
proposed labeling in the NDA submission.
Type of submission: Full report (NDA).
Rationale: The sponsor will use the test
results in the drug label (see Fig. 2.7).
2.7 Voluntary versus Required Submissions 61
Table 2.7 The effect of genotypes on the extent of drug interactions.
Substrate
(enzyme)
Inhibitor or inducer Outcome
(changes in plasma AUC or concentrations
of substrates)
Reference
Atomoxetine
(CYP2D6)
Fluoxetine Paroxetine AUC increase 6- to 8-fold in EM; no change
PM expected
21
Metoprolol
(CYP2D6)
Diphenhydramine Higher inhibition in EM vs PM (fold vs fold) 76
Tamoxifen
(CYP2D6)
Paroxetine Greater reduction in plasma levels of
endoxifen (active metabolite of tamoxifen
formed via CYP2D6) in homozygous EM as
compared to patients with at least one variant
allele
77
Diazepam
(CYP2C19)
Omeprazole No inhibition in PM 78
Omeprazole
(CYP2C19)
Fluvoxamine AUC increased 3- to 6-fold in EM;
no changes in PM
79
Omeprazole
(CYP2C19)
Gingko Boloba Higher induction in EM 80
2.7.3
Scenario 3
A sponsor conducts a Phase III clinical trial
of a NME in patients with the target indica-
tion. The NME is metabolized primarily by
CYP2D6 to an active metabolite equipotent
to the parent molecule. After the trial is com-
pleted, the sponsor genotypes a randomly se-
lected subset of the patients for their
CYP2D6 alleles in order to explore the asso-
ciation between genotype and plasma clear-
ance values. The sponsor has not proposed
to include the results in the labeling.
Type of submission: Abbreviated report
(IND or NDA/BLA).
Rationale: Although the test results may
not be used in decision-making about
Fig. 2.6 Submission of pharmacogenetics (PG) data to an Investigational New Drug (IND) report [7, 85].
Fig. 2.7 Submission of pharmacogenetics (PG) data to a New Drug Application (NDA)/BLA [7, 85].
drug dosing in the drug label, CYP2D6
is a known valid biomarker, therefore,
the test results need to be submitted as
an abbreviated report.
2.7.4
Scenario 4
A sponsor conducts a drug interaction
study in healthy volunteers of their NME,
a CYP3A substrate, co-administered with
ketoconazole as an enzyme inhibitor. Sub-
sequent to the study, the subjects are geno-
typed for their CYP3A5 alleles to deter-
mine the relative contribution of this poly-
morphism to inter-individual variability in
AUC.
Type of submission: For submissions
under IND, these data would be eligible
for a VGDS. For submissions under
NDA/BLA, these data would be required
to be submitted as a synopsis and a
VGDS is encouraged.
Rationale: The test results are not being
used in decision-making or scientific ar-
guments or in the drug label or as part
of the scientific database. In addition,
polymorphism of CYP3A5 is not widely
studied and is therefore neither a prob-
able or known valid biomarker. The in-
formation on this genotype is considered
to be exploratory.
2.8
Labeling Implications
Labeling for drug products in the US
needs to be in the format per the Code of
Federal Regulations (21 CFR 201.56). In a
proposed revision of physician labeling,
new content and format requirements are
described for the labeling of human pre-
scription drug and biological products
[83, 84]. Pharmacogenomic data and re-
lated information can be described in the
following sections as appropriate: Indica-
tions and Usage; Dosage and Adminis-
tration; Contraindications; Warnings
and Precautions; Adverse Reactions;
Drug Interactions; or Use in Specific
Populations. When different pharmacoge-
nomic subgroups show significantly differ-
ent responses (in safety, efficacy, pharma-
cokinetic, or pharmacodynamic profiles or
dose requirement), the information may
be included in the labeling. Depending on
the risk/benefit, the information may be
placed in different sections of the labeling.
When the genomic test must be conducted
prior to dosing (for patient selection and/
or dose selection), it may be stated in the
Indications and Usage section (see Table
2.8, Herceptin as an example; see also Part
I, Chapter 5) with relevant information
placed in other sections such as Clinical
Studies, HER2 testing, and HER2 de-
tection. When dose reduction may be im-
portant for specific genotypes, the infor-
mation can be placed in the Dosage and
Administration and Warnings sections
(see Table 2.8, Purinethol as an example)
with relevant information in other sections
such as Clinical Pharmacology, Labora-
tory test, and Adverse Reactions. When
the adverse events are serious (e.g., tor-
sades de pointes) and appropriate dose ad-
justments cannot be determined, the infor-
mation may be included in Contraindica-
tions (see Table 2.8, thioridazine, Mellaril)
and relevant information placed in other
sections as appropriate. When there are no
serious adverse events, however, the geno-
type information could be helpful in re-
ducing less serious adverse events, the in-
formation may be placed in various sec-
tions, such as Clinical Pharmacology,
Drug Interactions, Adverse Events,
Laboratory test, Special Populations,
etc. (see Table 2.8, Strattera).
2.8 Labeling Implications 63
2.9
Conclusion
Pharmacogenomic data can facilitate our
understanding of the sources of variability
in drug response, and can potentially lead
to improved safety and efficacy of drug
therapy for individual patients. Through
various initiatives [5, 6], the FDA is
encouraging that drug developers apply
the rapidly evolving pharmacogenomic
tools and integrate this data to the evalua-
Table 2.8 Examples of pharmacogenomic information in the drug label [21,81].
Brand name
(generic name)
Labeling section Labeling statement
Herceptin*
(trastuzumab)
August 2002
Indications
and Usage
Herceptin* should be used in patients whose tumors have been
evaluated with an assay validated to predict HER2 protein over-
expression (see PRECAUTIONS: HER2 Testing and CLINICAL
STUDIES: HER2 Detection).
Purinethol
(6-Mercapto-
purine)
July 2004
Warnings
Dosage and
administrations
Individuals who are homozygous for an inherited defect in the
TPMT (thiopurine-S-methyltransferase) gene may be unusually
sensitive to the myelosuppressive effects of mercaptopurine and
prone to developing rapid bone marrow suppression following
the initiation of treatment. (see Dosage and Administration).
Patients with inherited little or no thiopurine S-methyltransferase
(TPMT) activity are at increased risk for severe purinethol toxicity
from conventional doses of mercaptopurine and generally require
substantial dose reduction. The optimal starting dose for homo-
zygous deficient patients has not been established (see CLINICAL
PHARMACOLOGY WARNINGS and PRECAUTIONS sections)
Mellaril
(Thioridazine)
July 2003
Contraindica-
tions
Thioridazine is contraindicated in patients, comprising about 7%
of the normal population, who are known to have a genetic defect
leading to reduced levels of activity of P450 2D6 (see WARNINGS
and PRECAUTIONS).
Strattera
(atomoxetine)
March 2003
Drugdrug
interactions
Laboratory
tests
In EMs, inhibitors of CYP2D6 increase atomoxetine steady-state
plasma concentrations to exposures similar to those observed in
PMs. Dosage adjustment of STRATTERA in EMs may be necessary
when coadministered with CYP2D6 inhibitors, e.g., paroxetine,
fluoxetine, and quinidine (see Drug Interactions and PRECAU-
TIONS). In vitro studies suggest that coadministration of cyto-
chrome P450 inhibitors to PMs will not increase the plasma
concentrations of atomoxetine.
CYP2D6 metabolism: Poor metabolizers (PMs) of CYP2D6 have a
10-fold higher AUC and a 5-fold higher peak concentration to a
given dose of STRATTERA compared with extensive metabolizers
(EMs). Approximately 7% of a Caucasian population are PMs.
Laboratory tests are available to identify CYP2D6 PMs. The blood
levels in PMs are similar to those attained by taking strong inhibi-
tors of CYP2D6. The higher blood levels in PMs lead to a higher
rate of some adverse effects of Strattera (see ADVERSE
REACTIONS).
tion of patient variability. The FDA has
clarified when these data are required sub-
missions, and when they are exploratory
data that can be shared via a newly estab-
lished process (voluntary genomic submis-
sion) [7, 85].
Increasingly, pharmacogenetics and
genomic information is being included in
the labeling prior to market approval (e.g.,
Herceptin, Strattera) [21] or after approval,
when new information becomes available
(e.g., Purinethol, thioridazine) [21, 81] so
that healthcare providers and patients have
updated information on how pharmacoge-
nomics, along with other factors (age, gen-
der, hepatic, renal impairment, concomi-
tant medications, etc.) can influence indi-
vidual response.
There are many challenges to the effec-
tive translation of pharmacogenomic infor-
mation to clinical practice, and these must
be addressed before the full potential of
pharmacogenomics can be realized to opti-
mize patient therapy. The challenges in-
clude:
The education of healthcare providers
and patients.
Insurance coverage of pharmacogenomic
tests.
Pharmacogenomic test availability and
validation.
Past clinical practices (standard-of-care).
The need for an interdisciplinary team
approach to address complex issues.
Many individuals and organizations are
working to remove these barriers in order
to fully utilize pharmacogenomics to im-
prove biopharmaceuticals and hence pub-
lic health.
References
1 Food and Drug Administration Mission State-
ment. 2004 http://www.fda.gov/opacom/more-
choices/mission.html [accessed September 4
2004].
2 DiMasi JA, Hansen RW and Graboswki HG.
The price of innovation: new estimates of
drug development costs J Health Econ 2003;
22: 151185.
3 Gilbert J, P Henske, and A Singh. Rebuilding
Big Pharmas Business Model, In vivo: the
Business & Medicine Report, Windhover In-
formation, Vol. 21, No. 10, November 2003.
4 Jenkins, JK. Improving Efficiency of Drug De-
velopment and Review, presented at the Inter-
national Conference on Drug Development,
February 25, 2004, Texas.
5 FDA whitepaper, 2004: Innovation or Stagna-
tion, challenge and opportunity on the critical
path to new medical products. http://www.fda.
gov/oc/initiatives/criticalpath/whitepaper.html
[accessed September 4 2004].
6 FDA strategic plan: FDA Commissioner,
McClellan, MB, FDA Strategic Action Plan;
August 2003; http://www.fda.gov/oc/mcclellan/
strategic.html
7 Guidance for industry: pharmacogenomic data
submission, posted on March 2005, http://
www.fda.gov/cder/guidance/6400fril.pdf
8 Salerno RA, Lesko LJ. 2004b. Pharmacoge-
nomics in drug development and regulatory
decision-making: the Genomic Data Submis-
sion (GDS) proposal. Pharmacogenomics
2004; 5(1):2530.
9 Salerno RA, Lesko LJ. Pharmacogenomic data:
FDA voluntary and required submission guid-
ance. Pharmacogenomics 2004; 5(5): 503505.
10 Leighton JK, DeGeorge J, Jacobson-Kram D,
MacGregor J, Mendrick D, Worobec A. Phar-
macogenomic data submissions to the FDA:
non-clinical case studies, Pharmacogenomics
2004; 5(5): 507511.
11 Ruao G, Collins JM, Dorner AJ, Wang S-J,
Guerciolini R, Huang S-M. Pharmacogenomic
data submissions to the FDA: clinical pharma-
cology case studies. Pharmacogenomics 2004;
5(5): 513517.
12 Trepicchio WL, Williams GA, Essayan D, Hall
ST, Harty LC, Shaw PM, et al. Pharmacoge-
nomic data submissions to the FDA: clinical
case studies. Pharmacogenomics 2004; 5(5):
519524.
References 65
13 Co-Development of Drug, Biological and De-
vice Products, 29 July 2004, Arlington, VA.
Washington, DC: U.S. Food and Drug Admin-
istration/Horsham, PA: Drug Information As-
sociation. Available: http://www.diahome.org/
Content/Events/04040.pdf [accessed August 31
2004].
14 Frueh FW, Huang SM, Lesko LJ., 112-12 edi-
torial: regulatory acceptance of toxicogenomics
data. Environ Health Perspect 2004;
112(12):A663664.
15 Lesko L, Woodcock J. Translation of pharma-
cogenomics and pharmacogenetics: a regula-
tory perspective, Nat Rev Drug Discov 2004; 3:
763770.
16 ICH E5 (1998) Guidance on Ethnic Factors in
the Acceptability of Foreign Clinical Data,
http://www.fda.gov/cder/guidance/index.htm, E5
Questions and Answers (2004) http://www.fda.-
gov/cder/guidance/6200fnl.pdf [accessed Sep-
tember 4, 2004].
17 Cross J, Lee H, Westelinck A, Nelson J, Grud-
zinskas C, Peck C. Postmarketing drug dosage
changes of 499 FDA-approved new molecular
entities, 19801999. Pharmacoepidemiol Drug
Safety 2002; 11(6): 439446.
18 Lazarou J, Pomeranz BH, Corey PN. Inci-
dence of Adverse Drug Reactions in Hospita-
lized Patients, A Meta-analysis of Prospective
Studies, JAMA 1998; 279: 12001205.
19 Huang, S-M, Miller M, Toigo, T, Chen, M, Sa-
hajwalla, C, Lesko, LJ, Temple, R. Evaluation
of Drugs in Women: Regulatory Perspective
in Section 11, Drug Metabolism/Clinical phar-
macology (section editor: Schwartz, J), pp.
848859, in: Principles of Gender-Specific
Medicine, Ed., Legato M, Academic Press,
2004
20 Clinical Pharmacology and Biopharmaceutics
NDA Review Template. (Issued 4/27/2004,
Posted 6/24/2004). http://www.fda.gov/cder/
mapp/4000.4.pdf (last accessed August 31,
2004).
21 Labeling information from http://www.fda.gov/
cder/approval/index.htm or Physicians Desk
Reference at http://pdrel.thomsonhc.com/pdrel/
librarian
22 Phillips KA, Veenstra DL, Oren E, Lee JK, Sa-
dee W. Potential Role of Pharmacogenomics
in Reducing Adverse Drug Reactions: A Sys-
tematic Review, JAMA 2001; 286: 22702279.
23 Evans WE, Relling MV, Pharmacogenomics:
Translating Functional Genomics into Ra-
tional Therapeutics, Science 1999; Oct 15:
487491.
24 Human cytochrome P450 (CYP) nomenclature
committee web http://www.imm.ki.se/cypalleles/
[last accessed September 4, 2004]
25 Xie HG, Kim RB, Wood AJ, Stein CM. Molec-
ular basis of ethnic differences in drug dis-
position and response. Annu Rev Pharmacol
Toxicol 2001; 41: 81550.
26 Evans WE, McLeod HL, Pharmacogenomics
drug disposition, drug targets, and side ef-
fects. N Engl J Med 2003; 348(6): 538549.
27 Weinshilboum R, Inheritance and drug re-
sponse, N Engl J Med 2003; 348(6): 529537.
28 Pauli-Magnus C, Kroetz DL, Functional impli-
cations of genetic polymorphisms in the mul-
tidrug resistance gene MDR1 (ABCB1),
Pharm Res 2004; 21(6): 904913.
29 Evans WE, Relling MV, Moving towards indi-
vidualized medicine with pharmacogenomics.
Nature 2004; 429(6990): 464468.
30 Weinshilboum R, Wang L. Pharmacoge-
nomics: bench to bedside, Nat Rev Drug Dis-
cov 2004; 3(9): 739748.
31 Hill MA, Wilke RA, Caldwell MD, Berg RL,
Glurich I, Burmester JK. Relative impact of
covariates in prescribing warfarin according to
CYP2C9 genotype. Pharmacogenetics 2004;
14(8): 539547.
32 Peyvandi F, Spreafico M, Siboni SM, Moia M,
Mannucci PM. CYP2C9 genotypes and dose
requirements during the induction phase of
oral anticoagulant therapy, Clin Pharmacol
Ther 2004; 75(3): 198203.
33 Higashi MK, Veenstra DL, Kondo LM, et al.
Association between CYP2C9 genetic variants
and anticoagulation-related outcomes during
warfarin therapy. JAMA 2002; 287: 16901698.
34 Furuta T, Shirai N, Sugimoto M, Ohashi K,
Ishizaki T. Pharmacogenomics of proton
pump inhibitors, Pharmacogenomics 2004;
5(2): 181202.
35 Eckhardt K, Li S, Ammon S, Schanzle G, Mi-
kus G, Eichelbaum M. Same incidence of ad-
verse drug events after codeine administration
irrespective of the genetically determined dif-
ferences in morphine formation, Pain 1998;
76(1-2): 2733.
36 Rouits E, Boisdron-Celle M, Dumont A, Guer-
in O, Morel A, Gamelin E. Relevance of differ-
ent UGT1A1 polymorphisms in irinotecan-in-
duced toxicity: a molecular and clinical study
of 75 patients, Clin Cancer Res 2004; 10(15):
51515159.
37 Sai K, Saeki M, Saito Y, Ozawa S, Katori N,
Jinno H, Hasegawa R, Kaniwa N, Sawada J,
Komamura K, Ueno K, Kamakura S, Kitakaze
M, Kitamura Y, Kamatani N, Minami H, Oht-
su A, Shirao K, Yoshida T, Saijo N. UGT1A1
haplotypes associated with reduced glucuroni-
dation and increased serum bilirubin in irino-
tecan-administered Japanese patients with
cancer, Clin Pharmacol Ther 2004; 75(6): 501
515.
38 Iyer L, Das S, Janisch L, Wen M, Ramirez J,
Karrison T, Fleming GF, Vokes EE, Schilsky
RL, Ratain MJ. UGT1A1*28 polymorphism as
a determinant of irinotecan disposition and
toxicity, Pharmacogenomics J 2002; 2(1): 43
47.
39 Evans WE. Pharmacogenetics of thiopurine S-
methyltransferase and thiopurine therapy,
Ther Drug Monit 2004; 26(2): 186191.
40 Weinshilboum R, Thiopurine pharmacoge-
netics: clinical and molecular studies of thio-
purine methyltransferase, Drug Metab Disp
2001; 29(4 Pt 2): 601605.
41 Evans WE, Hon YY, Bomgaars L, Coutre S,
Holdsworth M, Janco R, Kalwinsky D, Keller
F, Khatib Z, Margolin J, Murray J, Quinn J,
Ravindranath Y, Ritchey K, Roberts W, Rogers
ZR, Schiff D, Steuber C, Tucci F, Kornegay N,
Krynetski EY, Relling M. Preponderance of
thiopurine S-methyltransferase deficiency and
heterozygosity among patients intolerant to
mercaptopurine or azathioprine, J Clin Oncol
2001; 19(8): 22932301.
42 Dai D, Tang J, Rose R, Hodgson E, Bienstock
RJ, Mohrenweiser HW, Goldstein JA. Identifi-
cation of variants of CYP3A4 and characteriza-
tion of their abilities to metabolize testoster-
one and chlorpyrifos, J Pharmacol Exp Ther
2001; 299(3): 825831.
43 Haufroid V, Mourad M, Van Kerckhove V,
Wawrzyniak J, De Meyer M, Eddour DC, Mal-
aise J, Lison D, Squifflet JP, Wallemacq P.
The effect of CYP3A5 and MDR1 (ABCB1)
polymorphisms on cyclosporine and tacroli-
mus dose requirements and trough blood lev-
els in stable renal transplant patients, Phar-
macogenetics 2004; 14(3): 147154.
44 Zheng H, Zeevi A, Schuetz E, Lamba J,
McCurry K, Griffith BP, Webber S, Ristich J,
Dauber J, Iacono A, Grgurich W, Zaldonis D,
McDade K, Zhang J, Burckart GJ. Tacrolimus
dosing in adult lung transplant patients is re-
lated to cytochrome P4503A5 gene poly-
morphism, J Clin Pharmacol 2004; 44(2):
135140.
45 Tsuchiya K, Gatanaga H, Tachikawa N, Teruya
K, Kikuchi Y, Yoshino M, Kuwahara T, Shira-
saka T, Kimura S, Oka S. Homozygous
CYP2B6 *6 (Q172H and K262R) correlates
with high plasma efavirenz concentrations in
HIV-1 patients treated with standard efavir-
enz-containing regimens, Biochem Biophys
Res Commun 2004; 319(4): 13221326.
46 Niemi M, Leathart JB, Neuvonen M, Backman
JT, Daly AK, Neuvonen PJ, Polymorphism in
CYP2C8 is associated with reduced plasma
concentrations of repaglinide, Clin Pharmacol
Ther 2003; 74(4): 380387.
47 Xu C, Rao YS, Xu B, Hoffmann E, Jones J,
Sellers EM, Tyndale RF. An in vivo pilot study
characterizing the new CYP2A6*7, *8, and
*10 alleles, Biochem Biophys Res Commun
2002; 290(1): 318324.
48 Hoffmeyer S, Burk O, von Richter O, Arnold
HP, Brockmoller J, Johne A, Cascorbi I, Gerl-
off T, Roots I, Eichelbaum M, Brinkmann U.
Functional polymorphisms of the human mul-
tidrug-resistance gene: multiple sequence var-
iations and correlation of one allele with P-gly-
coprotein expression and activity in vivo, Proc
Natl Acad Sci USA 2000; 97(7): 34733478.
49 Kim RB, Leake BF, Choo EF, Dresser GK,
Kubba SV, Schwarz UI, Taylor A, Xie HG,
McKinsey J, Zhou S, Lan LB, Schuetz JD,
Schuetz EG, Wilkinson GR. Identification of
functionally variant MDR1 alleles among Eu-
ropean Americans and African Americans,
Clin Pharmacol Ther 2001; 70(2): 189199.
50 Fellay J, Marzolini C, Meaden ER, Back DJ,
Buclin T, Chave JP, Decosterd LA, Furrer H,
Opravil M, Pantaleo G, Retelska D, Ruiz L,
Schinkel AH, Vernazza P, Eap CB, Telenti A.
Swiss HIV Cohort Study, Response to anti-
retroviral treatment in HIV-1-infected indivi-
duals with allelic variants of the multidrug re-
sistance transporter 1: a pharmacogenetics
study, Lancet 2002; 359(9300): 3036.
51 Siddiqui A, Kerb R, Weale ME, Brinkmann U,
Smith A, Goldstein DB, Wood NW, Sisodiya
SM. Association of multidrug resistance in
epilepsy with a polymorphism in the drug-
transporter gene ABCB, N Engl J Med 2003;
348(15): 14421428.
References 67
52 Ruano G, et al., XIV International Sympo-
sium on Drugs Affecting Lipid Metabolism,
July 12, 2003; abstract in 52nd Annual Ameri-
can College of Cardiology (Scientific Session)
March 30April 2, 2003, Chicago, IL.
53 Nishizato Y, Ieiri I, Suzuki H, Kimura M, Ka-
wabata K, Hirota T, Takane H, Irie S, Kusu-
hara H, Urasaki Y, Urae A, Higuchi S, Otsubo
K, Sugiyama Y. Polymorphisms of OATP-C
(SLC21A6) and OAT3 (SLC22A8) genes: con-
sequences for pravastatin pharmacokinetics,
Clin Pharmacol Ther 2003; 73(6): 554565.
54 Priori SG, Napolitano C, Schwartz PJ, Grillo
M, Bloise R, Ronchetti E, Moncalvo C, Tulipa-
ni C, Veia A, Bottelli G, Nastoli J. Association
of long QT syndrome loci and cardiac events
among patients treated with beta-blockers.
JAMA 2004; 292(11): 13411344.
55 Yang P, Kanki H, Drolet B, Yang T, Wei J, Vis-
wanathan PC, Hohnloser SH, Shimizu W,
Schwartz PJ, Stanton M, Murray KT, Norris K,
George AL Jr, Roden DM. Allelic variants in
long-QT disease genes in patients with drug-
associated torsades de pointes. Circulation
2002; 105(16): 19431948.
56 Israel E, Drazen JM, Liggett SB, Boushey HA,
Cherniack RM, Chinchilli VM, Cooper DM,
Fahy JV, Fish JE, Ford JG, Kraft M, Kunsel-
man S, Lazarus SC, Lemanske RF Jr, Martin
RJ, McLean DE, Peters SP, Silverman EK,
Sorkness CA, Szefler SJ, Weiss ST, Yandava
CN. National Heart, Lung, and Blood Institu-
tes Asthma Clinical Research Network, Effect
of polymorphism of the beta(2)-adrenergic re-
ceptor on response to regular use of albuterol
in asthma. Int Arch Allergy Immunol 2001;
124(1-3): 183186.
57 Chasman DI, Posada D, Subrahmanyan L,
Cook NR, Stanton VP, Jr, Ridker PM., Phar-
macogenetic study of statin therapy and cho-
lesterol reduction. JAMA 2004; 291(23): 2821
2827.
58 Zheng HX, Burckart GJ, McCurry K, Webber
S, Ristich J, Iacono A, Dauber J, McDade K,
Grgurich W, Zaldonis D, Pillage G, Griffith
BP, Zeevi A. Interleukin-10 production geno-
type protects against acute persistent rejection
after lung transplantation, J Heart Lung
Transplant 2004; 23(5): 541546.
59 Hetherington S, Hughes AR, Mosteller M,
Shortino D, Baker KL, Spreen W, Lai E, Da-
vies K, Handley A, Dow DJ, Fling ME, Sto-
cum M, Bowman C, Thurmond LM, Roses
AD. Genetic variations in HLA-B region and
hypersensitivity reactions to abacavir, Lancet
2002; 359(9312): 11211122.
60 Lynch TJ, Bell DW, Sordella R, Gurubhagava-
tula S, Okimoto RA, Brannigan BW, Harris
PL, Haserlat SM, Supko JG, Haluska FG,
Louis DN, Christiani DC, Settleman J, Haber
DA. Activating mutations in the epidermal
growth factor receptor underlying responsive-
ness of non-small-cell lung cancer to gefiti-
nib., N Engl J Med. 2004; 350(21): 21292139.
Epub 2004 Apr 29.
61 Louis E, El Ghoul Z, Vermeire S, DallOzzo S,
Rutgeerts P, Paintaud G, Belaiche J, De Vos
M, Van Gossum A, Colombel JF, Watier H.
Association between polymorphism in IgG Fc
receptor IIIa coding gene and biological re-
sponse to infliximab in Crohns disease, Ali-
ment Pharmacol Ther 2004; 19(5): 511519.
62 Chow, M, Huang, S, Sahajwalla, C, Lesko, L.
An informal survey of pharmacogenetics/
pharmacogenomics (PGTX) in a sample of
INDs and NDAs, Clin Pharmacol Ther 2003;
73 (2)33.
63 Lesko LJ, Salerno RA, Spear BB, Anderson
DC, Anderson T, Brazell C, Collins J, Dorner
A, Essayan D, Gomez-Mancilla B, Hackett J,
Huang SM, Ide S, Killinger J, Leighton J,
Mansfield E, Meyer R, Ryan SG, Schmith V,
Shaw P, Sistare F, Watson M, Worobec A.
Pharmacogenetics and pharmacogenomics in
drug development and regulatory decision
making: report of the first FDA-PWG-PhRMA-
DruSafe Workshop, J Clin Pharmacol 2003;
43(4): 342358.
64 Huang S-M. Regulatory issues in genotyping
metabolizing enzymes, presentation at the
FDA/PhRMA/JHU educational workshop,
September 1314, 2004, Rockville, MD. http://
www.fda.gov/cder/offices/ocpb/
workshops.htm
65 Flockhart D. Special consideration of meta-
bolic biomarker: CYP2D6 presentation at the
workshops.htm
66 Gaedigk A, Bradford LD, Marcucci KA, Leeder
JS. Unique CYP2D6 activity distribution and
genotypephenotype discordance in black
Americans, Clin Pharmacol Ther 2002; 72(1):
7689.
67 Liou YJ, Wang YC, Bai YM, Lin CC, Yu SC,
Liao DL, Lin MW, Chen JY, Lai IC. Cyto-
chrome P-450 2D6*10 C188T polymorphism
is associated with antipsychotic-induced per-
sistent tardive dyskinesia in Chinese schizo-
phrenic patients, Neuropsychobiology 2004;
49(4): 167173.
68 Wan YJ, Poland RE, Han G, Konishi T, Zheng
YP, Berman N, Lin KM. Analysis of the
CYP2D6 gene polymorphism and enzyme ac-
tivity in African-Americans in southern Cali-
fornia, Pharmacogenetics 2001; 11(6): 489
499.
69 Furman KD, Grimm DR, Mueller T, Holley-
Shanks RR, Bertz RJ, Williams LA, Spear BB,
Katz DA. Impact of CYP2D6 intermediate me-
tabolizer alleles on single-dose desipramine
pharmacokinetics, Pharmacogenetics 2004;
14(5): 279284.
70 Andersson T. Special consideration of meta-
bolic biomarker: CYP2C19 presentation at
the FDA/PhRMA/JHU educational workshop,
workshops.htm
71 Blaisdell J, Mohrenweiser H, Jackson J, Fergu-
son S, Coulter S, Chanas B, Xi T, Ghanayem
B, Goldstein JA. Identification and functional
characterization of new potentially defective
alleles of human CYP2C19, Pharmacogenetics
2002; 12(9): 703711.
72 Milos P. Special consideration of metabolic
biomarker: CYP2C9 presentation at the FDA/
PhRMA/JHU educational workshop, Septem-
ber 1314, 2004, Rockville, MD. http://
workshops.htm
73 Ratain M. Special consideration of metabolic
biomarker: UGT1A1 presentation at the
workshops.htm
74 Sai K, Saeki M, Saito Y, Ozawa S, Katori N,
Jinno H, Hasegawa R, Kaniwa N, Sawada J,
Komamura K, Ueno K, Kamakura S, Kitakaze
M, Kitamura Y, Kamatani N, Minami H, Oht-
su A, Shirao K, Yoshida T, Saijo N. UGT1A1
haplotypes associated with reduced glucuroni-
dation and increased serum bilirubin in irino-
tecan-administered Japanese patients with
cancer, Clin Pharmacol Ther 2004; 75(6): 501
515.
75. Innocenti F, Undevia SD, Iyer L, Chen PX.
Das S, Kocherginsky M, Karrison T, Janisch L,
Ramirez J, Rudin CM, Vokes EE, Ratain MJ.,
Genetic variants in the UDP-glucuronosyl-
transferase 1A1 gene predict the risk of severe
neutropenia of irinotecan., J Clin Oncol 2004;
22(8): 13821388. Epub 2004 Mar 08.
76 Hamelin BA, Bouayad A, Methot J, Jobin J,
Desgagnes P, Poirier P, Allaire J, Dumesnil J,
Turgeon J. Significant interaction between the
nonprescription antihistamine diphenhydra-
mine and the CYP2D6 substrate metoprolol
in healthy men with high or low CYP2D6 ac-
tivity, Clin Pharmacol Ther 2000; 67(5): 466
677.
77 Stearns V, Johnson MD, Rae JM, Morocho A,
Novielli A, Bhargava P, Hayes DF, Desta Z,
Flockhart DA. Active tamoxifen metabolite
plasma concentrations after coadministration
of tamoxifen and the selective serotonin reup-
take inhibitor paroxetine, J Natl Cancer Inst
2003; 95(23): 17581764.
78 Andersson T, Cederberg C, Edvardsson G,
Heggelund A, Lundborg P. Effect of omepra-
zole treatment on diazepam plasma levels in
slow versus normal rapid metabolizers of
omeprazole, Clin Pharmacol Ther 1990; 47(1):
7985.
79 Yasui-Furukori N, Takahata T, Nakagami T,
Yoshiya G, Inoue Y, Kaneko S, Tateishi T., Dif-
ferent inhibitory effect of fluvoxamine on
omeprazole metabolism between CYP2C19
genotypes, Br J Clin Pharmacol 2004; 57(4):
48794.
80 Yin OQP, Tomlinson B, Waye MMY, Chow
AHL, Chow MSS. Pharmacogenetics and
herbdrug interactions: experience with Gink-
go biloba and omeprazole, Pharmacogenetics,
2004; 14(12): 841850.
81 Purinenthol labeling. http://www.fda.gov/cder/
foi/label/2004/09053s024lbl.pdf [last accessed
September 18, 2004].
82 Huang S-M, Lesko, LJ. Drug-drug, drug-diet-
ary supplement and drug-citrus fruit and
other food interactions what have we
learned? J Clin Pharmacol 2004; 44(6): 559
569.
83 FR notice (2000): Labeling guideline (Federal
Register 65: 247; 8108281131; December 22,
2000).
84 Draft guidance for industry Labeling for Hu-
man Prescription Drug and Biological Prod-
ucts Implementing the New Content and
References 69
Format Requirements (will be published at
http://www.fds.gov/cder/regulatory/
default.htm).
85 Attachment to guidance on pharmacogenomic
data submission, examples of voluntary sub-
missions or submissions required under
21cFR 312, 314, or 601, March 2005, http://
www.fds.gov/cder/guidance/6400fnlattch.pdf
86 Drug-diagnostic co-development concept pa-
per, April 2005, http://www.fda.gov/cder/geno-
mics/pharmacoconcept.fn.pdf
Abstract
The human reference genome has been
sequenced, and progressing research in
the post-genome era is revealing the im-
pact of genetic variation on fitness. Hu-
man genomes are more than 99% identi-
cal, but less than 1% variation determines
genetic differences between individuals.
Over 80% of this variation is due to single
nucleotide polymorphisms (SNPs). These
alter the sequence in the genetic code by
changing single bases. Two out of three
SNPs are cytosine to thymine (C?T) tran-
sitions. More than 11 million SNP posi-
tions are believed to be present in the en-
tire human population. Of these, about 3
million differ between any two given indi-
viduals. A portion of these SNPs is located
in exons and/or regulatory elements,
which can lead to changes in the amino
acid sequence of resulting proteins, or af-
fect gene activity. This is a major cause for
different individual responses to drugs
and environmental substances. Further-
more, SNPs can determine the presence of
susceptibility alleles and, therefore, genetic
predisposition for hereditary diseases.
Most SNPs do not influence cell function,
but those that do are of high value for bio-
medical research. Currently, the focus of
academic medical research and the phar-
maceutical industry is on metabolic en-
zymes which control drug activity and on
structural proteins which change disease
susceptibility (see also Part I, Chapter 2).
To determine the large amount of base vari-
ation in the human genome reproducible,
fast, and economical techniques are re-
quired. Array-On, an innovative German
biotechnology enterprise, has developed an
extremely precise high-throughput DNA
chip-based technology for SNP typing. Two
patents were granted: a novel hybrid spot-
ting technique for microarrays and appen-
dant areal arrays that are part of a solid-
phase primer extension approach for auto-
mated SNP detection. The main advantage
of the new technology is that numerous in-
dividuals can be screened for various SNPs
on a single DNA chip without crosstalk be-
tween individual probes and samples. The
ability to examine the same genes in a large
number of individuals in one miniaturized
reaction chamber leads to great savings in
materials and time. The number of indivi-
duals and SNPs analyzed on a single chip
can be combined in a most flexible manner,
and up to 50000 simultaneous allele calls
are possible. Even orthologous genes of dif-
ferent species may be analyzed and com-
pared on the same microarray. The so-called
polydimensional SNP chip will, among
other techniques, contribute to the develop-
71
ISBN: 3-527-31184-X
3
Large-scale Detection of Genetic Variation:
The Key to Personalized Medicine
ment of safer and more effective medicines
which will address unmet medical needs
and be available faster with enormous sav-
ings. The creation of medicines with ap-
proved risk-benefit ratios, in particular
with reduced unwanted side effects or ad-
verse drug reactions (ADRs) and higher
personalized efficacy, seem to be reachable
with pharmacogenetic approaches. Major
research efforts are still necessary to fulfill
the promises of pharmacogenetic testing
in the future. It is already expected that
regulatory authorities will ask for SNP
genotyping not only to reduce clinical
trials in size and time but also to reduce
the risks for participants. The goal is to
generate genetically associated drug targets
with a break-through for the development
of first personalized medicines, and the
better control of generalized drugs bearing
high risks for patients with certain genetic
backgrounds.
This article provides an overview: 1) of
the latest pharmacogenetic findings; 2) of
validated SNPs ready for the implementa-
tion in pharmacogenetic programs; 3) of
state-of-the-art SNP technologies and detail
about the Array-On technology; and 4)
the future potential of pharmacogenetics
in the drug developmental process.
Abbreviations
5-LO 5-lipoxygenase
ACE angiotensin-converting
enzyme
ALL acute lymphoblastic leukemia
APOE apolipoprotein E
BDNF brain-derived neurotrophic
factor
CETB cholesteryl ester transfer
protein
dHPLC denaturing high-performance
liquid chromatography
dsDNA double-stranded DNA
EGFR epidermal growth factor
receptor
FLAP 5-lipoxygenase activating
protein
GVPs genome-wide variant pat-
terns
HLA human leukocyte antigen
LD linkage disequilibrium
MALDI-TOF matrix-assisted laser desorp-
MS tion ionization time of
flight mass spectrometry
MTHFR methylenetetrahydrofolate
reductase
NCBI National Centre of Biotech-
nology Information
OLA oligonucleotide ligation
assays
PCR polymerase chain reaction
PGRN Pharmacogenetics Research
Network
RA rheumatoid arthritis
REC DNA repair system
RET rearranged during transfec-
tion
RFLP restriction fragment length
polymorphism
SMA spinal muscle dystrophy
SMN survival motor neuron
genes
SNPs single nucleotide poly-
morphisms
Strength Statin response examined
by genetic haplotype
TAU s Parkinson related gene/
protein
TPMT thiopurine methyltrans-
ferase
TSER thymidilate synthase
enhancer region
VNTR variable number of tandem
repeats
WNK4 (with-no-lysine) kinase 4
3.1
Genetic Variation, Disease Susceptibility
and Drug Response
Common complex diseases with a genetic
component in etiology and pathogenesis
are a widely proclaimed focus of medical
research in the genomics or post-genomics
era. Promises are vast, but so are the chal-
lenges in identifying gene variants which
influence susceptibility to disease, disease
progression, or predict treatment efficacy
and safety. So far, the genetics of a num-
ber of monogenic disorders have been
solved. In these disorders, usually a single
gene variant influences the disease pheno-
type with high penetrance. However, inter-
actions with the genetic background and
with the environment still play an impor-
tant role. This is exacerbated in diseases
with more complex genetics, where vari-
ants of two or more genes contribute to
the phenotype. In general, these contribu-
tions individually have a lower penetrance
and the distinction between disease
genes and genetic background becomes
blurred. Due to the lower impact on phe-
notype, the contribution of each gene is
much harder to detect. A number of dis-
eases with presumably complex genetics
have been investigated in the past years.
Some progress has been made, and several
genes associated with disease susceptibili-
ty, disease progression and treatment suc-
cess have been identified. So far, however,
there are no diseases with complex genet-
ics which are considered to be solved ex-
haustively. Many aspects of complex dis-
eases are still unclear. It is unknown how
many genes and their variations are inter-
acting strongly in any given disease and
how these specific genes interact with the
general genetic background in a certain in-
dividual.
The penetrance of genetic factors is an-
other parameter in complex diseases. If
the penetrance of these factors is indeed
very low, as studies to date indicate, new
approaches will be necessary to detect
them. It seems already clear that the rela-
tive genetic risk of single variants is very
small in complex diseases. Therefore, pre-
dictions can only provide risk probabilities
but not risk certainties. On the other hand,
in cases of low complexity and single nu-
cleotide polymorphisms (SNPs) with high
penetrance, the predictive power of tests
reaches close to certainty as shown for he-
mophilia.
If the genetic bases of phenotypes are
known, disease risks may be assessed
before disease onset. This is desirable only
if preventive or therapeutic action can be
taken. Otherwise, the psychological burden
for the individual may be unwarranted.
Ideally, preventive steps are undertaken in
due time. Pre-symptomatic screening is
applied, if disease prevails in families
(for example, breast cancer and variation
in estrogen receptor subunits or Hunting-
tons disease). The same is true for prena-
tal and newborn screening, if family his-
tories indicate the need (for example
phenylketonuria, galactosemia, hypothyr-
oidism).
One of the problems to be addressed is
how to bring SNP diagnostics to the bed-
side, in due time contributing to patient
safety in a cost-effective way. This means,
to find out which testing is valid, informa-
tive and useful in which situations. The
awareness of the impact for medical genet-
ics is a prerequisite for the modernization
of healthcare. Unforeseen side effects and
ADRs are thought to be the fourth to sixth
leading cause in mortality in industrialized
nations. In this context, it is useful to dis-
tinguish between four categories:
early (pre-symptomatic) disease detection
3.1 Genetic Variation, Disease Susceptibility and Drug Response 73
dissecting complex disease mechanisms
predicting drug safety
predicting drug efficacy.
This promises to be fruitful fields of inves-
tigation with important clinical impact.
3.2
Pharmacogenetics and Pharmacogenomics
3.2.1
Terms, History and Definitions
The term pharmacogenetics has already
been formed in 1959 by Vogel [1], and
may even date back to 1931 when A. L.
Fox reported on taste blindness in the abil-
ity to taste phenylthiocarbamide, which is
regarded as the first pharmacogenetic find-
ing. The first online record is from 1963
about the design of pharmacogenetic stud-
ies of drug metabolism [2]. The term phar-
macogenomics appeared more recently,
and was first published in 1997 in the wake
of the human genome project, when the
complete sequence of the complex human
genome was already expected to be realized
[3, 4]. The use of both terms should be in
accordance to the traditional definitions
that, in genomics whole genomes and inter-
acting traits or genes therein are studied,
while individual genes, their alleles and dif-
ferential expression are the objectives in ge-
netics research. Both terms are used with
regard to the influence of genetics and
genomics on pharmacology kinetics and dy-
namics in response to medicines.
3.2.2
Pharmacogenetics
The pharmacogenetics approach is very
well suited to solve the problems of single-
gene (Mendelian or monogenic) disorders,
in which variation in a single gene has a
large effect on disease susceptibility (i.e., a
large penetrance). A historic example of
drug response and ADRs is the muscle re-
laxant succinylcholine, of which many pa-
tients died in the 1950s when undergoing
anesthesia. Important examples of modern
pharmacogenetics to discover susceptibility
genes are cystic fibrosis, Huntingtons dis-
ease and Duchenne muscular dystrophy.
In the contexts of low-complexity genetic
disorders, the candidate gene approach is
a basic tool to identify and isolate genes. It
is based on testing specific hypotheses to
elucidate the role of genes in susceptibility
and drug response and to identify, for ex-
ample, key enzymes in drug-metabolizing
pathways. Proteins belonging to the same
pathway can be identified and potentially
serve as new drug targets. In this func-
tional approach a selected subset of (candi-
date) genes is screened. These genes are
potentially relevant for drug absorption,
distribution, metabolism and excretion, or
are known to prevail in family histories
and have genetic map-based linkage infor-
mation.
So far, successful research has been car-
ried out in pharmacogenetics. Approxi-
mately 500 human gene products are un-
der development as targets for todays
medicines, and it is estimated that the pro-
gressing analysis of the human genome
will yield 5000 to 10000 additional targets
[5] (see Part I, Chapter 4).
3.2.3
Pharmacogenomics
Pharmacogenomics is applied in cases of
multi-factorial (complex or polygenic) situ-
ations such as cancer, heart disease, or dia-
betes. Quantitative traits composed of dif-
ferent loci are involved. Complexity is
caused by multiple genegene interactions
of which single genes may occur in several
variants (alleles). Involved gene products
may vary in their response to compounds
added to the bodys metabolism. SNPs
mainly contribute to genetic variability and
diversity in the human gene pool. Expres-
sion of individual genes may be influenced
by SNPs affecting regulatory sequence mo-
tifs in the DNA.
Since the one-geneone-protein para-
digm has fallen, we see much more pro-
teins than genes. These extra proteins
arise most probably from differential spli-
cing and varying activity of transcription
factors in different tissues under different
physiological conditions. This is mainly
caused by SNPs which alter binding or
splice sites in the DNA. As well as the ge-
netic components, environmental factors
(geneenvironment interactions) also play
a role.
In addition to single gene and multi-fac-
torial diseases, the genetic status or indi-
vidual genetic background can be even
more complex, if base changes in the
DNA sequence are not only chromosomal
or do not refer to single bases only. Inser-
tions or deletions (InDel-mutations) can
change amino acids in the resulting pro-
teins or cause frame shifts in the open
reading frames (ORFs) in coding se-
quences, resulting in totally different pro-
teins or no expression at all. Non-chromo-
somal changes occur in the mitochondrial
genome, where important metabolizing
enzymes are encoded. Further mentioned
may be the epigenome that silences or en-
hances gene expression via methylation
patterns of cytosine residues. Somatic mu-
tations occur in specific locations of an or-
ganism, for instance in many cancers
where solely tumor tissue is affected. On a
different level, phenotypes are strongly af-
fected by protein interactions and regula-
tion, as well as morphology. Stochastic ef-
fects also play a role. Taking into consid-
eration all of the factors that possibly con-
tribute to the determination of a pheno-
type, it is clear that the association of a
single base pair difference is more easily
detected the fewer other factors are in-
volved that is, the higher its penetrance.
3.2.4
Environmental Factors
It is thought that clinical outcome or ad-
verse drug reaction events are not alone
influenced by the personal genetic make-
up but also by gender and age, weight and
health status, and to a high degree by en-
vironmental factors. This results, for exam-
ple, from behavioral components (diet, al-
cohol intake, tobacco smoke, sports), abio-
tic stresses such as radiation, heat, cold, or
noise, and even mental factors (e.g., the
placebo effect). All of these factors are add-
ing additional layers to complexity. Today,
we see emerging research fields in nutri-
genomics and envirogenomics. To obtain
an overview of interacting factors, certain
aspects of systems biology such as genetic
variation, epigenetics, gene expression,
protein regulation and turn-over, protein
interactions, cell interactions, and tissue
morphology must be considered. The great
challenge to understand complex disease
and polygenic disorders in the light also
of non-genetic factors can most probably
not be solved without considering genetic
variation, especially SNPs, and well-de-
scribed populations in distinct environ-
ments (population association studies).
Studies should be designed in a way that
carrier families can be identified and seg-
regation tracked in subsequent studies
with considerably smaller sample sizes.
Even interspecies comparisons, on the ba-
sis of sequence homologies, can be useful
to discover important SNP candidates.
3.2 Pharmacogenetics and Pharmacogenomics 75
Such a comparative evolutionary genomics
study is now intended by Celera Diagnos-
tics, which compares human, chimpanzee
and mouse sequences. Future potential
can be imagined by considering that, in
the next few years, 16 mammalian ge-
nomes are expected to be completed.
In conclusion, pharmacogenetics is to
study the impact of single gene variations
on drug response or disease susceptibility,
while pharmacogenomics covers a broader
field, taking interactions of genes and their
variants into account. However, the aim of
both systems is the same to predict sus-
ceptibility or drug response and to select
appropriate drugs and doses for each pa-
tient on the basis of his or her genetic
background.
3.3
Personalized Medicine
3.3.1
Low-complexity Disorders
One of the oldest and best-known exam-
ples is the variation in the drug-metaboliz-
ing enzymes of the cytochrome P450 fami-
ly (see also Part VII, Chapter 2). It is the
metabolic pathway of choice for many fre-
quently prescribed drugs. In the CYP2D6
gene alone, more than 70 allelic variants
have been detected (www.imm.ki.se.CYPal-
leles/cyp2d6.htm). Variability in patients
reaches from very poor metabolizers up to
ultra-rapid metabolizers. Poor metabolizers
have a high chance of accumulating toxic
concentrations of drugs when conventional
doses are prescribed. Currently, investiga-
tions are under development to persona-
lize the dosing for individual patients, or
groups of them.
The CYP2C9 gene is known to cause
ADRs with drugs used to treat cardiovas-
cular disease [6]. Carriers of this poly-
morphism require lower doses of the
drugs digoxin and warfarin; indeed, in the
latter case the dose can vary up to 20-fold
among individuals. The CYP2C19 gene is
important for the pharmacokinetics of a
wide variety of antidepressive agents [7].
Of special interest in psychiatry are the re-
ceptors of neurotransmitters (e.g. seroto-
nin) and their transporter proteins respon-
sible for distribution or re-uptake of neuro-
transmitter substances [8].
Numerous examples exist from oncology
research. In the folate metabolism path-
way, gene products of MTHFR, REC and
TSER are known to cause ADRs in re-
sponse to chemotherapeutic agents such
as methotrexate and 5-fluorouracil [9]. Of
high impact are insufficiently functioning
gene products from the DNA repair sys-
tems [10] and mutations in the epidermal
growth factor receptor (EGFR) gene [11].
This factor is targeted for example by the
drug gefitinib (Iressa
), which acts as an
inhibitor of the EGFR kinase and is used
as a cancer therapeutic agent. Patients car-
rying at least one out of two important
mutations in the EGFR gene respond ex-
tremely well to Iressa.
Further research has also been con-
ducted in patients with asthma, infectious
diseases (HIV, meningitis and hepatitis C)
and last, but not least, the well-regarded
TPMT (thiopurine methyltransferase) stud-
ies, which provide one of the best exam-
ples in predictive pharmacogenetics. Mer-
captopurines are used for the treatment of
autoimmune diseases, organ transplanta-
tions and acute lymphoblastic leukemia
(ALL), the most common form of child-
hood cancer. If not metabolized correctly,
life-threatening concentrations of these
agents can be accumulated. The alleles
TPMT 2, 3A, 3C cover more than 95% of
the variations, and pharmacogenetic tests
for this monogenic trait on chromosome 6
are available. Poor metabolizers receive a
dosage reduced by as much as 95%. The
rare alleles 3B, 6 and 8 are currently under
investigation [12].
Not only poor metabolizers, but also
non-responders are known. Some 30% of
schizophrenics do not respond to anti-psy-
chotics.
Examples in which altered gene expres-
sion (but not protein structure or function)
influences drug response are known from
breast and pancreatic cancer tissues that
eventually over-express the HER2 gene.
The drug trastuzumab (Herceptin
) is
only effective in patients who over-express
the receptor. Trastuzumab binds to HER2
and, by blocking signaling, slows down tu-
mor growth. The therapy shows positive
results in about 2530% of all cases. The
HercepTest test kit, which measures
gene expression levels, but not the DNA
directly, is marketed along with the drug
(Roche and Genentech). This is currently a
unique combination, and a spearhead of
future developments (see also Part I,
Chapter 5).
3.3.2
Complex Disease
Local SNP fine scans in defined genomic
regions have been increasingly performed
during the past few years, whilst whole ge-
nome scans are still mostly carried out
with microsatellite genetic markers (vari-
able number of tandem repeats, VNTR
markers). The latter are well established in
genetics research, but lack relative genetic
instability and show an uneven distribu-
tion throughout the genome. SNPs are
quite stable genetic markers that are rela-
tively evenly distributed across the ge-
nome. They may serve as landmarks in fu-
ture genome studies and thereby enhance
the discovery of genes which are important
for drug response or susceptibility. Many
SNP markers have been discovered in a
variety of international projects (SNP Con-
sortium, HapMap, Human Genome Pro-
ject) and through contributions by specia-
lized networks such as the Pharmacoge-
netics Research Network (PGRN). By Sep-
tember 2004, a total of 693255 genotyped
SNPs had been released, and 62393760
genotypes had been detected by the Interna-
tional HapMap Consortium (www.hapma-
p.org). Further mention should also be
made of the National Centre of Biotechnol-
ogy Information (NCBI) variation database
dbSNP and the Seattle SNPs Project that
has evaluated gene-specific SNP poly-
morphisms in over 20 Caucasian and more
than 20 Afro-American individuals in order
to predict polymorphic sites and allele fre-
quencies. Further information sources for
reference include the pharmacogenomics
knowledge base and others (pharmgkb.org,
snp.cshl.org and rsi.ilsi.org). The latter web
address especially integrates pharmacoge-
netic microarray data.
More narrow SNP scans were very suc-
cessfully performed for candidate suscepti-
bility loci. Examples of strong association
exist for ischemic stroke, migraine, psoria-
sis, rheumatoid arthritis or Crohns dis-
ease. The investigation of known loci by
following the haplotype strategy has al-
ready been fruitful for brain-derived neuro-
trophic factor (BDNF) and obsessive com-
pulsive disorder [13], for RET and Hirsch-
sprungs disease [14], for apolipoprotein E
(APOE) and Alzheimers disease [15], and
for TAU and late-onset Parkinsons disease
[16].
In aiming to further improve population
association studies, Millennium Pharma-
ceuticals is building a database containing
large-scale registers of patients suffering
from rheumatoid arthritis, multiple myelo-
3.3 Personalized Medicine 77
ma and multiple sclerosis. All available in-
formation is filed for example, gene ex-
pression, genotype and phenotype data.
This dataset will be of invaluable benefit
for clinical trial design. The National Insti-
tute of Health (NIH) intends to launch a
similar project as an open access resource.
3.4
SNPs in Clinical Applications
3.4.1
Tests in Use
Several companies offer CYP gene pharma-
cogenetic testing kits, based on genotyping,
for subject inclusion or exclusion from clin-
ical trials. Gentris is marketing tests for the
five most predominant CYP alleles, while
Genelex is offering tests for the three major
alleles directly to the public. Roche is also
focusing on this issue, and offers the
CYP450 AmpliChip to estimate individual
dosing. Roche further claims to have chips
for cancer and chemotherapeutics as well
as leukemia in the pipeline.
Genaissance focuses on the Statin/APOE
system with a test called Strength (Statin
response examined by genetic haplotype).
The first ever high-density SNP map was
constructed around the APOE locus in
1997, and published the following year
[17]. Genaissance wishes to establish a
point-of-care system for the most important
cholesterol-lowering drugs. Statins also
have an anti-inflammatory potential that
targets the cholesteryl ester transfer protein
(CETB), for which the encoding gene exists
in two alleles (B1 and B2). The company
has also announced a test for individual re-
sponses to asthma drugs as being in the pi-
peline. The B2 adrenergic receptor and 5-li-
poxygenase (5-LO) show genetic variants,
and are each targets of asthma drugs; 5-
LO is an example of an SNP altering the
5' promoter sequence that regulates the ac-
tivity of a drug-related gene.
3.4.2
Candidates for Pharmacogenetic Testing
The APOE locus is not only associated with
poor response to cholesterol-lowering
drugs, but is also associated with a higher
risk of lower age of onset in Alzheimers
disease [18]. More differentiated pharmaco-
genetic tests will be offered on APOE. More-
over, neurodegenerative diseases are in the
focus of many pharmaceutical companies.
For angiotensin-converting enzyme
(ACE), a test can be expected in the near
future. The I and D alleles are most prob-
ably associated with ADRs in b-blocker
therapies. Likewise, a test can be expected
for the reninangiotensin pathway in-
volved in hypertension. A small deletion
in the (with-no-lysine) kinase 4 (WNK4)
gene causes bad regulation of the critically
balanced renal potassium/sodium excre-
tion system [19]. DXS Ltd. have announced
the development of tests for EGFR vari-
ants in order to predict the efficacy of can-
cer drugs.
Possible candidates are also the survival
motor neuron genes (SMN) in spinal mus-
cle dystrophy (SMA). In homozygous ab-
sence of the SMN1 gene (the primary
cause of SMA), the SMN2 genes (appear-
ing in different copy numbers) compen-
sate for the missing activity of SMN1. A
splice-site mutation in SMN2 is responsi-
ble for the only 10% production of correct
transcripts. Valproic acid compensates the
mutation by enhancing gene expression
and influencing the alternative splicing
factor Htra2-beta1. Evidence was provided
in a cell culture model with increasing
full-length transcripts under valproic acid
treatment [20].
3.4.3
Identification of New Candidates
There are several ways to detect genetic
contributions to a specific phenotype. Ge-
nome-wide linkage scans presume that the
genetic components of a phenotype segre-
gate linked to nearby markers in the ge-
nome due to the lower chance of homolo-
gous recombination between loci with a
small, rather than large, distance between
them. The advantage of genome-wide link-
age scans is that new loci can be detected
without prior knowledge or hypotheses. A
disadvantage is the generally low sensitiv-
ity of the method for low-penetrance gene
variants. Another drawback is that the re-
sults of a linkage scan usually comprise
rather large regions of the genome which
may contain a great number of genes.
Thus, the challenge to pinpoint the actual
culprit(s) remains.
Candidate gene association studies de-
pend on prior knowledge or hypotheses.
These may be derived from functional
analysis of biological processes, or they
may be the result of a genome scan. The
advantage of the candidate gene approach
is that definite hypotheses are tested and
specific answers can be expected. The
drawbacks include false-positive or false-
negative results, and that it seems difficult
to find unexpected candidates. This issue
may be overcome by genome-wide studies,
which are currently becoming feasible.
Common to both, linkage scans and
candidate gene studies are the problems of
multiple testing. When many hypotheses
are tested at once, the power of a given
study to detect true results decreases,
while the risk for false-positive results in-
creases. This is especially true for low-pe-
netrance gene variants. Sample sizes
which are large enough to balance this ef-
fect are usually difficult to obtain; there-
fore, it is of utmost importance that re-
sults are verified in independent studies
and that advantage is taken of intelligent
study design. Generally, efficient methods
for the detection of complex gene variant
interaction patterns are still lacking. The
new polydimensional microarray technol-
ogy, where many genes can be observed si-
multaneously in many samples, might of-
fer an opportunity to generate data as a ba-
sis for solving this problem.
In the investigation of diseases with
complex genetics, a number of questions
remain. It is not clear, how to estimate the
number of genes which are expected to
make significant contributions, though
this also depends on what is considered a
significant contribution. In general, this
would be determined by the ability to de-
tect the contribution. So far, however, this
has been limiting to an unacceptable de-
gree. The main reasons for this limitation
appear to be the number of samples in
family or case-control studies, and the reli-
ability, cost and speed of genotyping, as
well as the lack of efficient analysis meth-
ods to detect complex gene variant interac-
tions. The general lack of innovative study
designs to detect complex genetic contribu-
tions to diseases with more sensitivity and
a lower false-positive rate is a limiting fac-
tor. It appears prudent to study diseases
with complex genetics to gain more in-
sight into these points.
Rheumatoid arthritis (RA), for example,
is a common complex autoimmune dis-
ease with a strong support for a genetic
component from twin studies [21] and ge-
nome-wide linkage scans [2226]. The ge-
netics of susceptibility, pathogenesis, and
treatment outcome are presumably multi-
genic and complicated. RA is an inflam-
matory disease of the joints (but also of
connective tissue in general) which affects
about 1% of individuals in populations
3.4 SNPs in Clinical Applications 79
worldwide [27]. The results of genome-
wide scans in RA indicate that there may
be approximately 15 large regions in the
genome that may contain an unknown
number of associated genes. The identifi-
cation of their number and identity poses
a major challenge. Variants of many func-
tional candidate genes have been studied
[28], and several appear to be associated,
but many associations have not been veri-
fied. Most prominently associated are hu-
man leukocyte antigen (HLA) [29, 30] and
tumor necrosis factor (TNF) [31] alleles.
This information about RA makes it a
good test case for the analysis of diseases
with a complex genetic component.
A project at the University of Leipzig
(Germany) aims to investigate the genetics
of complex diseases in general, and strives
to contribute to the elucidation of the ge-
netic component in RA, in particular. The
working hypothesis is that the genetics of
RA are based on gene variant interactions
in genome-wide variant patterns (GVPs).
Together with several collaborators, the
project follows a strategy to identify GVPs.
Based on functional knowledge and on ge-
nome scan data, genes are selected which
are considered to be candidates for partici-
pation in RA-related GVPs. For these can-
didate genes SNPs are selected which at
the same time are good linkage markers
and have a high chance of influencing
gene structure or activity.
Samples available for the study are cases
and controls, and a set of family trios
which should reduce the genetic degrees
of freedom between affected and non-af-
fected individuals. This should improve
the power to detect association with single
SNPs or SNP variant patterns. The focus
is on avoiding false-negative results, and
deliberately accepting false positives. Vali-
dation studies on additional samples will
be carried out to verify positive results.
Considering limits on sample size and
requirements for detection power, the aim
is to achieve close to complete genotype
data sets, requiring very robust and reli-
able genotyping technologies. The Geno-
link
TM
single-base extension system with
mass spectrometry detection affords very
reliable results with excellent error track-
ing at medium sample throughput and
medium numbers of assays. The Array-On
single-base extension system with fluores-
cence detection on polydimensional arrays
promises similar quality data with in-
creased throughput which may allow the
project to be carried out more comprehen-
sively, and in a shorter time.
For data analysis to detect GVPs (or
parts thereof), multivariate testing and ma-
chine learning algorithms are employed.
The analysis promises to bring us a step
forward in understanding the pathome-
chanisms of RA, to provide marker sets
for predicting disease risks, and generally
to improve our intuition and knowledge
about diseases with complex genetics.
3.5
Strategies in SNP Discovery
The majority of genetic variation between
humans is due to SNPs. Some of these
change coding or regulatory sequences,
and thereby alter proteins in structure or
concentration. Although there is no strict
consent, a single base pair change in a
population is referred to as a mutation,
if the allele frequency is below 1%, above
that value as a SNP [32].
3.5.1
SNPs and Haplotypes
One method to handle the large amounts
of SNP data and extract information from
them in an effective way is linkage dise-
quilibrium (LD). This approach exploits
the observation that many SNPs are asso-
ciated to each other over long sequence
stretches, which are inherited in a block-
wise structure (haplotype blocks). SNPs
within blocks are in strong LD due to rela-
tively reduced recombination rates in af-
fected chromosomal regions. Haplotypes
are thought to be a useful tool for the rap-
id detection of high-penetrance SNPs. A
strong point is that haplotypes represent
inherited groups of SNPs that statistically
may influence drug response or suscepti-
bility risks more than individual SNPs do.
The HapMap Consortium is on the way to
genotype 1 million SNPs in a variety of
populations worldwide. The aim is to
establish haplotype tag (ht) SNPs which
identify certain conserved haplotype
blocks, and to use them for grouping the
population due to genotype. Most prob-
ably, htSNPs cannot represent all genetic
variation, but it seems to be possible for
allele frequencies greater than 56%. Be-
low that threshold, a direct approach is
needed [33].
3.5.2
Population Genetics
The pattern of DNA sequence variation
among humans is greater within popula-
tions than between populations. Popula-
tion genetics theory suggests that rare vari-
ants are more likely to be recently derived
as compared to common variants and are,
therefore, more likely to be population-spe-
cific. This is in concordance with the ob-
servation that the majority of human ge-
netic variation occurs among individuals
of local populations. Nevertheless, the con-
cept that there is one predominant or
wild-type allele and various rare or mutant
forms could not be proved. Instead, there
are multiple haplotypes, each of which is
observed in multiple populations. The
study of 313 genes in 82 unrelated indivi-
duals from four populations showed that
2% of the variation was globally distrib-
uted, while population-specific variants
were present for at least 5%, if single
SNPs were investigated. When haplotypes
were observed, almost 82% occurred glob-
ally, and 8% were specific for one popula-
tion, 4% for two populations, and 6% for
three populations [34]. So, there is a gener-
ally low variation abundance which is diffi-
cult to detect by conventional techniques.
One study showed that about 20% more
SNPs are needed to cover two populations
(Japanese and European) rather than one
population [33]. However it is also known
that, for example, Sub-Saharan Africans
show a much larger diversity than Cauca-
sians. The haplotype map must be filled to
higher density to be most useful at a very
fine resolution. This is thought to be a
helpful tool if it is not clear how muta-
tions have segregated and where, through
natural selection, they have already accu-
mulated.
Screening for heterozygosity in popula-
tions may be helpful for estimating the
frequencies of recessive alleles which do
not affect carriers, but it will cause exten-
sive problems in the homozygous state.
Population genetics and natural selection
research is also useful to identify gene
variants conferring variation in reproduc-
tive fitness. More importantly, SNPs affect
the germline of an organism and therefore
can spread in a population. SNPs with an
allele frequency above 7% can be used
very effectively as genetic markers. High-
throughput technologies for population-
based SNP mass screenings are therefore
urgently required.
To succeed, however, population studies
must have appropriate sample sizes, and
3.5 Strategies in SNP Discovery 81
include well-characterized controls, well-de-
fined and documented phenotypes, a model
to study interactions, and independent
study replications to diminish the rate of
false-positive results. In this respect, statis-
tics and bioinformatics tools are currently
being improved to support the development
of pharmacogenetic research.
3.5.3
Large-scale Projects
Discovery platforms are usually not acces-
sible to research laboratories, because they
are tightly embedded in SNP discovery
programs of major pharmaceutical compa-
nies, who fill their pipelines with candi-
dates for drug development. For reasons of
competitiveness and capacity, these plat-
forms are not offered commercially, and
are therefore not accessible to most re-
searchers. An additional reason is that
companies which establish innovative re-
search technologies and accumulate so far
unknown biological content with their ex-
clusive technologies are preferred by inves-
tors, compared to companies that have
either only technology or biological con-
tent. Examples of joint efforts are Glaxo-
SmithKline and Perlegene, who like to
map 1.7 million SNPs on the HLA-B57
(and to some extent the TNF-alpha locus)
to detect associations with the hypersensi-
tivity syndrome to the anti-HIV drug Aba-
cavir (multiplex strategy, detection limit at
10% allele frequency). DXS Ltd., with its
ARMS technology, works together with As-
traZeneca to detect statistically under-re-
presented mutations. Affymetrix and Perle-
gene have specialized microarrays that fo-
cus on transcription factors and their bind-
ing sites, and the results are marketed to
pharmaceutical companies. Illumina fo-
cuses on multiple sclerosis and diseases
such as malaria or Salmonella infections,
and also genotypes viral pathogens. Decode
Genetics is working on stroke, heart attack
and osteoporosis, and has identified the
gene for the 5-lipoxygenase activating pro-
tein (FLAP) that confers additional risk to
heart attack and stroke. Sequenoms mass
spectrometry platform is mining
for biological contents in genes asso-
ciated with schizophrenia, osteoarthritis,
type II diabetes and breast cancer. Some
28000 highly validated genome-wide SNPs
are currently applied to association studies.
In order to evaluate strong functional
candidates throughout the genome, as
many as 5000 tests are necessary. Ge-
nome-wide scans without evidence of any
candidates may need between 250000 to
500000 analyses to be performed, depend-
ing on study design and accepted false-
positive rates [35, 36].
Some companies claim to perform SNP
analysis at costs below 10 cents per single
allele call, though currently this seems
possible only if sample preparation and
amplification is excluded from cost calcula-
tions. Such cost might be feasible for
large-scale SNP discovery, where always
the same samples from volunteer and pa-
tient populations are used for different
SNP targets, and immobilized on microar-
rays. This is a method mainly to yield can-
didate SNPs that must be verified by tech-
niques with finer resolution in specialized
set-ups. For the daily scoring or screening
of individual patients in pharmacogenetic
testing, the whole workflow from DNA
isolation and amplification, experimental
set-up (hybridization or enzymatic) until
signal detection must be carried out. In
addition, the results must be refereed by
medical advisers who will also take quality
management and control into account.
Even sample preparation (DNA isolation)
is hardly performed at costs below 5 cents.
In a more realistic calculation, costs in this
scenario are around 50 cents per SNP
analysis. Since specific technical tools for
easy-to-measure SNP profiles are still un-
der development, costs for individualized
SNP analyses may fall further in the fu-
ture, but only time will tell if prices will
follow these developments.
A variety of large-scale SNP-based asso-
ciation studies, partly including haplotype
data, have contributed to narrowing down
and identifying the genes involved in Par-
kinsons disease [37], myocardial infarction
susceptibility [38], drug-induced cardiac ar-
rhythmia [39], and drug-induced morbidity
[40].
3.6
SNP Technologies
3.6.1
Overview of Currently Available Methods
Sophisticated molecular technologies are
needed to detect single base variations in
whole DNA sequences. There are essen-
tially three aspects which differentiate the
numerous SNP typing technologies that
are currently available commercially, and
are under constant development: 1) how
the sequence information in the genomic
DNA is translated into a detectable format;
2) whether signal amplification occurs be-
fore or after signal translation; and 3) the
method of signal detection. A fourth as-
pect is whether or not the analyte material
or the signal must be amplified at all.
Critical to the selection of an appropriate
genotyping technology are especially the
following considerations: specificity, repro-
ducibility, throughput, price per genotype
and, last but not least, the ease of assay
design and assay handling. Details of cur-
rent, most common SNP typing tech-
niques are summarized in Table 3.1.
3.6.2
Translation of Genomic Information
Since the pioneering studies of Sanger in
DNA sequencing and Mullis in DNA am-
plification (polymerase chain reaction,
PCR), most if not all DNA sequence
analysis consists of two basic steps. First, a
short synthetic primer molecule (2030
bases long) is hybridized to a denatured
DNA molecule. The primer binds during
this process to its complementary se-
quence in the DNA, and a small DNA du-
plex is formed at that point. Second, free
deoxynucleotides and DNA polymerase en-
zyme are added to the solution. Both use
the primer molecule as a starting point to
complement the DNA chain along the
template strand under investigation (prim-
er extension or primer elongation). When
chain-terminator nucleotides (dideoxynu-
cleotides) only represent a small percent-
age of nucleotides, the growing strands are
terminated randomly and thereby frag-
ments of all possible sizes are produced
(sequencing). If the terminators represent
100% of all nucleotides, the reaction will
simply be stopped after only one base has
been added to the primer (single-base
primer extension).
Techniques based solely on hybridization
circumvent the use of DNA polymerase
enzyme. Instead, the DNA is labeled
(radioactivity, fluorescence or lumines-
cence) and a distinct signal is detectable
when successful hybridization of comple-
mentary sequences has occurred. These
techniques were in the past mainly used
to detect longer sequence stretches in
DNA (Southern blotting), but were never
used to analyze single bases in a sequence.
Since miniaturization, immobilization,
automation, innovative dyes, and laser
technology have revolutionized technical
opportunities, it is in some cases possible
to detect even single sequence changes by
hybridization (sequencing by hybridiza-
tion). These methods suffer somewhat
from being highly sensitive with regard to
constant and standardized hybridization
conditions, and are not sufficiently robust
to work in all laboratories.
However, sequencing by hybridization is
not comparable to hybridization in gene
expression analyses, where longer DNA
stretches are observed. In hybridization se-
quencing, the hybridization conditions
must be able to discriminate between per-
fect match and a single mismatch, which
is statistically and chemically quite com-
plex. Nevertheless, the approach of com-
paring hybridization to perfect match and
mismatch probes is being used by Affyme-
trix (US) in microarray applications for
SNP detection and resequencing of geno-
mic regions. A newer approach (DASH)
monitors dehybridization, or melting, of
DNA probes from a target sequence. This
appears to be much more sequence-specif-
ic and robust in terms of reaction condi-
tions. Label-free approaches measure volt-
age differences in immobilized oligonu-
cleotides under hybridization sequencing
experiments. From statistical and technical
points of view, this is highly demanding,
and current research into these technolo-
gies is being conducted in Germany by
November and Directif.
3.6.3
Sample and Signal Amplification
No currently available technology appears
to be able for the detection of genotypes
directly in reasonable amounts (a few na-
nograms) of genomic DNA. PCR is the
method of choice to amplify target se-
quences from the sample DNA to make
them detectable among the vast amount of
Table 3.1 Common single nucleotide polymorphism technologies
available commercially.
(Trade) Name Instrumentation Experimentation Detection Reference
Invader Plate reader Cleavage Fluorescence 41, 42
RFLP Gel Cleavage Fluorescence 43
TaqMan Plate reader Cleavage Fluorescence 44
dHPLC HPLC Hybridization Absorbance 45
GeneChips Microarrays Hybridization Fluorescence 4648
Mass Array MALDI TOF MS Primer extension Mass spectrometry 49
TGGE Gel Hybridization Staining 50
Mol. Beacon Plate reader Hybrid. Quench. Fluorescence 51
SSCP Gel, Capillary Internal structure Fluorescence 52
Dash Fluor-Imager Dehybridization Fluorescence 53
Coded Spheres Flow Cytometer Primer extension Fluorescence 54, 55
OLA Gel, Plate reader Ligation Fluorescence 5658
APEX Microarray Primer extension Fluorescence 59
Sequencing Gel, Capillary Primer extension Fluorescence 60, 61
FP-TDI Plate reader Primer extension Fluorescence 62
SNaPshot Gel, Capillary Primer extension Fluorescence 44
SNP-IT Plate reader Primer extension Fluorescence 63
Pyrosequenc Pyrosequencer Primer extension Luminescence 64
ARMS Plate reader PCR, Quenching Fluorescence 65
genomic sequences that are not necessary
for the analyses (sample amplification).
PCR is a further developed primer exten-
sion reaction (earlier known as primer
elongation). The only differences com-
pared to primer extension are that two
primers are utilized instead of one, and
that the reaction is kept under a thermal
cycling regime instead of a constant tem-
perature. In up to 40 thermal cycles, the
DNA becomes repeatedly denatured and
renatured. In the renaturing step, the
primers bind to the genomic DNA and
fragments from the prior cycles. By cycling
the temperature, thermostable DNA poly-
merase repeatedly synthesizes and thereby
amplifies the sequence between the two
primers to high copy numbers. In theory,
millions of PCR fragments can be gener-
ated from a single target strand of geno-
mic DNA. The higher concentration of the
target sequence as compared to the non-
amplified remaining genomic sequences
makes detection possible. In the case of
SNP analyses, the polymorphic base is lo-
cated and amplified between the two prim-
ers, and becomes detectable in the huge
amount of copied PCR fragments. Sample
or target amplification is currently abso-
lutely necessary for almost all types of se-
quence analyses. A rare exception is that
of restriction fragment length polymorph-
ism (RFLP) analyses, in which high molec-
ular-weight genomic DNA is fragmented
by restriction enzymes, size-selected by gel
electrophoresis, and then hybridized to
specifically labeled probes.
Signal amplification is usually applied
after the genomic information has been
translated into a detectable format, and
mainly deals with labeling strategies, espe-
cially in amplifying the signals emitted
from the label. Labeling can be applied be-
fore or after target amplification. The main
signal amplification strategies include the
presence of primer-adapted binding se-
quences or catcher molecules. These func-
tion as targets for labeled molecules. Lin-
ear amplification or the early phase of ex-
ponential amplification is needed if abso-
lute amounts of the detected molecules
are of interest (quantitative analyses as, for
example, in gene expression analyses). Ex-
ponential signal amplification cascades are
used when only the presence or absence
of the signal is measured, while the con-
centration or intensity of the signal does
not contribute to the quality of results, as
in SNP analysis (qualitative analyses). Cas-
cades are usually initiated by antibodies,
where one antibody binds to its target mol-
ecule that was introduced in a prior assay
step. Each antibody typically carries two or
more binding sites for further antibodies,
thereby creating a branched structure
where more and more labeling is accumu-
lated. The biotinstreptavidin system is
frequently used to add label to a site,
where information was translated from
genomic sequences. Furthermore, dyes
may collect intensifying agents that are
utilized to enhance signal intensity.
3.6.4
Signal Detection
Once enzymatic and/or hybridization reac-
tions are performed for signal translation
and amplification, as experimental biologi-
cal prerequisites, the reaction outcomes
must be measured and the resulting sig-
nals detected. Sequencing reactions are
typically analyzed by gel electrophoresis,
which was further developed to capillary
gel electrophoresis and microfluidics sys-
tems. Real-time PCR measures amplifica-
tion products directly in the PCR tube by
using the TaqMan chemistry (Light Typer;
Roche or other real-time thermocyclers).
Alternatively, DNA-specific dyes are mea-
sured which change their fluorescence in-
tensity upon very selectively intercalating
between the two strands of double-
stranded (ds) DNA. This can be used to
monitor the synthesis of dsDNA in real-
time PCR, or the melting of DNA in dehy-
bridization. By elevating temperatures, the
double strand begins to melt, and the in-
tercalating dyes leave the DNA, and conse-
quently small differences in melting tem-
peratures can be measured. These are due
to mismatches in the double strand. If a
SNP is present (mismatch), the DNA helix
will melt at a lower temperature as com-
pared to a perfect match (comparable with
heteroduplex analyses).
One specialty in dehybridization detec-
tion is that of dHPLC (denaturing high-
performance liquid chromatography). Hy-
bridization products differ in melting tem-
perature and mobility on a specific support
if a mismatch is introduced by a SNP. In
the case of a SNP, a heteroduplex is
formed in the analyzed sequence, whereas
in the case of identical sequences (no SNP
present), a perfect homoduplex is formed.
These events can be discriminated by
dHPLC at a low to medium throughput
for SNP discovery or SNP detection.
Hybridization experiments are typically
performed on membranes or microarrays.
For signal detection on membranes fluor-
or phosphor-imagers are used, and the mi-
croarrays are analyzed by laser scanning
and fluorescence. For hybridization-based
SNP detection, the Affymetrix/Perlegene
system should be mentioned, where per-
mutated oligonucleotides are synthesized
photolithographically in very high density
directly on the chip (up to 1 million
probes per cm
2
). Patient DNA samples are
hybridized to the chip, and mismatch or
perfect match situations at the probes can
be detected by laser scanning. The data
must be processed by statistical and bioin-
formatics correction methods. Conven-
tional microarrays are produced by spot-
ting robots (up to 20000 probe spots per
cm
2
) that deposit DNA probes on the chip
surface (usually activated glass slides).
These probes are immobilized on the chip
and hybridized with patient DNA samples.
Only one patient can be analyzed per chip,
because by mixing or pooling patient sam-
ples, outcomes cannot be differentially de-
tected on one photolithographic or spotted
DNA chip.
Primer extension reactions can be ana-
lyzed using matrix-assisted laser desorp-
tion ionization time of flight mass spectro-
metry (MALDI-TOF MS). Specially pre-
pared single-base primer extension prod-
ucts are deposited on MALDI targets,
evaporated by a laser beam, and directed
into the time of flight mass spectrometer.
This sensitive, label-free method can dis-
criminate which base (A, C, G, or T) was
added to the primer at the very SNP posi-
tion, and affords excellent error tracking.
Allele-specific primer extension (four oli-
gonucleotides and one dye are needed to
analyze one SNP) or single-base primer ex-
tension (only one oligonucleotide probe,
but four dyes are needed per SNP) can be
performed on microarrays. Fluorescence-
labeled deoxynucleotides or dideoxynucleo-
tides are incorporated into the extension
primer (probe) in a template (sample) -de-
pendent manner, and are read out by laser
scanning the DNA chip. A specialty in
primer extension is to use microbeads in-
stead of microarrays as a solid phase. Mi-
crobead-bound SNP detection products are
analyzed by flow-sorting in cell or chromo-
some counters. Interesting options are
self-assembling bead arrays that are ran-
domly ordered on bundles of optic fibers,
as introduced by Illumina.
Oligonucleotide ligation assays (OLA)
are a combination of DNA ligation and
PCR. In a first step, two adjoining primers
are hybridized to the target DNA so that
the 3' terminal base of the upstream prim-
er is located directly over the SNP posi-
tion. There are two upstream primers,
each completely matching one of the two
SNP alleles. The two primers usually con-
tain a feature making them distinguish-
able, usually additional nucleotides as a
mass tag or zip code. In either single tube
reactions or separate reactions, only the
pairs of upstream and downstream prim-
ers are ligated which are hybridized to per-
fectly matching templates. The hybridiza-
tion products are amplified by PCR and
detected, for instance, by hybridization of
labeled oligonucleotides to the zip code tag
and capillary electrophoresis. The main ad-
vantage of this technique is a potential for
high multiplexing.
Label-free hybridization and detection by
measuring voltage changes in immobilized
oligonucleotides upon hybridization, as
well as label-free pyrosequencing (Pyro-
sequencing, Sweden) as example for an
enzymatic approach, request high end ma-
chinery.
3.6.5
Pooling and Multiplexing Strategies
The effectiveness of all methods is highly
influenced by pooling or multiplexing
strategies, as well as by miniaturization.
Such methods aim at savings in consum-
ables that, once the machinery is estab-
lished, represent the main operational
costs in subsequent SNP analyses. It is im-
portant which steps are multiplexed or
miniaturized to succeed in signal detec-
tion. All tube-based steps consume much
reaction components (label, enzyme, and
DNA). Volumes can be reduced by minia-
turization, from the microliter scale to
nano- or even picoliter amounts, thereby
saving drastically on resources and con-
sumables (factors of 1000 to 10000). For
example, multiplexing at the PCR level (at
present, samples must be amplified before
analysis) can only reduce costs of this step
by a factor of 10, because usually not more
than 810 primer pairs function together
in the same reaction solution. In very ex-
ceptional cases, up to 16 primer pairs to-
gether in one tube yield acceptable results
(amplification of microsatellite markers in
forensic and paternity applications). Pool-
ing at the hybridization step proves to be
almost impossible. Even if different pa-
tient samples are labeled with different
dyes, the amounts of each sample used in
the hybridization step must be adjusted
very precisely to eliminate competition ef-
fects during hybridization to the probes.
Hybridization times are prolonged to an
unacceptable extent (up to 72 h), and het-
erozygous samples complicate detection
even more.
Multiplexing the primer extension step
on DNA microarrays leads to enormous
savings. In almost all approaches, primer
extension is tube-based and requires at
least 10 lL of reaction volume per ana-
lyzed SNP, and is usually not multiplexed
at this stage. By performing multi-parallel
primer extensions on single DNA chips,
each extension requires only nanoliter
quantities of reaction mix (which contains
the most expensive additives such as the
label and the enzyme), and hence major
savings can be achieved.
3.6.6
Considerations on Gold Standards
With regard to quality, resequencing is still
the gold standard in SNP analyses. This
is followed by pyrosequencing, which also
bears high quality, due to the fact that not
only is the base in question investigated,
but that bases surrounding the SNP are
also delivered with the results. This pro-
vides high confidence about the composi-
tion of the locus and the localization and
identity of the very base to be analyzed.
Both techniques suffer from being neither
practical nor economic because of tube-
based approaches, the high demands for
expensive consumables, and the time-con-
suming procedure. Multi-parallel technolo-
gies are often also not affordable for small
or medium-sized pharmaceutical enter-
prises that are unlikely to be involved in
pharmacogenomics research, but may con-
duct well-defined pharmacogenetic pro-
jects, with strong evidence on functional
candidates that do not require extensive
testing. These findings may eventually be
reported to seek genotyping companies
with platforms open for any biological con-
tent as a partner.
3.7
Polydimensional SNP-Chips:
The Array-On Technology
3.7.1
A New Approach
Array-On offers a technology which is
highly competitive with currently used
SNP techniques. In particular, steps that
are time-consuming and suffer from low
reproducibility (as does differential hybri-
dization) have been either modified or cir-
cumvented. Multiplexing and pooling strat-
egies were shifted to points in the process
where they do not destabilize outcomes or
results. With these changes, the company
follows two objectives: 1) to establish an
open platform for rapid and reproducible
SNP detection; and 2) to develop ready-to-
use products for point-of-care diagnostics.
In order to attain these objectives, the
first aim was to assemble a streamlined
workflow that is optimized for practical
and rapid performance at each point of
the analytical process. Short, automated
steps contribute to reproducibility. By re-
viewing conventional assay designs, it be-
came clear that hybridizing pooled sam-
ples to multiple probes with the discrimi-
natory power of a single base pair causes
the main problems in currently used SNP
techniques. Time-demanding and error-
prone hybridization can cause unwanted
cross-hybridization events. Cross-hybridiza-
tion may also occur if the DNA of only
one patient is analyzed at multiple loci
with potentially high homology, and this
certainly occurs if many patient samples
are analyzed simultaneously for the same
SNP (high failure and false positive rates,
low reproducibility). Because of these
drawbacks, Array-On developed a new
approach for more convenient multi-paral-
lel SNP detection, as described in the fol-
lowing section.
3.7.2
Polydimensional SNP Platform
Array-On solved the problem of cross-hy-
bridization on arrays for the detection of
primer extension products by eliminating
the hybridization step from the microarray
platform. Conventional oligonucleotide mi-
croarrays, for hybridization or primer ex-
tension purposes, consist of many thou-
sands of microscopic oligonucleotide spots.
When bioinformatics was applied to de-
sign these specific sequences of the prob-
ing molecules, the required oligonucleo-
tides are purchased from DNA-synthesiz-
ing companies and spotted in a microarray
design onto the chips. Patient samples to
be probed on these microarrays are usually
present as genomic DNA or PCR-ampli-
fied genes that should be analyzed for the
occurrence of SNPs. In conventional ap-
proaches, all PCR fragments of one patient
are pooled, labeled and concentrated to be
hybridized to the chip. They must be incu-
bated for a long time at a critical tempera-
ture in order to form specific hydrogen
bonds with their corresponding immobi-
lized extension oligonucleotide, while hy-
bridization or dehybridization is monitored
or detected afterwards. Array-On invented
a new way of circumventing SNP precise
hybridization or dehybridization. This is
achieved through a direct strategy to guide
the extension oligonucleotides to their tar-
get sequences to perform signal transla-
tion and detection in a multiplexed set-up
(Fig. 3.1).
In the new approach, patient samples
are not pooled and concentrated, but in
contrast to all other microarray designs
are kept separate. This is a relevant advan-
tage that eliminates both competition at
the hybridizing probes and also false-posi-
tive results. After sample preparation and
amplification, pre-designed extension oli-
gonucleotides are (in automated fashion)
mixed individually with their correspond-
ing target amplicons in single microplate
wells. Fragments and oligonucleotides an-
neal inside the microplate wells within
30 min, and are transferred from there as
an microarray of stabilized hybrids onto
the chip surface. The hybrids are com-
posed of the synthetic probing molecule
(extension oligonucleotide) and the ampli-
fied DNA sample target strand. In a next
step, the extension reaction mix (contain-
ing DNA polymerase and labeled dideoxy-
nucleotides) is pipetted onto this array on
the chip and all probe/sample hybrids are
single-base extended in one multiplexed
step that lasts only about 45 min. The
primer extension reaction can process up
Fig. 3.1 Polydimensional SNP Microarray. Forty-
eight individuals at 48 SNPs with enlarged sub-
array of 48 individuals at two SNPs (A and B).
Every individual sample was spotted in three repli-
cates next to each other. The green color repre-
sents homozygous GG, red represents homo-
zygous AA, and yellow represents heterozygous
GA. At SNP B, the GG allele is rather rare (below
10%), with the appearance of one heterozygote
(yellow). At SNP A, alleles are distributed more
evenly (less then one-third AA, more then two-
thirds GG, no heterozygote in 48 individuals).
to 50000 SNP analyses at a time (patent
no. DE 102 451 45 / PCT/EP 03-10773).
The major advantage is that there is no
competition between different probes and
samples. Cross-hybridization is excluded
in terms of assay design, and multiple pa-
tients can be screened on a single chip at
the same multiple SNP loci (polydimen-
sional analysis). Previously, this was not
possible due to cross-reactions. The time
savings are tremendous: the process of sig-
nal translation and signal detection is fin-
ished within 1.5 h, and the chip surface
can be filled with a huge number of hybrids.
Even comparative genomics approaches
with spotted hybrids (e.g., from different
species) could be analyzed and compared
on the same polydimensional SNP chip.
The technology also affords the option to
perform replicate extension reactions on
the same chip, for increased reliability.
The processes prior to signal translation
and detection namely sample preparation
and signal amplification are comparable
with commonly used techniques. A PCR-
grade DNA must be extracted from the sam-
ple (blood or tissue material). SNP loci are
amplified from the sample DNAs with spe-
cific PCR primer pairs, and a crude purifica-
tion of PCR products is necessary.
Array-On has built a genotyping plat-
form on this technology which takes ad-
vantage of time and material savings, and
is used for service analyses. Customers
contribute sequence information and DNA
samples, while Array-On designs appropri-
ate extension oligonucleotides that are
mixed and annealed in individual sample
wells of microplates. After utilizing the
new hybrid microarray spotting technique,
extension primers are collectively extended
with fluorescence-labeled chain terminator
nucleotides on the chip in a template-de-
pendent manner. Results are detected by
laser scanning, and are highly secure and
guaranteed to meet 99.94% accuracy. SNP
information is returned to the customer in
due time.
3.7.3
Workflow Assembly
DNA isolation is automated on a Tecan
platform integrating Qiagen kits. The PCR
step is not multiplexed, but has two spe-
cial features: 1) an asymmetric set-up for
the favored production of the target tem-
plate strand for the subsequent single-base
primer extension; and 2) shortening PCR
time by utilizing a novel high-speed PCR
system from JenAnalytics (Germany). The
main innovations in the PCR system are
ultra-thin well walls and sophisticated Pel-
tier technology. One PCR cycle runs for
less than 1 min, and the whole process is
finished in about 25 min. This makes the
PCR very rapid and reproducible, because
PCR byproducts from mispriming or
primer dimers are avoided by very strin-
gent high-speed cycling.
Nevertheless, the quality of PCR prod-
ucts is checked to succeed in the remain-
ing steps. At this point, pooling is intro-
duced into the process. Small PCR ali-
quots are pooled with a 96-channel pipet-
ting robot (Rapid-Plate Liquid Handling
Instrument; Zymark), which fills a 384-
well plate in less than 2 min. In a next
step, PCR aliquots are checked with a Cali-
per microfluidics system, that works
through a 384-well plate within 2.5 h. The
system (LabChip 90) has a detection range
from 100 to 5000 base pairs (bp) at a 4-bp
resolution, and delivers fragment size and
concentration. When only 10 PCR frag-
ments, which differ by at least 4 bp in size
are pooled, 3840 fragments can be checked
within 2.5 h. Passed individual PCR frag-
ments are then purified in a Millipore 384
vacuum filtration station. The target
strands from asymmetric PCR are directly
resuspended in a solution that contains
the extension oligonucleotides and stabiliz-
ing agents for the subsequent hybrid spot-
ting onto microarrays. To anneal extension
oligonucleotides and target sequences, mi-
croplates are incubated at 508C for 5 min
only, and left at 208C for another 5 min to
cool down. The annealed and stabilized in-
dividual hybrids, a complex of extension
oligonucleotide and template strand, is
spotted and immobilized on activated glass
slides. Since there is no hybridization step
on the array, competition for the probing
molecules is excluded and the required
time reduced from 72 h to 30 min only.
After hybrid spotting, primer extension is
completed within 45 min, and is per-
formed in the Advalytics Slide-Booster-Sta-
tion which utilizes nanosound waves
(150 kHz) to ensure the steady and homo-
geneous mixing of all reaction compo-
nents. DNA polymerase access to the
spotted hybrids is enhanced by a three-di-
mensional chip surface, which is achieved
by a thin polymer layer on the activated
glass slide.
Results are read out with the LS4 scan-
ner (GenomicSolutions Ltd.). The scanner
provides four lasers (488, 532, 594 and
633 nm) for simultaneous detection of all
four labeled dideoxynucleotides. Data pro-
cessing to obtain easy to handle output for-
mats is performed with the GeneTAC LS4
software.
The streamlined setting of the Array-On
platform is very well suited for rapid and
high-throughput SNP analyses. Blackouts
are only observed in PCR-based sample
amplification, due to genomic primer site
mutations that prevent correct binding of
the PCR primers to the sample DNA. Only
in these cases can no PCR product be ob-
served and used for primer extension.
Primer extension will perform well, as
long as expected PCR fragments are ob-
tained. This is mainly due to the fact that,
whenever possible, one PCR primer is po-
sitioned directly adjacent to the SNP base,
which is also the target binding site of the
extension oligonucleotide. Perfect accuracy
of this binding site is granted through the
PCR primer sequence at this position, and
not necessarily through the genomic se-
quence. This ensures that extension oligo-
nucleotides will always match 100% with
the template strand. The annealing step
prior to spotting and primer extension is
therefore very effective, because two 100%
complementary DNA molecules associate
without disturbance by other molecules.
No mismatch situations can influence
DNA chip-based primer extension. In
cases of PCR amplification errors, the
primer sites may be shifted in the se-
quence to obtain a fragment in a second
trial. If the primer pair generally works
and failures are only observed in very few
samples, a null allele will be reported.
The methodology is highly compatible
with existing formats, and if the laboratory
already possesses the spotting robot and
laser scanner, implementation is possible
at very low cost. It should be noted that
the spotting and scanner facilities, as well
as other instrumentation, are not limited
only to SNP analyses but may also be used
for other purposes such as gene expres-
sion analyses.
3.7.4
Advantages and Applications
The common aim of all currently devel-
oped SNP technologies is to bring eco-
nomic and applied solutions to the mar-
ket. The contribution of Array-On offers a
straightforward technology with tremen-
dous time savings and cost reductions. All
steps of the analytical process are opti-
mized to meet these demands. Sample
preparation and PCR amplification were
miniaturized and developed from standard
protocols. After PCR, which is common to
most methods, the costs of operation come
from labeled dideoxynucleotides, DNA
polymerase, and only one extension oligo-
nucleotide. Due to primer extension min-
iaturization on the microarray platform,
extension nucleotides from one synthesis
batch are sufficient for thousands of ana-
lyses. Only nanoliter quantities of reaction
mixture are needed per SNP cell, and so
dideoxynucleotides and enzymes are used
very economically. The consumption of
plastic-ware is reduced by pooling at the
PCR quality control step, and by microar-
ray usage in the final detection step. Due
to its polydimensional design, the entire
surface of the DNA chip can be used, and
sample density is only limited by spotting
techniques, which show a steady increase
in density. The more samples are spotted
onto one chip, the lower the costs per gen-
otype. The technology is compatible with
all common formats, and has a cumulative
character in the sense that newly discov-
ered SNPs can easily be implemented in
ongoing projects or offered services.
Application fields are very diverse, and
include not only pharmacogenetics and
pharmacogenomics research but also ge-
netic mapping projects, population genet-
ics, mass-screenings, and molecular plant
and animal breeding or screenings for bio-
diversity. Disposable, easy-to-use SNP diag-
nostics can be developed from polydimen-
sional assay design.
3.7.5
Ready-to-use Products
As a consequence of the polydimensional
SNP technology, considerable opportu-
nities for ready-to-use disposable products
were developed. The idea of keeping pre-
pared samples separated initiated the de-
sign of so-called area SNP chips for par-
allel individualized primer extension. The
product is a conventionally sized plastic
carrier (glass slide format) that contains
up to 96 separate extension areas or
fields which are pre-coated with extension
oligonucleotides for defined SNP loci.
These chips will be offered in a kit-like de-
sign, together with pre-mixed asymmetric
PCR primers and extension mixtures opti-
mized for the loci in question. The target
sequences must be amplified from the
patients sample DNAs. Deoxynucleotides
that would disturb the subsequent single-
base primer extension, where only dide-
oxynucleotides should be present, are
eliminated by filtration or shrimp alkaline
phosphatase digests. In this way, purified
target strands are mixed with the exten-
sion reaction solution (containing DNA
polymerase and labeled dideoxynucleo-
tides) and are, either by hand or a pipet-
ting robot, applied to the corresponding
separate substrate area on the plastic car-
rier (patent no. DE 103 250 98 / PCT/EP
04-006002).
The extension substrate is an only
0.5 mm-thick composite structure of sili-
con and silica crystals forming regular cap-
illaries with tube diameters below 10 lm.
Extension oligonucleotides are immobi-
lized within these capillaries, and are spe-
cific for each separate area on the chip.
Once the extension solution containing
the target strand is applied to one special
SNP chip area, the sample is driven into
the crystal by capillary forces and exposed
to its complementary extension oligonu-
cleotides. Annealing takes place inside the
capillaries, and extension occurs due to
the presence of dideoxynucleotides and
DNA polymerase. Results can be read out
with laser scanners at a resolution of about
50 lm. Low-resolution requirements as
well as high effective signal intensity de-
pend mainly on a light-guiding property of
the silica tubes, which function like micro-
scopic mirrors and direct the emitted fluo-
rescence onto the detector inside the scan-
ner.
With these devices, pharmacogenetic
testing can be completed within 46 h,
and represents real point-of-care diagnos-
tics, as it can be performed prior to medi-
cation being administered. Array-On is
well prepared for in-license validated SNPs
to develop point-of-care testing for a variety
of pharmacogenetic needs. Relevant phar-
macogenetic SNPs will be detectable with-
in hours, directly at the location where the
information is needed.
The technology has been developed in
the framework of a genome project of the
German government at the IPK-Gatersle-
ben, the German gene bank and center for
biodiversity (http://www.ipk-gatersleben.
de). To develop prototypes of area SNP
chips, Array-On is working together with
three partners: Infineon Technologies (Ger-
many) which invented and further devel-
ops the crystalline primer extension sub-
strate by means of efficient coating with
extension oligonucleotides. For the biologi-
cal content that will be applied to area
SNP chips, Array-On is working together
with the previously mentioned project
on rheumatoid arthritis at the University
of Leipzig (Germany). A second partner
for discovering valuable biological content
is the European Nutrigenomics Organisa-
tion (http://www.nugo.org). Pre-sympto-
matic tests are planned in new born
screening for obesity and, most impor-
tantly, for galactose hypersensitivity with
an incidence of about 1520% in Euro-
peans.
3.8
Outlook
A huge amount of recent research find-
ings has indicated the enormous potential
for future pharmacogenetic testing. It is
expected that testing will reduce the over-
all costs in healthcare systems. Trial and
error prescription leads to more physician
visits and ADRs. Moreover, non-response
is reported for 2040% of people receiving
pharmacological agents, and even todays
most effective drugs do not work in about
20% of patients [66]. Therefore, persona-
lized medicines offer an opportunity to
make prescriptions more effective. The
concept of personalized drugs is unlikely
to mean that specific medicines will be de-
signed for each individual. Instead, medi-
cines for general application will be
adapted to certain groups of patients and,
after SNP analyses, be applied on a perso-
nalized basis. Group standards for the pre-
scription of medicines based on genetic
testing will be established. The real cost
(and life) savings will be in increased treat-
ment success and fewer ADRs.
Even low frequencies of ADRs can cause
regulatory authorities to withdraw medi-
cines from the market. Genetic testing is
one tool to help reduce ADRs, and will be
developed for promising substances. ADRs
are most likely the first area of benefit in
pharmacogenetic testing. Today, diagnos-
tics represents only 4 cents of each dollar
spent on healthcare. This proportion will
most probably change in favor for testing,
because the more often precise diagnostics
are applied, the more will failures and
costs be reduced.
In terms of drug development by phar-
maceutical industries, even higher regula-
tory standards in drug assessment and val-
idation are expected in the future. This
may cause delays in new drug launches,
3.8 Outlook 93
until molecular technologies have suffi-
cient throughput, precision, and pricing to
be integrated routinely into the drug devel-
opment process. Moreover, positive devel-
opments can be expected. The numbers of
participants needed in case-control studies
and clinical Phase II and III trials may be
reduced by 50% and by 10%, respectively.
In addition, the time required could be re-
duced by 20% if participants were selected
according to their genotype [67]. Geneti-
cally pre-screened volunteers in Phase I
would also be less endangered. In this
context, it should be noted that 45% of
Phase I compounds fail because of toxicity
concerns. More effective research is
needed: it is estimated that only 10% of
investigated compounds reach the market,
but the genotyping of trial participants
would improve these rates. When routine
testing is widely applied, cost savings can
be valuable, since ADRs occur in 510%
of all medical treatments, and the average
costs per case are US$ 20003000 [68, 69].
It can be expected that pharmacokinetic
and pharmacodynamic modeling, with the
help of molecular diagnostics, will make it
possible to administer the correct drug at
a safe dose.
The combination of a decrease in ADRs
and failed drug trials, time for drug ap-
proval, time on medication, and the num-
ber of medications taken, will promote a
decrease in healthcare costs as a whole.
There is no doubt that molecular disease
diagnostics and predictive testing will
change the face of healthcare in the near
future. Skeptics of high long-term invest-
ments in pharmacogenetic research have
several arguments in the realms of appli-
cability, ethics, and economics.
It is still unclear how widely applicable
the results of pharmacogenetics and phar-
macogenomics will be in the near future.
Awareness and acceptance are still quite
low. The time until pharmacogenetic find-
ings reach the markets in form of com-
mercial tests is still too long, as illustrated
by the example of the genetic predisposi-
tion for hemochromatosis, which was
known for more than 10 years to be
caused by two major and one minor SNP.
Currently, more than 50% of human genes
are of unknown function, and for most
genes the involvement in particular dis-
ease genotypes is also unknown. Research
on these topics is hampered by intellectual
property rights and patents.
Genetics training for all physicians is a
prerequisite to make sure the right test is
ordered, and that the results are properly
interpreted. If not properly educated, clini-
cians and physicians may be a limiting
factor for acceptance and growth in phar-
macogenetics and pharmacogenomics. The
healthcare system is poorly prepared for
pharmacogenetic testing and to handle
complicated genetic issues which should
influence the decisions which medication
and dose to choose (see also Part VIII,
Chapter 1). Many health professionals have
problems in making sense of probabilistic
information on likelihood, and doctors need
to know the science of the drugs and the
science of the tests to work efficiently.
Therefore, to deliver adequate information
will be an important challenge, to be met
not only by the medical community but also
by the industry involved in pharmacoge-
netics and pharmacogenomics.
A number of ethical concerns have been
raised in the past. In general, genetic test-
ing only makes sense if a clear benefit for
the patient can be achieved and outweighs
possible misuse of genetic information.
But even then, problems of disadvantages
and discrimination arise. Patient groups
may be identified by health insurance
companies as difficult or expensive to
treat, and could be excluded from cover-
age. Similarly, pharmaceutical research
may avoid patient subgroups in certain
diseases if there is little prospect of recov-
ering costs. However, since these patient
groups likely do not benefit much from
current medical development, this may
turn into an advantage. Once a genetic
subgroup of patients is identified, the
problem may be ameliorated by orphan
drug programs.
A very difficult question arises when a
decision is needed whether or not to pre-
scribe an expensive treatment to an indi-
vidual who is less likely to benefit from it
than others. In most cases, the genetic test
will give a different likelihood for treat-
ment success, but not absolute answers.
In particular cases, a patient with a non-re-
sponder genotype may still benefit from
the treatment, whereas a patient with the
responder genotype may not. Where
should the line be drawn?
Confidentiality of genotype data also de-
serves consideration. Violation of privacy
and data security, safe data banking and
data protection, anxiety and fears of mis-
use and discrimination because of geno-
type are issues that are presently dis-
cussed. Severely ill patients will be much
more interested in health benefits from
genotype data, and probably care less
about potential misuse. In many other
cases of less severe disease, these issues
may be more important. Currently, there
is a discussion about ethical considerations
in pharmacogenetics and pharmacoge-
nomics between scientists in academia,
biotechnology and pharmaceutical compa-
nies, health insurance companies, patient
organizations, and governmental and non-
governmental organizations. The resulting
societal consensus will be a moving target
for some time.
On the economics side, the overall aim
must be to develop therapeutics which are
better than existing medical treatments
and reduce morbidity and mortality. Subdi-
viding patient populations by genetic test-
ing means subdivided markets with smal-
ler volumes. This causes a problem in to-
days blockbuster strategies in pharmaceu-
tical development. Some argue that these
smaller markets will be more exclusive,
and todays competitors will enjoy better
co-existence in the future. In one scenario
it can be expected that, in the future,
blockbuster strategies may not work any-
more since treatment concepts are becom-
ing more biologically complex and there-
fore more specific to patient subgroups.
Also, regulatory bodies and competition
will call for safer, more specific and effec-
tive therapies. On the other hand, pharma-
cogenetic testing allows for the develop-
ment of profitable and safe pharmaceuti-
cals for application in broad parts of the
population, since companies have the tools
to effectively minimize the risk of ADRs
prior to drug release or even in early
phases of research on compounds and
drug development.
Yet another option to reduce drug devel-
opment costs may be to keep more drug
targets alive by including pre- and post-
marketing surveillance data paired with
genetic know-how. This may lead to
genetically limited indications, but since
more therapeutics will reach the market
more quickly, overall development costs
may decline. In extension, this could even-
tually lead to the re-submission of promis-
ing drug candidates which previously
failed in the assessment process. A promi-
nent example for the re-evaluation of a
drug, while not linked to genetic testing,
is thalidomide which is now approved for
treatment in certain severe cases of le-
prosy, despite causing birth defects.
For the development of genetically asso-
ciated drugs, tremendous numbers of
3.8 Outlook 95
DNA samples must be analyzed. DNA mi-
croarrays promise to fulfill this need and,
since polydimensional analyses are possi-
ble, this renders SNP-based genomic ap-
proaches on microarrays broadly applicable.
The Array-On platform is built on stream-
lined, optimized steps with consequent
elimination of error sources. Strong bioin-
formatics for primer and oligonucleotide
design strengthen and stabilize success
rates, even at SNP loci which are difficult
to observe, as known from conventional
techniques. The effectiveness of the plat-
form is further supported by miniaturiza-
tion and automation as well as multiplexing
and pooling strategies. SNPs associated
with increased risk to ADRs or susceptibili-
ty to disease will be examined in research
projects. The size of current studies can be
increased due to the effective miniaturized
technology. Better statistics through en-
larged sample sizes, lower failure rates
and holistic molecular portraits will bring
new strong functional SNP candidates into
focus of research and economics.
Disposable diagnostics products in the
form of manually or automatically handled
area SNP chips that are compatible with
all common formats are under develop-
ment. A point-of-care system is planned,
that is able to fulfill up to 96 parallelized
tests from DNA isolation to data interpre-
tation in less than 6 h. These should be
available for central laboratories or hospi-
tals to perform overnight testing prior to
the start of a therapy, both to increase the
benefits for patients and to foster the suc-
cess of modern biopharmaceuticals.
References
1 Vogel, F. Moderne Probleme der Humangen-
etik 1959, 12, 52125, in German.
2 Nelson, E. J. Theor. Biol. 1963, 5, 493496.
3 Hedgecoe, A. M., Martin, P. Soc. Stud. Sci.
2003, 33, 327364.
4 Gulcher, von J., Stefansson, K. Clin. Chem.
Lab. Med. 1998, 36, 523527.
5 Kurth, J. H. Rev. Gastroenterol. Disord. 2003,
3, 38.
6 Anderson, J. L., Carlquist, J. F., Horne, B. D.,
Muhlestein, J. B. J. Cardiovasc. Pharmacol.
Therapeut. 2003, 8, 7183.
7 Kirchheiner, J., Brsen, K., Dahl, M. L., Gram,
L. F., Kasper, S., Roots, I., Sjqvist, F., Spina,
E., Brockmller, J. Acta Psychiatr. Scand.
2001, 104, 173192.
8 Blanpain, C., LePoul, E., Parma, J., Knoop, C.,
Detheux, M., Parmentier, M., Vassard, G.,
Abramowicz, M. J. Cardiovas. Res. 2003, 60,
518528.
9 Ulrich, C. M., Robien, K., McLeod, H. L. Nat.
Rev. Cancer 2003, 3, 19.
10 Lenz, H. J. Semin. Oncol. 2003, 30, 4753.
11 Lynch, T. J., Bell, D. W., Sordella, R., Gurubha-
gavatula, S., Okimoto, R. A., Brannigan, B. W.,
Harris, P. L., Haserlat, S. M., Supko, J. G., Ha-
luska, F. G., Louis, D. N., Christiani, D. C.,
Settleman, J., Haber, D. A. N. Engl. J. Med.
2004, 21, 350.
12 Krynetski, E., Evans, W. E. Oncogene 2003, 22,
74037413.
13 Hall, D., Dhilla, A., Charalambous, A., Gogos,
J. A., Karayiorgou, M. Am. J. Hum. Genet.
2003, 73, 370436.
14 Borrego, S., Wright, F. A., Fernndez, R. M.,
Williams, N., Lpez-Alonso, M., Davuluri, R.,
Antiolo, G., Eng, C. Am. J. Hum. Genet.
2003, 72, 88100.
15 Martin, E. R., Lai, E. H., Gilbert, J. R., Rogala,
A. R., Afshari, A. J., Riley, J., Finch, K. L., Ste-
vens, J. F., Livak, K. J., Slotterbeck, B. D., Slifer,
S. H., Warren, L. L., Conneally, P. M., Schme-
chel, D. E., Purvis, I., Pericak-Vance, M. A.,
Roses, A. D., Vance, J. M. Am. J. Hum. Genet.
2000, 67, 383394.
16 Martin, E. R., Scott, W. K., Nance, M. A., Watts,
R. L., Hubble, J. P., Koller, W. C., Lyons, K.,
Pahwa, R., Stern, M. B., Colcher, A., Hiner,
B. C., Jankovic, J., Ondo, W. G., Allen, F. H.,
Goetz, C. G., Small, G. W., Masterman, D.,
Mastaglia, F., Laing, N. G., Stajich, J. M., Rib-
ble, R. C., Booze, M. W., Rogala, A., Hauser,
M. A., Zhang, F., Gibson, R. A., Middleton,
L. T., Roses, A. R., Haines, J. L., Scott, B. L.,
Pericak-Vance, M. A., Vance, J. M. J. Am. Med.
Assoc. 2001, 286, 22452250.
17 Lai, E., Riley, J., Purvis, I., Roses, A. Geno-
mics 1998, 54, 3138.
18 Goldstein, D. B., Tate, S. K., Sisodiya, S. M.
Nat. Rev. Genet. 2003, 4, 937947.
19 Kahle, K. T., Wilson, F. H., Leng, Q., Lalioti,
M. D., OConnell, A. D., Dong, K., Rapson,
A. K., McGregor, G. G., Giebisch, G., Herbert,
S. C., Lifton, R. P. Nat. Genet. 2003, 35, 372
376.
20 Brichta, L., Hofmann, Y., Hahnen, E., Sieb-
zehnrubl, F. A., Raschke, H., Blumcke, I., Eyu-
poglu, I. Y., Wirth, B. Hum. Mol. Genet. 2003,
12, 24812489.
21 MacGregor, A. J., Snieder, H., Rigby, A. S.,
Koskenvuo, M., Kaprio, J., Aho, K. Arthritis
Rheum. 2000, 43, 3037.
22 Jawaheer, D., Seldin, M. F., Amos, C. I., Chen,
W. V., Shigeta, R., Monteiro, J. Am. J. Hum.
Genet. 2001, 68, 927936.
23 Shiozawa, S., Hayashi, S., Tsukamoto, Y.,
Goko, H., Kawasaki, H., Wada, T. Int. Immu-
nol. 1998, 10, 18911895.
24 Cornelis, F., Faure, S., Martinez, M., Prud-
homme, J. F., Fritz, P., Dib, C. Proc. Natl.
Acad. Sci. USA 1998, 95, 1074610750.
25 Jawaheer, D., Seldin, M. F., Amos, C. I., Chen,
W. V., Shigeta, R., Etzel, C. Arthritis Rheum.
2003, 48, 906916.
26 MacKay, K., Eyre, S., Myerscough, A., Milicic,
A., Barton, A., Laval, S. Arthritis Rheum.
2002, 46, 632639.
27 Gabriel, S. E. Rheum. Dis. Clin. North Am.
2001, 27, 269281.
28 Haukim, N., Bidwell, J. L., Smith, A. J., Keen,
L. J., Gallagher, G., Kimberly, R., Huizinga, T.,
McDermott, M. F., Oksenberg, J., McNicholl,
J., Pociot, F., Hardt, C., DAlfonso, S. Genes
Immun. 2002, 3, 313330.
29 Reveille, J. D. Curr. Opin. Rheumatol. 1998,
10, 187200.
30 Gorman, J. D., Criswell, L. A. Rheum. Dis.
Clin. North Am. 2002, 28, 5978.
31 Cvetkovic, J. T., Wallberg-Jonsson, S., Steg-
mayr, B., Rantapaa-Dahlqvist, S., Lefvert, A. K.
J. Rheumatol. 2002, 29, 212219.
32 Kirk, B. W., Feinsod, M., Favis, R., Kliman,
R. M., Barany, F. Nucleic Acids Res. 2002, 31,
32953311.
33 Sevilla, C., Bourret, P., Nogues, C., Moatti,
J. P., Sobol, H., Julian-Reynier, C. Med. Sci.
(Paris) 2004, 20, 788792, in French.
34 Stephens, J. C., Schneider, J. A., Tanguay,
D. A., Choi, J., Acharya, T., Stanley, S. E.,
Jiang, R., Messer, C. J., Chew, A., Han, J. H.,
Duan, J., Carr, J. L., Lee, M. S., Koshy, B.,
Kumar, A. M., Zhang, G., Newell, W. R.,
Windemuth, A., Xu, C., Kalbfleisch, T. S.,
Shaner, S. L., Arnold, K., Schulz, V.,
Drysdale, C. M., Nandabalan, K., Judson, R. S.,
Ruano, G., Vovis, G. F. Science 2001, 293,
489493.
35 Buckland, P. R. Alcohol 2001, 36, 99103.
36 Dahlman, I., Eaves, I. A., Kosoy, R., Morrison,
V. A., Heward, J., Gough, S. C. L., Allahabadia,
A., Franklyn, J. A., Tuomilehto, J., Tuomilehto-
Wolf, E., Cucca, F., Guja, C., Ionescu-Tirgo-
viste, C., Stevens, H., Carr, P., Nutland, S.,
McKinney, P., Shield, J. P., Wang, W., Cordell,
H. J., Walker, N., Todd, J. A., Concannon, P.
Nat. Genet. 2002, 30, 149150.
37 Scott, W. K., Nance, M. A., Watts, R. L., Hub-
ble, J. P., Koller, W. C., Lyons, K., Pahwa, R.,
Stern, M. B., Colcher, A., Hiner, B. C., Janko-
vic, J., Ondo, W. G., Allen, F. H., Goetz, C. G.,
Small, G. W., Masterman, D., Mastaglia, F.,
Laing, N. G., Stajich, J. M., Slotterbeck, B.,
Booze, M. W., Ribble, R. C., Rampersaud, E.,
West, S. G., Gibson, R. A., Middleton, L. T.,
Roses, A. R., Haines, J. L., Scott, B. L., Vance,
J. M., Pericak-Vance, M. A. J. Am. Med. Assoc.
2001, 286, 22392244.
38 Ozaki, K., Ohnishi, Y., Iida, A., Sekine, A., Ya-
mada, R., Tsunoda, T., Sato, H., Sato, H.,
Hori, M., Nakamura, Y., Tanaka, T. Nat. Gen-
et. 2002, 32, 650654.
39 Splawski, I., Timothy, K. W., Tateyama, M.,
Clancy, C. E., Malhotra, A., Beggs, A. H., Cap-
pucino, F. P., Sagnella, G. A., Kass, R. S., Keat-
ing, M. T. Science 2002, 297, 13331336.
40 Kammerer, S., Burns-Hamuro, L. L., Ma, Y.,
Hamon, S. C., Cnaves, J. M., Shi, M. M., Nel-
son, M. R., Sing, C. F., Cantor, C. R., Taylor,
S. S., Braun, A. Proc. Natl. Acad. Sci. USA
2003, 100, 40664071.
41 Griffin, T. J., Hall, J. G., Prudent, J. R., Smith,
L. M. Proc. Natl. Acad. Sci. USA 1999, 96,
6301.
42 Mein, C. A., Barrat, B. J., Dunn, M. G., Sieg-
mund, T., Smith, A. N., Esposito, L., Nutland,
S., Stevens, H. E., Wilson, A. J., Pillips, M. S.,
Jarvis, L. S., deArruda, M., Todd, J. A. Genome
Res. 2000, 10, 330343.
43 Parsons, B. L., Heflich, R. H. Mutat. Res. 1997,
387, 97121.
44 Livak, K., Marmaro, J. A. Nat. Genet. 1995, 9,
341342.
References 97
45 Underhill, P. A., Jin, L., Zemans, R., Oefner,
P. J., Cavalli-Sforza, L. L. Proc. Natl. Acad. Sci.
USA 1996, 93, 196200.
46 Edman, C. F., Raymond, D. E., Wu, D. J., Tu,
E., Sosnowski, R. G., Butler, W. F., Nerenberg,
M., Heller, M. J. Nucleic Acids Res. 1997, 25,
49074914.
47 Sosnowski, R. G., Butler, W. F., OConnell,
J. P., Heller, M. J. Proc. Natl. Acad. Sci. USA
1997, 94, 11191123.
48 Wang, D. G., Fan, J.-B., Siao, C.-J., Berno, A.,
Young, P., Sapolsky, R., Ghandour, G., Perkins,
N., Winchester, E., Spencer, J., Kruglyak, L.,
Stein, L., Hsie, L., Topaloglou, T., Hubbell, E.,
Robinson, E., Mittmann, M., Morris, M. S.,
Shen, N., Kilburn, D., Rioux, J., Nusbaum, C.,
Rozen, S., Hudson, T. J., Lipshutz, R., Chee, M.,
Landers, E. M. Science 1998, 280, 10771082.
49 Shchepinov, M. S., Chalk, R., Southern, E. M.
Nucleic Acids Symp. Ser. 1999, 42, 107108.
50 Day, I. N., Spanakis, E., Palamand, D., Wea-
vind, G. P., ODell, S. D. Trends Biotechnol.
1998, 16, 287290.
51 Tyagi, S., Bratu, D. P., Kramer, F. R. Nat. Bio-
technol. 1998, 16, 4953.
52 Gonen, D., Veenstra-VanderWeele, J., Yang,
Z., Leventhal, B., Cook, E. H. Jr. Mol. Psychia-
try 1999, 4, 339343.
53 Jobs, M., Howell, W. M., Stromqvist, L., Mayr,
T., Brookes, A. J. Genome Res. 2003, 13, 916
924.
54 Iannone, M. A., Taylor, J. D., Chen, J., Li,
M.-S., Rivers, P., Slentz, K. A., Weiner, M. P.
Cytometry 2000, 39, 131140.
55 Chen, J., Iannone, M. A., Li, M.-S., Taylor, D.,
Rivers, P., Nelsen, A. J., Slentz-Kesler K. A.,
Roses, A., Weiner, M. P. Genome Res. 2000,
10, 549557.
56 Samiotaki, M., Kwiatkowski, M., Parik, J.,
Landegren, U. Genomics 1994, 20, 238242.
57 Baron, H., Fung, S., Aydin, A., Bahring, S.,
Luft, F. C., Schuster, H. Nat. Biotechnol. 1996,
14, 12791282.
58 Khanna, M., Park, P., Zirvi, M., Cao, W., Pi-
con, A., Day, J., Paty, P., Barany, F. Oncogene
1999, 18, 2738.
59 Shumaker, J. M., Metspalu, A., Caskey, C. T.
Hum. Mutat. 1996, 7, 346354.
60 Sanger, F., Nicklen, S., Coulson, A. R. Proc.
Natl. Acad. Sci. USA 1977, 74, 5463.
61 Huang, X. C., Stuart, S. G., Bente, P. F., Bren-
nan, T. M. J. Chromatogr. 1992, 29, 289295.
62 Chen, X., Levine, L., Kwok, P. Y. Genome Res.
1999, 9, 492498.
63 Johnson, C. L., Warren, J. H., Giles, R. C.,
Staub, R. W. J. Forensic Sci. 2003, 48, 1260
1268.
64 Ronaghi, M., Nygren, M., Lundeberg, J.,
Nyren, P. Anal. Biochem. 1999, 267, 6571.
65 Whitcombe, D., Theaker, J., Guy, S. P., Brown,
T., Little, S. Nat. Biotech. 1999, 17, 804807.
66 Roses, A. D. Hum. Mol. Genet. 2001, 10,
22612267.
67 Ross, J. S., Ginsburg, G. S. Drug Discov.
Today 2002, 7, 859864.
68 Wilkins, M. R. Toxicology Letters 2002, 127,
245249.
69 Abbott, A. Nature 2003, 425, 760762.
Abstract
Human chronic diseases represent eight
of the ten leading causes of death in the
developed world, and account for over
86% of the deaths occurring among its cit-
izens. Chronic diseases, such as cardiovas-
cular disease, diabetes and cancer, are
characterized by: 1) multifactorial webs of
causality that include complex interactions
between environmental and genetic deter-
minants; and 2) a long latency period be-
tween first exposure to the risk factor and
clinical presentation. To address this medi-
cal need, the pharmaceutical and biotech-
nology industries must identify novel
points in these causal pathways that are
amenable to therapeutic intervention. As
of 1995, the human pharmacopoeia con-
sisted of 1200 approved therapeutic com-
pounds directed against 277 human drug
targets and 61 microbial targets. Despite
the flurry of genomic research in the late
1990s, the anticipated rush of novel drug
target identification, validation and subse-
quent drug development has not been real-
ized. A 2002 updated census of the entire
human drug development industry re-
vealed that the increase from the 1995 sta-
tistic is quite small. Of the 263 non-anti-
microbial new molecular entities approved
in the United States between 1995 and
2003, only 50 act on new drug targets not
in the 1995 list. With the human genome
sequence now available, new strategies are
necessary to efficiently mine the genome
for the set of human druggable targets. In
our work, we have systematically identified
6273 potential drug targets defining for
the first time a complete Pharmaceutically
Tractable Genome (PTG). This chapter will
describe both laboratory and computa-
tional strategies used to identify the PTG
and discuss its subdivision into three dis-
tinct, but non-exclusive, categories: Protein
Therapeutics; Antibody Targets; and Small
Molecule Targets. Finally, the chapter will
present an integrated systems biology
strategy that combines the first large-scale
expression studies of the PTG and the first
whole-organism proteomic pathway ana-
lyses with traditional in vitro and in vivo
assays to streamline target validation and
to identify the subset of the PTG useful
for specific chronic diseases.
Abbreviations
AMPA alpha-amino-3-hydroxy-5-
methyl-4-isoxazolepropionate
ATs antibody therapeutic targets
BAC bacterial artificial chromosome
BLAST Basic Local Alignment Search
Tool
99
4
A Systems Biology Approach to Target Identification and Validation
for Human Chronic Disease Drug Discovery
ISBN: 3-527-31184-X
CC combinatorial chemistry
EGFR epidermal growth factor
receptor
ESTs expressed sequence tags
FGF fibroblast growth factor
FLIP-R fluorescence imaging plate
reader
GLU glutamate
GPCR G-protein coupled receptor
HTS high-throughput screening
IHC immunohistochemistry
mAbs monoclonal antibodies
NHRs nuclear hormone receptors
NK-1 neurokinin-1
NMDA N-methyl-D-aspartate
ORFs open reading frames
PAC phage P1-based artificial chro-
mosomes
PDGF-D platelet-derived growth factor-D
PPAR peroxisome proliferator-activated
receptor
PTG Pharmaceutically Tractable
Genome
PTs protein therapeutics
RACE Rapid Amplification of cDNA
Ends
RefSeq The Reference Sequence
Database
RNAi ribonucleic acid interference
RTQ real-time quantitative
SMTs small molecule therapeutic
targets
SPDI secreted protein discovery initia-
tive
TATA promoter with a TATA sequence
Y2H yeast-two-hybrid
4.1
Limitations in the Chronic Disease Drug
Discovery Process
4.1.1
Addressing the Burden of Chronic Diseases:
Progress Through 1995
Within the developed world, the cost of
chronic diseases, illnesses characterized by
a prolonged course with little chance for
spontaneous resolution, is tremendous.
Chronic diseases (e.g., cardiovascular dis-
ease, cancer, chronic obstructive lung dis-
ease) represent eight of the 10 leading
causes of death in the developed world [1],
and account for over 86% of the deaths oc-
curring among its citizens [2]. Nine of the
ranking 10 causes of disability (measured
as the sum of the years of life lost due to
premature mortality or severity-adjusted
disability [3]) are chronic diseases, and this
list extends to less-fatal medical conditions
such as unipolar depression, alcohol de-
pendence and deafness [4]. Chronic dis-
eases can also impact healthy life expectan-
cy. Current statistics indicate that while, in
the developed world, the average overall
life expectancy is about 80 years [5], the
average healthy life expectancy is 70 years
[6]. These data suggest that the last 10
years of ones life will be spent coping
with disability due to chronic disease. The
developing world is similarly affected. Car-
diovascular disease is already their leading
cause of mortality, and these countries are
experiencing a significant annual increase
in both the number of deaths and disabil-
ity-adjusted life years lost due to chronic
diseases [7]. Addressing the unmet medi-
cal needs presented by chronic diseases of-
fers significant challenges and opportu-
nities for the pharmaceutical industry.
Chronic diseases are, by nature, com-
plex. Epidemiologic advances of the past
4 A Systems Biology Approach to Target Identification and Validation 100
decade have established that chronic dis-
eases are characterized by multifactorial
webs of causality that include complex in-
teractions between both environmental
and genetic determinants [8]. Moreover,
the clinical onset of chronic disease symp-
toms typically occurs after a substantial la-
tency period following first exposure to the
risk factor [9] and may require continued
exposure over many years to accrue sub-
stantial clinical effect [10, 11]. Offering
medical relief from chronic illness will re-
quire identifying novel points in relevant
causal pathways that are amenable to ther-
apeutic intervention.
The pharmaceutical industry began pur-
suing chronic diseases in earnest during
the decades immediately following the
Second World War [12]. Seminal work by
R. P. Ahlquist which defined two major
types of adrenergic receptors [13] initiated
a quest to identify other physiologically
active cell-based receptors as well as both
natural and synthetic compounds that
could modulate their activity. By the 1970s,
pharmaceutically active compounds target-
ing other protein classes including meta-
bolic enzymes and ion channels emerged
from anti-microbial programs (e.g., lovas-
tatin [14, 15], rapamycin [16]) or from re-
finements of natural products (e.g., risperi-
done from lysergic acid diethylamide
(LSD) [17] and taxotere from paclitaxel
[18]). During this same period, the inven-
tion of recombinant DNA technologies led
to the birth of the biotechnology industry.
In 1982, recombinant human insulin be-
came the first genetically engineered prod-
uct to gain regulatory approval [19] fol-
lowed by growth hormone (1985) and al-
pha-interferon (1986) [20]. OKT3, the first
approved monoclonal antibody therapeutic,
introduced the second major biotechnolo-
gical innovation, targeted immunothera-
peutics, in 1986 [21]. In 1995, the human
marketed pharmacopoeia consisted of
1200 compounds [22] directed against 277
unique human drug targets and 61 micro-
bial proteins (redacted from [23]). Repre-
sented among this set of targets are 87 G-
protein coupled receptors, 46 ion channel
modifiers, 80 metabolic enzymes and eight
nuclear hormone receptors (Fig. 4.1a). Of
these targets, 30 are the receptors for pep-
tidergic hormones or other recombinant
human proteins used as therapeutics.
4.1.2
Evolving a New Paradigm
for Chronic Disease Drug Discovery
In the mid-1990s, the pharmaceutical in-
dustry introduced a series of technological
and strategic innovations, including high-
throughput screening (HTS), combinato-
rial chemistry (CC) and cell-based assays,
that were expected to both streamline the
discovery phase and yield more discoveries
worth pursuing as research efforts [21].
Contemporaneously, US legislation was
enacted to streamline regulatory approval
and stimulate pharmaceutical innovation
for orphan indications [24]. Ten years later,
however, fewer new therapeutic targets
were identified and developed than antici-
pated. This result is more startling when
considering that the period of 19961999
yielded one of the highest historical new
product approval levels [24]. A 2002 up-
dated census of the global drug develop-
ment industry counted 374 unique drug
targets for all reasonable small molecule
compounds in development from the late
preclinical stages through the post-market
(Fig. 4.1b). As this accounting includes all
compounds in development in addition to
those on the market, the new total is a
small increase from the 1995 statistic.
More tellingly, of the 263 non-antimicro-
bial new molecular entities approved in
the United States from 19952003, only 50
act on new drug targets not enumerated
in the 1995 list.
Pundits are quick to dismiss this appar-
ent lack of productivity as the consequence
of having set unrealistic goals for these
technologies back in the mid 1990s [25].
For example, the application of HTS tech-
nology on the large libraries generated by
CC methods was expected to inundate
drug development pipelines with structur-
ally innovative compounds [26]. However,
after 12 years of utilizing these methods,
these expectations have clearly not been
Fig. 4.1 The distribution of human drug targets
by protein family from (A) the 277 unique proteins
enumerated as drug targets against the 1995 mar-
keted human pharmacopoeia (adapted from
[Drews and Ryser pullout]) and (B) the 374 unique
proteins determined by the 2002 revised census
to be the targets of all Lipinski rule-of-five com-
pliant small molecule compounds in all stages of
drug development from preclinical phases through
the post-market (adapted from online Table 1 in
[Hopkins & Groom]). The increased number of Ja-
nus kinases in 1995 is due to the inclusions of
receptor targets for the set of biological therapeu-
tics which are not counted in 2002. The increased
number of ion channels in 1995 is explained by
the inclusion of all isoforms of ligand gated ion
channels (e.g., AMPA, kainate GLU 1-4, GLU 5-7
and NMDA 1, 2ad glutamate/aspartate isoforms
to total 14 entries) even if some are either only
weak or theoretical binders of known drugs.
met. To date, no drug candidate emerging
from HTS-CC has been approved for mar-
keting, although inhibitors to the cathe-
psins and p38 MAP kinase, among others,
are in clinical trials [26].
We feel, however, that the pharmaceuti-
cal industry is poised at its third historical
inflection point. Perhaps, the gap between
expected and realized innovation is be-
cause the standards of chronic disease
drug discovery and development have un-
dergone a paradigm shift such that pre-
viously successful best practices can no
longer prevail in the current environment.
In order to stay current, pharmaceutical
innovators must acknowledge this shift
and update their operations by integrating
methods and concepts evolving from this
new paradigm. The requisite first steps are
to understand the nature of this new para-
digm and the inevitable consequences it
precipitates.
A framework defining this new para-
digm can be derived from epidemiology.
Epidemiologic historians recognize three
specific eras of modern epidemiology
that display temporal parallels with the
history of drug discovery. The 1880s transi-
tion between the Sanitation Era to the In-
fectious Disease Era [27] occurred, conve-
niently, at the same time that the pharma-
ceutical industry emerged from the chemi-
cal dye industry [12]. Similarly, the post-
Second World War onset of the Chronic
Disease Era when, in the developed world,
rising chronic disease mortality overtook
mortality from infectious causes [27], coin-
cided with the watershed developments in
receptor biology that led to the first inno-
vations in chronic disease therapeutics
[12]. The Chronic Disease Eras dominat-
ing philosophy was the black box para-
digm, where a disease outcome was
viewed as a self-contained unit, the inner
processes of which were often unknown
and considered of little relevance to the in-
vestigator. Similarly, during this time peri-
od, therapeutics were discovered, devel-
oped and approved empirically, with little
or no knowledge of the mechanism of ac-
tion involved [25]. During its time, the
black box paradigm was successful. A
survey of 12 top pharmaceutical firms esti-
mated the average pre-tax out-of-pocket
cost per approved drug during the 1970s
to be $114 million US dollars (in 1987)
[28], a figure that could sustain industry
growth and support innovation.
The past 15 years development of new
large-scale wet-lab and information tech-
nologies facilitating experiments to un-
cover the pathophysiologic and genetic ba-
sis of chronic diseases, as well as the
mechanisms of action and toxicity for
drugs designed to treat these chronic dis-
eases, has replaced the black box para-
digm. Epidemiologists have already ac-
knowledged this by ending the Chronic
Disease Era and inaugurating a new era
that embraces these technologies [27]. We
believe that a similar paradigm shift must
occur in drug discovery and development.
The current model of appending genomic
and other high-throughput technologies
onto the existing drug development scaf-
fold is not sustainable. The amortized
costs of developing a single drug have ri-
sen to over $800 million US dollars (in
2000) [29], which has prompted many
drug development programs to limit their
efforts towards developing blockbusters
in order to recuperate costs [30]. As many
chronic disease markets represent smaller
population niches [7], their medical needs
would remain unmet in this current envi-
ronment.
Our approach at CuraGen has been to
practice an updated drug development
paradigm that specifically uses large-scale
systems biology innovations to drive and
prioritize our drug discovery pipeline. We
believe that this updated method is more
efficient and promises to be more cost-ef-
fective than the current model. This chap-
ter will describe our successes in utilizing
this schema for novel target discovery and
subsequent validation in pursuit of specific
chronic disease indications. We will pres-
ent our methods for leveraging the output
of the Human Genome Project, and our
unique efforts to first characterize the
transcriptome (the set of genes expressed
in a cell) to define the Pharmaceutically
Tractable Genome, the set of all potential
druggable targets. We will then describe
our integrated systems biology approaches
for systematically validating these potential
targets for selected disease indications. Fi-
nally, by example, we will demonstrate
how these processes have allowed us to
pursue development projects for orphan
and other unmet chronic disease indica-
tions including oral mucositis, ulcerative
colitis, and glomerulonephropathy.
4.2
Creating the Pharmaceutically Tractable
Genome
We define the Pharmaceutically Tractable
Genome (PTG) as the complete set of hu-
man genes with the potential to serve as tar-
gets for human chronic disease drug discov-
ery and development using the set of currently
available drug development technologies [30,
31]. These include genes and their encoded
proteins that can serve as small molecule
therapeutic targets (SMTs), as monoclonal
antibody therapeutic targets (ATs) or as
genes that can be manufactured as recombi-
nant protein therapeutics (PTs) for subse-
quent administration. Although products
from newer technologies such as gene ther-
apy or antisense are in clinical trials, these
are still unproven, and as such, suitable tar-
gets for these are not considered here.
Small molecule drugs, for the most part,
interact with the catalytic site of an en-
zyme, the ligand binding site of a receptor
or ion channel or an allosteric site, and re-
sult in either inhibition or stimulation of
the target. Consequently, protein families
amenable to small molecule drug develop-
ment include certain classes of transmem-
brane receptors (e.g., G-protein-coupled re-
ceptors, receptor tyrosine kinases), ion
channels and transporters/ion pumps,
kinases, phosphatases, proteases, nuclear
hormone receptors and all classes of meta-
bolic enzymes (e.g., dehydrogenases, iso-
merases, reductases) with a focus on those
that have chemical families already known
to interact with the target. Monoclonal an-
tibodies (mAbs), as therapeutics, bind to
their targets and either neutralize the tar-
gets activity, stimulate an antibody-depen-
dent cell cytotoxicity or complement-de-
pendent cytotoxicity immune response to
kill cells bearing its target, or deliver an
antibody-conjugated radioisotope, drug or
toxin to target-bearing cells [32]. A subset
of mAbs (e.g., tositumomab [33]) has activ-
ity through several of these mechanisms.
Due to their biophysical properties, mAbs
are typically confined to the extracellular
space; for this reason, valid ATs are limited
to cell-surface and secreted proteins on tis-
sues that are accessible to the blood
stream. Protein therapeutics are proteins
for which a recombinant form with sys-
temic therapeutic qualities can be pro-
duced. The two most readily considered
classes of PTs are hormones and growth
factors; however, this class can be extended
to include not only all secreted proteins
but also the extracellular domains of cell-
surface proteins, if these domains have the
potential to act as ligands for a second re-
ceptor (e.g., transmembrane semaphorins).
An important consequence of this classi-
fication schema is that the PT, AT, and
SMT subclasses are not mutually exclusive.
Protein families can simultaneously be-
long to two, or even all three, druggable
classes. Cell-surface receptors with extra-
cellular ligand binding domains and intra-
cellular catalytic domains can function as
both ATs and SMTs. For example, the epi-
dermal growth factor receptor (EGFR) is
the target for both geftinib, an approved
small molecule therapeutic, and cetuxi-
mab, an approved mAb therapeutic. Extra-
cellular proteases, like the tissue plasmino-
gen activator, are pharmaceutically tract-
able in all three categories. A recombinant
protein, alteplase, is an approved protein
therapeutic. Moreover, as a secreted en-
zyme, either small molecule or mAb thera-
peutics are potential inhibitors of the pro-
teins function.
4.2.1
Mining the Pharmaceutically Tractable
Genome
The majority of large-scale gene identifica-
tion efforts, including those undertaken
for the publication of the Human Genome
Project [34, 35], those undertaken at Gen-
entech [36] and those undertaken here at
CuraGen, have utilized highly complex, in-
tegrated and systematic approaches. As in-
tensive gene identification efforts have
spanned several years, the quality, compo-
sition, and availability of sequence data
has changed considerably with the comple-
mentary development of mining strategies
to fully utilize the ever-evolving data. All
approaches begin with roughly equivalent
sequence data, and most intertwine and
build upon a few basic bioinformatic
methods for gene identification. This sec-
tion presents the methods we used for
first defining and then mining the PTG,
as well as techniques used to overcome
some of the inherent limitations.
All gene identification efforts begin with
sequence data. Comprehensive mining
analyses utilize a combination of genomic
sequence data (e.g., genomic clones and
assembled genomic scaffold) and mRNA
sequence data (e.g., expressed sequence
tags (ESTs), cDNAs, CuraGens internally
generated SeqCalling database [37]). To ex-
tract the predicted gene sequences, these
data are subjected to one or more of the
bioinformatic sequence mining strategies
of homology/orthology mining, transcript
mining and algorithm-based de-novo gene
prediction mining. These are described be-
low and presented in Fig. 4.2.
4.2.1.1 In silico Gene Mining Methods
Homology/orthology mining Homology/
orthology mining seeks to identify novel
genes with similar sequence to known
genes or proteins. This approach leverages
Basic Local Alignment Search Tool
(BLAST) algorithms [38], where a known
gene or protein is used as the seed with
which to search the novel sequence space.
The query element can be a sequence
from the same organism (i.e., homology
mining) or from a different organism (i.e.,
orthology mining). This approach identi-
fies genes likely to be family members of
the query gene, and thus extend the num-
ber of genes belonging to protein families
previously proven to be druggable (e.g.,
growth factors and G-protein coupled re-
ceptors (GPCRs)). When mining from
genomic DNA, intron/exon boundaries
must be defined using consensus splice
site information [39, 40]. As this approach
requires a seed sequence, a limitation is
that novel gene families cannot be identi-
fied de novo. Caution must also be exerted
not to erroneously mine artifact genes.
Artifacts are created by either mining open
reading frames (ORFs) that do not occur
or are not expressed in nature from geno-
mic DNA or by misidentifying intron/exon
boundaries.
Transcript mining Transcript mining
searches expressed sequence databases cre-
ated from the assemblies of ESTs and
other mRNA- or cDNA-based sequences
for novel ORFs. To qualify as a novel gene,
an ORF must exhibit a Kozak sequence
[41, 42], start and stop codons, and either
possess similarity to a known gene or con-
tain a functional domain. As identifying
an ORF in a transcript does not depend
on comparison to a known gene or pro-
tein, both entirely novel gene families and
novel splice variants of known genes can
be identified. This method is largely lim-
ited by the low quality typical of many
EST sequences. A significant portion of
EST-based assemblies do not span the en-
tire coding region of a gene, which leads
to the incomplete mining of a partial gene
sequence. Sequence errors including stop
codons, frame shifts and other artifactual
changes can be introduced by the polymer-
ase or by misreads during sequencing [43].
Apparent insertions can result from the se-
quencing of partially spliced transcripts,
and apparent single-exon genes may result
from sequencing contaminating genomic
DNA. However, categorically discounting
singleton ESTs and sequences that appear
to be unspliced or only partially spliced in-
troduces the risk of missing genuine vari-
ants or single exon genes. Single-exon
gene contigs assembled from overlapping
sequence fragments that represent two or
more tissue types prompt increased confi-
dence in their existence. However, sparse
database coverage yields many single-exon
candidates without this level of confidence.
Finally, EST database representation is
biased towards highly expressed tran-
scripts [44]; thus, transcripts correspond-
ing to low-abundance transcripts may not
be detected using this method.
Algorithm-based de novo gene prediction
mining To identify genes from PAC- and
BAC-clone-derived genomic sequences, a
series of in silico gene prediction algorithms
have been developed. These algorithms use
established knowledge of gene structure to
predict the location of novel genes. Genscan
[45] searches both strands of a double-
stranded DNA sequence for both TATA-
based and TATA-less potential promoters,
translation initiation Kozak sequences, and
donor and acceptor splice sites. Exon and in-
tron structures are estimated based on em-
pirically observed length distributions, and
separate parameters are applied for internal,
initial, and terminal exons, and for single-
exon genes. Translation termination signals
are assessed through observed stop codon
frequency and potential polyadenylation sig-
Fig. 4.2 A schematic representation of the inte-
grated gene mining methods used to extract the
Pharmaceutically Tractable Genome (PTG). The
set of publicly available, purchased proprietary
and internally generated DNA sequences repre-
senting both genomic contigs and expressed se-
quences is uploaded into our in silico analysis sys-
tem. Novel gene and splice variant identification
is then accomplished using our integrated
mining approach which combines homology/
orthology mining, expressed sequence mining and
de novo gene prediction algorithms. Full-length
cloning methods both validate predicted se-
quences as well as supplement the gene discovery
process. All identified genes are then character-
ized in silico and sorted according to their phar-
maceutically tractable gene families. Targets are
then queued for chronic disease assignment.
3
nals. Finally, as gene density and structure
varies depending on G/C content, Genscan
categorizes the G/C content of a sequence
into one of four quartiles, and uses a sepa-
rate set of parameters for each quartile. Gen-
scans robustness was validated on the fin-
ished sequence of Chromosome 22. Some
94% of all annotated genes were at least par-
tially predicted by Genscan, and 20% of
genes had all exons correctly predicted
[46]. The FirstEF algorithm [47] was created
to optimize identification of promoters and
first exons, a weakness of other gene predic-
tion programs. FirstEF algorithm was
trained on the commonalities identified
among *2000 experimentally confirmed
first exons, and identifies first exons based
on predicted splice sites, CpG windows,
and promoters. In a test set of 121 genes,
FirstEF predicted 86% of confirmed first
exons with a 17% false positive rate.
The principal drawback of gene predic-
tion programs is their inaccuracy. Not only
do these programs overlook gene se-
quences whose structures do not conform
to the rules applied in the training data,
but they can also produce incorrect exon
predictions. As a result, in our experience,
intensive human quality control is re-
quired to review all predictions and elimi-
nate these errors.
Integrated mining approaches As none of
these gene identification techniques is suffi-
ciently robust, we have optimized novel
gene prediction by integrating several strat-
egies in various combinations superim-
posed upon a proprietary database compris-
ing the most complete set of human tran-
scripts. For example, we have identified sets
of pharmaceutically tractable genes by: 1)
beginning with sets of proteins of interest
(e.g., growth factors), using these to per-
form a TBlastN algorithm search of geno-
mic DNA, then supporting and completing
putative genes and their predicted exon
boundaries using expressed sequences and
Genscan processing; 2) beginning with
genomic scaffold, systematically using Gen-
scan to predict exons de novo, and then sub-
jecting each putative exon to BlastX analysis
to identify similar proteins for completing
the gene by homology mining; 3) beginning
with expressed sequences, assembling them
into longer virtual transcripts, subjecting
the transcripts to BlastX analysis, and
further extending partial genes using geno-
mic scaffold homology and Genscan
mining; and 4) beginning with genes or
proteins of interest from other species and
performing orthology mining supported
by expressed sequences and Genscan
mining. Note that many other scenarios
(and other species) besides those described
here have been used by CuraGen; this set
serves to illustrate what has been and can
be constructed using the three basic mining
techniques described.
Mining splice variants While the human
genome contains approximately 50000
genes [34, 35], further proteomic diversity
is achieved through alternative splicing of
multi-exon genes. Current estimates sug-
gest that alternative splicing occurs in 30
60% of human genes [34, 4850]. Though,
at present, no drugs have been developed
based on the products of alternatively
spliced genes, the potential impact of alter-
natively spliced gene products on drug dis-
covery is large. Druggable domains may
be alternatively spliced to alter a proteins
activity and/or targetability. Similarly, alter-
native splicing can alter a proteins subcel-
lular localization which can affect, in a dis-
ease state- or tissue-dependent manner, its
amenability for drug development. Some
10 to 30% of splice variants are expressed
in a tissue-specific manner [51], raising
the possibility that drugs that leverage
splice variation may have fewer effects on
non-targeted tissues.
The most effective method for mining
splice variants superimposes transcript
gene sequences on the genomic scaffold
and identifying exons added to or removed
from known sequences [52]. In one pub-
lished study, 2000 genes with predicted al-
ternative splicing were identified from
8429 scaffold-mapped, multi-exon tran-
script clusters [52]. A second study exam-
ined 171 genes for alternative splicing
using ESTs and identified splice variants
for 48.5% of these [40]. To avoid false posi-
tives, only exons with appropriate consen-
sus splice sites are considered and human
quality control is required to eliminate the
miscalling of retained introns as legitimate
variants.
4.2.1.2 Wet-lab Experimental Approaches
Experimental approaches complement the
bioinformatic methods in that they provide
empirical data of a genes existence and/or
function. However, experimental ap-
proaches are less amenable to high-
throughput scale-up and thus have been
used more sparsely. These approaches are
most successful when integrated with and
can complement the bioinformatic algo-
rithms. Two well-validated examples of ex-
perimental high-throughput gene identifi-
cation methods are the Yeast Signal Trap
and Full-length gene cloning.
The yeast signal trap This approach an-
chored Genentechs SPDI program de-
signed to identify secreted and transmem-
brane proteins. The method involves screen-
ing for sequences from a cDNA library that
directed the secretion of a reporter protein
from yeast [36]. Libraries of cDNA frag-
ments were cloned upstream of a reporter
gene, and inserts from yeast colonies secret-
ing the reporter protein were amplified, se-
quenced and then fully mined using a com-
bination of bioinformatic techniques. Gen-
entechs approach has led to the successful
subcloning of 47 novel gene loci and 209
variants of known genes representing 256
potential protein therapeutics [36].
Full-length gene cloning At CuraGen, we
took advantage of the fact that full-length
cloning, while critical of the generation of
intellectual property, can itself be a valu-
able mining tool for the set of genes that
display robust bioinformatics predictions
for only part of the gene sequence. For
those genes with strong 5' and/or 3' pre-
dictions, cloning the gene from cDNA
using primers designed in the high-confi-
dence regions can be used to identify the
missing middle region(s). Conversely, if a
middle section of the gene is predicted
with high confidence but not the ends,
Rapid Amplification of cDNA Ends
(RACE) is used to clone the missing ends.
Cloning from cDNA can also lead to the
identification of novel splice variants. To
confirm a predicted splice variant, se-
quencing primers must be designed to
bridge the alternate splice site or reside
within the novel insertion. Lastly, promis-
cuous cloning primers can also lead to the
fortuitous cloning of novel genes. As the
possession of a physical clone successfully
extracted from a cDNA library serves as
the ultimate experimental confirmation of
a predicted genes existence, these full-
length cloning strategies are also used to
generate these confirmatory clone sets.
4.2.2
Annotating and Organizing the Contents
of the PTG
The PTG contains 6273 genes. In the
PTG, known genes are named according
their recognized alpha-numeric abbrevia-
tions that are used in The Reference Se-
quence Database (RefSeq) [53]. Novel
genes with no previous annotation prior to
their mining by CuraGen were assigned
preliminary names according to their
homology with known proteins or func-
tional domains (e.g., semaphorin D-like,
IgG domain-containing protein). Each nov-
el gene is also assigned an internal alpha-
numeric abbreviation beginning with the
prefix CG.
The scope of the PTG is best appreciated
when juxtaposed to the contents of the en-
tire human genome. To date, in addition
to the two complementary versions of the
genome published in 2001 [34, 35], com-
pleted sequence has been published to
date for six of the autosomes [46, 5458].
Current best estimates suggest that the
human genome contains 35000 to 50000
genes. Our PTG, therefore, represents ap-
proximately 15% of all human genes.
Given the PTGs size, a strict organiza-
tion strategy is required. The 30 classes of
druggable protein families have been
sorted according to their suitability as PTs,
ATs, and SMTs. Each PTG entry is then as-
signed to one of these protein families. To
date, we have counted 4075 SMTs, 2933
ATs, and 1254 PTs. Similar to the 1995
and 2002 drug target censuses, GPCRs re-
present the largest single protein family in
our PTG with 865 counted members. Of
these, over 75 were single-exon GPCRs lo-
cated in several genomic clusters. The set
of metabolic enzymes, grouped in our ac-
counting into the intracellular proteases,
lipases, dehydrogenases, transferases and
other intracellular enzymes include 1251
unique targets. Hormones, cytokines, che-
mokines, and growth factors four protein
families that have previously been success-
fully introduced as recombinant protein
therapeutics comprise 402 entries in the
PTG. Finally, the PTG includes several
classes of molecules whose therapeutic po-
tential is only now being realized, such as
both intracellular and extracellular kinases
and phosphatases, as well as cell adhesion
molecules. These classes of targets repre-
sent over 1000 potential novel points of in-
tervention for chronic disease manage-
ment. The size of each protein family, and
their distribution according to target type,
is presented in Fig. 4.3.
Further understanding of the PTG can
be gained by organizing the member
genes of each protein family. To accom-
plish this, the predicted DNA sequences of
all members of a protein family was ana-
lyzed using the ClustalW algorithm [59] to
produce dendrograms of each protein fam-
ily where the distance between any two
members is proportional to the evolution-
ary divergence between the two sequences.
The dendrograms generated from the hor-
mone and chemokine protein therapeutic
families, as an example, are presented in
Fig. 4.4.
4.3
Integrated Systems Biology Approaches to
Drug Target Validation for Specific Clinical
Indications
4.3.1
Causal Inference and Target Validation
The precipitating step for any drug discov-
ery program is the successful validation of
the relationship between a candidate drug
target and the intended disease, thus indi-
cating that the target is a crucial and effec-
tive point of intervention for drug therapy
[30]. At the highest level, validating a target
involves invoking criteria of a causal rela-
tionship between a proposed target and a
selected disease (an exception to this rule
4.3 Integrated Systems Biology Approaches to Drug Target Validation for Specific Clinical Indications 111
F
i
g
.
4
.
3
V
e
n
n
d
i
a
g
r
a
m
r
e
p
r
e
s
e
n
t
i
n
g
t
h
e
d
i
s
t
r
i
b
u
t
i
o
n
o
f
t
h
e
6
2
7
3
m
e
m
b
e
r
s
o
f
t
h
e
P
h
a
r
m
a
c
e
u
t
i
c
a
l
l
y
T
r
a
c
t
a
b
l
e
G
e
n
o
m
e
a
c
c
o
r
d
i
n
g
t
o
t
h
e
i
r
p
r
o
t
e
i
n
f
a
m
i
l
i
e
s
a
n
d
d
r
u
g
g
a
b
l
e
c
l
a
s
s
(
i
.
e
.
,
P
T
,
A
T
,
S
M
T
)
.
A
s
d
r
u
g
g
a
b
l
e
c
l
a
s
s
a
s
s
i
g
n
m
e
n
t
i
s
n
o
t
e
x
c
l
u
s
i
v
e
,
g
e
n
e
s
c
a
n
b
e
l
o
n
g
t
o
o
n
e
,
t
w
o
,
o
r
t
h
r
e
e
o
f
t
h
e
d
r
u
g
g
a
b
l
e
c
l
a
s
s
e
s
.
is drug-conjugated mAbs where the target
only needs to be associated with the dis-
ease and not specifically on the causal
pathway). The standard epidemiologic defi-
nition of a cause would indicate that a can-
didate drug target lies in a diseases causal
pathway if its modulation precedes the on-
set of disease and, if in the absence of this
modulation, the disease would have oc-
curred at a later time, or not at all [60]. Tar-
get validation, a form of causal inference, is
based upon educated judgments using data
accumulated from the target discovery pro-
cess and subsequent integration of informa-
tion describing the targets expression, the
molecular pathways it participates in as well
as other knowledge of the targets biological
function derived from genetic, biochemical
and other physiologic studies (see also Part
V, Chapter 2 and Part III, Chapter 3).
Current standards in pharmaceutical re-
search list four essential criteria a target
must meet to obtain a validated status
[30]:
Inducing/suppressing the target by
either genetic or pharmacologic means
should lead reproducibly to an altered
physiologic state that is consistent with
the desired therapeutic goal.
This observed effect should be dose-de-
pendent.
The desired phenotypic change must
also be inducible in at least one relevant
animal model.
The targets role in the metabolic, signal-
ing or regulatory pathway in which it is si-
tuated should be defined in both the tar-
get tissue and other key organs where tar-
get manipulation may lead to side effects.
These bullets and their antecedent nine
Hill Criteria for Assigning Causality which
include strength of association, consis-
tency, specificity, temporality, biological
gradient, plausibility, coherence, experi-
ment and analogy [61], imply that a binary,
linear relationship links the candidate
drug target and the identified disease. Es-
sentially, the pathophysiology of disease Y
can be impacted by manipulating target X.
However, the literature demonstrates how
this simple view of causality breaks down
when applied to the relationship between
real compounds and their indicated dis-
eases [62]. Pharmacologic research has
strongly implicated the neurokinin-1 (NK-
1) receptor in both human and animal
pain states [63]. Yet, in the subsequent de-
velopment of selective NK-1 receptor an-
tagonists, despite convincing preclinical ef-
ficacy in animals, the compounds failed as
analgesics in human trials [64, 65]. This
target validation rubric also stumbles
when trying to assimilate drugs like cloza-
pine and carbamazepine compounds
that are clinically effective despite lack of
data supporting their modulation of a sin-
gle receptor that, in isolation, is causative
of their respectively targeted diseases [62].
The sufficient-component cause model [66]
is an increasingly sophisticated schema
that is more compatible with a post-geno-
mic paradigm for target validation. Under
this definition, a sufficient cause the com-
plete causal mechanism that inevitably
produces disease is not a single element
but is made up from unique and specific
component causes. While component
causes, in isolation, cannot cause disease,
Fig. 4.4 Protein Therapeutic protein families den-
drograms displaying the evolutionary relationships
among family members. Known genes are named
according the their standard gene name; novel
genes are assigned an arbitrary alpha-numeric ab-
breviation and are named using their RefSeq an-
notation standard.
3
all components of a single sufficient caus-
al pie must fall into place in order for dis-
ease to develop. Finally, most chronic dis-
eases have multiple sufficient causal con-
stellations that each independently can
cause the selected disease; any component
cause that is found in all independent suf-
ficient causes is designated a necessary
cause [60]. The ramifications for target vali-
dation are as follows. Drug therapy that in-
terrupts a targets function as a component
cause of disease should mitigate the dis-
ease if and only if no other intact suffi-
cient causes are present. Clozapine is ef-
fective because its dirty polypharmacy si-
multaneously interrupts several sufficient
causes for schizophrenia. Applying this
paradigm to target validation implies un-
derstanding which sufficient causes a can-
didate target is a component of and to rule
out (or acknowledge) the possibility that
the target is not a necessary cause of dis-
ease. Integrated systems biology is ideal
for efficiently addressing this issue.
Our validation strategy uniquely le-
verages a set of integrated systems biology
methods including high-throughput cellu-
lar assays, gene/protein expression profil-
ing, proteinprotein interaction mapping,
genetic mining, expression pharmacoge-
nomics and gene knockdown/knockout.
We will present only briefly in this chapter
our experience with this integrated plat-
form, citing specific examples from our
own drug development program.
4.3.2
Validation Strategies for Assigning Targets
to Selected Chronic Diseases
Our validation process begins with assign-
ing PTG members to diseases in which they
may contribute to a causal pathway. While
this step seems trivial for the set of known
PTG candidates for which protein function
and disease associations have been charac-
terized in the literature, these methods are
crucial for initiating work on novel PTG
elements, the sequences of which were first
determined as a result of the mining strate-
gies listed above, as well as for identifying
novel therapeutic uses for previously
thought to be well-understood targets. For
ATs and SMTs, one expectation is to identify
the PTG subset which is overexpressed in
tissues from one or more specific disease
states, while being minimally expressed
if at all in normal tissues. This safeguards
against the development of adverse events.
Identifying chronic disease associations for
novel targets also facilitates the assignment
of intellectual property for both the compo-
sition of matter of the target itself and first-
in-class method of use applications for the
target (as a protein therapeutic) and for
any monoclonal antibody or small molecule
modulators of the target. A flow-chart of our
target validation process is presented in
Fig. 4.5.
Disease assignment is initiated by assay-
ing levels of target mRNA expression in
both normal and representative diseased
tissues. RTQ-PCR (reviewed in [67]) and
Fig. 4.5 A schematic representation of the meth-
ods used to triage and validate targets from the
Pharmaceutically Tractable Genome (PTG). Ana-
lyses of a targets expression pattern, pathway in-
volvement, and genetics are used to determine if
the target may be associated with disease (target
qualification). Disease-associated targets are then
subjected to further study in vitro, followed by
study in animal models in order to demonstrate
that modulation of the target affects the disease
process (target validation).
"
microarray technologies (reviewed in [68])
are equally successful in this regard. Cura-
Gen was the first to recognize the advan-
tage of developing a PTG-specific gene
chip, and has used this chip to situate new
targets in biological context. RTQ-PCR, in
addition to confirming microarray discov-
eries, is used to define more specifically
the expression of a gene of interest across
a larger set of diseased and normal tissues.
To date, we have hybridized mRNA from
over 1200 normal and diseased tissues as
well as cell lines to our PTG gene chip,
and subsequently have performed over
20000 RTQ-PCR studies.
Our RTQ-PCR methods involve creating
96- and 384-well tissue panels which
contain an array of representative normal
and disease-specific tissues against which
PTG target mRNA is amplified. For exam-
ple, an oncology tissue panel contains
samples of: 1) primary tumors, represent-
ing the gamut of stages and grades, from
selected organ sites; 2) metastatic tumors
derived from the same primary site (e.g.,
for prostate cancer, both liver and bone
metastases would be represented); 3) nor-
mal adjacent tissue at the margins of sur-
gical resection; 4) normal tissue from an
individual with no history of cancer; and
5) established cell lines derived from that
cancer type. For inflammatory diseases, we
use a series of primary immune system
cells both stimulated with known cyto-
kines and unstimulated controls as well as
primary tissues from important target or-
gans of autoimmune and other inflamma-
tory diseases (e.g., articular cartilage from
rheumatoid arthritis patients). The micro-
array approach involves creating a 35-mer
oligonucleotide-based chip with a unique
probe for each PTG entry and iteratively
hybridizing mRNA from diseased and nor-
mal tissue selections that parallel the
RTQ-PCR panel selections.
An initial round of qualitative data anal-
ysis scans the panel output for gene candi-
dates that display disease tissue expression
compared to normal controls. More so-
phisticated analysis strategies involve mul-
tivariate statistical methods including hier-
archical and k-means clustering, principal
components analysis and multivariate
analysis of variance [69, 70] to identify sys-
tematically those targets that display statis-
tically significant up-regulated levels of ex-
pression across a set of diseased samples
compared to the normal control tissue
sampling. For example, scientists mining
a Phase 3 genomic clone from chromo-
some 11 discovered CG50595, a novel ace-
tylglucosaminyltransferase with identical
amino acid sequence to the LARGE pro-
tein. However, unlike LARGE, which
spans over 660kb of chromosome 22 [71],
CG50595 covers only 7kb on chromosome
11. RTQ-PCR demonstrated that while
CG50595 had relatively minimal expres-
sion on most normal tissues except for
placenta and pancreas, its expression was
significantly up-regulated in a wide spec-
trum of cancer cell lines including colon
cancer SW620 and lung cancer LX-1
(Fig. 4.6). Further RTQ-PCR studies show
that CG50595 is significantly up-regulated
in samples from primary lung, prostate,
breast and colon cancers with minimal ex-
pression in matched normal adjacent tis-
sue samples. These findings support the
anti-neoplastic potential for LARGE small-
molecule antagonists previously suggested
in the literature [72, 73].
To verify that levels of mRNA expression
correlate with a meaningful clinical pheno-
type, tissue immunohistochemistry (IHC)
is applied to evaluate this putative associa-
tion. IHC not only identifies the specific
cell type within the diseased tissue that
bears the PTG target but also allows con-
firmation of the targets subcellular local-
Fig. 4.6 RTQ-PCR panel results for CG50595, a
novel acetylglucosaminyltransferase across a spec-
trumof normal tissues and cancer cell lines. The cell
line with the highest expression level (i.e., SW620)
is arbitrarily set to 100%, and the relative expres-
sion of all other data points are presented as per-
cent expression to this reference point.
ization (i.e., membranous, cytoplasmic,
nuclear). Additionally, it is at the IHC step
that we determine a statistically robust es-
timate of the targets prevalence across a
representative sample of cases. The long-
term stability of paraffin blocks guarantees
a broader case base that fuels our high-
throughput operation. IHC can be per-
formed using standard single slide itera-
tive methods or by leveraging newer tissue
microarray technology [74].
The set of ATs that circulate in the
bloodstream require a modification to this
IHC protocol. These are situations where
the disease stimulates abnormal produc-
tion of a hormone/growth factor by one
tissue that has deleterious effects on a sec-
ond organ and a therapeutic neutralizing
monoclonal antibody that eliminates the
function of this ectopic factor is sought. To
verify the suspected elevated circulating
levels of these factors, sera from patients
with the clinical condition is compared to
a set of valid controls using immunochem-
istry.
Platelet-derived growth factor-D (PDGF-
D) is a novel member of the PDGF growth
factor family mined from a unique com-
plementary cDNA initially sequenced at
CuraGen [75]. Initial in vitro and biochem-
ical experiments demonstrated that PDGF-
D, like PDGF-B, a well-established onco-
genic protein [76], forms homodimers that
activate both PDGF receptors a and b and
stimulates the growth of CCD 1070sk pri-
mary human foreskin fibroblast cells and
primary human smooth muscle cells as
well as several human cancer cell lines
[75]. We suspected that PDGF-D may con-
tribute to cancer progression. To support
that neutralizing anti-PDGF-D monoclonal
antibodies may be useful in treating cer-
tain cancers, our scientists assayed levels
of PDGF-D in both the serum and pri-
mary tumor tissues from cancer patients.
A variety of primary tumor types were re-
presented. Where control subjects with no
history of cancer had mean serum levels
of PDGF-D below the detection limit of
4.0ngmL
1
(n=50), ovarian cancer pa-
tients had mean serum levels of
10.8ngmL
1
(n=43) and lung cancer pa-
tients displayed mean serum levels of
5.7ngmL
1
(n=32). Overall, PDGF-D was
expressed at concentrations >10ngmL
1
in 69 of 245 cancer patients compared to 3
of 50 normal controls [77].
Assigning protein therapeutics to dis-
ease indications requires a different per-
spective. Here, the goal is to identify dis-
eases where the application of an exoge-
nous hormone or growth factor will ame-
liorate a pathologic situation. Our process
has established a systematic battery of cell-
based assays using a broad spectrum of
primary cells and established cell lines to
determine whether PTs can, for example,
inhibit or stimulate epithelial, mesenchy-
mal or hematopoeitic cell proliferation,
mesenchymal cell migration, angiogenesis,
apoptosis, immune cell differentiation, or
chemotaxis. The approach was designed
both to be systematic both in terms of
chronic diseases and to capture the ability
to modulate specific mechanisms underly-
ing chronic diseases. By processing FGF-
20, a novel member of the fibroblast
growth factor family, through this assay
set, we found that FGF-20 induced DNA
synthesis in CCD1070sk primary human
fibroblasts, CCD-1106 KERTr human kera-
tinocytes, and human breast epithelial
cells [78]. In contrast, FGF-20 had no ef-
fect in the immune response modulation
assay set and the angiogenesis assay set.
We have leveraged the finding that FGF-20
is one of the few FGFs to be mitogenic on
both epithelial and mesenchymal cells to
advance programs in oral mucositis and
inflammatory bowel disease [79, 80], where
accelerating the repair of gastrointestinal
mucosa and stroma may offer substantial
clinical benefit. Similarly, this cell-based in
vitro activity screening panel revealed that
angioarrestin, a novel angiopoeitin-like
molecule that does not bind the Tie1 or
Tie2 receptors, inhibited multiple angio-
genic processes including endothelial cell
proliferation, migration, adhesion and tu-
bular network formation [81], suggesting a
possible use as a protein therapeutic for
the treatment of cancer.
A smaller subset of the PTG can be as-
signed to a disease following the discovery
that genetic variants within the gene (e.g.,
single nucleotide polymorphisms, small in-
sertions or deletions) are associated with a
specific disease process (see also Part I,
Chapter 2). Despite the fact that recent ad-
vances in large-scale genotyping methods
have made this approach technically feasi-
ble [82], the epidemiologic challenges asso-
ciated with recruiting the appropriate study
population for conducting these experi-
ments (e.g., the C981T polymorphism in
the protein tyrosine phosphatase 1B gene
is associated with type 2 diabetes onset in
an Oji-Cree community [83], a finding that
has not been replicated in other type 2 dia-
betes populations) makes this approach less
popular. It is our belief, however, that once
human whole genome sequencing can be
accomplished in real time, genetic methods
will become one of the most powerful meth-
ods of new target validation.
PTG elements that fit into multiple
druggable target classifications are pro-
cessed, in parallel, through each of the tar-
get-class preliminary validation processes.
4.3.3
Validation Strategies for Confirming Disease
Association and Determining the Targets
Specific Role in the Disease Process
Once a PTG target has received a prelim-
inary disease assignment, more rigorous,
systematic experimentation is required to
both confirm the targets association with
the selected disease as well as to deter-
mine the precise role the target plays its
pathogenesis. At this stage, our process
combines focused cellular and biochemical
assays, expression pharmacogenomics,
proteinprotein interaction mapping, gene
knockdown methods (e.g., RNAi) and ani-
mal knockout models (see also Part I,
Chapter 10, Part III, Chapter 3, and Part
III, Chapter 4), as well as, for PTs, assay-
ing the gene product in relevant genetic,
pharmacologic and other (e.g., xenograft
models for oncologic indications) models
of the candidate disease.
For novel drug targets, determining their
specific role in disease pathogenesis often
requires de-orphanizing the target as part
of the process. The de-orphanization of
GPCRs by the calcium ionophore-based
FLIP-R technology is one example [84]. A
more recent innovation is the application
of expression pharmacogenomics for the
de-orphanization of the nuclear hormone
receptors (NHRs). NHRs are intracellular
receptors that, when bound to their ligands,
bind DNA and act as transcription factors
[85]. NHR ligands include the steroids, ster-
ols, prostanoids and polyunsaturated fatty
acids. In several cases, synthetic NHR li-
gands were discovered as therapeutics
either before discovery of the NHRs natural
ligand (e.g., development of thiazolidene-
diones as PPAR-c agonists [86]) or discovery
of the receptor itself (e.g., the development
of fibrates for lipid management [87]). Fif-
teen of the 55 NHR family members are es-
tablished drug targets. Through the applica-
tion of differential gene expression profiling
to elucidate the gene set responsive to
NHR-regulated transcription activation, we
have learned that not only does each NHR
induce a unique gene response signature
within its target tissue but also that this sig-
nature often includes entire metabolic, sig-
naling or regulatory pathways that define
its mechanism of action. While the PPAR-
a receptors role in regulating lipid metabo-
lism is well established, our expression
pharmacogenomics analysis was the first
to reveal the systematic up-regulation of
over 20 metabolic enzymes and transporters
associated with all three pathways of fatty
acid oxidation in hepatic tissue following
treatment with a potent PPAR-a ligand
(Fig. 4.7) [88]. In a similar fashion, the far-
nesoid-X-activated receptor was discovered
to be a regulator of bile acid metabolism
and homeostasis with the potential to clini-
cally modulate cholestasis, the liver-X-recep-
tors a and b as key hepatic sensors of dietary
cholesterol and regulators of feed-forward
pathways in cholesterol catabolism as well
as regulators of adipose tissue glucose utili-
zation, and the pregnane-X-receptor and
constitutive-active-receptor as a principal
pathway regulators of xenobiotic metabo-
lism through CYP3A and CYP2B families,
respectively [85, 89, 90]. CuraGen pharma-
cogenomicists have demonstrated that spe-
cific GPCRs and CNS-related ligand-gated
ion channel receptors, when either stimu-
lated or inhibited, also generate specific ex-
pression signatures [91]. These data support
an alternate mechanism for de-orphanizing
GPCRs and ion channels.
For the set of gene families that play
roles in intracellular signal transduction,
additional insight into their specific dis-
ease associations can be gained from situ-
ating them within protein interaction net-
works generated by yeast-two-hybrid (Y2H)
[92] and other similar technologies. While
specific drug target candidates can be indi-
vidually cloned into Y2H vectors and
assayed for their interaction with other
genes contained in a specific cDNA library,
we have pioneered a more comprehensive
systematic approach to situating PTG ele-
ments within the proteome. CuraGen col-
leagues have generated whole-proteome
Y2H-based interaction maps for the first
eukaryotic organism, S. cerevisiae, and D.
melanogaster, the latter being the first such
attempt in a multi-cellular organism [93,
94]. Since many elements in the PTG are
reasonably well conserved in Drosophila
[95], situating the Drosophila ortholog of a
PTG candidate within the interaction map
and identifying the networks of genes
emanating from it can suggest specific
physiologic roles for the candidate [93]. In
this manner, putative disease assignments
can be made for a substantial section of
the PTG. Fig. 4.8 displays the protein in-
teraction network for the Drosophila PTG.
We rely on traditional validation methods
including biochemical assays to confirm the
receptor profiles for novel PTs [75, 78] and
on treating both pharmacologic and genetic
animal models of targeted diseases with the
Fig. 4.7 Modulation of fatty acid oxidation by
PPAR-a ligands. Genes up-regulated in a single
differential gene expression profiling experiment
are represented by red circles. Genes not found to
be modulated in this study are represented by
black circles. Purple triangles represent genes with
known PPAR-a transcription factor binding sites.
Aqua circles represent genes previously identified,
prior to this study, as being PPAR-a responsive.
[Reprinted with permission from Gould Rothberg,
et al., Functional and Integrative Genomics 1:294
304 (2001).]
3
novel therapeutic to confirm that modulat-
ing the target impacts the disease [79, 80,
96]. Where possible, we couple these animal
model experiments with specific biomarker
assays to confirm that the desired target was
indeed impacted regardless of the experi-
ments outcome. Our validation strategies
also include in vitro gene knockdown assays
(see also Part I, Chapter 10) and, in certain
cases, animal knockout models (see also
Part III, Chapter 4). As these strategies are
more comprehensively addressed else-
where, we will not cover them here.
4.4 Conclusion
Adopting a new drug discovery and devel-
opment paradigm that places newer large-
scale laboratory and informatic technolo-
gies at the forefront to drive chronic dis-
ease drug target identification and valida-
tion is essential to ensure the pharmaceu-
tical industrys continued sustainability.
CuraGens approach does just that. We
specifically recognize that in order to pro-
duce breakthrough drugs for unmet
chronic diseases we had to discover the
complete set of potential drug targets, elu-
cidate the remaining intervention points
in known pathways as well as construct
novel pathways and identify the interven-
tion points in these. When we started this
approach, it was our belief that a mechan-
istic understanding of human disease
would lead to the discovery and develop-
ment of innovative and effective new
drugs. To streamline target identification,
we have systematically mined the output
of the Human Genome Project as well as
the large volume of proprietary cDNA se-
quencing to define the Pharmaceutically
Tractable Genome, the comprehensive set
of 6273 targets that can be fed into current
small molecule, monoclonal antibody and
recombinant protein therapeutic drug
discovery programs. We have brought
CG53135 (a.k.a., FGF-20), one of the first
drugs to be mined from the genome using
the above-described methods, into the
clinic. Moreover, our approach has been
validated by the successes of other prod-
ucts whose clinical efficacy is due to their
targeting a root cause of a disease (e.g.,
the treatment of chronic myelogenous leu-
kemia and gastrointestinal stromal tumor
by imatinib due to the directed inhibition
of pathogenic receptor tyrosine kinases
[97, 98]). By integrating valuable concepts
from epidemiology, we have argued that
an integrated systems biology approach
that uses high-throughput genomic, tran-
scriptomic and proteomic technologies to
anchor a drug development process can ef-
fectively and efficiently nominate, prose-
cute, validate and ultimately develop tar-
gets and their relevant drugs as useful in-
tervention points for both common and
rare chronic diseases.
Note
In 2004, Dr. J. M. Rothberg was elected to
the United States National Academy of
Engineering for his pioneering work in
mining the human genome.
Note 123
Fig. 4.8 Protein family/disease ortholog view of the
Drosophila protein-interaction map. Proteins are
color-coded according to protein family as anno-
tated by the Gene Ontology hierarchy. Proteins
orthologous to human disease proteins have a
jagged, starry border. Interactions were sorted ac-
cording to their interaction confidence score and
the top 3000 interactions are shown with their
corresponding 3522 proteins. [Reprinted with per-
mission from Giot, et al., Science 302:17271736
(2003).]
3
References
1 C. D. Mathers, C. Stein, D. M. Fat, et al., Glo-
bal Burden of Disease 2000: Version 2 Methods
and Results. Global Programme on Evidence
for Health Policy Discussion Paper No. 50.
World Health Organization, 2002.
2 C. J. L. Murray, A. D. Lopez, The Global Burden
of Disease: a Comprehensive Assessment of Mor-
tality and Disability from Diseases, Injuries and
Risk Factors in 1990 and Projected to 2020. Har-
vard University Press, USA, 1996.
3 C. J. L. Murray, A. D. Lopez, Lancet 1997, 349,
143642
4 C. J. L. Murray, A. D. Lopez, The World Health
Report 2000: Reducing Risks, Promoting Healthy
Life. World Health Organization, 2002.
5 C. M. Michaud, C. J. L. Murray, B. R. Bloom,
JAMA 2001, 285, 535539.
6 C. D. Mathers, R. Sadana, J. A. Salomon, et al.,
Lancet 2001, 357, 16851691.
7 D. Yach, C. Hawkes, C. L. Gould, K. J. Hof-
man, JAMA 2004, 291, 26162622
8 D. Clayton, P. M. McKeigue, Lancet 2001, 358,
13561360.
9 I. J. Selikoff, J. Churg, E. C. Hammond, JAMA
1964, 188, 2226.
10 J. M. Samet, G. R. Eradze, Environ. Health Per-
spect. 2000, 108(suppl 4), 635641.
11 A. T. Chan, E. L. Giovannucci, E. S. Schern-
hammer, et al., Ann. Intern. Med. 2004, 140,
157166.
12 J. Drews, Science 2000, 287, 19601964.
13 R. P. Ahlquist, Am. J. Physiol. 1948, 153, 586
600.
14 A. Endo, M. Kuroda, Y. Tsujita, J. Antibiot.
(Tokyo) 1976, 29, 13461348.
15 A. Endo, Y. Tsujita, M. Kuroda, K. Tanzawa,
Eur. J. Biochem. 1977, 77, 3136.
16 S. N. Sehgal, H. Baker, C. Vezina, J. Antibiot.
1975, 38, 727732
17 F. C. Colpaert, Nat. Rev. Drug Discov. 2003, 2,
315320.
18 I. Ringel, S. B. Horwitz, J. Natl Cancer Inst.
1991, 83, 28891.
19 I. S. Johnson, Nat. Rev. Drug Discov. 2003, 2,
747751.
20 Approved Biotechnology Drugs, http://
www.bio.org/speeches/pubs/er/approveddrugs.asp,
2004.
21 J. Drews, Drug Discov. Today 1998, 3, 491494.
22 The 1996 Tarascon Pocket Pharmacopoeia,
Tarascon Publishing, USA, 1996.
23 J. Drews, S. Ryser, Nature Biotechnol. 1997, 15,
special pullout.
24 J. M. Reichert, Nat. Rev. Drug Discov. 2003, 2,
695702
26 L. J. Gershell, J. H. Atkins, Nat. Rev. Drug Dis-
cov. 2003, 2, 321327.
27 M. Susser, E. Susser, Am. J. Public Health
1996, 86, 668673.
28 J. A. DiMasi, R. W. Hansen, H. G. Grabowski,
L. Lasagna, J. Health Econ. 1991, 10, 107142
29 J. A. DiMasi, R. W. Hansen, H. G. Grabowski,
J. Health Econ. 2003, 22, 151185.
31 A. L. Hopkins, C. R. Groom, Nat. Rev. Drug
Discov. 2002, 1, 727730.
32 D. E. Milenic, E. D. Brady, M. W. Brechbiel,
Nat. Rev. Drug Discov. 2004, 3, 488498.
33 A. J. Davies, A. Z. S. Rohatiner, S. Howell, et
al., J. Clin. Oncol. 2004, 8, 14691479.
34 International Human Genome Sequencing
Consortium, Nature 2001, 409, 860921.
35 J. C. Venter, M. D. Adams, E. W. Myers, et al.,
Science 2001, 291, 130451.
36 H. F. Clark, A. L. Gurney, E. Abaya, et al., Ge-
nome Res. 2003, 13, 22652270.
37 K. E. Quinn-Senger, R. Ramachadran, J. A. Ri-
ninger, et al., Curr. Opin. Chem. Biol. 2002, 6,
418426.
38 S. F. Altschul, W. Gish, W. Miller, et al., J.
Mol. Biol. 1990, 215, 403410.
39 S. M. Mount, Nucleic Acids Res. 1982, 10, 459
472
40 F. Clark and T. A. Thanaraj, Hum. Mol. Genet.
2002, 451464.
41 M. Kozak, J. Mol. Biol. 1987, 196, 947950.
42 M. Kozak, Nucleic Acids Res. 1987, 15, 8125
8148.
43 A. Lindlof, Appl. Bioinformatics 2003, 2, 123129.
44 M. Cariaso, P. Folta, M. Wagner, et al., Bioin-
formatics 1999, 15, 965973.
45 C. Burge, S. Carlin, J. Mol. Biol. 1997, 268,
7894.
46 I. Dunham, N. Shimizu, B. A. Roe, et al., Na-
ture 1999, 402, 489495.
47 R. V. Davuluri, I. Grosse, I., M. Q. Zhang, Na-
ture Genet. 2001, 29, 412417.
48 L. Croft, S. Schandorff, F. Clark, et al., Nature
Genet. 2000, 24, 340341.
49 A. A. Mironov, J. W. Fickett, M. S. Gelfand, Ge-
nome Res. 1999, 9, 12881293.
50 Z. Kan, E. C. Rouchka, W. R. Gish, D. J. States,
Genome Res. 2001, 11, 889900.
51 Q. Xu, B. Modrek, C. Lee, Nucleic Acids Res.
2002, 30, 37543766.
52 B. Modrek, A. Resch, C. Grasso, C. Lee, Nu-
cleic Acids Res. 2001, 29, 28502859.
53 K. D. Pruitt, D. R. Maglott, Nucleic Acids Res.
2001, 29, 137140.
54 A. J. Mungall, S. A. Palmer, S. K. Sims, et al.,
Nature 2003, 425, 805811.
55 S. J. Humphray, K. Oliver, A. R. Hunt, et al.,
Nature 2004, 429, 369374.
56 P. Deloukas, M. E. Earthrowl, D. V. Grafham,
et al., Nature 2004, 429, 375381.
57 P. Deloukas, L. H. Matthews, J. Ashurst, et al.,
Nature 2001, 414, 865871.
58 M. Hattori, A. Fujiyama, T. D. Taylor, et al.,
Nature 2000, 405, 311319.
59 A. Aiyar, Methods Mol. Biol. 2000, 132, 221241.
60 A. Aschengrau, G. R. Seage, Essentials of Epide-
miology and Public Health, Jones and Bartlett
Publishers, USA, 2003.
61 A. B. Hill, Proc. R. Soc. Med. 1965, 58, 295300.
62 M. Williams, Curr. Opin. Pharmacol. 2003, 3,
571577.
63 Q.-P. Ma, R. G. Hill, Curr. Opin. CPNS Invest.
Drugs 1999, 1, 6571.
64 R. A. Dionne, Curr. Opin. CPNS Invest. Drugs
1999, 1, 8285.
65 R. G. Hill, Trends Pharmacol. Sci. 2000, 21,
244246.
66 K. J. Rothman, Epidemiology: An Introduction,
Oxford University Press, USA, 2002
67 J. C. Willey, E. L. Crawford, C. R. Knight, et al.,
Methods Mol. Biol. 2004, 258, 1341.
68 M. Schena, Microarray Analysis, Wiley-Liss,
USA, 2003.
69 H. K. Hamadeh, P. R. Bushel, S. Jayadev, et
al., Toxicol. Sci. 2002, 67, 219231.
70 H. Zhang, C. Y. Yu, Front. Biosci. 2002, 7, c6367.
71 M. Peyrard, E. Seroussi, A.-C. Sandberg-
Nordqvist, et al., Proc. Natl Acad. Sci. USA
1999, 96, 598603.
72 H.-B. Guo, I. Lee, M. Kamar, et al., Cancer
Res. 2002, 62, 68376845.
73 T. Saito, E. Miyoshi, K. Sasai, et al., J. Biol.
Chem. 2002, 277, 1700217008.
74 R. L. Camp, G. G. Chung, D. L. Rimm, Nat.
Med. 2002, 8, 13231327.
75 W. J. LaRochelle, M. Jeffers, W. F. McDonald,
et al., Nat. Cell Bio. 2001, 3, 517521.
76 C.-H. Heldin, B. Westermark, Physiol. Rev.
1999, 79, 12831316.
77 W. J. LaRochelle, M. Jeffers, J. R. F. Corvalan,
et al., Cancer Res. 2002, 62, 24682473.
78 M. Jeffers, R. Shimkets, S. Prayaga, et al.,
Cancer Res. 2001, 61, 31313138.
79 E. Alvarez, E. G. Fey, P. Valax, et al., Clin. Can-
cer Res. 2003, 9, 345434561.
80 M. Jeffers, W. F. McDonald, R. A. Chiilakuru,
et al., Gastroenterology 2002, 123, 11511162.
81 M. Dhanabal, W. J. LaRochelle, M. Jeffers, et
al., Cancer Res. 2002, 62, 38343841.
82 K. Tang, P. Oeth, S. Kammerer, et al., J. Pro-
teome Res. 2004, 3, 218227.
83 A. Mok, H. Cao, B. Zinman et al., J. Clin En-
docrinol. Metab. 2002, 87, 724727.
84 A. Wise, S. C. Jupe, S. Rees, Annu. Rev. Phar-
macol. Toxicol. 2004, 44, 4366.
85 J. T. Moore, B. Goodwin, T. M. Willson, S. A.
Kliewer, Crit. Rev. Eukaryot. Gene Expr. 2002,
12, 119135.
86 S. A. Kliewer, J. M. Lenhard, T. M. Willson, et
al., Cell 1995, 83, 813819.
87 M. M. Best, C. H. Duncan, Arch. Intern. Med.
1966, 118, 97102
88 B. E. Gould Rothberg, S. S. Sundseth, V. A. Di-
Pippo, et al., Funct. Integr. Genomics 2000, 1,
294304.
89 T. M. Stulnig, K. R. Steffensen, H. Gao, et al.,
Mol. Pharm. 2002, 62, 12991305.
90 J. M. Rosenfeld, R. Vargas, W. Xie, R. M.
Evans, Mol. Endocrinol. 2003, 17, 12681282
91 E. C. Gunther, D. J. Stone, R. W. Gerwien, P.
Bento, M. P. Heyes, Proc. Natl Acad. Sci. USA
2003, 100, 96089613.
92 E. Phizicky, P. I. Bastiaens, H. Zhu, M. Sny-
der, S. Fields, Nature 2003, 422, 208215.
93 L. Giot, J. S. Bader, C. Brouwer, et al., Science
2003, 302, 17271736.
94 P. Uetz, L. Giot, G. Cagney, et al., Nature
2000, 403, 623627.
95 L. T. Reiter, L. Potocki, S. Chien, M. Gribskov,
E. Bier, Genome Res. 2001, 11, 111425.
96 T. Ostendorf, C. R. C. Van Roeyen, J. D. Peter-
son, et al., J. Am. Soc. Nephrol. 2003, 14,
22372247.
97 B. J. Druker, M. Talpaz, D. J. Resta, et al.. N.
Engl. J. Med. 2001, 344, 10311037.
98 A. T. van Oosterom, I. Judson, J. Verweij, et al.
2001 Lancet 2001, 358, 14211423.
References 125
Abstract
Human epidermal growth factor receptor 2
(HER2) is a trans-membrane receptor with
tyrosine-kinase activity encoded by a proto-
oncogen (her-2) which is amplified in 20
30% of breast cancer patients. The discovery
that overexpression of HER2 is a negative
prognostic factor led to the concept of devel-
oping a therapy specifically targeted against
this molecule. To this end, several murine
monoclonal antibodies (mAbs) to the extra-
cellular domain of HER2 were developed,
some of which showed growth inhibition
of cell lines overexpressing the HER2 recep-
tor. The most potent of these mAbs, 4D5,
was found to markedly inhibit proliferation
of cell lines overexpressing HER2, but had
little or no effect on cells without elevated
HER2 levels. As 4D5 was also demonstrated
to be a potent inhibitor of growth of human
breast cancer xenografts, it was selected for
further development and was subsequently
humanized. The resulting antibody, Her-
ceptin
(trastuzumab), retained the high af-

finity for the HER2 epitope and showed
similar promising preclinical tumor inhibi-
tion as the parental antibody.
During early clinical development, Her-
ceptin has shown to comprise activity as a
single agent in metastatic breast cancer
even in heavily pretreated patients. Subse-
quently, a large pivotal Phase III combina-
tion trial has demonstrated that its use
with the chemotherapeutic agent paclitaxel
results in a significant improvement in
survival, time to progression, and re-
sponse. This has recently been reinforced
by another randomized trial in combina-
tion with docetaxel. More than 14000 pa-
tients were enrolled in early breast cancer
trials (adjuvant treatment) with Herceptin.
Interim analyses demonstrated an approxi-
mately 50% reduction of the risk of recur-
rence.
The co-development of a companion di-
agnostic test for assessment of the pa-
tients HER2 status was critically impor-
tant as the HER2 status of a tumor turned
out to be an essential determinant of re-
sponse to Herceptin-based treatment. Pa-
tients that express HER2 at high levels as
assessed by immunohistochemistry (IHC
3+) or show HER2 gene amplification de-
rive the greatest benefit from treatment
with Herceptin. Hence, the pre-selection
of HER2-positive patients throughout Her-
ceptins clinical development was a prere-
quisite for the positive results seen in the
pivotal trials and in todays routine clinical
practice. Nowadays, three different meth-
ods for HER2 status determination are ap-
plied in clinical practice: IHC, which is
the most widely used method, and fluores-
127
5
The Development of Herceptin
:
Paving the Way for Individualized Cancer Therapy
ISBN: 3-527-31184-X
cence in situ hybridisation (FISH) or chro-
mogenic in situ hybridization (CISH), both
measuring quantitatively the amount of
HER2 gene amplification. In the majority
of laboratories, the use of FISH and CISH
is mostly restricted for re-testing IHC
equivocal (IHC 2+) cases. The use of stan-
dardized and validated testing protocols
for all three methods is critical to ensure
that the right patients are identified and
selected for Herceptin therapy.
With this successful co-development,
Herceptin has established a new paradigm
in cancer drug development and repre-
sents a cornerstone in paving the way for
individualized cancer therapy.
Abbreviations
CISH chromogenic in situ hybridiza-
tion
CTA Clinical Trials Assay
DR duration of response
EGF(R) epidermal growth factor
(receptor)
FISH fluorescence in situ hybridization
HER2 human epidermal growth factor
receptor-2
IHC immunohistochemistry
ISH in situ hybridization
MBC metastatic breast cancer
ORR overall response rate
OS overall survival
PST primary systemic therapy
QoL quality of life
REC Response Evaluation
Committee
TTP time to progression
TTF time to failure
5.1
Introduction
Cancer development is the result of cumu-
lating genetic alterations. A recent major
advance in the treatment of cancer is the
emergence of therapies aimed specifically
at altered gene products or distorted gene
expression, often referred to as targeted
therapy. As these modifications are not
present in normal cells, these new antican-
cer drugs very specifically target the tumor
cells and, to a large extent, avoid damage
to normal cells. Reliable detection of the
altered gene or its protein product to iden-
tify patients that may benefit from these
targeted therapies is therefore often indi-
cated. This article outlines the clinical
development of Herceptin
and develop-
ment of the diagnostic tests that identify
those patients that are most likely to bene-
fit from Herceptin-based therapy, and ex-
plains why this co-development may serve
as a prime example for individualized can-
cer therapy.
Herceptin is a humanized monoclonal
antibody (mAb) targeted to the human epi-
dermal growth factor receptor 2 (HER2),
which is overexpressed in 2030% of hu-
man breast cancers [13]. In the vast ma-
jority of cases, HER2 overexpression is
caused by amplification of the HER2 gene
[2], which occurs early in the development
of breast tumors and is seen frequently in
ductal carcinoma in situ [4, 5]. HER2 gene
amplification results in increased HER2
mRNA levels and concomitant overexpres-
sion of the HER2 receptor on the cell sur-
face [6, 7] (Fig. 5.1). HER2 protein levels
are consequently several orders of magni-
tude greater on the surface of HER2-posi-
tive cells than on adjacent normal breast
epithelium [8].
The high incidence of HER2 overexpres-
sion on the surface of breast cancer cells
: Paving the Way for Individualized Cancer Therapy 128

and the recognized prognostic value of
HER2 was the basis for the development
of Herceptin. Subsequently, the develop-
ment of Herceptin has marked the begin-
ning of a new era of rationally designed,
targeted drugs, which was accompanied
with the development of predictive diag-
nostic tests. This example has set a new
standard for drug development in oncol-
ogy and paved the way for individualized
therapy.
5.2
HER2
HER2 belongs to a family of four homolo-
gous human EGFRs designated as HER1,
HER2, HER3 and HER4 (or ErbB1, ErbB2,
ErbB3 and ErbB4) (Fig. 5.2). The HER re-
ceptors are transmembrane tyrosine ki-
nases with growth-stimulating activity in-
volved in the regulation of normal tissue
growth, cell survival and differentiation [9].
Under normal conditions, HER2 receptors
on the cell surface function as the preferred
heterodimerization partner with other HER
proteins and enhance and/or initiate ligand-
dependent signal transduction.
The underlying mechanism of HER2
protein overexpression in human breast
cancer is of genetic nature, i.e., HER2
overexpression is triggered by the genetic
event of HER2 gene amplification. The
HER2 gene is located on chromosome 17.
HER2 gene amplification leads to in-
creased transcription and consequently to
an overexpression of HER2 receptor pro-
teins on the cell surface. High HER2 re-
ceptor density on these cancer cells facili-
tates formation of HER2 homodimers, re-
sulting in constitutive, ligand-independent
receptor signaling.
Overexpression of HER2 protein has a
number of effects that result in carcino-
genesis:
1. Preferential formation of HER2-contain-
ing heterodimers, which show prolonged
and enhanced downstream signaling due
to increased stability compared with other
HER-containing heterodimers [10, 11],
reduced ligand dissociation [9, 11] and a
decreased rate of endocytosis (compared
with EGFR homodimers) [9, 11].
5.2 HER2 129
Fig. 5.1 Indicators of HER2 status: HER2 gene amplification and
HER2 protein overexpression.
2. Increased recycling of EGFR to the cell
surface, resulting in increased signaling
[1214].
3. Formation of constitutively active HER2
homodimers, resulting in initiation of
downstream signaling pathways [15].
4. Activation and suppression of numerous
signal transduction pathways with po-
tential roles in tumor development and
growth [16].
The clinical relevance of HER2 was first
noted in 1987, when a subset of breast can-
cer patients whose tumors contained an
amplified version of the HER2 gene was
identified [9]. These patients were reported
to have a more aggressive form of breast
cancer [17]. Since that time, the prognostic
and predictive significance of HER2 ampli-
fication and overexpression in breast cancer
has been extensively investigated [18].
HER2 positivity correlates with poor breast
cancer prognosis, including reduced re-
lapse-free and overall survival (OS) [17, 19
24]. The association between HER2 amplifi-
cation/overexpression and poor clinical out-
come suggests that HER2 has a key role in
pathogenesis. An increasing body of evi-
dence also supports the role of HER2 as
an important predictive factor of response
to chemotherapy and hormonal therapy in
breast cancer (reviewed in [2528]).
5.3
Herceptin Mechanism of Action and Effects
on Cellular Processes
While data clearly indicate that Herceptin
markedly inhibits breast tumor growth
[2931], the underlying mechanism that
mediates the antitumor effects of anti-
HER2 mAbs has not yet been fully eluci-
dated. The potential mechanisms by which
Herceptin exerts its antitumor effects are
multiple and currently under investigation.
Available data suggest that major mecha-
nisms for Herceptin include:

Fig. 5.2 The HER family of tyrosine kinase receptors. Known ligands
for each receptor are indicated. No ligand is identified for the HER2
receptor. The HER3 receptor lacks adequate intrinsic tyrosine kinase
activity.
1. Accelerating the internalization and deg-
radation of HER2 receptors from the cell
membrane [3236].
2. Recruiting immune cells to attack and
kill target tumor cells via antibody-de-
pendent cellular cytotoxicity [3740].
3. Inhibition of cleavage of the HER2 extra-
cellular domain by metalloproteinase,
preventing homodimerization of HER2
remnants and therefore antagonizing
the constitutive growth signaling [16, 41,
42].
4. Interaction with other signaling path-
ways [43].
5.4
Preclinical Evidence
Initial preclinical studies have indicated
that anti-HER2 mAbs are able to stop the
growth of HER2-overexpressing tumor
cells [34, 40, 44, 45]. Importantly, numer-
ous subsequent in vitro studies have dem-
onstrated that Herceptin exerts antiprolif-
erative activity and antitumor effects
against a variety of HER2-overexpressing
cancer cell lines including human breast,
gastric and ovarian cancer [8, 30, 37, 46
48], but not against those cells that do not
overexpress HER2.
This anti-HER2 activity has also been
demonstrated in human xenografts models
[30, 4648]. Inhibition of tumor growth by
muMAb 4D5, the murine parent antibody
to Herceptin, was observed in vivo in a
xenograft model with human Murray
breast tumors and human Paxton ovarian
tumors that were implanted into the sub-
renal capsule of nude mice. Since these
original studies with muMAb 4D5, similar
in vitro and in vivo efficacy studies have
been repeated with Herceptin. The efficacy
of Herceptin has been demonstrated in
vivo in a xenograft model using both
MCF7-HER2 cells, a cell line that was
engineered to overexpress HER2 and a
HER2-overexpressing breast cancer cell
line, BT-474 (Roche, data on file).
Importantly, HER2 antibodies consider-
ably enhance the effect of conventional
chemotherapy, as demonstrated in numer-
5.4 Preclinical Evidence 131
Table 5.1 Mean combination index values for chemotherapeutic drug/
Herceptin combinations in vitro (SK-BR-3 cell line)
Drug Combination
index
p value Interaction
Vinorelbine [49] 0.24 <0.001 synergy
Docetaxel [49] 0.30 <0.001 synergy
4-Hydroxycyclophosphamide [49] 0.38 <0.001 synergy
Carboplatin [49] 0.42 <0.001 synergy
Etoposide [30] 0.54 0.0003 synergy
Cisplatin [30] 0.56 0.001 synergy
Thioepa [30] 0.67 0.0008 synergy
Paclitaxel [49] 0.87 0.381 addition
Doxorubicin [49] 0.88 0.284 addition
Epirubicin [49] 0.88 0.297 addition
Vinblastine [30] 1.09 0.26 addition
Methotrexate [30] 1.36 0.21 addition
5-Fluorouracil [30] 2.87 0.0001 antagonism
ous in vitro studies. Additive and synergistic
effects have been noted with Herceptin plus
a number of agents commonly used in the
treatment of breast cancer (Table 5.1) [30,
49]. In-vivo studies using human HER2-pos-
itive breast cancer xenografts in estrogen-
primed athymic mice confirm the synergis-
tic activity of Herceptin in combination with
cisplatin [30]. Herceptin also enhances the
tumoricidal effects of paclitaxel, docetaxel,
doxorubicin alone, carboplatin plus doxoru-
bicin, cyclophosphamide, methotrexate, eto-
poside and vinblastine in other in vivo stud-
ies of various human HER2-overexpressing
breast cancer xenografts (Fig. 5.3a and b)
[29, 30, 50].
The additive or synergistic therapeutic
effects demonstrated in these studies high-
light the potential clinical significance of
combining various anticancer agents with
Herceptin, providing the rationale to inves-
tigate these combinations in clinical trials.

Fig. 5.3 (a) Antitumor effect of Herceptin in combination with paclitaxel
against human breast cancer xenografts in athymic mice [29]. (b) Antitumor
effect of Herceptin in combination with docetaxel against human breast
cancer xenografts in athymic mice [49].
5.5
HER2 Testing as a Prerequisite for Herceptin
Therapy: Development of Commercially
Available and Validated Testing
Methodologies
As outlined above, already early on in the
development of Herceptin it was evident
that the target of this therapeutic antibody,
HER2, must be present on the cell surface
of the tumor cells in relatively high
amounts so that a relevant therapeutic ef-
fect could be achieved. Thus, it was appar-
ent for the early clinical trials that a diag-
nostic test assessing the HER2 status of
the breast tumor had to be performed in
order to detect and select patients with
HER2 overexpressing disease. The deci-
sion to preselect those patients with
HER2-overexpressing breast tumors was
highly critical for the further development
and a prerequisite for the demonstration
of Herceptins clinical efficacy.
Testing for HER2 status in the two initial
Herceptin monotherapy and combination
registration trials was performed using an
investigational immunohistochemistry
(IHC) assay developed specifically for the
trials and known as the Clinical Trials Assay
(CTA). IHC employs antibodies specifically
directed against an epitope of the HER2
protein in the tumor tissue, thereby detect-
ing HER2 on the cell surface. HER2 expres-
sion in fixed breast tumor samples is recog-
nized by a typical IHC staining pattern of
tumor cells and is interpreted semi-quanti-
tatively by the observer, applying a 03+
scale, where IHC 3+ indicates the strongest
staining intensity (Fig. 5.4a). Only patients
with tumors showing an IHC 2+ and IHC
3+ result were allowed to participate in
these initial pivotal trials. In these studies,
benefit from Herceptin was seen mainly
in the IHC 3+ patient population ([5153],
see Section 5.7 for details). Hence, Hercep-
5.5 HER2 Testing as a Prerequisite for Herceptin Therapy 133
Fig. 5.4 Determination of HER2 positivity: exam-
ples of IHC 3+ (a), FISH-positive (b) and CISH-
positive (c) breast cancer cases.
tin was approved in Europe in 2000 for me-
tastatic breast cancer (MBC) patients with
tumors showing HER2 overexpression with
an IHC score of 3+.
As the CTA was too impractical for com-
mercialization and widespread clinical use,
a standardized and validated IHC assay kit
(HercepTest
TM
) was developed by DakoCy-
tomation. Following approval of Herceptin,
the testing for HER2 status has drawn in-
creasing interest, and has led to the devel-
opment and commercial marketing of sev-
eral anti-HER2 antibodies and assays for
IHC testing. IHC assays are optimized for
usage on fixed, paraffin-embedded tumor
tissues, which is the commonly used ma-
terial in clinical practice.
Over time, it became evident that HER2
gene amplification is the genetic event
which leads to overexpression of HER2 on
the cell surface, and already in initial stud-
ies investigating the relationship between
HER2 gene amplification and HER2 over-
expression, gene amplification without
protein overexpression and protein overex-
pression without gene amplification was
extremely rare [17, 5456]. Many subse-
quent studies have further investigated the
relationship between HER2 gene amplifi-
cation and HER2 protein expression status
and a high correlation has been observed
(see Table 5.2ac).
Therefore, several methods have been
investigated to detect HER2 gene amplifi-
cation including Southern blot analysis,
polymerase chain reaction (PCR) and in
situ hybridization (ISH) methodologies,
such as fluorescence ISH (FISH) and
chromogenic ISH (CISH).
FISH is a DNA-based methodology that
directly assesses the HER2 gene copy num-
ber. Interpretation of the testing results is
numeric and more quantitative than IHC
(Fig. 5.4b). As with IHC, FISH is per-
formed on formalin-fixed paraffin-em-

Table 5.2 (a) IHC/FISH, (b) IHC/CISH and
(c) FISH/CISH concordance data
Study/Reference No.
of
cases
Overall
con-
cordance
(a) IHC/FISH concordance data
Anderson et al., 2004 [74] 1296 92
Yaziji et al., 2004 [75] 4111 91
Yaziji et al., 2004 [76] 2913 91
Dowsett et al., 2003 [77] 426 92
Hofmann et al., 2003 [78] 289 93
Vincent-Salomon et al.,
2003 [79]
116 91
Cianciulli et al., 2002 [80] 66 70
McCormick et al., 2002 [81] 198 87
Paik et al., 2002 [82] 104 94
Roche et al., 2002 [83] 119 92
Birner et al., 2001
a)
[84] 207 98
202 93
207 92
Lebeau et al., 2001
a)
[85] 78 95
79 86
79 95
Maas et al., 2001 [59] 529 90
Tsuda et al., 2001
a)
[86] 215 95
101 95
Tubbs et al., 2001
a)
[87] 145 90
145 90
Hoang et al., 2000 [88] 100 97
Kakar et al., 2000 [89] 112 92
Ridolfi et al., 2000 [90] 116 87
Tanner et al., 2000 [91] 157 92
(b) IHC/CISH concordance data OC
Bilous et al., 2004
b)
[92] 50 82
Hofmann et al., 2004 [93] 86 87
Peiro et al., 2004
a)
[94] 59 93
59 92
Arnould et al., 2003 [95] 75 76
Kournelis et al., 2003 [96] 66 85
Muller et al., 2003 [97] 73 85
Sapino et al., 2003
a)
[98] 106 85
106 80
Van de Vijver et al., 2003
[99]
199 85
Wixom et al., 2003 [100] 81 89
Dandachi et al., 2002 [101] 171 92
Zhao et al., 2002
a)
[102] 62 92
62 95
62 92
bedded tumor tissue samples, and assesses
HER2 levels on a cell-by-cell basis. Nowa-
days, several validated FISH assays are com-
mercially available and FISH is broadly ap-
plied in routine clinical practice. Subse-
quently, another ISH hybridization method-
ology, i.e., CISH, has been developed. With
CISH, the HER2 gene is detected using a
peroxidase enzyme-labeled probe with a
chromogenic detection instead of using a
fluorescent dye (Fig. 5.4c). CISH is nowa-
days also a commercially available assay, va-
lidated and applied in clinical practice.
Testing of HER2 amplification using
standardized and validated FISH and
CISH test kits has been proven to provide
alternative and reliable methodology for
assessing HER2 status in breast cancer
specimens. This is demonstrated by an in-
creasing number of comparative studies
that have reported high concordance (80
100%) between all three HER2 testing
methodologies (see Table 5.2ac).
The reliability and high concordance be-
tween IHC, FISH and CISH and the cur-
rent routine HER2 testing practice in the
laboratories are reflected in the diagnostic
information of the Herceptin SmPC which
was approved in 2004 by the European
Commission. Since then patients are eligi-
ble for Herceptin therapy whose tumors
either have HER2 overexpression or HER2
gene amplification.
5.6
HER2 Testing Algorithm
Although the HER2 testing methodologies,
IHC, FISH and CISH, are all specific and
highly reliable, IHC is used in the vast
majority of cases as the first test for as-
sessing the HER2 status of breast cancer
specimens. This is mainly due to its sim-
plicity, its relatively low costs and the rela-
tively short time to obtain results. How-
ever, the development of FISH and CISH
and clinical evidence (see Section 5.7) led
to the recommendation by many national
testing guidelines to apply a HER2 testing
algorithm (Fig. 5.5; adapted from [57, 58]).
The HER2 testing algorithm reflects that
Herceptin has been shown to provide most
benefit in women whose tumors overex-
press the HER2 protein at the 3+ level or
demonstrate amplification of the HER2
gene. Furthermore, many trials clearly
demonstrated that clinical outcomes are
similar in IHC 3+ or FISH-positive patient
populations [52, 5961]. A proportion of
samples defined as equivocal by IHC (i.e.,
IHC 2+) also demonstrate amplification of
the HER2 gene. For example, in the pivo-
tal phase III Herceptin combination trial
(H0648g), 24% of IHC 2+ tumors also
showed HER2 gene amplification [62].
Therefore, it is recommended that IHC 2+
tumor samples are assessed for amplifica-
5.6 HER2 Testing Algorithm 135
Table 5.2 (continued)
Study/Reference No.
of
cases
Overall
con-
cordance
Tanner et al., 2001 [103] 94 100
Tanner et al., 2000 [91] 157 98
(c) FISH/CISH concordance data
Bilous et al., 2004
b)
[92] 50 94
Hofmann et al., 2004 [93] 86 90
Arnould et al., 2003 [95] 75 96
Park et al., 2003 [104] 188 94
van de Vijver et al., 2003
b)
[99]
208 90
Zhao et al., 2002 [102] 62 100
Tanner et al., 2000 [91] 157 94
Concordance was calculated with IHC 0, 1+, and 2+
as negative and IHC 3+ as positive.
a) Study used different antibodies for IHC, therefore
concordance data presented per antibody.
b) Inter-laboratory concordance, i.e., IHC and CISH
were performed in different laboratories.
tion of the HER2 gene by FISH or CISH
to ensure that all patients who may benefit
from Herceptin are identified.
Guidelines from the American Society
of Clinical Oncology recommend that
HER2 overexpression should be evaluated
in every primary breast cancer patient,
either at diagnosis or at the time of recur-
rence [63]. National HER2 testing guide-
lines also recognize early determination of
HER2 status [57]. HER2 testing at first di-
agnosis has a number of distinct advan-
tages over later testing:
1. The prognostic value of HER2 status
can help direct therapy choices.
2. Knowledge of HER2 status gives pa-
tients an informed choice about possible
participation in (neo)adjuvant trials.
3. Upfront knowledge of HER2 status al-
lows immediate first-line use of Hercep-
tin at the time of diagnosis of metastatic
disease.
4. HER2 testing at first diagnosis may be
more accurate than testing later at the
time of recurrence, as it overcomes po-
tential issues resulting from long-term
storage of specimens and handling in
different laboratories.
5.7
Herceptin in Clinical Use
Based on the knowledge of HER2 patho-
genesis and the accumulating data from
preclinical experiments, the clinical devel-
opment of Herceptin was initiated in the
early 1990s, when the hypervariable anti-
gen-binding regions of a potent murine
anti-HER2 mAb (muMAb 4D5) were
grafted into a human immunoglobulin
framework without loss of specificity [37].
The efficacy and safety of this humanized
mAb, now called trastuzumab or Hercep-
tin, were investigated in a series of clinical
trials for the treatment of HER2-positive
breast cancer, which resulted in the license
approval (US in 1998 and EU in 2000).
Fig. 5.6 outlines some important mile-
stones in the development of Herceptin.
Three phase I clinical trials investigating
1517 patients each assessed the safety,
pharmacokinetics and signs for activity of
Herceptin as a single agent (H0407g,
H0452g) and in combination with cisplatin
(H0453g). Selection of cisplatin was based
on available nonclinical data demonstrat-
ing activity of the combination.

Fig. 5.5 HER2 testing algorithm for identifying patients eligible for
Herceptin therapy [58].
The initial license was based on data
from two pivotal trials of first-line Hercep-
tin plus paclitaxel (H0648g) [52] and sec-
ond/third-line monotherapy (H0649g) [51].
Although these trials included women
with IHC 2+ and IHC 3+ disease, sub-
group analyses showed that women with
IHC 3+ or FISH-positive disease (now
classified as HER2-positive disease) gained
the greatest clinical benefits from Hercep-
tin. Details of this development and the
major Herceptin trials are outlined in the
following sections.
5.7.1
Pivotal Trial H0649g:
Herceptin Monotherapy
as Second/Third-line Treatment
Following initial, small clinical trials indi-
cating that Herceptin was an active and
well-tolerated drug in HER2-overexpress-
ing MBC [31, 46], a larger pivotal, phase II
trial was initiated to confirm the efficacy
of Herceptin monotherapy and to further
characterize the safety profile of the drug.
A total of 222 women with HER2-overex-
pressing (IHC 2+ or IHC 3+) MBC were en-
rolled into this multicenter, open-label, sin-
gle-arm trial of Herceptin as second/third-
line monotherapy [51]. All patients had pre-
viously received therapy for MBC, with 68%
of patients having received two or more
lines of therapy. A total of 94% patients
had received prior anthracyclines and 67%
prior taxanes. Primary endpoints of the trial
were overall response rate [ORR; assessed
by an independent Response Evaluation
Committee (REC)] and characterization of
the safety profile of Herceptin. Secondary
endpoints included duration of response
(DR), time to progression (TTP), time to
treatment failure (TTF), OS, quality of life
(QoL) assessment and investigation of Her-
ceptin pharmacokinetics.
The intent-to-treat ORR turned out to be
15%, which is remarkable in view of the
poor prognosis of this patient population
and their extensive pretreatment. The
median DR (9.1 months) in the intent-to-
treat population was significantly higher
than the one achieved with previous che-
motherapy regimens (5.2 months). Median
survival was 13 months, median TTP 3.1
months and median TTF 2.4 months. In
patients responding to Herceptin (n=34)
the median TTF was 11 months compared
with 5.4 months for the prior regimens of
chemotherapy, highlighting the clinical
benefits offered by Herceptin in patients
with HER2-overexpressing disease (Ta-
ble 5.3 [51]).
Fig. 5.6 Hallmarks of Herceptins development.
H0649g
Cobleigh
H0648g
Slamon
US approval
H+ Paclitaxel
H monotherapy
EU approval
H+ Paclitaxel
H monotherapy
M77001
Marty
EU approval
H+ Docetaxel
The IHC 3+ patient subgroup demon-
strated higher response rates than the
overall (IHC 2+/3+) patient population (18
versus 15%, respectively) as well as a long-
er median OS (16.4 versus 13 months).
Retrospective analysis indicated similar
clinical benefits in patients with FISH-pos-
itive and IHC 3+ tumors, including a 19%
response rate in patients with FISH-posi-
tive disease [61]. Significantly, all patients
in the IHC 3+ and IHC 2+ subgroups
who responded to Herceptin tested FISH
positive (Table 5.3 [51]).
5.7.2
Herceptin Monotherapy Trial H0650g:
An Effective Treatment Option
in the First-line Setting
A total of 114 patients with HER2-overex-
pressing (IHC 2+/3+) MBC were enrolled
into this phase II multicenter trial of first-
line Herceptin monotherapy [64]. The
study investigated the utility of Herceptin
monotherapy as first-line treatment in
general but looked also at a higher dose of
Herceptin (8 mg kg
1
loading dose fol-
lowed by 4 mg kg
1
week
1
) in addition to
the standard weekly dose (4 mg kg
1
load-
ing dose, followed by 2 mg kg
1
week
1
).
Patients were randomly allocated to the
high- or standard-dose regimens. Hercep-
tin was continued until disease progres-
sion. The primary endpoints of the study
were ORR and safety, with DR, TTP and
OS as secondary endpoints.
In total, 111 patients were assessable for
response at final analysis 18 months after
enrolment of the last patient (Table 5.4).
At longer than 12 months of follow-up, 17
of 30 responding patients and 22 of 43 pa-
tients obtaining clinical benefit (intent-to-
treat population) had not progressed, pre-
cluding accurate estimates of TTP and DR
in the responding group. Median OS for
all enrolled patients was 24.4 months.

Table 5.3 Comparison of outcomes in all patients and IHC and FISH subgroups in trial H0649g [51, 61]
Population ORR (%) TTP (months) OS (months)
All patients 15 3.1 13
IHC 3+ 18 3.3 16.4
IHC 2+ 6 1.9 NR
FISH-positive 19 3.2 14.2
FISH-negative 0 1.9 8.8
NR=not recorded
Table 5.4 Efficacy of first-line Herceptin monotherapy
in patients with HER2-positive MBC at final analysis
[64]
Percentage of patients
responding
ORR
(CR+PR)
Clinical
benefit
rate
a)
All assessible patients
(n=111)
26 38
Herceptin 2 mg kg
1
weekly (n=58)
24 34
Herceptin 4 mg kg
1
weekly (n=53)
28 42
IHC 3+ (n=84) 35 48
IHC 2+ (n=27) 0 7
FISH-positive (n=79) 34 48
FISH-negative (n=29) 7 10
a) Clinical benefit =complete (CR), partial (PR) or
minor response, or stable disease >6 months.
There was no statistically significant dif-
ference in ORRs obtained with the stan-
dard- versus higher-dose regimens of Her-
ceptin (Table 5.4). Median TTP and OS
were also similar in the two arms (TTP
3.8 and 3.5 months, respectively; OS 22.9
and 25.8 months, respectively) [64].
Subset analyses on the basis of IHC
score revealed that all of the responses
were seen in patients with IHC 3+ tumors
(ORR in patients with IHC 3+ disease was
35%); no patient with an IHC 2+ tumor
responded to Herceptin. Retrospective
FISH analysis showed a similar response
rate in the FISH-positive and IHC 3+ sub-
groups (Table 5.4). Median TTP was longer
in patients with FISH-positive versus
FISH-negative disease (4.9 versus 1.7
months; p>0.0001) [64].
5.7.3
Pivotal Trial H0648g: First-line Herceptin
plus Paclitaxel Improves Survival
Different chemotherapy agents have differ-
ent mechanisms of action and, conse-
quently, different effects on tumor cells.
Preclinical and early clinical data demon-
strated that combining Herceptin with
chemotherapy has superior activity over
chemotherapy alone. Paclitaxel is com-
monly used and effective in MBC [65], and
was investigated in combination with Her-
ceptin as part of the pivotal combination
therapy trial (H0648g).
A total of 469 patients were enrolled into
a randomized, multicenter, phase III trial
of chemotherapy with or without Hercep-
tin [52]. All patients had previously un-
treated, IHC 2+/3+ MBC. Patients who
had previously received anthracyclines in
the adjuvant setting were randomized to
receive paclitaxel (175 mg m
2
3-weekly)
alone (n=96) or with Herceptin (n=92).
All other patients were randomized to
receive anthracycline (doxorubicin 60 mg
m
2
or epirubicin 75 mg m
2
) plus cyclo-
phosphamide (600 mg m
2
) alone (n=138)
or with Herceptin (n=143) (Fig. 5.7). The
primary endpoint was TTP. Secondary
endpoints were ORR, DR, TTF, OS and 1-
year survival. An independent REC deter-
mined disease progression and response.
After a median 30-month follow-up, the
combination of Herceptin plus paclitaxel
improved all clinical endpoints compared
with paclitaxel alone (Table 5.5). ORR dra-
matically improved from 17 to 41%. TTP
and DR more than doubled from 3.0 to
6.9 and 4.5 to 10.5 months, respectively.
Fig. 5.7 Design of pivotal phase III Herceptin combination trial (H0648g) [52].
Of particular note, the addition of Hercep-
tin to paclitaxel also improved survival
from 18.4 to 22.1 months. A survival ad-
vantage of this magnitude is impressive,
especially in light of the poor prognosis of
patients with HER2-positive disease, given
the fact that this subgroup had received
prior anthracyclines and despite the notion
that 72% of the patients initially random-
ized to chemotherapy alone subsequently
received Herceptin following disease pro-
gression. This type of crossover design
normally biases against observing a surviv-
al advantage and would be expected to di-
minish true differences between treatment
regimens [52]. At the same time, Hercep-
tin added little to the toxicity of paclitaxel
which was important for the overall risk-
benefit assessment of this drug.
As seen in the weekly monotherapy
trials (H0649g and H0650g), subset analy-
sis according to HER2 overexpression level
revealed that Herceptin plus paclitaxel
therapy improved outcomes in patients
with IHC 3+ disease relative to the overall
patient population (IHC 2+ and IHC 3+;
Table 5.5) [53]. Of note, the addition of
Herceptin to paclitaxel improved median
survival in patients with IHC 3+ disease
by 7 months (18 to 25 months; Fig. 5.8).
Retrospective FISH testing showed that
92% of IHC 3+ samples were FISH posi-
tive [52, 53].
5.7.4
Pivotal Trial M77001: First-line Herceptin
Added to Docetaxel Confirms Survival
Benefit in Combination with Taxanes
One hundred and eighty-eight patients with
previously untreated MBC were enrolled in
this randomized trial [66]. Entry criteria sti-
pulated that all patients should have IHC 3+
and/or FISH-positive disease. However,

Table 5.5 Efficacy of Herceptin plus paclitaxel in all patients
(IHC 2+ and IHC 3+) compared with patients with
IHC 3+ disease [52, 53]
Herceptin+paclitaxel Paclitaxel alone
All (n=92) IHC 3+ (n=68) All (n=96) IHC 3+ (n=77)
ORR (%) 41 49 17 17
Median TTP (months) 6.9* 7.1* 2.7 3.0
Median DR (months) 10.5 10.9 4.5 4.6
Median TTF (months) 5.3* 6.7 2.9 2.8
Median OS (months) 22 25 18 18
*p<0.05.
Fig. 5.8 Overall survival in patients with IHC 3+
disease who received Herceptin plus paclitaxel or
paclitaxel alone in the pivotal trial (H0648g) [52, 53].
eight patients had IHC 2+/FISH-negative
disease and one patient had IHC 0/1+/
FISH-unknown disease. Ninety-four pa-
tients were randomized to receive docetaxel
alone (100 mg m
2
3-weekly) and 94 to the
same chemotherapy regimen plus weekly
Herceptin (4 mg kg
1
loading dose, fol-
lowed by 2 mg kg
1
week
1
). Two patients
in the combination arm did not receive
study medication. Patients in the docetax-
el-alone arm could cross over to receive Her-
ceptin upon disease progression. ORR was
the primary endpoint. Clinical responses
were assessed radiologically and reviewed
by an independent committee. Secondary
endpoints of the trial were safety, TTP,
TTF, DR, OS and 1-year survival.
At a follow-up of 24 months after the last
patient entered the trial, the addition of
Herceptin to docetaxel was shown to signif-
icantly improve all clinical outcomes inves-
tigated (Table 5.6). Of particular note, med-
ian OS was increased from 22.7 to 31.2
months on addition of Herceptin to docetax-
el (p =0.0325; Fig. 5.9). Patients in the doce-
taxel-alone arm known to have crossed over
to receive Herceptin (at least 57%) appeared
to survive longer than those who did not re-
Table 5.6 Efficacy summary for trial M77001 [66]
Outcome Herceptin+docetaxel (n=92) Docetaxel alone (n=94) p value
ORR (%) 61.0 34.0 0.0002
Median DR (months) 11.7 5.7 0.009
Median TTP (months) 11.7 6.1 0.0001
Median OS
a)
(months) 31.2 22.7 0.0325
a) Kaplin-Meier estimates.
Intent-to-treat population, 24-month cut-off.
Fig. 5.9 Overall survival in patients treated with Herceptin plus docetaxel or
docetaxel alone [66].
ceive subsequent Herceptin suggesting that
using Herceptin upfront provides greater
benefit than sequential use of Herceptin
only after initial chemotherapy [66].
In all subgroups analyzed, Herceptin
plus docetaxel produced higher response
rates compared with docetaxel alone, as in-
dicated in the Forest plot in Fig. 5.10.
5.7.5
Conclusions from the Pivotal Herceptin
Trials
Data from two randomized trials prove un-
equivocally that Herceptin plus a taxane is
associated with significantly improved clin-
ical outcomes, including response rates,
TTP and most importantly survival, com-
pared with taxane alone. In addition, the
crossover occurring within these trials pro-
vides evidence that for optimal clinical
benefits, Herceptin should be given up-
front for the treatment of HER2-positive
MBC. These data on treatment benefit in
combination with chemotherapy have been
accompanied with positive data from a
large number of usually small, mostly sin-
gle-arm phase II trials of different Hercep-
tin chemotherapy regimens (i.e., combina-
tions with vinorelbine, platins, capecita-
bine, triple combinations and others).
In patients who are not eligible to re-
ceiving chemotherapy (i.e., age, concomi-
tant diseases, wish of patient) Herceptin
monotherapy has been shown to offer effi-
cacy as first-line and even as subsequent-
line therapy. Response rates to first-line
monotherapy in patients with HER2-posi-
tive disease were in the same range as
those reported for first-line, single-agent
standard chemotherapy in non-HER2 pre-
selected populations [6770].
Clearly, appropriate selection of the pa-
tients as well as of the therapy regimen that
is applied in the clinical trials are critical
during the development of targeted agents.
Historically, the efficacy of Herceptin has
been questioned during the early phase I
trials due to initial disappointing results.
For instance, one of these trials (H0452g)

Fig. 5.10 Subgroup analysis in trial M77001: forest plot of adds ratio95% confidence intervals [66].
investigated Herceptin monotherapy in re-
fractory cancer patients. Out of 17 patients
only 14 had breast cancer and all patients
have had extensive chemotherapy pretreat-
ment. In addition, the tumors of some of
the patients expressed HER2 only on a nor-
mal/low level (IHC 1+). Overall, this phase
I trial did not show objective responses
(complete/partial responses) and led the
clinical development of Herceptin go on
critical path. In contrast to these results, in-
vestigating Herceptin monotherapy in ear-
lier disease stages demonstrated good effi-
cacy (18 to 35%; see above trials H0650g
in the first-line setting and H0649g in the
second/third-line setting).
5.8
Future Prospects for Herceptin
and other Targeted Therapies
A number of other anticancer agents tar-
geted to specific molecular characteristics
of tumor cells are currently in develop-
ment or undergoing clinical trials. These
include the HER1 tyrosine kinase inhibitor
Tarceva
TM
(erlotinib) and the anti-vascular
endothelial growth factor agent Avastin
TM
(bevacizumab). Combining Herceptin with
other targeted agents may enable tailoring
of therapy to match various molecular
characteristics of individual patients tu-
mors. Such targeted therapy combinations
may enhance efficacy and tolerability com-
pared with cytotoxic regimens.
5.9
Herceptin in Early Breast Cancer
New drugs for the treatment of breast can-
cer are generally introduced into clinical
practice in the metastatic setting. However,
it is accepted that therapeutic response
usually improves when drugs are used ear-
lier in the disease. Therefore, once agents
have shown a therapeutic impact in meta-
static disease, investigation of their
(neo)adjuvant use often follows. The ratio-
nale for introducing Herceptin into early
breast cancer is based on several factors:
1. HER2-positive disease is associated with
poor prognosis, aggressive disease and
high risk of recurrence and metastasis
[3, 17, 1921, 24].
2. HER2-positivity is an early event in
breast cancer development [5].
3. Herceptin specifically targets HER2-posi-
tive breast cancer and is associated with
significant clinical benefits in the meta-
static setting [52, 61, 66].
4. Herceptin has been shown to have few
side-effects and an overall favorable
safety profile [51, 61, 71].
5.10
Herceptin Adjuvant Trials
Four large randomized multicenter trials
that have started in 2000/2001 are investi-
gating Herceptin as adjuvant therapy [72].
The ongoing adjuvant Herceptin trial pro-
gramme is extensive, involving more than
14000 women. Each trial is examining
Herceptin in different regimens and using
different dose schedules (Fig. 5.11). To-
gether, these trials will:
1. Investigate the efficacy and safety of
Herceptin in the adjuvant setting.
2. Compare different adjuvant approaches.
3. Help to determine the optimal duration
of adjuvant Herceptin therapy.
Two trials (NSABP B-31 and Intergroup
N9831) conducted in North America were
examining Herceptin with the standard
US adjuvant regimen of AC followed by a
taxane. The Breast Cancer International
5.10 Ongoing Herceptin Adjuvant Trials 143
Research Group (BCIRG) 006 trial is being
conducted globally. In addition to investi-
gating Herceptin added to a taxane follow-
ing AC, the BCIRG 006 trial is also inves-
tigating the triple combination of Hercep-
tin, docetaxel and carboplatin. The HERA
Trial involves centers outside the US and
uses an alternative approach of administer-
ing Herceptin. Here, the administration of
Herceptin alone following a prior adjuvant
regimen chosen by the investigator as-
sesses the benefit of Herceptin indepen-
dently from previous chemotherapy.
Furthermore, the HERA Trial is investi-
gating the impact of Herceptin given for 1
and 2 years.
All four trials currently ongoing have
successfully passed preplanned interim
safety analyses. Intergroup and NSABP
trials were granted FDA approval for a
joint interim efficacy analysis including
3351 patients. The HERA trial performed
a pre-planned interim efficacy analysis at
475 events including 3387 evaluable pa-
tients. These analyses, presented during
ASCO 2005, showed a reduction in the
risk of recurrent disease of 52% (joint in-
terim analysis) and 46% (HERA interim
analysis), respectively. In addition, the
joint analysis with a median follow-up of 2
years showed a significant benefit in OS
(HR: 0.67; 2p=0.015) whilst the median
observation interval in HERA with only 1
year appeared to be too short to see signifi-
cant benefits in OS at that time (HR: 0.76;
p=0.26).
Furthermore, all analysed trials demon-
strated a positive benefit:risk ratio. Cardiac
safety, which had been an initial concern,
turned out to be within predefined limits
and was as low as 0.5% when Herceptin
was given after chemotherapy (HERA
trial).
Primary systemic therapy (PST), also
known as neoadjuvant therapy, is being in-
creasingly used in early breast cancer to
reduce the size of the tumor, allowing
greater opportunity for breast-conserving
surgery, and decrease the number of posi-
tive nodes. Several phase II PST Herceptin
studies have been conducted using a vari-
ety of different regimens. Results so far
are promising and Herceptin PST has
been shown to be well tolerated and asso-
ciated with a high pathological complete
response rate [73].

52
52
12 H qw40 "
12 H q3w* 14
18 H q3w* 12
"
"
Fig. 5.11 Summary of Herceptin adjuvant trials.
5.11
Conclusion
The promise of targeting a specific and
disease-causing genetic event (HER2 am-
plification) with an anti-HER2-directed
antibody has led to tremendous progress
in the treatment of breast cancer patients
and for the understanding of tumor biol-
ogy. The pitfalls, specifically around pa-
tient selection, that have emerged during
the clinical development program have
shed light on the issue of individualized
medicine and helped to evolve the concept
of proper patient selection for optimal
treatment benefit.
These findings have been of importance
not only for the successful development of
Herceptin itself but are likely to serve as a
prime example and to provide guidance
for the development of future innovative
(anticancer) drugs overall. Ideally, future
development of targeted therapies should
involve concomitant diagnostic and thera-
peutic strategies to circumvent develop-
ment failures and to allow us to maximize
clinical benefit.
References
1 Owens MA, Horten BC, Da Silva MM, et al.
HER2 amplification ratios by fluorescence in
situ hybridization and correlation with immu-
nohistochemistry in a cohort of 6556 breast
cancer tissues. Clin Breast Cancer 2004, 5, 63
69.
2 Ross JS, Fletcher JA, Bloom KJ, et al. Targeted
therapy in breast cancer: the HER-2/neu gene
and protein. Mol Cell Proteomics 2004, 3, 379
398.
3 van de Vijver MJ, Mooi WJ, Peterse JL, et al.
Amplification and over-expression of the neu
oncogene in human breast carcinomas. Eur J
Surg Oncol 1988, 14, 111114.
4 Barnes DM, Bartkova J, Camplejohn RS, et al.
Overexpression of the c-erbB-2 oncogene: why
does this occur more frequently in ductal car-
cinoma in situ than in invasive mammary car-
cinoma and is this of prognostic significance?
Eur J Cancer 1992, 28, 644648.
5 Chong D, Cooke TG, Reeves JR, et al. Quanti-
tation of EGFR and c-erbB-2 expression in pre-
invasive compared to invasive breast cancer. Eur
J Cancer 1999, 35(suppl 4), S203 (abstr 792A).
6 Hynes NE, Gerber HA, Saurer S, et al. Over-
expression of the c-erbB-2 protein in human
breast tumor cell lines. J Cell Biochem 1989,
39, 167173.
7 Hynes NE. Amplification and overexpression
of the erbB-2 gene in human tumors: its in-
volvement in tumor development, significance
as a prognostic factor, and potential as a target
for cancer therapy. Semin Cancer Biol 1993, 4,
1926.
8 Lewis GD, Figari I, Fendly B, et al. Differen-
tial responses of human tumor cell lines to
anti-p185HER2 monoclonal antibodies. Cancer
Immunol Immunother 1993, 37, 255263.
9 Yarden Y, Sliwkowski MX. Untangling the
ErbB signaling network. Nat Rev Mol Cell Biol
2001, 2, 127137.
10 Alroy I, Yarden Y. The ErbB signaling network
in embryogenesis and oncogenesis: signal
diversification through combinatorial ligand-
receptor interactions. FEBS Lett 1997, 410,
8386.
11 Yarden Y. The EGFR family and its ligands in
human cancer: signaling mechanisms and
therapeutic opportunities. Eur J Cancer 2001,
37, S38.
12 Baulida J, Kraus MH, Alimandi M, et al. All
ErbB receptors other than the epidermal
growth factor receptor are endocytosis im-
paired. J Biol Chem 1996, 271, 52515257.
13 Hendriks BS, Opresko LK, Wiley HS, et al.
Coregulation of epidermal growth factor re-
ceptor/human epidermal growth factor recep-
tor 2 (HER2) levels and locations: quantitative
analysis of HER2 overexpression effects. Can-
cer Res 2003, 63, 11301137.
14 Levkowitz G, Waterman H, Zamir E, et al. c-
Cbl/Sli-1 regulates endocytic sorting and ubi-
quitination of the epidermal growth factor re-
ceptor. Genes Dev 1998, 12, 36633674.
15 Weiner DB, Liu J, Greene MI. A point muta-
tion in the neu oncogene mimics ligand in-
duction of receptor aggregation. Nature 1989,
339, 230231.
References 145
16 Sliwkowski MX, Lofgren JA, Lewis GD, et al.
Nonclinical studies addressing the mechanism
of action of trastuzumab (Herceptin). Semin
Oncol 1999, 26 (suppl 12), 6070.
17 Slamon DJ, Godolphin W, Jones LA, et al.
Studies of the HER2/neu proto-oncogene in
human breast and ovarian cancer. Science
1989, 244, 707712.
18 Ross JS, Fletcher JA. The HER-2/neu onco-
gene in breast cancer: prognostic factor, pre-
dictive factor, and target for therapy. Stem Cells
1998, 16, 413-428.
19 Gusterson BA, Gelber RD, Goldhirsch KN, et
al. Prognostic importance of c-erbB-2 expres-
sion in breast cancer. J Clin Oncol 1992, 10,
10491056.
20 Hynes NE, Stern DF. The biology of erbB-2/
neu/HER-2 and its role in cancer. Biochim Bio-
phys Acta 1994, 1198, 165184.
21 King CR, Kraus MH, Aaronsen SA. Amplifica-
tion of a novel c-erbB-related gene in a human
mammary carcinoma. Science 1985, 229, 974
976.
22 Mnard S, Fortis S, Castiglioni F, et al. HER2
as a prognostic factor in breast cancer. Oncol-
ogy 2001, 61(suppl 2), 6772.
23 Press MF, Bernstein L, Thomas PA, et al.
HER-2/neu gene amplification characterized
by fluorescence in situ hybridization: poor
prognosis in node negative breast carcinomas.
J Clin Oncol 1997, 15, 28942904.
24 Slamon DJ, Clark GM, Wong SG, et al. Hu-
man breast cancer: correlation of relapse and
survival with amplification of the HER-2/neu
oncogene. Science 1987, 235, 177182.
25 Hanna W, Gelmon KA. Review of the litera-
ture on HER2/neu testing and the role of
HER2/neu as a prognostic and predictive fac-
tor in breast cancer update December 2000
to July 2001. Curr Oncol 2002, 9, S217.
26 Lohrisch C, Piccart M. HER2/neu as a predic-
tive factor in breast cancer. Clin Breast Cancer
2001, 2, 129135.
27 Piccart M, Lohrisch C, Di Leo A, Larsimont
D. The predictive value of HER2 in breast
cancer. Oncology 2001, 61 (suppl 2), 7382.
28 Piccart MJ, Di Leo A, Hamilton A. HER2, a
predictive factor ready to use in the daily man-
agement of breast cancer patients? Eur J Can-
cer 2000, 36, 17551761.
29 Baselga J, Norton L, Albanell J, et al. Recom-
binant humanized anti-HER2 antibody (Her-
ceptin
TM
) enhances the antitumor activity of
paclitaxel and doxorubicin against HER2/neu
overexpressing human breast cancer xeno-
grafts. Cancer Res 1998, 58, 28252831.
30 Pegram M, Hsu S, Lewis G, et al. Inhibitory
effects of combinations of HER-2/neu anti-
body and chemotherapeutic agents used for
treatment of human breast cancers. Oncogene
1999, 18, 22412251.
31 Pegram MD, Lipton A, Hayes DF, et al. Phase
II study of receptor-enhanced chemosensitivity
using recombinant humanized anti-
p185HER2/neu monoclonal antibody plus cis-
platin in patients with HER2/neu-overexpres-
sing metastatic breast cancer refractory to che-
motherapy treatment. J Clin Oncol 1998, 16,
26592671.
32 De Santes K, Slamon D, Andersen SK, et al.
Radiolabeled antibody targeting of the HER-2/
neu oncoprotein. Cancer Res 1992, 52, 1916
1923.
33 Drebin JA, Link VC, Stern DF, et al. Down-
modulation of an oncogene protein product
and reversion of the transformed phenotype
by monoclonal antibodies. Cell 1985, 41, 697
706.
34 Hudziak RM, Lewis GD, Winget M, et al.
p185HER2 monoclonal antibody has antipro-
liferative effects in vitro and sensitizes human
breast tumor cells to tumor necrosis factor.
Mol Cell Biol 1989, 9, 11651172.
35 Klapper LN, Waterman H, Sela M, et al. Tu-
mor-inhibitory antibodies to HER-2/ErbB-2
may act by recruiting c-Cbl and enhancing
ubiquitination of HER-2. Cancer Res 2000, 60,
33843388.
36 Sarup JC, Johnson RM, King KL, et al. Char-
acterization of an anti-p185HER2 monoclonal
antibody that stimulates receptor function and
inhibits tumor cell growth. Growth Reg 1991,
1, 7282.
37 Carter P, Presta L, Gorman CM, et al.
Humanization of an anti-p185HER2 antibody
for human cancer therapy. Proc Natl Acad Sci
USA 1992, 89, 42854289.
38 Clynes RA, Towers TL, Presta LG, et al. Inhi-
bitory Fc receptors modulate in vivo cytotoxic-
ity against tumor targets. Nat Med 2000, 6,
443446.
39 Gennari R, Menard S, Fagnoni F, et al. Pilot
study of the mechanism of action of preopera-
tive trastuzumab in patients with primary
operable breast tumors overexpressing HER2.
Clin Cancer Res 2004, 10, 56505655.

40 Hancock MC, Langton BC, Chan T, et al. A
monoclonal antibody against the c-erbB-2 pro-
tein enhances the cytotoxicity of cis-diaminedi-
chloroplatinum against human breast and
ovarian tumor cell lines. Cancer Res 1991, 51,
45754580.
41 Burgess AW, Cho HS, Eigenbrot C, et al. An
open-and-shut case? Recent insights into the
activation of EGF/ErbB receptors. Mol Cell
2003, 12, 541552.
42 Molina MA, Codony-Servat J, Albanell J, et al.
Trastuzumab (Herceptin), a humanized anti-
HER2 receptor monoclonal antibody, inhibits
basal and activated HER2 ectodomain cleavage
in breast cancer cells. Cancer Res 2001, 61,
47444749.
43 Holbro T, Beerli RR, Maurer F, et al. The
ErbB2/ErbB3 heterodimer functions as an on-
cogenic unit: ErbB2 requires ErbB3 to drive
breast tumor cell proliferation. Proc Natl Acad
Sci USA 2003, 100, 89338938.
44 McKenzie SJ, Marks PJ, Lam T, et al. Genera-
tion and characterization of monoclonal anti-
bodies specific for the human neu oncogene
product, p185. Oncogene 1989, 4, 543548.
45 Stancovski I, Hurwitz E, Leitner O, et al. Me-
chanistic aspects of the opposing effects of
monoclonal antibodies to the ERBB2 receptor
on tumor growth. Proc Natl Acad Sci USA
1991, 88, 86918695.
46 Baselga J, Tripathy D, Mendelsohn J, et al.
Phase II study of weekly intravenous recombi-
nant humanized anti-p185HER2 monoclonal
antibody in patients with HER2/neu-overex-
pressing metastatic breast cancer. J Clin Oncol
1996, 14, 73744.
47 Baselga J, Mendelsohn J. The epidermal
growth factor receptor as a target for therapy
in breast carcinoma. Breast Cancer Res Treat
1994, 1, 12738.
48 Tokuda Y, Ohnishi Y, Shimamura K, et al.
In vitro and in vivo anti-tumor effects of a
humanised monoclonal antibody against
c-erbB-2 product. Br J Cancer 1996, 73, 1362
1365.
49 Pegram MD, Konecny GE, OCallaghan C, et
al. Rational combinations of trastuzumab with
chemotherapeutic drugs used in the treatment
of breast cancer. J Natl Cancer Inst 2004, 96,
739749.
50 Pietras RJ, Pegram MD, Finn RS, et al. Re-
mission of human breast cancer xenografts on
therapy with humanized monoclonal antibody
to HER-2 receptor and DNA-reactive drugs.
Oncogene 1998, 17, 22352249.
51 Cobleigh MA, Vogel CL, Tripathy D, et al.
Multinational study of the efficacy and safety
of humanized anti-HER2 monoclonal anti-
body in women who have HER2-overexpres-
sing metastatic breast cancer that has pro-
gressed after chemotherapy for metastatic dis-
ease. J Clin Oncol 1999, 17, 26392648.
52 Slamon DJ, Leyland-Jones B, Shak S, et al.
Use of chemotherapy plus a monoclonal anti-
body against HER2 for metastatic breast can-
cer that overexpresses HER2. N Engl J Med
2001, 344, 783792.
53 Baselga J. Herceptin alone or in combination
with chemotherapy in the treatment of HER2-
positive metastatic breast cancer: pivotal trials.
Oncology 2001, 61, 1421.
54 Naber SP, Tsutsumi Y, Yin S, et al. Strategies
for the analysis of oncogene overexpression.
Studies of the neu oncogene in breast carcino-
ma. Am J Clin Pathol 1990, 94, 125136.
55 Kallioniemi OP, Kallioniemi A, Kurisu W, et
al. ERBB2 amplification in breast cancer ana-
lyzed by fluorescence in situ hybridization.
Proc Natl Acad Sci USA. 1992, 89, 53215335.
56 Pauletti G, Godolphin W, Press MF, Slamon
DJ. Detection and quantitation of HER-2/neu
gene amplification in human breast cancer ar-
chival material using fluorescence in situ hy-
bridization. Oncogene 1996, 13, 6372.
57 Bilous M, Dowsett M, Hanna W, et al. Cur-
rent perspectives on HER2 testing: a review of
national testing guidelines. Mod Pathol 2003,
16, 173182.
58 Hanna W and Kwok K. Can CISH be an alter-
native to FISH in the HER2/neu testing algo-
rithm? The Breast 2005, 14 (suppl 1), S17
(abstr P10).
59 Mass RD, Press M, Anderson S, et al. Im-
proved survival benefit from Herceptin (trastu-
zumab) in patients selected by fluorescence in
situ hybridization (FISH). Proc Am Soc Clin
Oncol 2001, 20, 22a (abstr 85).
60 Smith IE. Efficacy and safety of Herceptin in
women with metastatic breast cancer: results
from pivotal clinical studies. Anticancer Drugs
2001, 12 (suppl 4), S310.
61 Vogel CL, Cobleigh M, Tripathy D, et al. Supe-
rior outcomes with Herceptin (trastuzumab)
(H) in fluorescence in situ hybridization
(FISH)-selected patients. Proc Am Soc Clin
Oncol 2001, 20, 22a (abstr 86; poster update).
References 147
62 Mass RD, Sanders C, Charlene K, et al. The
concordance between the clinical trials assay
(CTA) and fluorescence in situ hybridization
(FISH) in the Herceptin pivotal trials. Proc
Am Soc Clin Oncol 2000, 19, 75a (abstr 291).
63 Bast RC Jr, Ravdin P, Hayes DF, et al. 2000
update of recommendations for the use of tu-
mor markers in breast and colorectal cancer:
clinical practice guidelines of the American
Society of Clinical Oncology. J Clin Oncol
2001, 19, 18651878.
64 Vogel CL, Cobleigh MA, Tripathy D, et al. Effi-
cacy and safety of trastuzumab as a single
agent in first-line treatment of HER2-overex-
pressing metastatic breast cancer. J Clin Oncol
2002, 20, 719726.
65 Di Leo A, Piccart MJ. Paclitaxel activity, dose,
and schedule: data from phase III trials in
metastatic breast cancer. Semin Oncol 1999, 26
(suppl 8), 2732.
66 Marty M, Cognetti F, Maraninchi D, et al. Ef-
ficacy and safety of trastuzumab combined
with docetaxel in patients with HER2-positive
metastatic breast cancer administered as first-
line treatment: results of a randomized phase
II trial by the M77001 study group. J Clin On-
col 2005, 23, published ahead of print.
67 Norris B, Pritchard K, James J, et al. Phase III
comparative study of vinorelbine combined
with doxorubicin versus doxorubicin alone in
disseminated metastatic/recurrent breast can-
cer: National Cancer Institute of Canada Clini-
cal Trials Group Study MA8. J Clin Oncol
2000, 18, 23852394.
68 OShaughnessy JA. The evolving role of cape-
citabine in breast cancer. Clin Breast Cancer
2003, 4 (suppl 1), S2025.
69 Parideans R, Biganzoli L, Bruning P, et al. Pa-
clitaxel versus doxorubicin as first-line single-
agent chemotherapy for metastatic breast can-
cer: a European organization for research and
treatment of cancer randomized study with
cross-over. J Clin Oncol 2000, 18, 724733.
70 Sledge G, Neuberg D, Ingle J, et al. Phase III
trial of doxorubicin, paclitaxel, and the combi-
nation of doxorubicin and paclitaxel as front-
line chemotherapy for metastatic breast can-
cer: an Intergroup Trial (E1193). J Clin Oncol
2003, 21, 588592.
71 Cook-Bruns N. Retrospective analysis of the
safety of Herceptin
immunotherapy in meta-
static breast cancer. Oncology 2001, 61 (suppl
2), 5866.
72 Tan-Chiu E, Piccart M. Moving forward with
Herceptin
in the adjuvant setting. Oncology

2002, 63 (suppl 1), 5763.
73 Buzdar AU, Ibrahim NK, Francis D, et al. Sig-
nificantly higher pathologic complete remis-
sion rate after neoadjuvant therapy with tras-
tuzumab, paclitaxel, and epirubicin chemo-
therapy: Results of a randomized trial in hu-
man epidermal growth factor receptor 2-
positive operable breast cancer. J Clin Oncol
2005, 23 (published ahead of print).
74 Anderson S, Reddy JC, Rai S, et al. Concor-
dance between central and local lab IHC and
FISH HER2 testing in a community-based
trial of first-line trastuzumab plus a taxane in
HER2+ metastatic breast cancer (MBC).
ASCO annual meeting proceedings. J Clin
Oncol 2004, 22 (abstract 9580) 14S.
75 Yaziji H, Barry TS, Goldstein LC, et al. Discre-
pancies between immunohistochemistry
(IHC) and fluorescence in situ hybridisation
(FISH) for HER-2 testing in breast cancer: the
role of quality assurance program on a cohort
of 3564 patients. ASCO annual meeting pro-
ceedings. J Clin Oncol 2004, 22 (abstract 679)
14S.
76 Yaziji H, Goldstein LC, Barry TS, et al. HER-2
testing in breast cancer using parallel tissue-
based methods. JAMA 2004, 291, 19721977.
77 Dowsett M, Bartlett J, Ellis IO, et al. Correla-
tion between immunohistochemistry (Hercep-
Test) and fluorescence in situ hybridization
(FISH) for HER-2 in 426 breast carcinomas
from 37 centres. J Pathol 2003, 199, 418423.
78 Hofmann M, Gaiser T, Gross C et al. Predic-
tive value of Herceptest IHC and PathVysion
FISH data: analysis from a Herceptin phase II
monotherapy study. Proc Am Soc Clin Oncol
2003, 22 (abstract 740) 185.
79 Vincent-Salomon A, MacGrogan G, Couturier
J, et al. Calibration of immunohistochemistry
for assessment of HER2 in breast cancer: re-
sults of the French multicentre GEFPICS
study. Histopathology 2003, 42, 337347.
80 Cianciulli AM, Botti C, Coletta AM, et al. Con-
tribution of fluorescence in situ hybridization
to immunohistochemistry for the evaluation
of HER-2 in breast cancer. Cancer Genet Cyto-
genet 2002, 133, 6671.
81 McCormick SR, Lillemore TJ, Beneke J, et al.
HER2 assessment by immunohistochemical
analysis and fluorescence in situ hybridization:
comparison of HercepTest and PathVysion

commercial assays. Am J Clin Pathol 2002,
117, 935943.
82 Paik S, Bryant J, Tan-Chiu E, et al. Real-world
performance of HER2 testing national surgi-
cal adjuvant breast and bowel project experi-
ence. J Natl Cancer Inst 2002, 94, 852854.
83 Roche PC, Suman VJ, Jenkins RB, et al. Con-
cordance between local and central laboratory
HER2 testing in the breast intergroup trial
N9831. J Natl Cancer Inst 2002, 94, 855857.
84 Birner P, Oberhuber G, Stani J, and the Aus-
trian Breast and Colorectal Cancer Study
Group. Evaluation of the United States Food
and Drug Administration-approved scoring
and test system of HER-2 protein expression
in breast cancer. Clin Cancer Res 2001, 7,
16691675.
85 Lebeau A, Deimling D, Kaltz C, et al. HER-2/
neu analysis in archival tissue samples of hu-
man breast cancer: comparison of immuno-
histochemistry and fluorescence in situ hybri-
dization. J Clin Oncol 2001, 19, 354363.
86 Tsuda H, Akiyama F, Terasaki H, et al. Detec-
tion of HER-2/neu (c-erb B-2) DNA amplifica-
tion in primary breast carcinoma. Interobserv-
er reproducibility and correlation with immu-
nohistochemical HER-2 overexpression. Can-
cer 2001, 92, 29652974.
87 Tubbs RR, Pettay JD, Roche PC, et al. Discre-
pancies in clinical laboratory testing of eligi-
bility for trastuzumab therapy: apparent im-
munohistochemical false-positives do not get
the message. J Clin Oncol 2001, 19, 2714
2721.
88 Hoang MP, Sahin AA, Ordonez NG, Sneige N.
HER-2/neu gene amplification compared with
HER-2/neu protein overexpression and inter-
observer reproducibility in invasive breast car-
cinoma. Am J Clin Pathol 2000, 113, 852859.
89 Kakar S, Puangsuvan N, Stevens JM, et al. Her-
2/neu assessment in breast cancer by immu-
nohistochemistry and fluorescence in situ hy-
bridization: comparison of results and correla-
tion with survival. Mol Diagn 2000, 5, 199207.
90 Ridolfi RL, Jamehdor MR, Arber JM. HER-2/
neu testing in breast carcinoma: a combined
immunohistochemical and fluorescence in situ
hybridization approach. Mod Pathol 2000, 13,
866873.
91 Tanner M, Gancberg D, Di Leo A, et al. Chro-
mogenic in situ hybridization: a practical alter-
native for fluorescence in situ hybridization to
detect HER-2/neu oncogene amplification in
archival breast cancer samples. Am J Pathol
2000, 157, 14671472.
92 Bilous M, Morey A, Armes J et al. The inter-
laboratory reproducibility of CISH testing
for HER2 and correlation with IHC and
FISH results. Eur J Cancer 2004, 2 (Suppl. 3,
abstract 160): 100.
93 Hofmann M, Gaiser T, Kneitz H et al. Com-
parison of chromogenic in situ hybridisation
(CISH) with FISH and IHC for assessment of
HER2 status and prediction of response to
therapy: analysis from a study of trastuzumab
monotherapy. Ann Oncol 2004, 15 (Suppl 3).
94 Peiro G, Mayr D, Hillemanns P, et al. Analy-
sis of HER-2/neu amplification in endome-
trial carcinoma by chromogenic in situ hybri-
dization. Correlation with fluorescence in
situ hybridization, HER-2/neu, p53 and Ki-
67 protein expression, and outcome. Mod
Pathol 2004, 17, 111.
95 Arnould L, Denoux Y, MacGrogan G, et al.
Agreement between chromogenic in situ hy-
bridisation (CISH) and FISH in the determi-
nation of HER2 status in breast cancer. Br J
Cancer 2003, 88, 15871591.
96 Kounelis S, Kapranos N, Malamos et al.
Evaluation of HER-2/neu gene status in
breast cancer by chromogenic in situ hybridi-
sation: comparison with immunohistochem-
istry. USCAP 2003 (abstract 154).
97 Muller V, Thomssen C, Karakas C, et al.
Quantitative assessment of HER-2/neu pro-
tein concentration in breast cancer by en-
zyme-linked immunosorbent assay. Int J Biol
Markers 2003, 18, 1320.
98 Sapino A, Coccorullo Z, Cassoni P, et al.
Which breast carcinomas need HER-2/neu-
gene study after immunohistochemical anal-
ysis? Results of combined use of antibodies
against different c-erbB2 protein domains.
Histopathology 2003, 43, 354362.
99 Van de Vijver MJ, Rueschoff J, Penault-Llor-
ca F, et al. Chromogenic in situ hybridisa-
tion (CISH) compared with FISH and IHC
for detection of HER2 gene amplification:
an international validation ring study. Breast
Cancer Res Treat 2003 (abstract).
100 Wixom CR, Albers EA, Weidner N, et al.
HER2 amplification: correlation of chromo-
genic in situ hybridisation (CISH) with im-
munohistochemistry (IHC) and fluorescence
in situ hybridisation (FISH). USCAP 2003
(abstract 219).
References 149
101 Dandachi N, Dietze O, Hauser-Kronberger
C. Chromogenic in situ hybridization: a nov-
el approach to a practical and sensitive
method for the detection of HER2 oncogene
in archival human breast carcinoma. Lab In-
vest 2002, 82, 10071014.
102 Zhao J, Wu R, Au A, et al. Determination of
HER2 gene amplification by chromogenic in
situ hybridization (CISH) in archival breast
carcinoma. Mod Pathol 2002, 15, 657665.
103 Tanner M, Jarvinen P, Isola J. Amplification
of HER-2/neu and topoisomerase II alpha in
primary and metastatic breast cancer. Cancer
Res 2001, 61, 53455348.
104 Park K, Kim J, Lim S, et al. Comparing fluo-
rescence in situ hybridization and chromo-
genic in situ hybridization methods to deter-
mine the HER2/neu status in primary
breast carcinoma using tissue microarray.
Mod Pathol 2003, 16, 937943.

Abstract
Therapeutic angiogenesis is a novel treat-
ment strategy for patients with chronic
myocardial ischaemia (stable angina). The
goal of therapeutic angiogenesis is to stim-
ulate the formation of collateral vessels to
restore blood flow to ischemic regions of
the heart. It was hypothesized that local
production of an angiogenic growth factor
would result in increased collateral forma-
tion, myocardial perfusion, and improved
contractile function. The rationale for
choosing viral gene therapy for therapeutic
angiogenesis is reviewed in the context of
other therapeutic options (see also Part I,
Chapters 12 and 13). The choice of adeno-
virus as the delivery vehicle, the options
for virus delivery, and selection of FGF-4
as the therapeutic gene to be delivered is
discussed. Preclinical studies in pigs with
myocardial ischemia showed that intracor-
onary injection of Ad5FGF-4 is well toler-
ated and effective. Intracoronary infusion
of Ad5FGF-4 increased myocardial perfu-
sion to control levels and restored normal
heart wall motion. The effect, after single
injection, lasted for at least 3 months. GLP
toxicology and biodistribution studies re-
vealed no product-related adverse effects.
Based on these data, development (con-
struction, scale-up, manufacturing, purifi-
cation, characterization) of this unique bio-
pharmaceutical and clinical trials were in-
itiated. With regard to product develop-
ment, the construction of the adenoviral
vector is discussed, in addition to the cell
culture optimization necessary for scale-up
and clinical manufacturing. Several novel
purification and engineering steps are
described. For product characterization,
the viral vector was first defined at the lev-
el of the Virus Bank, by complete double-
stranded genome sequencing. Nucleic
151
ISBN: 3-527-31184-X
6
Adenovirus-based Gene Therapy: Therapeutic Angiogenesis
with Adenovirus 5 Fibroblast Growth Factor-4 (Ad5FGF-4)
in Patients with Chronic Myocardial Ischemia
Elisabeth Lehmberg, Mei Tan, Jacob Kung, Bruce Mann, Erno Pungor Jr., and Gabor M. Rubanyi
siRNA the Magic Bullet and Other Gene Therapeutical Approaches
acid- and protein-based testing for the
virus, and PCR-based confirmation of an
intact transgene with appropriate vector
junction sequences were applied. This was
followed by infection of test cell lines to
demonstrate viral infectivity, growth factor
production and its biological activity
(growth promotion). Clinical studies were
supported by adenoviral-specific bioassays
(to assess virus shedding, stability, and bio-
distribution), detection of FGF-4 protein in
patient blood, and the measurement of
serum antibodies titers (both total and
neutralizing) against adenovirus. To date,
two clinical trials have been completed
employing intracoronary gene transfer of
Ad5FGF-4. The Angiogenic GENe Therapy
(AGENT) trial was the first multicenter,
randomized, double-blind, placebo-con-
trolled, dose-escalation trial of intracoro-
nary gene therapy on 79 chronic stable an-
gina patients. A subsequent trial (AGENT-
2) of intracoronary Ad5FGF-4 at a single
dose of 10
10
virus particle was undertaken
to study the effect on myocardial perfusion
on 52 patients with chronic stable angina.
Safety, efficacy and pharmacokinetic data
of this modern biopharmaceutical obtained
in the trials are briefly reviewed.
Abbreviations
Ad5 adenovirus serotype 5
CABG coronary artery bypass grafting
cGMP current good manufacturing
practice
CMV cytomegalovirus
CZE capillary zone electrophoresis
DLS dynamic light scattering
EPCs endothelial progenitor cells
ETT exercise treadmill test
FFF field flow fractionation
GLP good laboratory practice
HCP host cell proteins
HGF hepatocyte growth factor
HHFS hydrohermatic feed system
HIF hypoxia-inducible factor
MALS multiangle light scattering
MCB master cell bank
MVB master virus bank
MWCB manufacturers working cell bank
MWCO molecular weight cut-off
NS not significant
ORF open reading frame
PBS phosphate-buffered saline
PDGF platelet-derived growth factor
PDS perfusion defect size
RPDS reversible perfusion defect size
SPECT single-photon emission computed
tomography
VEGF vascular endothelial growth factor
6.1
Introduction
Patients with severe coronary artery dis-
ease and subsequent chronic myocardial
ischemia (stable angina pectoris) may not
have satisfactory outcomes, despite maxi-
mal medical therapy and multiple revascu-
larization procedures. The consequences
of inadequate therapy may include myocar-
dial infarction and sudden death, or living
with anginal pain and physical activity lim-
itations. Current coronary revascularization
therapies (e.g., coronary angioplasty or by-
pass graft surgery) act by restoring blood
flow through preexisting coronary vessels
[1]. These procedures do not promote the
growth of new collateral vessels (angiogen-
esis or arteriogenesis). Therefore, there is
a major medical need to identify novel
approaches that facilitate the growth of
new collateral vessels in patients with
chronic myocardial ischemia, termed ther-
apeutic angiogenesis [2].
6 Adenovirus-based Gene Therapy 152
6.2
Therapeutic Angiogenesis
and the Importance of Collateral Vessels
Therapeutic angiogenesis is based on the
concept that myocardial ischemia can be
alleviated by stimulating coronary collat-
eral vessel development from the existing
microvasculature. New blood vessel growth
may be achieved by the local delivery of
angiogenic growth factor into the myocar-
dium, either through protein or gene ther-
apy. Protein therapy involves the delivery
of a growth factor directly into the isch-
emic myocardium, whereas gene therapy
involves the delivery of DNA coding for a
growth factor. Gene therapy has the poten-
tial advantage of enabling expression of
the angiogenic factor over a period long
enough and at a sufficient local concentra-
tion to stimulate effective angiogenesis
from a single administration [3].
Collateral arteries protect the myocar-
dium from ischemia by forming natural
micro-bypasses that provide alternative
routes for myocardial blood flow [4]. Pa-
tients with extensive collateral formation
are less likely to experience myocardial in-
farction, ST-segment depression, abnormal
exercise tests, or heart failure [5]. Most im-
portantly, survival after 10 years is higher
with well-formed collaterals [5]. Collateral
vessel growth following the administration
of angiogenic gene or growth factor pro-
teins has already been demonstrated in ani-
mal models of coronary artery disease [68].
This chapter will review the rationale for
the use of Ad5FGF-4 for therapeutic angio-
genesis, the composition of the gene thera-
py product, the manufacturing and analy-
sis of the product, preclinical animal data
and initial clinical results in the largest
population of patients to date who have
undergone intracoronary administration of
gene therapy for myocardial ischemia.
6.3
Designing an Intervention Suitable
for Therapeutic Angiogenesis
The treatment strategies employed in ther-
apeutic angiogenesis for coronary artery
disease patients have included intramyo-
cardial injection, both epicardial and endo-
cardial, and intravascular administration;
both intracoronary and via saphenous vein
graft conduits [914]. A variety of thera-
peutic genes have been employed in cardi-
ovascular gene therapy trials, including
vascular endothelial growth factor (VEGF),
fibroblast growth factor (FGF), and plate-
let-derived growth factor (PDGF), with vec-
tors divided between adenoviral and non-
viral vectors including naked plasmid
DNA [3, 15]. Angiogenic factor gene up-
take and resulting angiogenesis have al-
ready been demonstrated in human myo-
cardium [16, 17]. However, optimal gene
therapy treatment techniques and thera-
peutic implications of this novel approach
to myocardial ischemia still remain un-
clear. Therefore, several choices had to be
made whether to apply protein or gene de-
livery and about the best growth factor,
vector and vector delivery technology be-
fore initiating product development.
6.3.1
Protein versus Gene Delivery
Fibroblast growth factor protein has been
infused into the coronary arteries, or in-
jected directly into the heart muscle. Be-
cause of poor FGF protein pharmacoki-
netics (i.e., low retention by the heart) [18],
stimulation of myocardial angiogenesis re-
quires high doses of FGF protein intracor-
onary infusion over a long period of time.
However, hypotension and tachycardia
may result from circulating concentrations
of 48 lg kg
1
or more of FGF-2. Nausea
and leukocytosis have also been observed
after intracoronary administration of
36 lg kg
1
FGF-2 [19]. In the FIRST trial,
intracoronary doses of recombinant
(r)FGF-2 protein were reduced below the
toxic level, but no convincing efficacy was
observed. Exercise tolerance was not in-
creased at any time point, but hypotension
was observed more often in the rFGF-2-
treated patients [20].
Injection of a sustained-release prepara-
tion of rFGF-2 directly into the heart mus-
cle during coronary artery bypass grafting
(CABG) has been proposed as an alterna-
tive method for improving FGF protein lo-
calization and retention. While this limits
the eligible patient population to those un-
dergoing CABG, a small study (n=8 per
group) showed that perfusion defects were
smaller in patients treated with rFGF-2
and followed for ~ 32 months [21].
By delivering the gene coding for a pro-
tein instead of the protein itself, high
amounts of the protein can be locally pro-
duced continuously from the DNA over an
extended period of time. This approach re-
duces systemic effects, increases the dura-
tion of therapy, and avoids the costs of
large-scale protein production. Indeed, in
2000, 17% of US trials of gene transfer in-
volved cardiovascular gene therapy, pre-
dominantly directed at therapeutic angio-
genesis [22]. Therefore, a gene therapy
approach was chosen to deliver the angio-
genic growth factor to the myocardium.
6.3.2
Growth Factor
The formation of new blood vessels is a
complex process in which a wide variety of
genes participate by activating endothelial
cells, recruiting monocytes, degrading the
existing extracellular matrix, stimulating
migration and division of smooth muscle
cells, or stabilizing nascent vessels [23, 24].
A partial list of pro-angiogenic molecules
involved in this process includes VEGFs,
FGFs, hepatocyte growth factor (HGF),
and hypoxia-inducible factor (HIF)-1a [3,
15] (see also Part I, Chapter 10 and Part V,
Chapters 4 and 6).
Most investigational therapies that have
progressed to clinical trials to date use a
form of either VEGF or FGF to stimulate
formation of new blood vessels [25]. Nu-
merous experimental studies have demon-
strated differences between the physiologi-
cal functions of VEGF and those of FGF.
While VEGF is a survival factor for endo-
thelial cells and stimulates the release of
endothelial progenitor cells (EPCs) from
the bone marrow, the resultant vessels can
be permeable or leaky [26], or may re-
gress when VEGF levels decrease [25]. In
contrast, FGF stimulates VEGF production
[27], nitric oxide release [28] and HGF ex-
pression [29], and also increases the den-
sity of PDGF receptors [23]. Additionally,
FGF stimulates the proliferation of endo-
thelial cells, smooth muscle cells, and fi-
broblasts [25]. The increased cellularity of
the FGF-stimulated vessels may promote
stability and produce more mature vessels.
FGF, but not VEGF, was shown to partici-
pate in arteriogenesis, a key biological pro-
cess of collateral formation [30]. The re-
sults of recent studies also showed that, in
addition to therapeutic angiogenesis, FGF
is also involved in cardioprotection [31],
which may contribute to its therapeutic
benefits in patients with chronic myocar-
dial ischemia. These were the main rea-
sons for choosing FGF-4 as the growth
factor for our angiogenic gene therapy
product.
6.3.3
Gene Therapy Vector
A variety of vectors for delivering thera-
peutic genes are currently available, rang-
ing from naked plasmid DNA to various
viral vectors. Gene transfer with plasmid
DNA is generally inefficient and produces
low levels of protein. Transfer efficiency
can be increased by addition of various
liposomes, carrier compounds, or using
high-volume/high-pressure intravascular
injection (see also Part I, Chapter 7 and
Part VI, Chapters 1, 3, 6, and 7). Alterna-
tively, genetically modified viruses have
been created that transfer DNA of interest
but are not capable of replication. Modi-
fied adenovirus, adeno-associated virus,
herpes simplex virus, lentivirus, retrovirus,
and Sendai virus are some of the available
options [25]. To date, most clinical trials in
gene therapy have used a retrovirus, DNA-
liposome combination, or adenovirus [3].
Adenovirus lacking its E1 region is a
highly efficient vector for cardiovascular
gene transfer. It has lost its ability to repli-
cate, but it is capable of high-level transfer
of DNA to non-dividing, terminally differ-
entiated cells such as those found in the
heart. Among the more than 50 known
serotypes, Adenovirus Serotype 5 (Ad5) is
by far the most widely used to date (for ex-
ample, thousands of healthy young sub-
jects were immunized with Ad5 in the US
during the 1950s). After delivery by Ad5,
the DNA does not incorporate into the ge-
nome and is gradually lost from the cells
over time. Although local inflammation
can occur at high concentrations of adeno-
virus and anti-adenoviral antibodies usual-
ly form, the lack of chromosomal integra-
tion and transient expression increase vec-
tor safety [3]. Furthermore, high concentra-
tions of adenovirus can be produced
reproducibly and are stable in storage,
which is practicable for routine use. Thus,
an E1-deleted, replication-incompetent ade-
novirus of serotype 5 (Ad5) was chosen as
the vector for transferring the gene of
FGF-4 to the myocardium.
6.3.4
Delivery Method
Numerous surgical or catheter-based tech-
niques have been developed that permit
access to the heart and can be adapted for
gene transfer. The major approaches used
for myocardial gene transfer are epicardial,
endocardial, and intracoronary delivery. Ac-
cessing the hearts surface for epicardial
techniques requires a thoracotomy, which
may be performed during CABG or re-
quire a separate surgical procedure. This
is an invasive procedure that may result in
complications. However, the vector can be
injected specifically near ischemic areas.
In contrast, percutaneous endocardial in-
jection is appealing because it is easier for
the patient, though finding the ischemic
sites can be challenging. Currently, electro-
mechanical (NOGA) mapping is used to
localize ischemic areas, but this procedure
requires specialized catheters, instrumen-
tation, and skills [32]. Furthermore, direct
injection may also cause complications
and restrict vector distribution to a few
millimeters around the injection site with-
in the myocardium.
Other possible modes of administering
adenovirus include intracoronary infusion
[8, 13], infusion during stenting [33] or retro-
grade infusion into the coronary veins [34].
Intracoronary infusion adapts the tech-
niques used in angiography to infuse the
vector throughout the coronary vascular
tree. This procedure, which is used rou-
tinely in catheterization laboratories by in-
terventional cardiologists, accesses all of
the coronary arteries with less risk of com-
plications than a direct myocardial injec-
tion. A concern is that not all of the ade-
novirus may be taken up by the heart. If
adenovirus is systemically disseminated,
some toxicity may result depending on the
dose injected [13]. However, a survey of
clinical studies on over 2000 patients
treated with replication-incompetent ade-
noviral vectors has not yet revealed any
significant short- or long-term safety con-
cerns [15, 22]. For the reasons listed above,
we have chosen percutaneous intracoro-
nary infusion for Ad5FGF-4 delivery.
6.4
Production and Characterization
of the Ad5FGF-4 Vector
Converting a small-scale production
scheme for Ad5FGF-4 drug product into a
GMP-compliant process that is scalable for
clinical and commercial manufacture re-
quired substantial process and analytical
optimizations that often would not be dis-
cussed in scientific journals. The intent of
this section is to share some of these ac-
complishments and to show how each
contributed to the successful advancement
of this gene therapy project.
6.4.1
Construction and Isolation of Ad5FGF-4
The construction of the recombinant ade-
novirus expressing human FGF-4 required
three components: the FGF-4 transgene; a
plasmid shuttle vector to carry the trans-
gene as well as 5' Ad5 sequences; and a
second plasmid carrying the bulk of the
Ad5 genome.
6.4.1.1 Transgene
The full-length cDNA for human FGF-4
was isolated from a cDNA library, which
was constructed from mRNA of Kaposis
sarcoma DNA-transformed NIH3T3 cells
[35]. The cDNA that encodes the FGF-4
peptide is approximately 1.2 kB, and the
FGF-4 protein has 206 amino acids, in-
cluding a 33-amino acid signal peptide at
the N-terminus.
6.4.1.2 Shuttle Vector
Plasmid pACCMVpLpASR() contains a
cytomegalovirus (CMV) promoter, a poly-
linker, and SV40 polyadenylation se-
quences flanked by partial human adeno-
virus-5 sequences [36]. The FGF-4 cDNA
was subcloned, as an EcoR1 fragment, into
this adenovirus shuttle vector at its single
EcoR1 site. For this construct, a clone was
selected in which the transcription of the
FGF-4 gene would occur in a 3'?5' orien-
tation with respect to the Ad5 sequences;
in other words, the CMV promoter was in-
terior or 3' to the transgene. This shuttle
vector is identified as pACSR/FGF-4.
6.4.1.3 Adenovirus Plasmid
Plasmid pJM17 contains required adeno-
virus-5 sequences, except that the E1 re-
gion is disrupted by the insertion of
pBR322 sequences [37]. A unique feature
of this plasmid is the presence of unex-
pected non-viral sequences in the E3 re-
gion [38] that will be mentioned later in
the context of viral vector identity testing.
6.4.1.4 Viral Vector Generation
In order to generate the Ad5FGF-4 virus,
plasmids pACSR/FGF-4 and pJM17 were
then co-transfected into HEK 293 cells
using a calcium phosphate method. The
cells were overlaid with nutrient agarose.
Homologous recombination between the
vectors created E1-deleted, FGF-4 gene-
containing vector genomes capable of rep-
lication in HEK 293 cells. The predicted
genome structure is shown in Fig. 6.1.
6.4.2
Virus Bank
The replicating virus gave rise to virus pla-
ques that were picked 1012 days later. Six
clones were isolated from one round of
plaque purification and screened for pro-
tein expression. One clone was selected,
amplified on HEK 293 cells, purified by
cesium chloride ultracentrifugation, and
used for preparation of a Master Virus
Bank (MVB).
A purified Ad5FGF-4 virus seed was
propagated through three consecutive
rounds of plaque purification. A final
plaque was then expanded and purified by
anion-exchange chromatography, sterilized
by filtration and stored at 708C108C.
This virus stock was checked for sterility
and absence of measurable replication
competent adenovirus (RCA). A MVB was
then created by one additional cycle of pro-
pagation in serum-free suspension culture
of HEK 293 cells, and this virus was puri-
fied, aliquoted, and frozen. The MVB was
tested and confirmed to be free of RCA as
well as adventitious agents. The MVB was
stored at 708C108C.
6.4.3
Cell Lines and Cell Banks
A scaleable manufacturing process, able to
meet clinical as well as commercial needs
for recombinant adenovirus, would only be
attractive if it could be performed without
using bovine serum in the culture medi-
um (for both cost and safety reasons, e.g.,
BSE). For these reasons, attachment and
serum-dependent HEK 293 cells were first
adapted to suspension culture by direct
transfer into shake flasks in a modified
Williams Essential medium containing
2% fetal bovine serum (Hyclone; gamma-
irradiated). The cultures were passed con-
tinually for over 6 weeks in the same envi-
ronment until the cells exhibited a consis-
tent growth pattern of less than 48 h dou-
bling time and greater than 90% viability.
These suspension-adapted cells were sub-
sequently adapted to serum-free medium
(modified IS293; Irvine Scientific) by grad-
ual weaning. The serum content of the
medium was reduced with each passage
until the cells were in completely serum-
free medium. The weaning process lasted
approximately 3 weeks, and the cells were
passed for 6 additional weeks in serum-
free medium before preparing an inter-
mediate bank [39]. Cells from this bank
were thawed and expanded and then used
to generate a Master Cell Bank (MCB)
which was tested for growth and virus pro-
duction kinetics and yields. In addition,
culture identification as human cells,
Fig. 6.1 Predicted genome structure for Ad5FGF-4.
growth on soft agarose to assess tumori-
genicity, confirmation of absence of endog-
enous and screening for adventitious
agents was performed (following current
cell line testing guidelines). Starting from
the MCB, a Manufacturers Working Cell
Bank (MWCB) was also prepared and then
similarly tested before routine use.
6.4.4
Manufacture and Purification
of Recombinant, E1-deleted,
Replication Incompetent Ad5 Virus
The Ad5FGF-4 vector was manufactured
under current Good Manufacturing Prac-
tices (cGMP) in a state-of-the-art, validated
facility, following Biosafety Level 2 prac-
tices.
For the preparation of the Ad5FGF-4
drug substance, cells from the MWCB and
virus from the MVB were employed. The
virus was propagated in suspension cul-
tures of HEK 293 cells using serum-free
medium, and the virus progeny purified
using a combination of anion-exchange
chromatography and ultra-filtration (UF)
steps as shown in the flow diagram
(Fig. 6.2). The purified bulk drug sub-
stance was stored at 708C.
6.4.4.1 Small-scale Virus Purification
Medium components and extracellular
(non-viral) contaminants were largely re-
moved from adenoviral-infected HEK 293
cells using three successive phosphate-buf-
fered saline (PBS) washes and centrifuga-
tion of the cells at the time of harvest. The
harvested cells were then frozen in a solu-
tion of PBS with 2% sucrose and stored at
708C, until purification. The virus purifi-
cation process was initiated with two addi-
tional cycles of freeze and thaw steps. Fol-
lowing the final thaw, the ruptured cell
suspension was centrifuged to remove cell
debris. Virus purification was established
with a protocol based on traditional pro-
tein purification techniques, column chro-
matography and ultrafiltration. The initial
column purification, adapted from a pre-
viously published method [40] utilized a
Fractogel DEAE column chromatography,
but did not use Benzonase (or any other
nuclease) for the removal of nucleic acid
contaminants. In our purification, separa-
tion from nucleic acid contaminants was
achieved by combining strategic peak col-
lection and utilization of a tangential flow
ultrafiltration step with a 100 kDa molecu-
lar weight cut-off (MWCO) pore-size mem-
brane. Both of these purification proce-
dures may be readily and linearly scaled-
up, and thus can support the large-scale
production of adenovirus for gene therapy
for both clinical and commercial needs.
This recovery scheme is represented in
Fig. 6.2.
6.4.4.2 Fractogel-DEAE Chromatography
The Fractogel DEAE step was initially de-
veloped based on small-scale columns
(1.5 mL to 100 mL). This step was scaled
for our routine clinical production lots at
500 mL. The clarified cell lysate superna-
tant was decanted and loaded onto a Frac-
Fig. 6.2 Process schematic for Ad5FGF-4 produc-
tion.
togel-DEAE column and eluted during a
34-min linear gradient using 300 mM to
600 mM NaCl in a 50 mM Tris-HCl, pH
7.5, buffer with 2% sucrose and 2 mM
MgCl
2
. The elution fractions were col-
lected with a fraction collector and pooled
using the chromatogram as a guide to
minimize the amount of nucleic acid ma-
terial in the pool (Fig. 6.3). In the initial
chromatography cycles the nucleic acid
material was collected and identified as
primarily host cell RNA. Subsequently, an
automated peak collection method was de-
veloped and implemented. The virus prod-
uct peak collection initiated at a predeter-
mined slope of the 260 nm UV signal and
terminated at a fixed 260 nm absorbance
level or the minimum 260 nm absorbance
level between the virus and nucleic acid
peaks, whichever came first.
6.4.4.3 Ultrafiltration
The ultrafiltration step was developed with
smaller-scale disposable devices (Centra-
sette) and then expanded to a larger scale
using a multiple cassette system (the Cen-
traprep); both devices were from Pall-Fil-
tron. The DEAE column eluate, containing
the adenovirus and some remaining nu-
cleic acid contaminants, was further puri-
fied with an ultrafiltration step using a
100 kDa MWCO membrane. Intact adeno-
virus (at 150 mDa) was retained while resi-
dual RNA and DNA fragments (viral and
host cell in origin with an average molecu-
lar mass below 50 kDa) passed through
into the filtrate. Before filtration, the
DEAE eluate pool was diluted 100-fold into
the final formulation buffer. The ultrafil-
tration step was performed until the di-
luted eluate pool had been concentrated
back to the original volume, thus achiev-
ing the dual goals of removing nucleic
acid contaminants and achieving buffer ex-
change. As a last step, the purified, con-
centrated bulk virus was diluted into the
final virus dose range, filled into 5 mL
glass vials, frozen, and stored at 708C un-
til shipment.
Fig. 6.3 Chromatogram (260-nm detection) of a
typical 100-mL (5-cm diameter) preparative Frac-
togel DEAE column using the BioCad Vision
Chromatography system operated at a linear flow
rate of 76 cm h
1
. Approximately 150 mL of clari-
fied cell lysate containing 1e14 viral particles
was loaded onto the column.
6.4.5
Manufacture Scale-up Improvements
The adenovirus harvest process consisted
of three general steps: concentration; lysis;
and clarification. In the small-scale pro-
cess, the intact adenovirus-infected cells in
culture medium were aliquoted and batch-
centrifuged to achieve a 30-fold concentra-
tion; the supernatant was manually re-
moved and replaced with a smaller volume
of the freeze buffer. Whilst sufficing for
harvest volumes of no more than 10 L,
such a manual process would prove to be
unwieldy on a larger manufacturing scale.
Since these harvests involved infectious
virus, special care and precautions con-
cerning containment and material man-
agement were necessary during process
scale-up. Efforts to develop a scaleable har-
vest method based upon tangential flow
filtration and continuous centrifugation (to
allow simultaneous concentration and lysis
of virus within the hollow fiber) were not
successful due to fouling and clogging of
the hollow-fiber membranes and a resul-
tant drop in virus yield. Performing cell ly-
sis by mechanical means within the
growth medium followed by the use of a
hollow-fiber unit for clarification of the ly-
sate was too variable and necessitated the
handling of unnecessarily large volumes of
virus-containing medium.
A successful harvest method was identi-
fied utilizing a continuous centrifugation
in a Westfalia CSC-6 disc-stack device with
a hydrohermatic feed system (HHFS). A
continuous centrifuge allows the discharge
of the pelleted cellular material to be
ejected at precise time intervals while con-
tinuously and separately discharging the
supernatant. This feed system reduces the
Fig. 6.4 Mean cell size of adenovirus-infected
cells resulting from different processing methodol-
ogies. Particle size was measured using the Coul-
ter Multisizer II to determine cell sizes within the
suspension by measurement of changes in impe-
dance as cells pass through an aperture. Cell cul-
ture samples were diluted 1: 10 in isotonic buf-
fered saline diluent prior to measurement.
lysis of cells during the filling of the cen-
trifuge bowl throughout operation. Scale-
up of the adenovirus harvest process using
a continuous centrifuge with a HHFS has
been shown efficiently to separate and
concentrate an infected cell pellet while
allowing the relatively easy exchange of
growth medium for freezing buffer with-
out open, manual manipulation. By em-
ploying this unit operation, the separate
cell lysis step could be eliminated due to
the significant disruption of the cell pellet
occurring exclusively in the product
stream upon ejection of the concentrated
cell pellet. Mean particle size measure-
ments from harvest samples revealed the
highly efficient nature of cell lysis from
this harvest process (Fig. 6.4), with addi-
tional improved viral yields compared to
repeated freezethaw cycles. This patented
process step combined separation of the
cell pellet from culture medium, medium/
buffer exchange, cell concentration, and
then cell lysis in a single unit operation
step.
The resulting lysate can be clarified and
the virus purified as described.
6.4.6
Ad5FGF-4 Formulation Enhancement
and Stability
During the analytical assay development of
an HPLC assay for adenoviral protein fin-
gerprinting, the observation was made that
the recovery of a given sample was de-
creasing over time. Such losses were inves-
tigated and confirmed by non-destructive
LC methods such as size-exclusion and
ion-exchange chromatography. It was
shown that the losses were, at least par-
tially, due to binding of virus to surfaces
such as the autosampler vials. Virus bind-
ing to surfaces could further mediate the
precipitation of virus aggregates, thus in-
creasing the losses. Mechanical mixing, al-
ternative container surfaces, or changes in
excipient concentration (e.g., PBS, sucrose,
MgCl
2
, pH, or the addition of small mole-
cules such as alcohols, Tris, guanidine,
mannitol, serum albumin, or chelators)
did not prevent virus loss. However, non-
ionic detergents (Brij 35, Tween 20, or
Tween 80) in a concentration ranging from
0.05% to 0.08% were able effectively to
prevent the loss of virus. For example, the
recovery of a virus sample without Tween
after 24 h was 80% (as compared to the
first injection made immediately after
sample preparation), whereas in the pres-
ence of 0.05% Tween 20 recovery after the
same time period was 96% (ion-exchange
chromatography). Subsequently, Tween 20
at a concentration of 0.05% was used as
the preferred excipient for all stored virus
samples.
Ongoing stability studies of Ad5FGF-4
product, using a virus infectivity assay
showed a temperature-dependent, gradual
decay in infectivity. At 70108C, a decay
rate of 1.6% per month was calculated
from the stability data of multiple lots
placed on the stability program. At ambi-
ent temperature, infectivity of the drug
product lots decreased at a rate of 2.8%
per hour. For this reason, handling of the
drug product in the clinic was limited to
6 h post thaw at ambient temperature. Pre-
liminary information suggested that virus
particle aggregation could explain the
apparent loss of infectivity, since in the
same experiments there was no loss ob-
served in viral DNA or viral structural pro-
tein, and there was no sign of viral disinte-
gration. The benefits of adding 0.05%
Tween 20 to Ad5FGF-4 (as noted above
for HPLC samples) significantly improved
the product stability as measured by infec-
tivity assay under these conditions. The
addition of Tween 20 stabilized the infec-
tivity of Ad5FGF-4 drug product stored at
70108C over at least a 4-year period.
Several analytical approaches were ap-
plied for the direct evaluation of virus ag-
gregation. Small Ad5FGF-4 aggregates
were successfully separated by capillary
zone electrophoresis (CZE) [41]. The ag-
gregates were observed as a series of peaks
eluting before the main Ad5FGF-4 peak.
Each major peak contained functional
virus (as shown by EPD assay), as well as
intact viral DNA determined by PCR. Mod-
erately sized aggregates, up to a diameter
of about 200300 nm, could be detected
using field flow fractionation (FFF) with
multiangle light scattering (MALS). The
FFF separates particles based on their dif-
fusivity (related to the hydrodynamic ra-
dius); a typical FFF separation is shown in
Fig. 6.5. The system was calibrated using
polystyrene particle standards. With aggre-
gation, the virus peak loses its symmetry
Fig. 6.5 Field flow fractionation (FFF) of aggre-
gated adenovirus sample. The figure shows the
mass of adenovirus (cumulative and non-cumula-
tive) as a function of particle diameter as deter-
mined by FFF using a Focus AF2000 device (Post-
nova Analytics) with a nominal 250-lm channel
thickness and 1 PBS mobile phase. Particle di-
ameter was calibrated using spherical polystyrene
particles as reference.
Fig. 6.6 Dynamic light scattering of virus samples.
The figure shows a non-aggregated (a) and aggre-
gated (b) adenovirus sample as measured with a
DynaPro MS/XTC device (Proterion Corporation)
using 1 PBS buffer diluent. The bars on the
graph represent the scattered dynamic signal in
each size range shown on the x-axis.
"
and the tailing end contains the aggre-
gates.
As the aggregated particles become larg-
er than 300 nm, other techniques become
necessary for their detection. One useful
option is that of dynamic light scattering
(DLS), which also provides size informa-
tion based on the diffusivity (and can be
calibrated with polystyrene beads). The
DLS analysis of a non-aggregated and an
aggregated virus preparation is shown in
Fig. 6.6.
6.4.7
Analytical Techniques needed for Ad5FGF-4
6.4.7.1 General Testing of Drug Substance
The crude viral harvest was tested for ad-
ventitious agents as needed to comply with
current regulatory guidelines. Purified
virus was tested to confirm the absence of
RCAs, residual host cell proteins and
DNA, endotoxin, and to quantify total and
infectious vector particles and thus, the in-
fectivity ratio. Bulk virus lots were released
after meeting specifications which in-
cluded one or more endotoxin unit mL
1
and infectivity ratios of >2% (infectious ti-
ter/total viral particles).
6.4.7.2 Chromatographic Assays
One of the main challenges in the devel-
opment of the adenovirus purification pro-
cess was the quantitation of the intact viral
particles in crude samples, such as the
clarified cell lysate, for the optimization of
viral recovery at each step. Non-aggre-
gated, purified viral samples were easily
quantified by the reverse-phase (RP)-HPLC
method developed by Lehmberg et al. [42].
The RP-HPLC assay quantitates the indi-
vidual structural proteins with UV absor-
bance at 214 nm as the basis for the mea-
surement of the adenovirus concentration.
Unfortunately, this assay could not be used
to calculate the recoveries or yields of our
purifications as it could not reliably quan-
titate crude samples, due to column foul-
ing caused by such samples. To solve this
problem, an anion exchange (AIEX)-HPLC
procedure was developed utilizing a tri-
methyl anion exchange (TMAE) resin. The
TMAE-HPLC was suitable for the analysis
of all in-process samples ranging from
clarified cell harvest to final formulated,
purified recombinant adenovirus test arti-
cles. This assay allowed us to determine
the virus recovery at each step of the en-
tire purification process, thus allowing de-
velopment of the production process. The
linear range for the TMAE assay method
was 110
9
to 6.510
10
particles per injec-
tion (typically a 100 lL volume), thus pro-
viding a dynamic range of nearly two or-
ders of magnitude. For a clarified (but
crude) cell harvest test, care must be taken
not to overload the column. Small injec-
tion volumes in the range of 10 to 20 lL
were sufficient in most cases. The concen-
tration determination of the virus is based
on the calibration of a virus working stan-
dard by an orthogonal method, such as
the RP-HPLC assay. This TMAE-HPLC
method was capable of baseline separation
of the intact virus from contaminating nu-
cleic acids, free hexon, and other impuri-
ties (Fig. 6.7).
However, this assay has a limit of quan-
titation of 10
10
particles mL
1
, and for
more dilute samples we sought other
quantitative methods for example, using
amplification with PCR, to assess virus
particle count. A quantitative real-time
(TaqMan) PCR procedure was developed
utilizing amplification of a portion of the
Ad5 hexon gene. When samples were
treated with 0.05% sodium dodecyl sulfate
(SDS) and standard PCR reaction condi-
tions applied, this assay was shown to
Fig. 6.7 TMAE analytical HPLC chromatogram,
showing 260 nm absorbance traces of an non-in-
fected cell harvest control (A) and adenovirus in-
fected cell harvest sample (B). Chromatograms
were generated using a HP 1100 system with a
Fractogel EMD Tentacle Ion-Exchanger resin in a
4.675 mm Peak column (TMAE; EM Science,
Gibbstown, NJ, USA). The equilibration and load-
ing buffer was 50 mM HEPES at pH 7.5 and was
eluted with a gradient of NaCl. Data analysis was
performed with Chemstation (Agilent Technolo-
gies).
have a quantitative range of 510
2
to
510
6
particles in an input volume of
5 lL, corresponding to 10
5
to 10
9
parti-
cles mL
1
in the original sample.
6.4.7.3 End-point Dilution (EPD) Assay
for Infectivity
In order to determine infectivity, we devel-
oped an end-point dilution assay, as de-
scribed previously [43]. The precision of
this assay has a standard deviation of 0.2
when viral titers are expressed in log
10
.
6.4.7.4 Replication-Competent Adenoviruses
(RCA)
Recombination between the viral E1 gene
carried by HEK 293 cells and the E1-de-
leted recombinant adenovirus can, and
does, occur at low frequency to generate
RCA, essentially wild-type Ad5 lacking the
FGF4 transgene. A highly sensitive assay
involving viral amplification and cytopathic
effect (CPE) readout has been described
[43, 44]. In our laboratory, the limit of de-
tection is one RCA in 3.210
12
virus parti-
cles. Virus banks and bulk Ad5FGF-4 virus
testing positive for RCA were not used in
clinical studies. It should be noted that
while other cell lines and viral backbones
have been engineered to make an RCA-
free system, it has also been shown that
non-homologous recombination events can
still occur and have generated odd viral
progeny, the clinical effects of which are
unknown [44, 45].
6.4.7.5 Host Cell Proteins by ELISA
Typically, biotechnology products have
been produced in cell lines not of human
origin. In such cases, residual host cell
proteins (HCP) are perceived as a safety
concern to the patient. In the case of
Ad5FGF-4 produced in HEK 293 cells it is
expected that the patient may respond to
viral proteins with an immune response,
but it is less clear that human HCP will
be immunogenic or pose a serious safety
concern. To be cautious, we sought to as-
sure removal of most if not all HEK
293 HCP during preparation of the
Ad5FGF-4 product. One approach to the
detection of such HCP is by ELISA, using
a polyclonal antibody reagent such as the
HEK 293 HCP detection kit offered by
Cygnus Technologies (Renchesler, MD,
USA). However, an inherent limitation in
this approach is that the antibodies (gener-
ated in goats) do not recognize all HEK
293 proteins due to conserved protein
structures between mammalian species.
When we measured an HEK 293 cell ly-
sate by both chemical means, using BCA
reagents (Pierce Chemical) and immuno-
assay, the ELISA was found to report only
1% of the total protein actually in the sam-
ple. Thus, many HCPs cannot be mea-
sured by using ELISAs. Furthermore, the
complex signal generated by multiple anti-
gens (expected in the sample) and multi-
ple antibodies (in the test reagents) were
sufficiently uncharacterized that a simple
and quantitative doseresponse curve is
not readily generated. Variation in the im-
munogenicity of individual proteins and in
the distribution of HCP components in
samples obtained throughout a purifica-
tion process likely limit the interpretation
of the data. However, in the absence of a
better system, the ELISA was used to show
a consistent and high HCP titer in crude
cell lysates, and also showed that the
DEAE chromatography step is both suffi-
cient and reproducible for reducing HCP
to below the level of quantitation
(20 ng mL
1
). We interpret these data as
indicating that the chromatography con-
tributes more than 2 logs of HCP clear-
ance (greater than 100-fold signal reduc-
tion) based upon titration of sample versus
the reference HCP. Challenge of the ultra-
filtration step with HCP-rich samples con-
firmed that additional removal of HCP oc-
curred during ultrafiltration. Finally, dilu-
tion of the concentrated bulk virus to any
of the patient doses evaluated further re-
duces HCP exposure.
It should also be noted that the HPLC
methods used for virus testing (IEX and
RP) are also capable of detecting unex-
pected (non-viral) proteins, albeit with less
sensitivity than this ELISA. All impurity
measures are best applied to the most con-
centrated in-process samples, specifically
at the ultrafiltered bulk virus stage of the
process. Taken together, none of these
methods has shown any evidence of mea-
surable HCP in the ultra-filtered bulk sam-
ples.
6.4.7.6 Residual DNA
Two methods were used to quantitate resi-
dual host cell (HEK 293) DNA in Ad5FGF-
4 preparations. The first, utilizing mem-
brane hybridization of an oligonucleotide
probe followed by chemiluminescence de-
tection, was based on a previously de-
scribed procedure [46]. Extracted samples
were blotted onto a positively charged ny-
lon membrane using a slot-blot apparatus;
the blot was then hybridized with a bio-
tinylated oligonucleotide probe specific to
the primate alpha-satellite sequence
D17Z1, and bound probe was detected
with a streptavidinalkaline phosphatase
conjugate, followed by a chemilumines-
cent alkaline phosphatase substrate. The
second method utilized TaqMan PCR
quantitation of Alu repeat sequences in
virus samples extracted using the Qiagen
QIAAmp Viral RNA Mini Kit. In-process
sample testing showed that the two meth-
ods gave similar results, with both demon-
strating a greater than 3 log reduction in
the level of host cell DNA by the purifica-
tion process, and the final purified materi-
al containing <0.5 ng mL
1
DNA (see Ta-
ble 6.1).
6.4.7.7 Product Identity Tests by PCR
The adenoviral backbone of the Ad5FGF-4
genome is derived from dl309, an Ad5 mu-
tant, commonly used for the construction
of E1-deleted adenoviral vectors. This mu-
tant has a short stretch of foreign DNA in-
serted in place of a portion of the E3 re-
Table 6.1 Residual host cell DNA in Ad5FGF-4 samples
Sample Hybridization result
[ng mL
1
]
TaqMan PCR result
[ng mL
1
]
Process set 1
Column load <3045 539
Column eluate <1.9 1.9
Filtered UF bulk <0.5 (LOQ) <0.4 (LOQ)
Process set 2
Column load <1523 356
Column eluate <1.9 2.3
Filtered UF bulk <0.5 (LOQ) <0.4 (LOQ)
LOQ=Limit of quantitation.
gion [38]. PCR confirmation of the type of
adenovirus backbone (dl309 or wild-type
Ad5 E3 region) served as a test for product
identity and purity. The initial 15-min/
958C sample incubation necessary for acti-
vation of hotstart Taq DNA polymerase
also served to disrupt the virus capsid, al-
lowing the direct input of intact virus and
eliminating the requirement for prior
DNA purification. Amplified product from
Ad5FGF-4 was digested with the restric-
tion endonucleases AscI, XbaI, and NdeI,
thus generating a distinctive and product-
specific digestion profile. In this way it
was possible to confirm transgene orienta-
tion as well as to differentiate product di-
gest pattern from those obtained from
either wild-type Ad5, RCA, or vectors con-
taining different transgenes.
6.4.7.8 Potency and Identity:
Transgene Expression
and Production of FGF-4 Protein
Two assays were developed that measure
the potency of the FGF-4 transgene carried
by Ad5FGF-4. In the first case, a one-step
growth-promotion assay is conducted on
normal, human retinal pigment epithelial
cells (ARPE-19). The assay measures meta-
bolic activity (Alamar blue dye metabo-
lism) following infection of ARPE-19 cells
with a serial dilution of the virus. The in-
crease in metabolic activity was measured
in relation to a mock-infected control. This
increase correlates with FGF-4 production
determined by an FGF-4 ELISA, increased
de-novo DNA synthesis measured by BrdU
incorporation, and an increase in cell
number. This procedure is therefore an ap-
propriate in-vitro efficacy measure, indicat-
ing that the FGF-4 transgene product is
biologically active.
The second assay measures the produc-
tion of FGF-4 protein produced in and se-
creted by A549 cells (a non-small cell lung
carcinoma cell line) following infection
with different doses of Ad5FGF-4. An
FGF-4-specific ELISA (R & D Systems) ac-
complished quantitation of FGF-4 present
in the cell culture medium. The FGF-4
produced was confirmed as biologically ac-
tive by stimulating the growth of ARPE-19
cells (Fig. 6.8). Both assays show a dose-de-
pendent response and are indicating stabil-
ity for the Ad5FGF-4 product.
6.4.7.9 Parallel Line Analysis
Potency of a sample in both assays was
calculated by the parallel line analysis of
the absorbance data, performed with PLA
Parallel-Line Assay software (Stegmann
Systemberatung, Germany). All statistical
analyses were performed at 95% probabil-
ity level (p=0.05). The absorbance data of
the replicates at each dilution were tested
for outliers by a Dixon test. The remaining
dataset (in log base 2 of fold dilutions)
was analyzed by linear regression. The
slope of the linear regression line and line-
arity of the dataset were tested for statisti-
cal significance. The linear regression
lines of the positive control and test sam-
ple were tested against each other for par-
allelism. The statistical analysis process
was repeated until a range with a mini-
mum number of specified consecutive di-
lutions with the best statistical outcome is
independently identified for the positive
control and test sample (Fig. 6.9). The po-
tency of a sample was normalized against
the potency of the positive control and
reported as a relative potency. For the
growth-promotion assay, only relative po-
tency data, whose Fiducial limit (95% con-
fidence interval) results were within the
specified limit (based on the normal distri-
bution of a large number of data collected)
were reported.
Fig. 6.8 One-step growth-promotion assay. Mock-
infected ARPE-19 cells (A) and ARPE-19 cells in-
fected with the negative control Ad5LacZ (B) or
Ad5FGF-4 (C) on day 1 post infection. (Original
magnification, 4.)
Fig. 6.9 Parallel line analysis of a growth-promo-
tion assay dataset, using PLA software. This iden-
tified five consecutive dilutions of the positive
control, and six consecutive dilutions of the sam-
ple (solid lines) as regions with best statistical
outcome. A relative potency of 1.289 (Fiducial lim-
it: 1.072 to 1.552) was calculated.
6.4.8
Product Characterization Studies:
Molecular Biology
6.4.8.1 DNA Sequencing of the Viral Vector
As part of our product characterization
strategy we obtained the complete DNA
sequence of the Ad5FGF-4 genome, using
double-stranded DNA sequencing of
Ad5FGF-4 DNA extracted from a MVB ex-
pansion. A total of 166 oligonucleotide
primers was used to obtain overlapping
readings from both strands of DNA. The
observed sequence was compared with that
predicted for Ad5FGF-4 by insertion of the
FGF-4 transgene sequence into wild-type
Ad5 sequence [47] (Genbank Accession
Number M73260, originally deposited as
NC001406), into which the dl309-specific
mutations had been incorporated [38, 48].
The observed matched the predicted for
the left-hand 4.9 Kb of vector sequence, in-
cluding the FGF-4 transgene. The se-
quence data distal to the transgene con-
firmed that the Ad5FGF-4 backbone was
derived from the Ad5 mutant dl309. How-
ever, we found 24 discrepancies in the
backbone adenoviral sequence when com-
paring our vector to the GenBank se-
quence. Eight of the differences resulted
in single amino acid changes in protein
coding sequences, seven (all single-base
substitutions) resulted in silent mutations
within coding sequences, and nine (four
single-base substitutions, two single-base
insertions, and two single-base and one
six-base deletions) were located outside of
known coding sequences. Another recent
publication [49] identifies these same dis-
crepancies with the wild-type Ad5 se-
quence in GenBank, but also suggests sev-
eral more differences not indicated in our
study including one that would lead to
premature termination of the E3 10.4 K
protein. These authors sequenced a shot-
gun library of sheared adenoviral DNA
fragments, whereas our sequencing tem-
plate was purified viral DNA. Either varia-
tion may exist within Ad5 vectors in use
by various gene therapy laboratories, or
the sequencing methods themselves may
have contributed to these apparent differ-
ences in the DNA sequence assigned to
our vector.
6.4.8.2 Genome Analysis
As a further safety precaution, and as a
specific regulatory request, the DNA se-
quence of the FGF-4 transgene was ana-
lyzed for other possible open reading
frames (ORFs) that might encode unex-
pected foreign proteins upon infection of
target cells with the recombinant adeno-
virus. This analysis also included any read-
ing frames that crossed the transgene/
virus junctions and might therefore in-
volve viral sequences. While some poten-
tial ORFs were found, there was no evi-
dence for unexpected protein synthesis to
be directed by these sequences. Typically,
they lacked transcriptional control ele-
ments such as promoters, as well as trans-
lation initiation consensus sequences. A
total of 10 possible ORFs of more than 50
codons was identified in and around the
viral transgene. However, amino acid se-
quence alignments with SwissPro database
entries failed to detect any homology be-
tween these possible ORFs and any known
proteins. We conclude only FGF-4 protein
is made from the transgene carried in
Ad5FGF-4.
6.4.9
Clinical Assays
Additional assays were developed to sup-
port studies of this viral vector in a thera-
peutic context. For doses of virus given to
patients by intracoronary injection, we
sought to assess measurable infectious ti-
ters in blood samples and urine samples.
6.4.9.1 Blood Study
Measurement of circulating levels of infec-
tious Ad5FGF-4 following its administra-
tion to coronary arteries was used to assess
the extent of escape of virus from the site
of administration and its systemic clear-
ance. After determining that typical anti-
coagulants had no cytotoxic effect on the
HEK 293 indicator cell line when diluted
at least 10-fold, all blood collections were
performed with heparin, using green cap
tubes. A simple centrifugation of spiked
whole blood to remove the cellular compo-
nents found measurable virus in the cell
pellet after its resuspension, suggesting
that blood fractionation before assay might
misrepresent the true viral titer of the
sample.
The presence of red blood cells in the
microtiter plate wells containing HEK 293
cells made it quite difficult to observe lysis
of the HEK 293 cells unless each sample
was diluted at least 100-fold. In addition,
spiking undiluted blood with Ad5FGF-4
resulted in a 90% loss of the infectivity
within 5 min. After 45 min a further 5-fold
reduction in signal was observed.
An immediate 1000-fold dilution of blood
resulted in stabilization of the clinical sam-
ples for up to 48 h. A final clinical protocol
was established which included the 1000-
fold dilution of blood into DMEM medium
plus antibiotics, immediate cold shipment,
and assay within 24 h of collection. Using
the EPD infectivity assay, the limit of
detection for infectious virus in the
original blood sample was approximately
1000 IFU mL
1
. Clinical data collected from
patients receiving Ad5FGF-4 showed in-
creasing detectable infectious virus depen-
dent upon the dose given, ranging from
250 IFU mL
1
to 2.310
4
IFU mL
1
for
doses of 6.410
6
and 8.810
8
total IFU,
respectively [13].
6.4.9.2 Urine Study
As part of the clinical study with Ad5FGF-
4 it was also necessary to screen urine for
infectious virus as a means of ensuring
that no environmental contamination oc-
curred with the adenovirus. Infectivity was
again assessed using the EPD assay. Con-
trol urine samples were assayed for possi-
ble inhibition (cytotoxicity) within the EPD
over a range of initial dilutions (1: 2 to
1: 500) in DMEM culture medium, but no
cytotoxic effects were observed. Adenovirus
spiked into urine and then sterile filtered
showed no significant loss of infectivity for
up to 48 h, and no significant sample
dilution appeared necessary. Furthermore,
the addition of antibiotics, penicillin
(100 units mL
1
) and streptomycin
(100 lg mL
1
) did not inhibit the infectivity
assay, and thus were added to all clinical
samples as a precaution. In the clinic,
pooled urine was collected over a 6-h post-
treatment period and supplemented with
one-tenth volume of 10 virus stabilization
solution (10 PBS, 20% sucrose, 20 mM
MgCl
2
), shipped at 48C, and assayed im-
mediately upon receipt. The limit of detec-
tion of infectious virus in the original
urine sample was approximately 100 IFU
mL
1
. Clinical data collected from patients
receiving doses of up to 910
8
IFU of
Ad5FGF-4 showed no detectable infectious
virus in their urine [13].
6.4.9.3 Immunoassays
Total and neutralizing (anti-Ad5) antibody
titers in patient sera were tracked in order
to observe pre-existing antiviral antibody ti-
ters. Total anti-Ad5 antibody titers were
measured by ELISA with a serum dilution
series applied to microtiter plates on
which Ad5FGF-4 had been adsorbed and
inactivated by UV exposure. The neutrali-
zation titers were determined by pre-incu-
bation of reference virus in various serum
dilutions, followed by application to HEK
293 cells to observe cytopathic effects in
an EPD assay format. This study focused
on (neutralizing) antibodies that would
block initial viral entry to a target cell, the
required first step before transgene expres-
sion could occur. Several viral proteins
have been identified as targets of neutraliz-
ing antibodies in humans [50, 51]. No ef-
fort was made to detect an anti-FGF-4 anti-
body response independent of anti-viral re-
sponse.
Another safety assessment was achieved
with an FGF-4 serum ELISA for which a
detection limit target of <50 pg mL
1
of
FGF-4 in patient sera was achieved. How-
ever, reliable FGF-4 quantitation at such a
low level required sample concentration
before using a commercial ELISA, the lim-
it of detection of which was 100 pg mL
1
(R&D Systems). Serum was incubated
with heparin-conjugated Sepharose in or-
der to bind any FGF-4 and allow more
than a 10-fold concentration during elu-
tion. However, no FGF-4 was detected in
any patient sera with these procedures
[13].
6.5
Pre-clinical Efficacy and Safety
of Ad5FGF-4 in Pigs
Specific criteria need to be fulfilled for an-
giogenic gene therapy to be successful.
The gene selected should code for a pro-
tein with proven angiogenic activity, the
vector should provide high gene-transfer ef-
ficiency, the delivery technique should tar-
get the ischemic tissue, and the procedure
should be safe, both in the short and long
term [3]. In addition, appropriate cGMP
manufacturing and analytical processes
need to be established. Pre-clinical and ini-
tial clinical studies indicate that angiogenic
gene transfer with Ad5FGF-4, delivered by a
single intracoronary injection, may meet
these criteria [3, 8, 13, 14, 5254].
6.5.1
FGF-5
The basis of this approach was first re-
ported in 1996 by Giordano and co-work-
ers, who employed a pig ameroid model of
chronic myocardial ischemia and demon-
strated effective gene transfer of human
FGF5 following a one-time intracoronary
injection of Ad5FGF-5 [8]. Gene transfer,
but not placebo control (Ad5LacZ), was as-
sociated with angiogenesis, increased cap-
illary density, and improvement in pacing
stress-induced myocardial function deficit
and collateral perfusion deficit in the area
of chronic myocardial ischemia (Fig. 6.10).
6.5.2
FGF-4
Pre-clinical studies on the safety and effi-
cacy of Ad5FGF-4 were conducted in do-
mestic pigs, which have a coronary anato-
my similar in size and structure to that of
humans. This model system allowed vali-
dation of the delivery method and per-
mitted detailed study of Ad5FGF-4 efficacy,
distribution and toxicity [3, 8, 5254].
6.5.2.1 Proof-of-concept Study
Pre-clinical investigation of Ad5FGF-4 in-
volved demonstration of myocardial angio-
genesis in a porcine model of stress-in-
Fig. 6.10 Proof of principle pig studies. (A) Re-
gional contractile function. Basal wall thickening
in the ischemic region of the ameroid-equipped
pig heart was normal, but atrial pacing (200 bpm)
was associated with reduced wall thickening. At 2
weeks after intracoronary Ad5FGF-4 gene transfer
(10
12
v.p.), there was a 2.3-fold increase in wall
thickening in the ischemic region during pacing
(p<0.0001); this was similar to results observed
after Ad5FGF-5 gene transfer (see Ref. [8]). Pigs
receiving Ad5LacZ showed a similar degree of pa-
cing-induced deficit in the ischemic region before
and 2 weeks after gene transfer. (B) Regional myo-
cardial blood flow. Prior to gene transfer, animals
showed a deficit in blood flow in the ischemic re-
gion during pacing. At 2 weeks after Ad5FGF-4 gene
transfer, the animals showed homogeneous con-
trast enhancement in the two regions, indicating
improved flow in the ischemic region (p<0.0001);
this was similar to results observed after Ad5FGF-5
gene transfer (see Ref. [8]). Pigs receiving Ad5LacZ
showed a similar degree of pacing-induced deficit
in the ischemic region before and 2 weeks after
gene transfer. Columns represent mean values for
number of experiments shown below the columns;
error bars denote 1 SD.
A
B
duced myocardial ischemia similar to the
model where therapeutic angiogenesis
with Ad5FGF-5 was demonstrated. In this
experimental model, an ameroid constric-
tor is placed around the proximal left cir-
cumflex coronary artery, leading to gradual
closure of the artery over the following
weeks. The subsequent development of
collateral vessels allows normal myocardial
function and perfusion at rest, but blood
flow is insufficient to prevent ischemia
during periods of increased oxygen de-
mand. In the proof-of-concept studies,
stress-induced (atrial pacing) left ventricu-
lar function and blood flow changes were
evaluated by two-dimensional echocardiog-
raphy approximately 35 days after ameroid
placement. Ad5FGF-4 (10
11
viral particles
[v.p.] per animal) (n=6) was then adminis-
tered by intracoronary injection. Ventricu-
lar function and blood flow were reas-
sessed 14 days later, and the animals were
killed to quantify the presence of adenovir-
al DNA and FGF-4 mRNA and protein ex-
pression in the myocardium.
PCR and RT-PCR analysis demonstrated
the presence of Ad5 DNA and FGF-4
mRNA, respectively [53]. Immunoblotting
showed that pigs treated with Ad5FGF-4
expressed FGF-4 protein in heart tissue
but not in other organs, such as the liver
and eye. Contractile function, assessed as
the degree of wall thickening of the isch-
emic region during atrial pacing, had im-
proved significantly 2 weeks after FGF-4
gene transfer compared with preinjection
(Fig. 6.10A). This functional improvement
was associated with a normalization of re-
gional blood flow during atrial pacing
(Fig. 6.10B) [53]. Thus, this study provided
experimental evidence that a single intra-
coronary injection of Ad5FGF-4 amelio-
rates deficits in myocardial blood flow and
function in the setting of chronic myocar-
dial ischemia in pigs.
6.5.2.2 DoseResponse
and Persistence-of-effect Studies
The porcine ameroid model of myocardial
ischemia has also been used to evaluate
the relationship between the dose of
Ad5FGF-4 injected into the coronary ar-
teries and left ventricular function and
flow. Myocardial perfusion and function
were evaluated before and 2 weeks after
intracoronary injection of different
Ad5FGF-4 doses ranging from 10
9
to
1.610
12
v.p. Ad5FGF-4 doses of 10
10
v.p. were all associated with significant
and comparable improvements in blood
flow and function, but Ad5FGF-4 at the
10
9
v.p. dose had no significant effect on
these parameters [53].
The period over which improvements in
myocardial perfusion and function are sus-
tained after adenovirus-mediated FGF-4
gene transfer has also been studied. In
this experiment, Ad5FGF-4 was adminis-
tered as a single intracoronary injection to
pigs with ameroid-induced ischemia. Effi-
cacy evaluations were conducted before
administration and 2, 4, 8, and 12 weeks
after gene transfer. The significant im-
provements in blood flow and cardiac
function observed after 2 weeks in the
proof-of-concept study were sustained until
the final evaluation at week 12 (Fig. 6.11)
[53].
6.5.2.3 Toxicology Study
A GLP toxicology and biodistribution study
was performed to determine the potential
adverse effects of Ad5FGF-4 in pigs after
either intracoronary or left ventricular ad-
ministration (10
12
v.p. left ventricular, 10
10
to 10
12
v.p. intracoronary). Systemic biodis-
tribution of the product was also assessed.
This study found no significant test
article-related toxicologic effects [53]. PCR
analysis for adenoviral DNA and RT-PCR
analysis for the transgene (FGF-4) mRNA
were performed to assess vector biodistri-
bution. Adenoviral DNA was detected in
27 of 110 organs examined from animals
injected intracoronary with 10
12
v. p.
Ad5FGF-4. Adenoviral DNA was detected
in the lung, liver, spleen, and testis. De-
tectable adenoviral DNA typically de-
creased over time in most organs. The
presence of viral DNA in the testis was
seen at 5 days but not at 28 or 84 days. In
no case in which adenoviral DNA was
present in extracardiac sites was transgene
expression detectable at the mRNA level
after RT-PCR analysis [53].
6.6
Clinical Studies
There are several goals of Ad5FGF-4 an-
giogenic gene therapy for patients with
chronic myocardial ischemia. It should
promote new collateral vessel formation in
the heart. In turn, this should increase
perfusion of ischemic regions, leading to
improved myocardial oxygen delivery and
left ventricular function. From the patients
perspective, the ultimate goal of treatment
should be to reduce or ameliorate symp-
toms of angina, increase exercise capacity,
improve quality of life, and decrease the
long-term risk of coronary events (see also
Part I, Chapters 12 and 13).
Based on the positive efficacy and safety
outcomes of the pre-clinical studies (see
Section 6.4), a clinical program was ini-
tiated in 1998 to determine whether the
improvement in cardiac perfusion and
function (without any product-related ad-
verse effects) detected in animals trans-
lates into clinical therapeutic benefit in
patients with chronic stable angina. Two
studies, the Angiogenic Gene Therapy
(AGENT) trial and AGENT 2, involving a
total of 131 patients followed for 12
months, have been completed [13, 14].
6.6.1
The AGENT Trial
The AGENT trial was the first ever ran-
domized, double-blind, placebo-controlled
(12 sites) US clinical trial of angiogenic
gene therapy for myocardial ischemia. The
objectives of the trial were to evaluate the
safety and anti-ischemic effects of five as-
cending doses of Ad5FGF-4 (from 3.210
8
to 3.210
10
v.p.), randomizing in a ratio
of 1: 3 (placebo: active) in patients with
chronic stable exertional angina.
A total of 79 patients with chronic stable
angina was enrolled into the AGENT trial.
Patients could exercise for 3 min in an
exercise treadmill test (ETT) using the
modified Balke protocol. The adenovirus
vector was infused over 90 seconds
through subselective catheters into all
major patent coronary arteries and grafts
(40% into the right coronary distribution
and 60% into the left coronary distribu-
tion). Repeat ETTs were performed at 4
and 12 weeks after treatment [13].
Fig. 6.11 Duration of efficacy; pig study. After a
single intracoronary injection of 10
12
v.p. Ad5FGF-
4, pacing-induced left ventricular functional deficit
was reversed at 2 weeks after gene injection, and
maintained for up to 12 weeks. Numbers in col-
umns indicate mean number of experiments; error
bars denote 1 SD.
Weeks After Ad5FGF-4
The increase from baseline in treadmill
exercise duration at weeks 4 and 12 was
greater among patients receiving 10
10
v.p.
Ad5FGF-4 than among those receiving pla-
cebo (the difference at week 4 was statisti-
cally significant) (Fig. 6.12). The propor-
tion of patients achieving a >30% im-
provement in ETT time at week 12 (com-
pared with baseline) was also higher in
the Ad5FGF-4 treatment group (36%) than
in the placebo treatment group (21%).
Analysis of the subgroup of patients
with a baseline ETT time of <10 min
showed that gene therapy produced a sta-
tistically significant increase in exercise ca-
pacity both at 4 and 12 weeks than pa-
tients treated with placebo (Fig. 6.13). The
incidence at 12 months of worsening/un-
stable angina or revascularization (CABG
or PTCA) was lower in the Ad5FGF-4
group (17% and 10%, respectively) com-
pared with the placebo group (21% and
16%, respectively).
6.6.2
The AGENT 2 Trial
AGENT 2 was designed to assess whether
Ad5FGF-4 improved myocardial perfusion
in patients with stable angina. It was also
designed to evaluate safety. Based on the
results of the AGENT trial, a dose of 10
10
Fig. 6.12 Mean ( SEM) change
in total exercise treadmill test
(ETT) time for patients at 4 and
12 weeks after intracoronary infu-
sion of 10
10
v.p. of Ad5FGF4
(n=22) or placebo (n=19).
Fig. 6.13 Percentage increase in
exercise treadmill test (ETT) time
at 4 and 12 weeks after intracoro-
nary infusion of Ad5FGF-4 or pla-
cebo in patients with baseline
ETT of 10 min (n=50).
v.p. was selected. The primary end-point
was the change in stress-related (adeno-
sine-induced) reversible perfusion defect
size (RPDS) as assessed by single-photon
emission computed tomography (SPECT),
at 8 weeks after treatment [14].
A total of 52 patients underwent double-
blind randomization (35 to Ad5FGF-4, 17
to placebo). Their individual RPDS had to
be >9% of the left ventricle. Total perfu-
sion defect size (PDS) at baseline was 32%
and RPDS was 20%.
The mean reduction in RPDS from
baseline at 8 weeks post-treatment in the
Ad5FGF-4 group was 4.25.6% (p<0.001),
a 21% decrease from baseline, versus a re-
duction in the placebo group of only
1.65.4% (p=0.32), corresponding to a de-
crease of 8% from baseline (Fig. 6.14).
Similar results were seen for the change
in total PDS from baseline at 8 weeks
post-treatment: Ad5FGF-4 decreased total
PDS by a mean of 4.65.6% (p<0.001)
compared to a reduction of 2.46.5% with
placebo (NS).
More of the patients who received active
gene therapy than those who received pla-
cebo reported complete resolution of angi-
na (30% versus 13%) and no nitroglycer-
ine use (43% versus 17%) at 8 weeks. In
addition, the incidence of worsening/un-
stable angina and revascularization by
CABG or PTCA at 12 months was consid-
erably lower in the Ad5FGF-4 group (6%
and 6%, respectively) compared to those in
the placebo group (24% and 16%, respec-
tively).
6.6.3
Safety in AGENT and AGENT 2 Trials
In the AGENT and AGENT-2 trials com-
bined a total of 131 patients (95 on
Ad5FGF-4 and 36 on placebo) were treated
and followed for 12 months. Overall,
Ad5FGF-4 was well tolerated. There was
no rise in cardiac enzyme activities, elec-
trocardiographic change, or clinical evi-
dence of myocarditis associated with the
treatment. The in-hospital stay was also
uneventful [13, 14].
Fig. 6.14 Mean change in reversible perfusion de-
fect size (RPDS) compared to baseline for all ac-
tive (n=35) and placebo (n=17) patients at 4 and
8 weeks after intracoronary infusion of Ad5FGF-4
(10
10
v.p.) or placebo. RR = relative reduction in
RPDS compared to baseline. (Reprinted from Ref.
[14], with permission from Elsevier.)
Vector distribution after intracoronary
administration was examined in the
AGENT trial [13]. Adenovirus could be de-
tected by end-point dilution infectivity as-
say in the pulmonary artery during intra-
coronary infusion, and in the peripheral
venous blood 1 h later. The frequency of
positive samples increased with increasing
doses of Ad5FGF-4 (Fig. 6.15). Virus was
not detected in the urine collected over
6 hours. The neutralizing antibody titer to
Ad5 increased in most patients, but no
FGF-4 protein was detected in the circula-
tion at any time. Semen samples (n=8)
were tested by PCR (at 8 weeks post-treat-
ment) and found to be negative for
Ad5FGF-4 DNA. Ad5FGF-4 related adverse
effects included dose-related transient fe-
ver (n=8 [8%]), and transient increase in
liver enzymes (n=3 [3%]).
Overall, the safety profile was reassuring
and consistent with safety data in other
cardiovascular adenoviral gene therapy
trials [15,22]. There was no evidence of
myocarditis, retinal neovascularization, or
angioma formation.
6.7
Summary and Conclusions
The findings of pre-clinical studies and
initial clinical trials with Ad5FGF-4, admi-
nistered as a single intracoronary injec-
tion, suggest that this gene therapy prod-
uct is safe and well tolerated and may
represent a promising new treatment for
patients with chronic stable angina.
Although improvements were seen in exer-
cise capacity and myocardial perfusion in
the completed AGENT and AGENT 2
trials, these changes failed to reach statisti-
cal significance for the entire study popu-
lation because of the magnitude of the pla-
cebo response and the low number of pa-
tients in both trials. However, in the sub-
group of patients with a baseline ETT time
of <10 min, a significant difference was
detected between the Ad5FGF-4 and place-
bo, suggesting that the therapy may be
better suited to patients with more severe
disease.
The intracoronary administration proce-
dure was not associated with any adverse
events. Mild, transient fever and small
transient elevations in hepatic enzymes
were reported in a few patients on
Fig. 6.15 Detection of adenovirus in the pulmonary
artery during intracoronary infusion and in venous
blood at 1 h after infusion in patients treated with
increasing doses of Ad5FGF-4 in the AGENT trial
(n=60). (Adapted from Ref. [13]; reprinted with
permission from Lippincott, Williams & Wilkins.)
Ad5FGF-4, reactions that have been ob-
served in previous adenoviral gene therapy
studies. In addition, neither AGENT nor
AGENT 2 has indicated that Ad5FGF-4 is
associated with unwanted angiogenesis,
FGF-4 expression in extracardiac tissues,
or germ-line viral DNA transmission.
However, because only a total of 131 pa-
tients was studied in these two trials, a
larger clinical database must be accumu-
lated to provide additional confidence in
the safety of this gene therapy product.
Indeed, larger-scale, double-blind, place-
bo-controlled clinical trials with Ad5FGF-4
(AGENT 3 and AGENT 4) are currently
under way at centers throughout the
world. These trials will characterize further
the riskbenefit profile of the product, the
optimal dose that should be administered,
and the patient population likely to derive
greatest benefit from this promising new
biopharmaceutical therapy.
Acknowledgements
For all their hard work, we thank the fol-
lowing colleagues: Carlo Alesandrini, Mou-
tasem Elsheik, Dennis Lee, Tao Yu, Bolong
Zhao (CCBD, Berlex Biosciences, Rich-
mond, CA); Dr. Pran Marrott (Berlex Labo-
ratories, Montville, NJ), Dr. Ruprecht Zierz
(Schering AG, Berlin, Germany) and Dr.
Kirk Hammond (University California at
San Diego, CA).
References
1 European Society of Cardiology. Management
of stable angina pectoris: recommendation of
the Task Force of the European Society of Car-
diology. Eur Heart J 1997; 18:394413.
2 Isner JM. Angiogenesis and collateral forma-
tion. In: March KL, Ed. Gene Transfer in the
Cardiovascular System: Experimental Approaches
and Therapeutic Implications. Boston: Kluwer
Academic Publishers, 1997; pp. 307330.
3 Rubanyi GM. The future of human gene ther-
apy. Mol Aspects Med 2001; 22:113142.
4 Wustman K, Zbinden S, Windecker S, Meier
B, Seiler C. Is there functional collateral flow
during vascular occlusion in angiographically
normal coronary arteries? Circulation 2003;
107:22132220.
5 Hansen JF. Coronary collateral circulation:
clinical significance and influence on survival
in patients with coronary artery occlusion. Am
Heart J 1989; 117:290295.
6 Unger EF, Banai S, Shou M, Jaklitsch M,
Hodge E, Correa R, Jaye M, Epstein SE.
A model to assess interventions to improve
collateral blood flow: continuous administra-
tion of agents into the left coronary artery in
dogs. Cardiovasc Res 1993; 27:785791.
7 Harada K, Grossman W, Friedman M, et al.
Basic fibroblast growth factor improves myo-
cardial function in chronically ischemic por-
cine hearts. J Clin Invest 1994; 94:623630.
8 Giordano FJ, Ping P, McKirnan MD, et al. In-
tracoronary gene transfer of fibroblast growth
factor-5 increases blood flow and contractile
function in an ischemic region of the heart.
Nature Med 1996; 2:534539.
9 Losordo DW, Vale PR, Symes JF, et al. Gene
therapy for myocardial angio-genesis: initial
clinical results with direct myocardial injec-
tion of phVEGF165 as sole therapy for myo-
cardial ischemia. Circulation 1998; 98:2800
2804.
10 Rosengart TK, Lee LY, Patel SR, et al. Angio-
genesis gene therapy: phase I assessment of
direct intramyocardial administration of an
adenovirus vector expressing VEGF121 cDNA
to individuals with clinically significant severe
coronary artery disease. Circulation 1999;
100:468474.
11 Rosengart TK, Lee LY, Patel SR, et al. Six-
month assessment of a phase I trial of angio-
genic gene therapy for the treatment of coro-
nary artery disease using direct intramyocar-
dial administration of an adenovirus vector ex-
pressing the VEGF121 cDNA. Ann Surg 1999;
230:466470.
12 Vale PR, Losordo DW, Milliken CE, et al. Ran-
domized, single-blind, placebo-controlled pilot
study of catheter-based myocardial gene trans-
fer for therapeutic angiogenesis using left ven-
References 179
tricular electromechanical mapping in patients
with chronic myocardial ischemia. Circulation
2001; 103:21382143.
13 Grines CL, Watkins MW, Helmer G, et al. An-
giogenic Gene Therapy (AGENT) trial in pa-
tients with stable angina pectoris. Circulation
2002; 105(11):12911297.
14 Grines CL, Watkins M, Mahmarian J, et al.
A randomized double blind placebo-controlled
trial of Ad5FGF-4 gene therapy and its effect
on myocardial perfusion in patients with
stable angina. J Am Coll Cardiol 2003;
42:13391347.
15 Yl-Herttuala S, Martin JF. Cardiovascular
gene therapy. Lancet 2000; 355:213222.
16 Schumacher B, Pecher P, von Specht BU, et
al. Induction of neoangiogenesis in ischemic
myocardium by human growth factors: first
clinical results of a new treatment of coronary
heart disease. Circulation 1998; 97:645650.
17 Laitinen M, Hartikainen J, Hiltunen MO, et
al. Catheter-mediated vascular endothelial
growth factor gene transfer to human coro-
nary arteries after angioplasty. Hum Gene Ther
2000; 11:263270.
18 Khurana R, Simons M. Insights from angio-
genesis trials using fibroblast growth factor
for advanced arteriosclerotic disease. Trends
Cardiovasc Med 2003; 13:116122.
19 Bush MA, Samara E, Whitehouse MJ, et al.
Pharmacokinetics and pharmacodynamics of
recombinant FGF-2 in a phase I trial in coro-
nary artery disease. J Clin Pharmacol 2001;
41:378385.
20 Simons M, Annex BH, Laham RJ, et al. Phar-
macological treatment of coronary artery dis-
ease with recombinant fibroblast growth fac-
tor-2: double-blind, randomized, controlled
clinical trial. Circulation 2002; 105:788793.
21 Ruel M, Laham RJ, Parker JA, et al. Long-
term effects of surgical angiogenic therapy
with fibroblast growth factor 2 protein. J Thor-
ac Cardiovasc Surg 2002; 124:2834.
22 Isner JM, Vale PR, Symes JF, et al. Assess-
ment of risks associated with cardiovascular
gene therapy in human subjects. Circ Res
2001; 89:389400.
23 Carmeliet P. Angiogenesis in health and dis-
ease. Nature Med 2003; 9:653660.
24 Pugh CW, Ratcliffe PJ. Regulation of angio-
genesis by hypoxia: role of the HIF system.
Nature Med 2003; 9:677684.
25 Yl-Herttuala S, Alitalo K. Gene transfer as a
tool to induce therapeutic vascular growth.
Nature Med 2003; 9:694701.
26 Ferrara N, Gerber HP, LeCouter J. The biol-
ogy of VEGF and its receptors. Nature Med
2003; 9:669676.
27 Rissanen TT, Markkanen JE, Arve K, et al. Fi-
broblast growth factor-4 induces vascular per-
meability, angiogenesis and arteriogenesis in a
rabbit hindlimb ischemia model. FASEB J
2003; 17:100102.
28 Barbato JG, Tzeng E. Nitric oxide and arterial
disease. J Vasc Surg 2004; 40:187193.
29 Onimaru M, Yonemitsu Y, Tanii M, et al. Fi-
broblast growth factor-2 gene transfer can
stimulate hepatocyte growth factor expression
irrespective of hypoxia-mediated downregula-
tion in ischemic limbs. Circ Res 2002; 91:923
930.
30 Deindl E, Fernndez B, Hfer IE, et al. Arte-
riogenesis, collateral blood vessels, and their
development. In: Rubanyi GM, Ed. Angiogen-
esis in Health and Disease. New York: Marcel
Dekker; 2000; pp. 3145.
31 House SL, Bolte C, Zhou M, et al. Cardiac-
specific overexpression of fibroblast growth
factor-2 protects against myocardial dysfunc-
tion and infarction in a murine model of low-
flow ischemia. Circulation 2003; 108:3140
3148.
32 Losordo DW, Vale PR, Hendel RC, et al. Phase
1/2 placebo-controlled, double-blind, dose-es-
calating trial of myocardial vascular endotheli-
al growth factor 2 gene transfer by catheter
delivery in patients with chronic myocardial
ischemia. Circulation 2002; 105:20122018.
33 Hedman M, Hartikainen J, Syvanne M, et al.
Safety and feasibility of catheter-based local in-
tracoronary vascular endothelial growth factor
gene transfer in the prevention of postangio-
plasty and in-stent restenosis and in the treat-
ment of chronic myocardial ischemia: phase
II results of the Kuopio Angiogenesis Trial
(KAT). Circulation 2003; 107:26772683.
34 Boekstegers P. Perspectives on selective retro-
infusion of coronary veins as an alternative
approach for myocardial gene transfer and an-
giogenesis. J Invasive Cardiol 2001; 13:339
342.
35 Delli Bovi P, Curatola AM, Kern FG, et al. An
oncogene isolated by transfection of Kaposis
sarcoma DNA encodes a growth factor that is
a member of the FGF family. Cell 1987;
50:729737.
36 Gomez-Foix AM, Coats WS, Baque S, et al.
Adenovirus-mediated transfer of the muscle
glycogen phosphorylase gene into hepatocytes
confers altered regulation of glycogen metabo-
lism. J Biol Chem 1992; 267:2512925134.
37 McGrory WJ, Bautista DS, Graham FL. A sim-
ple technique for the rescue of early region 1
mutations into infectious human Adenovirus
Type 5. Virology 1988; 163:614617.
38 Bett AJ, Krougliak V, Graham FL. DNA se-
quence of the deletion/insertion in early re-
gion 3 of Ad5 dl309. Virus Res 1995; 39:7582.
39 Schoofs G, Monica T, Ayala J, et al. A high
yielding serum-free cell culture process to
manufacture recombinant adenoviral vectors
for gene therapy. Cytotechnology 1998; 28:81
89.
40 Huyghe BG, Liu X, Sutjipto S., et al. Purifica-
tion of a type 5 adenovirus encoding human
p53 by column chromatography. Human Gene
Ther 1995; 6:14031416.
41 Mann B, Traina J, Soderbloom C, et al. Capil-
lary zone electrophoresis of a recombinant
adenovirus. J Chromatogr A 2000; 895:329337.
42 Lehmberg E, Traina JA, Chakel JA, et al. Re-
versed-phase high performance liquid chroma-
tographic assay for the adenovirus type 5 pro-
teome. J Chromatogr B 1999; 732:411423.
43 Lehmberg E, McCaman M, Traina J, et al.
Analytical assays to characterize adenoviral
vectors and theoretic applications. In: Subra-
manian G, Ed. Manufacturing of Gene Thera-
peutics. Kluwer Academic/Plenum Press, 2004;
pp. 210225.
44 Murakami P, Pungor E, Files J, et al. A single
short stretch of homology between adenoviral
vector and packaging cell line can give rise to
cytopathic effect-inducing, helper-dependent
E1-positive particles. Human Gene Ther 2002;
13:909920.
45 Murakami P, Havenga M, Fawaz F, et al.
Common structure of rare replication-defi-
cient E1-positive particles in adenoviral vector
batches. J Virol 2004; 78:62006208.
46 Walsh PS, Varlaro J, Reynolds R. A rapid che-
miluminescent method for quantitation of hu-
man DNA. Nucleic Acids Res 1992; 20:5061
5065.
47 Chroboczek J, Bieber F, Jacrot B. The se-
quence of the genome of adenovirus type 5
and its comparison with the genome of ade-
novirus type 2. Virology 1992; 186(1):280285.
48 Thimmappaya B, Jones N, Shenk T. A muta-
tion which alters initiation of transcription by
RNA Polymerase III on the Ad5 chromosome.
Cell 1979; 18:947954.
49 Sugarman BJ, Hutchins BM, McAllister DL,
et al. The complete nucleic acid sequence of
the Adenovirus Type 5 Reference Material
(ARM) genome. Bioprocess J 2003; 2(5):2733.
50 Chirmule N, Propert K, Magosin S, et al. Im-
mune responses to adenovirus and adeno-as-
sociated virus in humans. Gene Ther 1999;
6:15741583.
51 Molinier-Frenkel V, Gahery-Segard H, Mehtali
M, et al. Immune response to recombinant
adenovirus in humans: capsid components
from the viral input are targets for vector-spe-
cific cytotoxic T lymphocytes. J Virol 2000;
74(16):76787682.
52 Watkins MW, Rubanyi GM. Gene therapy for
coronary artery disease: Preclinical and initial
clinical results with intracoronary administra-
tion of Ad5FGF-4. In: Rubanyi GM & Yl-
Herttuala S, Eds. Human Gene Therapy: Cur-
rent Opportunities and Future Trends. Berlin:
Springer, 2003; pp. 6180.
53 Gao MH, Lai NC, McKirnan MD, et al. In-
creased regional function and perfusion after
intracoronary delivery of adenovirus encoding
fibroblast growth factor-4: Report of preclinical
data. Human Gene Ther 2004; 15:574587.
54 Rubanyi GM. The design and preclinical test-
ing of Ad5FGF-4 to treat chronic myocardial
ischemia. Eur Heart J (Suppl. E) 2004; 6:E12
E17.
References 181
Abstract
MIDGE vectors are innovative, non-viral
expression constructs that are character-
ized by unique features, particularly by
their high safety. Their structure is linear,
covalently closed with single-stranded
loops at both ends. MIDGE expression vec-
tors only consist of a minimalistic genetic
content: the CMV promoter, the transgene,
and a poly-adenylation site, leading to the
characteristic small size of these vectors.
As a prominent feature, the nucleotides of
the loops are ideally suited to covalently at-
tach various molecules, leading to defined
properties of the vector such as tissue- and
cell-specific targeting an increase in trans-
fection rate or the expression of proteins.
Modification of MIDGE vectors with T
H
1-
peptide results in MIDGE-T
H
1 vectors
with increased potency. Besides this, they
induce a T
H
1-type of immune response,
and thus, are particularly suitable for DNA
vaccination. dSLIM immunomodulators
are non-coding, DNA-based molecules,
that are characterized by a double-stranded
stem and two single-stranded loops, result-
ing in a dumbbell-shaped structure. In ad-
dition, they contain non-methylated cyto-
sine-guanine dinucleotide motifs (CG-mo-
tifs). Due to their unique structure and se-
quence they exhibit defined immuno-
modulatory potential such as activation of
immune cells. Both molecules MIDGE
vectors and dSLIM immunomodulators
have successfully been applied in several
areas, particularly in immunization against
infectious diseases and in tumor therapies.
For DNA vaccination, MIDGE-T
H
1 vectors
are superior to plasmids, showing an in-
creased immunological response with an
advantageous T
H
1-based pattern. There-
fore, convincing results have been
achieved with MIDGE-T
H
1 vectors to pre-
vent leishmaniasis in dogs. Furthermore,
MIDGE-T
H
1 coding for HBsAg (antigene
from hepatitis B virus) induce high titers
of antibodies in mice. In tumor therapy,
either combined effects of MIDGE vectors
and immunomodulators dSLIM, or a com-
bination of dSLIM with chemotherapy, are
used to increase the patients immunologi-
cal response against tumor cells. In mice,
tumor diseases have been prevented by
immunization with ex vivo transfected tu-
mor cells (for cell-based therapy) or by
DNA vaccination with a tumor-associated
antigen (TAA). In both strategies, dSLIM
were used as immunomodulatory mole-
cules. In a human clinical trial, the appli-
cation of a therapeutic vaccine using
MIDGE vectors for ex-vivo transfection of
autologous tumor cells and combination of
this cell-based vaccine with dSLIM, re-
183
7
MIDGE Vectors and dSLIM Immunomodulators:
DNA-based Molecules for Gene Therapeutic Strategies
ISBN: 3-527-31184-X
sulted in a clinical response of 50% of the
patients. In addition, another clinical can-
cer trial showed the safety, benefit, and the
immunomodulatory potential of dSLIM in
combination with chemotherapy, which
will also be discussed in this chapter.
Abbreviations
ALL acute lymphoid leukemia
APC antigen presenting cell
CAP-1 carcinoembryonic antigen
peptide-1
CEA carcinoembryonic antigen
CFA complete Freunds adjuvant
CG-motif cytosine-guanine dinucleotide
motif
CMV cytomegalovirus
CTL cytotoxic T lymphocytes
DC dendritic cells
dSLIM double stem-loop immuno-
modulator
eGFP enhanced green fluorescent
protein
FIV feline immunodeficiency virus
GM-CSF granulocyte macrophage-colony
stimulating factor
GMP good manufacturing practice
IFN interferon
Ig immunoglobulin
LACK Leishmania homologue of recep-
tors for activated C kinase
IL Interleukin
LPS lipopolysaccharide
MHC major histocompatibility complex
MIDGE minimalistic, immunogenically
defined gene expression
ODN oligodeoxynucleotides
PAMPs pathogen associated molecular
patterns
PBMC peripheral blood mononuclear
cells
PO phosphorodiester
PT phosphorothioate
SCID severe combined immunodeficiency
TAA tumor associated antigen
TLR toll like receptor
7.1
Vectors for Gene Therapy
7.1.1
Requirements of Vectors for Gene Therapy
Gene therapy is a promising tool to treat a
variety of diseases, not only those caused
by a single gene defect (e.g., cystic fibrosis
or hemophilia) but also those demanding
a local or transient expression of therapeu-
tic proteins [1], such as the induction of
angiogenesis for the treatment of coronary
artery disease or therapies for tumors,
pain, or infectious diseases (see Part I,
Chapter 6 and Part VI, Chapter 6). In ad-
dition, DNA therapeutic strategies include
the broad area of vaccination [2], helping
the immune system to fight viruses,
pathogens, or tumors on its own (see Part
VI, Chapter 3). Vaccination can be divided
into prophylactic strategies to protect from
infectious diseases and therapeutic ap-
proaches to fight established diseases.
The success of gene therapy depends
largely on the availability of suitable vec-
tors that meet the following basic features:
Time course of expression of a therapeutic
protein: The recombinant protein should
be expressed in an adequate amount
and over a period of time that is able to
meet the therapeutic goal. Different appli-
cations may ask for different intensities
and durations of expression of protein.
Adjustment of monogenetic diseases re-
quires a long and stable expression of
the therapeutic protein, whereas some
disease-related applications (e.g., athero-
7 MIDGE Vectors and dSLIM Immunomodulators: DNA-based Molecules for Gene Therapeutic Strategies 184
sclerotic cardiovascular disease, arthritis,
pain) ask for transient (days to weeks) ex-
pression of protein. For the purpose of
vaccination, an even shorter time of ex-
pression of the recombinant protein will
lead to a proper immune response.
Targeting of vectors: Many applications
need the transfection of a certain cell
type to reach the therapeutic goal. For
example, endothelial cells for the secre-
tion of angiogenetic factors, kidney cells
for improving renal function, or tumor
cells for the induction of apoptosis. This
can be achieved either by the use of viral
vectors with an intrinsic ability to trans-
duce specialized types of cells or by ad-
dition of a targeting system to the vector,
motivating research and development of
targeting strategies [3, 4].
Induction of expression of a recombinant
protein: The possibility of inducible ex-
pression of recombinant protein is con-
nected to the issue of tissue targeting.
For example, the use of specific promo-
ters, which exhibit their highest activity
in the context of a defined tissue, helps
to target the expression of the therapeutic
protein to the designated tissue without a
direct targeting of the vector itself [5].
Size of insert: One aim of vector technol-
ogy is to offer a variable size of the cod-
ing insert, allowing the expression of
even large proteins or of multiple pro-
teins (>5 kb).
Safety of therapy: Besides the desired ex-
pression of a recombinant protein, no
further effects should be induced. Vector
technology must assure that vectors are
not transmissible, do not replicate auto-
nomically in the host, or revert to a viru-
lent form. Especially, immunologic reac-
tions towards constituents of the vector
should be desperately avoided, because
anti-vector immunity may lead to the
elimination of transfected cells or, even
worse, to severe immunological compli-
cations after repeated applications of the
vector. Additionally, pre-existing immu-
nity to the vector could compromise a
vectors efficacy. Furthermore, integra-
tion of vector DNA into the host ge-
nome should be circumvented while po-
tentially leading to a maligned pheno-
type of these cells.
Ease of application: In order to keep costs
of treatment low and to provide many
people with access to gene therapy, tech-
niques are required that allow easy ap-
plication, such as injection, inhalation,
or oral administration.
Stability of the vector: Vectors should ex-
hibit storage stability under common
conditions (e.g., frozen in buffer solu-
tion), which includes the putative ship-
ment.
Cost efficient production of vector: The
costs for the treatment of patients must
be in an economically justifiable range.
Therefore, simple biotechnological pro-
cesses are required for the bulk produc-
tion of a vector in order to keep the
costs for clinical trials and treatment of
patients within acceptable limits.
All vectors applied for a DNA immuniza-
tion strategy must address additional
needs:
Induction of humoral, cellular, and muco-
sal immunity: For prophylactic vaccina-
tion, all three types of immune response
should be addressed. Vaccination against
diseases caused by viruses or other intra-
cellular pathogens essentially requires a
strong induction of cellular immunity,
whereas bacteria-borne diseases need a
strong induction of the humoral im-
mune response. For therapeutic vaccina-
tion, and especially for tumor treatment,
a strong induction of cellular immunity
is desirable [6].
Induction of a long-lived immune response:
In order to achieve (life)-long protection
against the pathogen, the induction of
memory cells is necessary.
Easy manipulation at a molecular level:
Vector technology that allows a rapid re-
design of the insert is requested, in or-
der to achieve easy production of vectors
with different transgenes, or for various
vaccine candidates.
Potential for sequential immunization: The
vector system should possess the ability
to be used in the same person for multi-
ple indications.
7.1.2
Types of Vectors for Gene Therapy
Substantial advances have been made dur-
ing the past decades to develop high-class
vectors that are able to meet the individual
needs of different applications (gene thera-
py, prophylactic vaccination, therapeutic
vaccination) [79]. Vectors usually are di-
vided into either viral, bacteria-derived, or
naked DNA vectors.
7.1.2.1 Features of Virus-based Vectors
Viruses of mammalian cells have evolved
to abuse the cellular machinery of their
hosts to permit their own multiplication.
Viruses transfect specifically their target
cells, and may integrate into the hosts ge-
nome, achieve replication, gene amplifica-
tion, recombination and gene expression.
Therefore, a variety of virus-based vectors
was developed for gene therapeutic ap-
proaches, and these especially show high
efficiency of transfection and long-lasting
expression of therapeutic proteins. In addi-
tion, they are able to transfect dividing
cells as well as non-dividing cells, the lat-
ter being targeted via lentiviruses. This is
accompanied by the disadvantages of un-
wanted integration into the hosts genome
[10], the recombination and mobilization
of host genetic material. Each of these pro-
cesses may lead to severe safety concerns.
For example, two children that had been
successfully treated for severe combined
immunodeficiency (SCID) with a retroviral
vector developed leukemia two years after
gene transfer [11].
Another severe problem can be caused
by the induction of an immune response
(or pre-existing immune response) against
constituents of the vector. This was the
reason for the death of Jesse Gelsinger in
1999 [12].
To solve these problems, virus-based vec-
tors were adapted by the removal of infec-
tion-promoting genes, elimination of
genes responsible for recombination and
replication, as well as by reduction of im-
munoreactivity. Second and third genera-
tions of virus-based vectors have improved
safety marks, but have reduced efficiency
of expression of recombinant protein.
Moreover, they are expensive and difficult
to produce pharmaceutically as bulk ware
according to the Good Manufacturing
Practice (GMP) guidelines.
Prominent viral-based vectors are adeno-
[13], vaccinia- [14], lenti- [15], herpes- [16],
adeno-associated viruses [17], simian virus-
40 [18], and retroviral-based vectors [10, 19].
7.1.2.2 Features of Plasmid-based Vectors
Plasmid-based strategies have been devel-
oped to transfect cells, hopefully with fewer
side effects. In contrast to viruses, plasmids
did not evolve to transfer genetic material to
mammalian cells, and therefore no active
mechanisms of cellular uptake and trans-
port to the nucleus exist. In order to in-
crease the efficacy of gene transfer and pro-
tein expression, several modes of packing
and transfection have been developed. The
packing of plasmids as cationic lipids (as li-
posomes) [20] or their association with poly-
lysine [9, 21] are important approaches to
improve transfection (see Part VI, Chapter
6). A reduction in size would also improve
transfection rates (usual size of plasmids
400012000 bp). For DNA vaccination, the
addition of adjuvants to the plasmids may
be helpful (see Section 7.3.1).
Plasmids do not integrate into the ge-
nome of the host cell, thus being unquali-
fied for stable transfection and long-term
expression of therapeutic protein, but are
suited for all applications requiring transi-
ent gene expression [22].
In addition to the intrinsic high safety of
plasmids compared to viral vectors, several
characteristics of plasmids require improve-
ment, even with regard to safety aspects:
Immunologic aspects: Gene products from
DNA-sequences necessary for the replica-
tion of plasmids in bacteria (e.g., origin of
replication, antibiotic resistance markers)
are weakly expressed, but are sometimes
strong antigens and may lead to un-
wanted immunological responses or even
allergic reactions. Immune reactions
against these plasmid-based gene prod-
ucts may lead to the elimination of cells
expressing such antigens, or to immuno-
logical complications when plasmids are
used in repeated applications. All of these
undesired immunological processes must
be prevented carefully, especially in pro-
phylactic DNA vaccination, where a
healthy population is subjected to contact
with vectors. Additionally, the origin of re-
plication favors recombination processes,
accenting the need for technical improve-
ments of these vectors.
Inflammation: Non-methylated CG-mo-
tifs are present in bacterial DNA and ac-
tivate the innate immune system, lead-
ing to inflammation. Their presence
should be avoided, especially when gene
expression is desired for more than sev-
eral hours. Furthermore, systemic in-
flammation may be induced by cationic
lipid: plasmid DNA complexes [23].
Targeting: Plasmids themselves are
barely suited for targeting, because the
monotone chemistry of naked plasmid
DNA polymer does not provide unique
modification sites. For molecular target-
ing, the tools of viruses should be opti-
mized, reduced to the size of peptides,
and added to the liposomal cover.
Further aspects: Recently, improvements
of the properties of plasmids have been
made, such as the plasmid-based alpha-
virus replicons [24]. These vectors lead
to the production of replicon RNAs,
which exit the nucleus, amplify them-
selves and express heterologous genes
(e.g., antigens) to a high level.
7.1.3
MIDGE Vectors
Recently, a new vector system, MIDGE
(Minimalistic Immunogenically Defined
Gene Expression), was developed in order
to overcome the hurdles of plasmid-based
technologies and to solve the above-men-
tioned problems. These small and linear
expression vectors are superior in several
aspects to all other known vectors. In par-
ticular, they are characterized by their
excellent safety and their intrinsic and
unique structure that allows the chemical
attachment of targeting signals.
7.1.3.1 Structure of MIDGE Vectors
This new type of vector is derived from plas-
mids and contains a minimalistic expres-
sion cassette. It shows a linear covalently
closed topology, with single-stranded loops
at both ends (Fig. 7.1). The covalently closed
loops protect the vectors from degradation
by exonucleases and the transfected host
cells from fragment-induced apoptosis.
MIDGE expression vectors only consist
of a minimalistic genetic content, leading
to the characteristic small size of these
vectors: the cytomegalovirus (CMV) pro-
moter (regulating the intensity of tran-
scription of the transgene), the transgene,
and a poly-adenylation site from SV40
large T-antigen, preventing the rapid de-
gradation of mRNA. All three elements
are indispensable for the therapeutic aim.
The transgene can be chosen freely accord-
ing to the planned application. Addition-
ally, the insertion of more than one gene
into the vector is possible. MIDGE vectors
with a wide range of the size of the insert
(800 to more than 8000 base pairs (bp))
have been produced and used.
7.1.3.2 Production of MIDGE Vectors
MIDGE vectors are produced by a simple
and rapid biotechnological process, origi-
nating from plasmids (Fig. 7.2).
Plasmids, containing the expression cas-
settes of the corresponding MIDGE vectors
between two recognition sites for restric-
tion endonuclease Eco31I are produced
and purified. Plasmids are incubated with
Eco31I, resulting in two fragments. Oligo-
nucleotides (ODN), suited to form the sin-
gle-stranded loop are added and ligated.
T7-DNA-polymerase is used to degrade
Fig. 7.1 Schematic drawing of a MIDGE vector. Minimalistic, linear
vector with CMV-promoter, transgene and polyadenylation-site.
Fig. 7.2 Schematic drawing of the production process for MIDGE vectors.
Originating from plasmids, biotechnological techniques lead to the
purified linear MIDGE vector.
DNA molecules with open ends, mainly the
plasmid-backbone including for example
antibiotic resistance genes. MIDGE vectors
are further purified by a HPLC method,
leading to a pure product. MIDGE vectors
are about 23 kilobases (kb) smaller in size
than the corresponding plasmids. The
workflow can be performed in one vessel
with sequential addition of enzymes and
only one final purification step.
This manufacturing process is easily
adaptable to GMP standards, resulting in
vectors that can be used for example in
clinical trials.
7.1.3.3 Characteristics of MIDGE Vectors
MIDGE vectors show a variety of advan-
tages compared to the corresponding plas-
mids. First, MIDGE vectors can be de-
signed almost free from CG-motifs. These
sequences may cause objectionable inflam-
matory reactions and auto-immunity in
animals or humans [25, 26].
Second, and in contrast to plasmids,
MIDGE vectors are free from sequences
that are not necessary for the expression
of the transgene. These are sequences that
are essential for the multiplication of plas-
mids in bacteria (e.g., origin of replication,
conformational DNA, or genes that are
necessary for the regulation of gene ex-
pression, or for selection of transfected
bacteria, as resistance to an antibiotic).
Their absence helps to assure the genome-
integrity of the transfected cells. Further-
more, the use of MIDGE vectors prevents
the spread of antibiotic resistance genes to
ubiquitous bacteria, thereby helping not to
worsen one of the most pressing problems
in public health systems. Furthermore,
MIDGE vectors do not induce any im-
mune response against basic vector consti-
tuents and do not evoke allergic anti-vec-
tor-reactions in recipients.
Third, the single-stranded ends of linear
MIDGE vectors are easily accessible to
linkage chemistry, independent of the
chemical nature of the attached molecule
(protein, peptide, sugar residue, lipid, bio-
tin, etc.). Cross-linking of molecules to
MIDGE vectors combines the features of
the vector and the attached molecule (e.g.,
for targeting).
Fourth, MIDGE vectors exhibit a feasible
expression of recombinant proteins in
mammalian cells. Compared to plasmids,
expression rates are equal or even higher
(depending on the promoter employed,
the transfection method and the cell line
used), probably due to their small size. A
high level of transient gene expression is
achieved for several hours or days, and a
lower level of expression can be detected
for several weeks (for up to 80 days after
intramuscular injection).
Fifth, the biotechnological production
process of MIDGE vectors is easy and
cheap, starting from a corresponding plas-
mid. Simple techniques as restriction and
ligation and a purification method are
used to produce MIDGE vectors.
In summary, MIDGE vectors are stable,
non-super-coiled, non-replicating, non-inte-
grating, non-inflammation inducing and
are, therefore, extremely safe. This makes
them superior not only to plasmids but
also to some viral vectors.
7.1.3.4 Experimental Proof of the
Characteristics of MIDGE Vectors
Expression of a recombinant protein Com-
pared to plasmids, MIDGE vectors exhibit
similar expression of a recombinant protein
in vitro. The intensity of protein expression
depends on the cell line used (e.g., HeLa,
K562, DU-145) and the method of transfec-
tion (e.g., electroporation, lipofection). In-
terestingly, when interleukin-2 (IL-2) and
eGFP expression cassettes are used,
MIDGE vectors show an even two- to four-
fold higher expression than the correspond-
ing plasmids [27]. Thus, the expression level
depends on the conditions used and must
be evaluated for each application.
In vivo, various conditions yielded simi-
lar expression levels of protein for plas-
mids and MIDGE vectors (Fig. 7.3).
Kinetic of the expression of a recombinant
protein The use of MIDGE vectors in vivo
results in a typical biphasic expression ki-
netic of the recombinant protein. The ex-
pression is high for the first few hours
and declines over a period of about 80
days, showing a long, but transient trans-
formation of the cells (Fig. 7.4).
Similar results were obtained in vitro,
when cells were transfected with MIDGE
vectors encoding CD80/B7.1. The percent-
age of transfected cells was 75% one day
after transfection and 47% four days later.
According to the time course of expression
of recombinant protein, MIDGE vectors
Fig. 7.3 In-vivo expression of luciferase encoded by either MIDGE
vectors or plasmids. Mice were injected intramuscularly, and expres-
sion of luciferase was measured after extraction of the targeted tissue.
Fig. 7.4 In-vivo expression of luciferase after intramuscular trans-
fection. Mice were injected intramuscularly with MIDGE vectors or
plasmids, and expression of luciferase was determined after extrac-
tion of the targeted tissue after the respective time period.
are especially suited for DNA vaccination
and the transient expression of therapeutic
proteins.
7.1.3.5 Smart MIDGE Vectors
During the past two decades, many efforts
have been undertaken to engineer a vector
that combines the characteristics of trans-
fection with the characteristics of target-
ing or potentiation or selecting, lead-
ing to numerous technologies that fulfill
these aspects to different degrees [28].
Due to their unique structure with sin-
gle-stranded loops, MIDGE vectors are ide-
ally suited to become equipped with func-
tionally active molecules, directly bound to
the DNA moiety to superpose the vector
features with an additional function. Dur-
ing the biotechnological production of
MIDGE, oligonucleotides that are chemi-
cally activated for cross-linking processes
can be used, resulting in MIDGE vectors
that can easily be linked to otherwise acti-
vated molecules of different chemical
properties such as peptides, antibodies (or
parts thereof), antigens, ligands, biotin,
sugar residues, steroids, lipids and so
forth. Those molecules show a unique and
direct covalent bond between DNA and
the helper-molecule (Fig. 7.5).
MIDGE vectors with such modifications
can easily bind to appropriate partners, for
example tumor cells, antigen-presenting
cells (APC), or liver cells. Besides targeting
the modification of the vectors can either
additionally or exclusively be used to im-
prove intracellular transport of the vectors
to the nucleus or other intracellular com-
partments.
7.1.3.6 Immunomodulating Vectors
MIDGE-T
H
1
MIDGE technology was advanced to
MIDGE-T
H
1, being characterized by a cova-
lently bound T
H
1-peptide to the MIDGE
vectors at one of the single-stranded loops
(Fig. 7.6). The T
H
1-peptide is composed of
11 amino acids (PKKKRKEDPYC). When
used for DNA vaccination, MIDGE-T
H
1 vec-
tors induce a significantly improved humor-
al and cellular immune response, compared
to MIDGE vectors or plasmids. Additionally,
many further and potent immunomodulat-
Fig. 7.5 Schematic drawing of various smart MIDGE vectors.
Linkage of molecules (middle column) for defined purposes (right column).
ing characteristics can be attrubuted to
MIDGE-T
H
1 vectors, including the induc-
tion of CD4
+
interferon (IFN)-c producing
T cells in vivo and the secretion of IL-
12p40 in vitro in murine and human
systems. Therefore, T
H
1-peptide-modified
MIDGE vectors are ideally suited for DNA
vaccination (see Section 7.3.1.1).
7.1.3.7 Experimental Proof of the
Characteristics of MIDGE-T
H
1
Vectors
Comparison of the expression of a recombi-
nant antigen in vivo as well as the induction
of specific antibodies showed that MIDGE-
T
H
1 vectors are superior to pure MIDGE
vectors and plasmids (Figs. 7.7 and 7.8). This
was confirmed by numerous experiments,
using different animal models, different
antigens, and different routes of delivery.
The superior characteristics of MIDGE-
T
H
1 vectors for the purpose of immuniza-
tion were affirmed by determination of
amount of DNA necessary to induce anti-
bodies against HBsAg, an antigen from
hepatitis B virus, in a prime boost model
(Fig. 7.8). The threshold for the induction
of antibodies was 25 lg for plasmids, but
less than 1 lg for MIDGE-T
H
1 vectors.
Mice immunized with MIDGE-T
H
1 vec-
tors encoding HBsAg, showed a superior
induction of antibodies (compared to plas-
mids and MIDGE vectors) and a T
H
1-
based immune response, as measured by
the relation of IgG1 to IgG2a (Table 7.1)
[29] and IFN-c producing T cells.
In summary, MIDGE-T
H
1 vectors shift
the immune response to T
H
1 dominance,
resulting in a mixed response by cytotoxic
T lymphocytes and antibodies of the IgG2a
isotype [30].
Fig. 7.6 Schematic drawing of MIDGE-T
H
1 vector. A peptide
of 11 amino acids (T
H
1-peptide) is coupled to the single-
stranded loop via linker chemistry.
Fig. 7.7 Specific antibody response in vivo after
HBsAg immunization. Mice were immunized in-
tradermally (i.d.) with corresponding amounts
(50 lg) of MIDGE, MIDGE-T
H
1 vector or plasmid,
all encoding HBsAg. As a control, luciferase en-
coding plasmid (control-plasmid) was used. At 4
weeks after immunization, the presence of anti-
HBsAg antibodies was determined using ELISA.
7.1.4
Conclusion
Virus-based vectors are characterized by ef-
ficient gene transfer, but many concerns
have emerged about their safety. Plasmid-
based vectors are less efficient but are not
free from an intrinsic potency to harm the
recipient. In order to circumvent some of
these problems, MIDGE vectors were de-
veloped. These exhibit a linear, covalently
closed topology with single-stranded loops
of four bases at each end, and are com-
posed by a minimalistic genetic setup of
promoter, transgene, and poly-adenylation
site. Most impressive is their excellent
safety performance (no foreign genes, no
immune reactions towards vector constitu-
ents, repeated application in one individu-
al is possible). The expression of trans-
genes in vitro and in vivo is similar or su-
perior to that of plasmids. MIDGE-T
H
1
vectors, characterized by a T
H
1-peptide
coupled to MIDGE vectors, induce excel-
lent and T
H
1-based immune responses
when used in DNA vaccination.
7.2
Immunomodulatory Molecules
The innate immunity holds the key to the
initiation and primary activation of adap-
tive immune responses. Therefore, mole-
Fig. 7.8 Specific antibody response after in-vivo HBsAg immunization. Mice
received a prime and a boost injection (after 7 weeks) with either MIDGE,
MIDGE-T
H
1 vector or plasmid encoding HBsAg. At week 11, anti-HBsAg
antibodies in the blood were determined using ELISA.
Table 7.1 Specific IgG2a/IgG1 ratios after intradermal immuniza-
tion of mice with DNA vectors coding for HBsAg.
Immunizing vector IgG2a/IgG1
Plasmid 0.130.22
MIDGE 0.650.71
MIDGE-T
H
1 1.151.55
cules that activate the innate immune sys-
tem are perfectly suited to improve an effi-
cient, protective, and long-lasting immune
response of the adaptive system. Immuno-
modulating molecules are indispensable
constituents of most vaccines, especially
for those composed of recombinant pro-
teins or peptides. Their antigens often fail
to initiate an adaptive immune response,
but rather induce tolerance. Therefore,
these antigens are mixed with substances
known as adjuvants (latin; adiuvare, to
help), which facilitate and amplify the in-
duction of a protective immune response.
Adjuvants have been defined as agents
that act non-specifically to increase the
specific immune response or responses to
an antigen (see Part VI, Chapter 3).
7.2.1
Common Adjuvants
A wide range of unrelated substances have
adjuvant activity (oil emulsions, synthetic
surfactants, mineral gels, bacterial deriva-
tives). Despite their structural diversity,
they can be classified into bacterial-derived
compounds, host-derived adjuvants, and
synthetic substances.
During microbial infection, the patho-
gen provides antigens as well as pathogen-
associated molecular patterns (PAMPs),
the latter being activators of the innate im-
mune system, and therefore, function as
adjuvants. Examples of these adjuvants are
lipopolysaccharide (LPS), lipid A derivates,
bacterial DNA, or bacterial toxins. Some of
the PAMPs bind to receptors such as Toll-
like receptors (TLR) and, as a conse-
quence, the specific activation towards for-
eign antigens is increased. Host-derived
adjuvants can be subdivided into heat
shock proteins and cytokines, whereas the
third group synthetic substances is re-
presented by aluminum salts, microparti-
cles, and liposomes. The molecular basis
of their effects has been revealed during
the past few years, but some of them re-
main to be elucidated.
In addition to their origin, adjuvants can
be classified into T
H
1- or T
H
2-promoting
molecules according to the type of im-
mune response they trigger.
During the past few decades, numerous
unspecific stimulators of the immune sys-
tem have been evaluated for their adjuvant
effects, which may increase the potency of
vaccines in immunization strategies.
7.2.1.1 Bacterial-derived Compounds
These include the following:
Complete Freunds adjuvant (CFA): This
is composed of an oil-in-water emulsion
containing killed mycobacteria, and exhi-
bits a strong adjuvant effect but also se-
vere toxic side-effects [31].
LPS and lipid A derivatives: LPS is a com-
ponent of the membrane of Gram-nega-
tive bacteria, with lipid A, the central poly-
acylated disaccharide, as the main active
part. Since LPS is highly toxic, derivates
with reduced toxicity have been developed,
one of them is monophospholipid A
(MPL), a mucosal adjuvant supporting hu-
moral and cellular response [32].
Bacterial toxins: As an example, cholera
toxin is a strong mucosal adjuvant
which can act as adjuvant.
7.2.1.2 Host-derived Compounds
Heat shock proteins: These are released
after necrotic cell death and are able to
activate APCs, which secrete pro-inflam-
matory cytokines and up-regulate co-
stimulatory molecules.
Genetic adjuvants: Addition of vectors
encoding cytokines (i.e., granulocyte
(GM-CSF) (see Part VIII, Chapter 3), IL-
2, IL-12), co-stimulatory molecules or
other immunomodulating molecules
have a beneficial effect on DNA-based
immunization.
7.2.1.3 Synthetic Compounds
Aluminum salts (Alum): Aluminum com-
pounds have been used as adjuvants in
vaccination for more than 60 years [33],
and induce an early and efficient immu-
nity. They are the most widely used ad-
juvants in both veterinary and human
vaccines, their main disadvantage being
the stimulation of a T
H
2 immune re-
sponse [34].
Imidazoquinolines: These low molecular-
weight molecules are activators of the
adaptive and innate immunity, and in-
duce a T
H
1-type immune response (i.e.,
R848, Imiquimod) [35]. Interestingly,
they also induce IFN-a and are used in
the treatment of viral infections and
warts.
Oil emulsions: MF59 is an oil-in-water
emulsion of three components: squalene
(a cholesterol metabolite), sorbitan triole-
ate (an oil soluble surfactant), and poly-
sorbate 80 (a water-soluble surfactant).
MF59 is approved for use in humans,
and elicits high antibody titers when
used in combination with antigens [36].
Another family of adjuvants is built by
saponins (triterpenoid glycosides) and
their derivatives (i.e., QS21, Onjisapo-
nins). These are derived from Quil,
which is extracted from the bark of a
tree (Quillaja saponaria molina). Their
use in veterinary immunization results
in a T
H
1-type cytokine pattern and
IgG2a antibody response.
7.2.2
DNA-based Immunomodulatory Molecules
Bacterial DNA is recognized by the mam-
malian organism as foreign due to its dif-
ferent methylation pattern compared to
mammalian DNA. Therefore, its adminis-
tration leads to activation of the immune
system. Nucleotide sequences with non-
methylated CG-motifs (CpG-ODN) and
thus high similarity to bacterial DNA pos-
sess an immunomodulatory potency, and
can serve as danger signals or PAMPs
[37]. These CpG-ODN act via distinct sig-
naling pathways. After endosomal binding
to TLR-9, the signal is processed through
the MyD88-IRAK-NFjB pathways to acti-
vate the transcription of multiple genes
[38, 39]. The CpG-ODN not only direct the
production of the pro-inflammatory cyto-
kine IL-6, but also promote a T
H
1-re-
sponse by secretion of IL-12 and IFN-c,
and by activation of natural killer (NK)
cells, B cells, and dendritic cells (DC).
Thus, they have been successfully used as
T
H
1-promoting molecules. In addition,
and due to their T
H
1-based activity, they
have been applied to reduce allergic reac-
tions. Recent reports have shown that the
strength of the immune activation is af-
fected by number, site and nature of the
CG-motifs [40]. However, the response of
peripheral blood mononuclear cells
(PBMC) to specific CpG-ODN is heteroge-
neous and depends on the donor [41]. For
human application, distinct groups of
CpG-ODN were characterized which differ
in structure and function. The first group
(K-type, B-ODN) promotes B-cell prolifera-
tion, monocyte stimulation and secretion
of IgM and IL-6 [42, 43]. These CpG-ODN
are optimally active when longer than 12
bases, contain several CG-motifs, and are
completely phosphorothioate-modified.
The other group of CpG-ODN (D-type, A-
ODN) activates cdT-cells and natural killer
cells (NK) to express CD69 and secrete
IFN-c, which is dependent on IFN-a/-b
production [44, 45]. Their palindromic se-
quence flanked by poly-G-ends is supposed
to form a hairpin structure with the CG
dinucleotide at its apex [42]. A third group
of CpG-ODN was characterized which
combines the properties of both previous
groups [4648]. To circumvent the degrada-
tion of linear single-stranded phosphoro-
diester (PO)-based ODN, phosphorothioate
(PT) modifications were normally used,
thereby enhancing their stability in vitro
and in vivo [49]. In addition to the in-
creased stability, PT-ODN were more po-
tent in the stimulation of B-cell prolifera-
tion than PO-ODN. However, PT-modifica-
tions of ODN result in several side effects
when used in vivo. First, PT-ODN led to a
prolongation of the blood clotting time via
inhibition of the intrinsic tenase complex
[50]. Second, PT-ODN bind non-specifically
to various proteins (i.e., transcription fac-
tors) and thus may affect cell signaling
[51]. Third, PT-ODN can result in acute
toxicities in Rhesus monkeys via comple-
ment activation [52, 53]. Fourth, PT-based
CpG-ODN account for dramatically re-
duced functionality and definition of lym-
phoid organs in mice [54].
7.2.3
dSLIM Immunomodulators
Since linear PT-protected CpG-ODN re-
sulted in multiple side effects, the develop-
ment of immunomodulatory molecules
without PT-protection but with a similar
stability is required. One possibility of
avoiding PT-modification is to protect the
open ends of linear PO-ODN by generat-
ing a short, covalently-closed dumbbell-like
structure [55]. These dumbbell-shaped
molecules resemble natural DNA and,
thus, have an increased stability towards
cellular nuclease [56, 57]. Therefore, im-
munomodulatory dumbbell-shaped mole-
cules, termed double-stem loop immuno-
modulator (dSLIM), have been designed
(Fig. 7.9). These molecules contain non-
methylated CG-motifs in their loops, their
stem, or both. They are generated via a
simple production process, with the poten-
tial for introducing molecular modifica-
tions. A dumbbell-like molecule with CG-
motifs in its stem was used to increase the
tumor-protective effect of a cell-based vac-
cine in a murine leukemia model [58]. In
addition, dSLIM molecules with CG-motifs
in both of their loops, have been used for
clinical Phase I/II trials as part of a tumor-
specific therapy against various solid tu-
mors [59, 60].
Fig. 7.9 Schematic drawing of a
dSLIM molecule. The dumbbell-
shaped PO-based immunomodu-
latory DNA-molecule contains
non-methylated CG-motifs (light
boxes) in each loop.
7.2.3.1 Production of dSLIM
The production of dSLIM molecules is a
simple biotechnological process, starting
with PO-based ODN (Fig. 7.10) [32, 33].
Defined ODNs, suited to form a single-
stranded hairpin are heated and cooled for
annealing. After ligation of specific DNA
overhangs through the addition of T4
DNA ligase, the dumbbell shaped product
is purified via HPLC.
The quality of the purified product is an
important aspect, and must be ensured be-
cause it can influence the potency of
dSLIM. Therefore, integrity is tested via
exonuclease activity of T7-DNA-polymerase
degrading residual DNA molecules with
open ends and subsequent gel electrophor-
esis. Furthermore, the content of endotox-
in is determined by an end-point Limulus
amoebocyte lysate (LAL) test, and must
meet strict standards (<10 EU mg
1
DNA).
7.2.3.2 Properties of dSLIM
Some properties of this new class of im-
munomodulatory molecules, the dSLIM,
resembles those of regular PT-based CpG-
ODN. A broad range of dSLIM molecules
has been designed, combining variations
of stem- and loop-size, numbers of CG-
motifs, and their location in the mole-
cules. Some of their prominent features
are as follows:
Increased expression of immunological sur-
face marker: In vitro experiments with
B cells which belong to the profes-
sional APCs showed that dSLIM in-
crease the expression of immunological-
relevant surface markers. The presence
of major histocompatibility complex
class II (MHC II), important for antigen
presentation, the co-stimulatory factor
CD80/B7.1, and the immunorelevant
proteins CD40 and ICAM-1/CD54 were
amplified. This effect of dSLIM was
both dose- and time-dependent.
Enhanced production of cytokines: In ad-
dition to the upregulation of surface
marker expression, dSLIM enhance cyto-
kine production from a variety of cells.
dSLIM led to the increased production
of T
H
1 cytokines IL-12, IL-2, and IFN-c
and the inflammatory cytokines IFN-a,
IL-6, tumor necrosis factor (TNF) a from
PBMC both in vitro and in vivo. To be
precise, IL-12 was preferentially secreted
by B cells, IFN-c by NK cells, and IFN-a
by DC. This effect of dSLIM was also
dose- and time-dependent.
Influence of conformation: Intriguingly, a
change in structure or size, as well as
variation of number and localization of
CG-motifs within the molecule, signifi-
cantly reduced the stimulatory potential
of dSLIM regarding surface marker ex-
pression on B cells, as well as cytokine
production by PBMC. The influence of
the above-mentioned parameters was
verified with murine spleen cells in vitro.
These data also showed that dSLIM is
Fig. 7.10 Schematic drawing of
the dSLIM production process.
Originating from oligonucleotides
(ODN), biotechnological tech-
niques lead to the purified cova-
lently-closed dSLIM.
effective on both human and murine
cells.
Stability: dSLIM integrity was main-
tained after storage in PBS for 6 months
at 208C and 48C. No loss in its poten-
tial to stimulate cytokine production in
vitro was observed. Even pre-incubation
for 15 days with serum-containing me-
dia at 378C results in levels of cytokine
and surface marker expression compar-
able to those in PT-protected CpG-ODN.
Reduced side-effects compared to PT-based
ODN: In contrast to PT-based CpG-
ODN, dSLIM induced no liver and
spleen enlargements after 7 days of con-
secutive intraperitoneal injection into
mice. Both molecule groups were com-
pared regarding equal mass and molar-
ity. Furthermore, PO-based dSLIM do
not significantly prolong the activated
partial thromboplastin time (aPTT), as
reported previously for PT-based ODN.
Effect on immunoglobulins: dSLIM moder-
ately enhance IgM, IgA, and IgG pro-
duction from PBMC in vitro, but de-
crease IgE secretion.
The defined linkage of various molecules
to ODN, which was described for MIDGE
vectors (see Section 7.2.3.5), will give rise
to modified dSLIM molecules that cur-
rently are under development. For exam-
ple, the coupling of defined peptides to
dSLIM will generate new molecular com-
pounds which are expected to hold specific
and improved properties such as enhance-
ment of cellular antigen presentation.
7.2.4
Conclusion
A broad variety of substances has been
tested for their adjuvant and immunomo-
dulatory properties. One of the most pro-
mising groups is that of DNA-based im-
munomodulators, which exhibit potent
features of a broad activation of immune
cells resulting in proliferation or differen-
tiation, secretion of cytokines, and the ex-
pression of immune-relevant proteins on
their surface. In addition, these immuno-
modulators are easy to produce and store.
To circumvent the side effects of common
PT-protected ODN, covalently-closed,
dumbbell-shaped dSLIM molecules with-
out PT-protection have been developed.
These molecules combine broad immune
activation properties with negligible side
effects, and have been used successfully in
animal models and human clinical trials.
They are powerful molecular substances
with applications in the fields of infectious
diseases, allergy, and cancer.
7.3
Application of MIDGE Vectors and dSLIM
Immunomodulators
The development of biopharmaceuticals
suited to vaccination represents a major
challenge of modern medicine and espe-
cially to maintain the balance between
maximal functional activity and minimal
risks of side effects. To meet these require-
ments, MIDGE vectors and dSLIM immu-
nomodulators were developed and dedi-
cated for use in either prophylactic or ther-
apeutic vaccination.
In the field of prophylactic vaccination,
a variety of vaccines has reached the mar-
ket place during the past few decades.
However, many attempts to prevent infec-
tious diseases such as AIDS, malaria and
tuberculosis as well as infections caused
by hepatitis C virus (HCV), human papillo-
ma virus (HPV), herpes simplex virus or
CMV have failed. The reasons for this
failure are the devious methods used by
pathogens to outflank the immune system,
including antigen-drift, antigen-shift, or
the inhibition of processing or presenta-
tion of antigens. Indeed, the main goal of
prophylactic vaccination is to overcome
these problems.
By contrast, therapeutic vaccines need to
evoke a specific activation of the immune
system against present pathogens or tu-
mor cells. Therapeutic agents in this area
must modulate and enhance the molecular
interactions between cells, that specifically
activate and induce the molecular propaga-
tion of cytotoxic T lymphocytes (CTL) or
tumor-infiltrating cells (Fig. 7.11).
One possible means of fortifying the
molecular interactions between externally
added antigens and the immune system is
to express cytokines and co-stimulating
molecules at the site of injection, in order
to produce an immunoactive microenvir-
onment.
7.3.1
DNA Vaccination
A major challenge for immunologists has
been the development of vaccines de-
signed to emphasize the cellular immune
response and to generate high levels of T-
cell memory. This is especially needed to
prevent infectious diseases caused by
viruses or intracellular pathogens [6].
Many vaccines, especially recombinant
proteins, predominantly induce T
H
2-based
(humoral) immune responses, and there-
fore are not ideally qualified to prevent
virus-mediated diseases.
DNA vaccines are plasmids that are
used to transport genetic information of
selected and immunogenic antigens origi-
nating from a pathogen into the tissue of
the vaccinated individual. Transgenes are
expressed by using the host cells protein
expression machinery, and this leads to se-
cretion of the soluble foreign proteins or
expression of foreign antigens at the cell
surface. In this way, an immune response
towards these vector-encoded proteins is
induced. Consequently, DNA vaccines re-
present a promising tool to induce strong
cellular immune responses, namely CTL
and helper T lymphocytes [61, 62].
Although DNA vaccines are promising,
entry of the vaccine to the nucleus is sub-
ject to intrinsic barriers such as the cell
and nuclear membranes. Likewise, sub-
maximal expression of the antigene will
provide only limited efficacy. Therefore,
many improvements regarding the potency
of DNA vaccines have been developed [63],
particularly in the field of formulation of
plasmids and the method of delivery [64]
(see Part I, Chapter 6 and Part VI, Chapter
6), as well as the use of adjuvants [65] (see
Part VI, Chapter 3) and in respect of anti-
gen expression [66]. The latter includes op-
timized codon usage and the use of multi-
ple codons, as well as the choice of an ap-
propriate promoter, whereas delivery com-
prises the package of plasmids into
cationic liposomes (lipoplexes), DNA-poly-
mer complexes [67], or other nanoparticles
[9, 68] as well as different applications,
such as hydrodynamic delivery (intravascu-
lar) [69] (see Part VI, Chapter 1).
Vaccines based on DNA can be pro-
duced quickly and cheaply, and can be
transported and stored without cooling
due to the high structural stability of
DNA. No safety concerns have been ob-
served following their administration to
several hundred human volunteers. There-
fore, clinical trials with DNA vaccines are
under current investigation, and the first
registration of a DNA vaccine is expected
in the near future [70, 71].
In addition to plasmids, several other vec-
tors such as Mini-circles or MIDGE vectors
have been used for DNA vaccination [72].
MIDGE-T
H
1 vectors were especially devel-
F
i
g
.
7
.
1
1
S
c
h
e
m
a
t
i
c
d
r
a
w
i
n
g
o
f
i
m
m
u
n
i
z
a
t
i
o
n
p
a
t
h
w
a
y
s
u
s
i
n
g
M
I
D
G
E
v
e
c
t
o
r
s
a
n
d
d
S
L
I
M
.
D
N
A
-
b
a
s
e
d
v
a
c
c
i
n
a
t
i
o
n
w
i
t
h
M
I
D
G
E
v
e
c
t
o
r
s
e
n
c
o
d
-
i
n
g
a
n
t
i
g
e
n
s
/
c
y
t
o
k
i
n
e
s
o
r
c
e
l
l
-
b
a
s
e
d
t
h
e
r
a
p
i
e
s
w
i
t
h
t
r
a
n
s
f
e
c
t
e
d
t
u
m
o
r
c
e
l
l
s
i
n
c
o
m
b
i
n
a
t
i
o
n
w
i
t
h
d
S
L
I
M
i
m
m
u
n
o
m
o
d
u
l
a
t
o
r
s
r
e
s
u
l
t
e
d
i
n
b
r
o
a
d
a
c
t
i
v
a
t
i
o
n
o
f
i
m
m
u
n
e
c
e
l
l
s
.
oped for DNA vaccination, and have been
proven to show superior efficacy compared
to other vectors [73]. The covalent attach-
ment of one or two T
H
1-peptides to the vec-
tor induces the expression of T
H
1-promot-
ing interleukins (IFN-c, IL-12) in vitro, thus
supporting a cell-mediated immune re-
sponse. In vivo, the desired fortification of
cell-mediated immunity can also be at-
tested. A T
H
1-based immune response is
of advantage for protection against viruses
and other intracellular pathogens as well
as tumors.
The molecular mechanisms of these
phenomena are not fully understood.
However, one hypothesis is that the combi-
nation of linear DNA with a cross-linked
peptide may be interpreted as a danger
signal, evoking a strong immune re-
sponse.
7.3.1.1 Application of MIDGE Vectors
in Vaccination against Leishmaniasis
and other Diseases
Leishmaniasis presents as a variety of clin-
ical manifestations from cutaneous le-
sions to the visceral form, with the latter
being fatal if untreated. The disease is
caused by Leishmania species (e.g., L. ma-
jor or L. infantum); these intracellular
pathogens are widespread over Africa,
Asia, and South America, and worldwide
more than two million new cases of leish-
maniasis are reported each year. Leishma-
niasis also affects dogs, which serve as the
main reservoir host for L. infantum. At
present, there is no effective and safe vac-
cine against any form of this disease.
A good model to study the immune re-
sponse induced by a potential vaccine is
the infection of certain inbred strains of
mice with L. major. Protective immunity
against L. major is mediated by the expan-
sion of a T
H
1 subset of CD4
+
T lympho-
cytes secreting IFN-c and IL-12.
The LACK antigen, a 36 kDa protein, is
highly conserved among Leishmania spe-
cies, and is a preferential target for early
anti-parasite immune response; it is, there-
fore, a candidate for a vaccine [7476].
MIDGE-T
H
1 vectors encoding LACK were
used in a prime-boost vaccination regi-
men, and were shown to confer highly ef-
fective protection against Leishmania infec-
tion in susceptible Balb/c mice (Fig. 7.12).
The success of the vaccine using MIDGE-
T
H
1 in a prime-boost regimen was superi-
or to that of the present gold standard
(prime with plasmid and boost with re-
combinant Vaccinia virus encoding the
same protein) [74].
At present, the MIDGE-T
H
1-based vac-
cine is under investigation for use in dogs,
and will be further developed in order to
reach market approval. In parallel to the
Fig. 7.12 Immunization of mice
against Leishmania major. Specific
vaccination was performed as de-
picted via prime/boost scheme
using various vectors encoding
LACK antigen before challenge
with L. major. Protection was
measured using the size of foot-
path lesions.
vaccine for dogs, an effective vaccine for
humans is also under current develop-
ment partly in cooperation with the World
Health Organisation.
MIDGE-T
H
1 vectors are currently tested
in various other vaccines to prevent dis-
eases caused by intracellular pathogens
(e.g., Mycobacterium tuberculosis).
Feline immunodeficiency virus (FIV)
causes an infection in cats which resem-
bles human HIV infection in many as-
pects. Therefore, a vaccination strategy
against FIV is expected to be a suitable
model for a vaccine against infection with
HIV. Cats vaccinated with MIDGE (which
encoded the envelope protein of FIV) and
co-immunized with MIDGE (encoding
feline IL-12) showed a high rate of pro-
tective immunity after challenge with FIV
[77, 78].
7.3.2
DNA-based Tumor Therapies
In contrast to prophylactic DNA-based vac-
cination that is used to prevent infectious
diseases, therapeutic DNA-based vaccina-
tion is aimed to combat cancer [79]. Many
tumor cells express TAAs, which are not
expressed by normal cells (see Part V,
Chapter 6). Therefore, these TAAs may
serve as the molecular basis of recognition
of tumor cells by the immune system. The
DNA of TAAs is administered to the
patient (e.g., intradermally), whereupon
some of the patients cells are transfected
and thereafter express TAA. Recombinant
expression of TAA, especially from APCs
in the skin, induces a systemic immune
response towards the TAAs and, therefore,
to all tumor cells expressing those mole-
cules [80, 81]. Instead of using the DNA of
the whole TAA, the DNA of tumor-specific
peptides may also be used.
Although there are some limits to this
therapeutic approach, the application of
DNA from tumor-specific peptides or pro-
teins represents a promising means of im-
munizing mice efficiently against solid tu-
mors [8284], and is presently undergoing
intense investigation for non-solid tumors.
7.3.2.1 Application of MIDGE Vectors and
dSLIM in DNA-based Immunization
against Acute Lymphoid Leukemia
Patients with Philadelphia chromosome-
positive (Ph
+
) acute lymphoid leukemia
(ALL) exhibit a poor prognosis. Tumor
cells of these patients express BCR-
ABL
p185
, a fusion protein with intrinsic
and constitutive tyrosine kinase activity
that is responsible for the maligned phe-
notype of the affected cells.
One promising strategy to prevent re-
lapses is activation of the immune system
against BCR-ABL
p185
. The protective ef-
fects of DNA-based vaccination in mice
was shown in a prophylactic vaccination
strategy, followed by challenge with the
syngeneic Ph
+
ALL cell line BM185. Injec-
tion of these cells into non-pretreated mice
resulted in massive tumor growth within 3
weeks. Mice that were pretreated with a
combination of dSLIM and MIDGE vectors
encoding a peptide of 83 amino acids
from BCR-ABL
p185
and GM-CSF showed a
significant protection, with 26% of mice
not developing a tumor at all, whilst the
remaining mice developed a tumor at a
significantly later stage. Control experi-
ments revealed that all three components
(MIDGE encoding BCR-ABL
p185
, MIDGE
encoding GM-CSF, and dSLIM) were es-
sential for the protective effect. Gene
transfer was performed using the gene
gun.
Although the efficacy of this method
was only shown in a prophylactic vaccina-
tion model in mice, DNA-based tumor vac-
cination represents a promising therapeu-
tic strategy for the treatment of tumors in
humans.
A similar strategy was reported by Ren
et al., who used a prophylactic vaccination
strategy (plasmids encoding prostate-spe-
cific antigen (PSA) in combination with
CpG-ODN) to prevent tumor growth after
challenge of mice with B16 prostate tumor
cells [85].
7.3.3
Cell-based Tumor Therapies
Another strategy to treat tumors by stimu-
lation of the immune system is to use cell-
based therapies [86]. In addition to several
therapies based on cytotoxic T-cells (e.g.,
adoptive T-cell therapy [87] or tumor-
loaded, activated DCs [88]), one important
approach is the use of gene-modified tu-
mor cells. Cell-based tumor therapies can
be divided into either autologous and allo-
genous strategies:
The autologous approach uses tumor cells
of the patient for an individualized ther-
apeutic agent. This is a cost-intensive
procedure which requires the isolation
of appropriate tumor cells from tumor
material, as well as their cultivation and
transfection. The efficacy of the thera-
peutic agent depends on the quality of
the appointed cells (i.e., the expression
of TAA).
The allogenous strategy is based on an es-
tablished and well-characterized tumor
cell line that is characterized by multiple
tumor-associated attributes.
In order to improve the tumor-specific
stimulation of the immune system, the
autologous or allogenous tumor cells are
transfected with genes, encoding immune-
activating molecules, such as cytokines
(IL-2, IL-7, IL-12, IL-27, GM-CSF) or co-
stimulatory molecules (CD80/B7.1, CD86/
B7.2). These modified cells are irradiated
in order to disable further division, and
then applied to patients as a therapeutic
vaccination. At the injection site, TAA in
combination with immunomodulating
molecules are abundant within a local
area, and this leads to an increase in num-
ber, and activation of cytotoxic tumor-spe-
cific T-cells and tumor infiltrating T-cells.
The first-generation therapies of gene-
modified tumor cells used either GM-CSF
or IL-2 as the immune stimulating mole-
cule [89]. Several tumor models in ani-
mals, as well as clinical trials with tumor
patients, have shown promising results,
but further intensification of the induced
immune response appears to be necessary
[90, 91].
This has led to the development of a
second generation of cell-based therapies
using two or more different immune stim-
ulating molecules expressed by the tumor
cells.
dSLIM in Cell-based Tumor Therapy
of an ALL Murine Model
The efficacy of a cell-based tumor therapy
using MIDGE vectors for transfection of
tumor cells in combination with dSLIM
immunomodulators was demonstrated in
a mouse model of ALL [58]. In this tumor
model, the syngeneic pre-B-ALL leukemia
cell line BM185 was used to inoculate tu-
mors in mice, with nave mice developing
tumors of 20 mm diameter within 3
weeks.
To prevent tumor growth, gene-modified
BM185 cells were used as a vaccine.
BM185 cells were double transfected with
the co-stimulatory molecule CD80/B7.1
and the cytokine GM-CSF, both encoded
by MIDGE vectors, and irradiated. dSLIM
immunomodulators were added to the
transfected cells, and the mice were twice
vaccinated with the cell suspension. After
a challenge with BM185 cells, 60% of the
vaccinated mice did not develop any tumor,
while the remaining 40% had a longer tu-
mor-free interval and an increased overall
survival. Thus, tumor growth of BM185
cells can be efficiently prevented by vaccina-
tion with double-transfected BM185 cells
and dSLIM (Fig. 7.13). All animals used in
Fig. 7.13 Cell-based vaccination against acute lym-
phoid leukemia in a murine model. Syngeneic mice
were immunized via prime/boost scheme with
BM185 tumorigenic cells that had been transfected
with MIDGE vectors encoding CD80/B7.1 and GM-
CSF either with addition of dSLIM (blue circles) or
without dSLIM (red circles). Retrovirally CD80/
B7.1- and GM-CSF-transfected cells (black circles)
were used as a positive control. (A) KaplanMeier
plot of mice survival; (B) tumor growth in the
flank of mice.
the control groups and vaccinated with dif-
ferent regimens showed significant in-
creased tumor growth compared to the
CD80/B7.1, GM-CSF, and dSLIM group.
dSLIM in Cell-based Tumor Therapy
of Metastatic Tumors
A Phase I/II clinical trial was conducted in
10 patients with metastatic tumors (renal
cell carcinoma, colon carcinoma, melano-
ma). Tumor material from each patient
was isolated, and the cells were ex vivo
double-transfected using MIDGE vectors
encoding GM-CSF and IL-7. Thereafter,
the cells were irradiated and dSLIM added.
This autologous therapeutic agent was in-
jected into the patients four times at inter-
vals of 14 days. At the injection site, the
APCs were in contact with TAAs and
immunomodulating molecules, and this
led to a tumor-specific immune response.
In total, 50% of the patients responded
clinically to the therapy, exhibiting com-
plete or partial response, or stable disease
(Fig. 7.14) [60].
Further applications of allogenous cell-
based therapies are planned, using a pre-
cisely characterized master cell bank for
each application. These cells are trans-
fected by using MIDGE vectors encoding
four different genes of cytokines and co-
stimulatory molecules. The addition of
dSLIM immunomodulators will complete
the therapeutic agent for tumor vaccina-
tion. The intended field of application is
the therapy of metastatic tumors, as well
as an adjuvant therapy after surgical tu-
mor resection aimed at preventing meta-
static disease.
7.3.4
Peptide-based Tumor Therapies
Another strategy aimed at inducing an im-
mune response against TAAs is the appli-
cation of peptides thereof with an immu-
nogenic potential [92]. This soluble, rather
than cell-based, material has been used to
stimulate anti-tumor T cells and to destroy
cells that present the peptide epitopes on
their cell surface [9395]. This therapy has
several advantages, as peptide-based vac-
Fig. 7.14 Computed tomography of the lung from a patient
enrolled in a clinical trial, using transfected, autologous tumor
cells in combination with dSLIM. (A) Lung metastasis of renal
carcinoma (red circles) before vaccination treatment; (B) after
vaccination.
cines can be manufactured easily and
show few toxic effects. However, fortifica-
tion of the induced immune response is
required by the use of adjuvants [96, 97].
7.3.4.1 Application of dSLIM in a Peptide-
based Therapy of Metastatic Colon
Cancer
A Phase I/II clinical trial was conducted to
evaluate the safety and efficacy of three cy-
cles of standard chemotherapy followed by
vaccination with CEA-derived CAP-1 pep-
tide admixed with different adjuvants, in
particular GM-CSF/IL-2, dSLIM/IL-2 and
IL-2 alone. Patients received weekly vacci-
nations until progression of disease. The
highest rate of response (89%) was ob-
served in the dSLIM group, with 44% of
patients showing a complete response and
45% showing stable disease [59]. It is
hoped that these data will lead to further
clinical investigation of dSLIM in peptide-
based immunization strategies.
7.3.5
Conclusion
MIDGE vectors and dSLIM immunomodu-
lators have been applied to a variety of
models and diseases. In DNA vaccination,
MIDGE and MIDGE-T
H
1 vectors are supe-
rior compared to corresponding plasmids
in the induction of an immune response
to antigens from pathogens causing leish-
maniasis and FIV. This was confirmed in
a hepatitis model in mice. In addition,
MIDGE and MIDGE-T
H
1 vectors were
both shown to meet all demands of safety.
In cell-based tumor therapies, MIDGE
vectors have been used in a clinical trial
and in murine models for ex vivo transfec-
tion of tumor cells. At the injection site,
transfected cells created a surrounding
with the combined presence of all stimuli
necessary to induce an efficient immune
response. These are: 1) the presentation of
antigens (TAA); 2) the presence of co-stim-
ulatory molecules; and 3) a microenviron-
ment characterized by the presence of
danger signals (induced by dSLIM) and
the multiplication of the signals by the
cytokines.
In DNA-based tumor therapies, MIDGE
vectors encoding TAA have been used to
directly transfect cells of animals in vivo
(via ballistic transfer).
For both cell-based and DNA-based strat-
egies of tumor therapy, dSLIM immuno-
modulators are added in order to further
increase the immune response by CTL.
The combination of the effects of dSLIM
and MIDGE-encoded molecules might be
a successful approach for the treatment of
tumors. In addition, the use of dSLIM im-
munomodulators in combination with a
peptide-derived vaccine represents a pro-
mising option for the treatment of meta-
static cancer.
MIDGE vectors have also been tested in
a variety of other fields; for example, to cir-
cumvent the rejection of corneal trans-
plants in mice [98101] and to reduce pain
in a mouse model of chronic inflamma-
tion. Clearly, it is hoped that this technol-
ogy will develop further into the produc-
tion of mature biopharmaceuticals in the
foreseeable future.
Acknowledgments
The authors thank J. Alfken, G. Glowacz,
C. Juhls, C. Junghans, S. A. Knig-Mere-
diz, E. Liebeherr, L. Lopez, S. Moreno, D.
Oswald, F. Sack, G. Schmiedeknecht, M.
Schroff, C. Smith, M. Timon, A. Vila-Coro,
and I. Westfehling for their remarkable
dedication to these studies with MIDGE
vectors and dSLIM immunomodulators.
References
1 Rubanyi, G. M. (2001) The future of human
gene therapy. Mol Aspects Med 22:113142.
2 Wolff, J. A., R. W. Malone, P. Williams, W.
Chong, G. Acsadi, A. Jani, and P. L. Felgner
(1990) Direct gene transfer into mouse muscle
in vivo. Science 247:14651468.
3 Kaneda, Y. (2003) New vector innovation for
drug delivery: development of fusigenic non-
viral particles. Curr Drug Targets 4:599602.
4 Muller, O. J., F. Kaul, M. D. Weitzman, R. Pas-
qualini, W. Arap, J. A. Kleinschmidt, and M.
Trepel (2003) Random peptide libraries dis-
played on adeno-associated virus to select for
targeted gene therapy vectors. Nat Biotechnol
21:10401046.
5 Chyung, Y. H., P. D. Peng, and M. A. Kay
(2003) System for simultaneous tissue-specific
and disease-specific regulation of therapeutic
gene expression. Hum Gene Ther 14:1255
1264.
6 Rocha, C. D., B. C. Caetano, A. V. Machado,
and O. Bruna-Romero (2004) Recombinant
viruses as tools to induce protective cellular
immunity against infectious diseases. Int Mi-
crobiol 7:8394.
7 Gill, D. R., L. A. Davies, I. A. Pringle, and S. C.
Hyde (2004) The development of gene therapy
for diseases of the lung. Cell Mol Life Sci
61:355368.
8 Baker, A. H. (2004) Designing gene delivery
vectors for cardiovascular gene therapy. Prog
Biophys Mol Biol 84:279299.
9 Miller, A. D. (2003) The problem with cationic
liposome/micelle-based non-viral vector sys-
tems for gene therapy. Curr Med Chem
10:11951211.
10 Anson, D. S. (2004) The use of retroviral vec-
tors for gene therapy what are the risks? A
review of retroviral pathogenesis and its rele-
vance to retroviral vector-mediated gene deliv-
ery. Genet Vaccines Ther 2:9.
11 Hacein-Bey-Abina, S., C. von Kalle, M.
Schmidt, F. Le Deist, N. Wulffraat, E. McIn-
tyre, I. Radford, J. L. Villeval, C. C. Fraser, M.
Cavazzana-Calvo, and A. Fischer (2003) A ser-
ious adverse event after successful gene thera-
py for X-linked severe combined immunodefi-
ciency. N Engl J Med 348:255256.
12 Somia, N. and I. M. Verma (2000) Gene thera-
py: trials and tribulations. Nat Rev Genet 1:91
99.
13 Jooss, K. and N. Chirmule (2003) Immunity
to adenovirus and adeno-associated viral vec-
tors: implications for gene therapy. Gene Ther
10:955963.
14 Guo, Z. S. and D. L. Bartlett (2004) Vaccinia as
a vector for gene delivery. Expert Opin Biol
Ther 4:901917.
15 Poeschla, E. M. (2003) Non-primate lentiviral
vectors. Curr Opin Mol Ther 5:529540.
16 Calderwood, M. A., R. E. White, and A. White-
house. (2004) Development of herpesvirus-
based episomally maintained gene delivery
vectors. Expert Opin Biol Ther 4:493505.
17 Lu, Y. (2004) Recombinant adeno-associated
virus as delivery vector for gene therapy a
review. Stem Cells Dev 13:133145.
18 Vera, M. and P. Fortes (2004) Simian virus-40
as a gene therapy vector. DNA Cell Biol
23:271282.
19 McTaggart, S. and M. Al-Rubeai (2002) Retro-
viral vectors for human gene delivery. Biotech-
nol Adv 20:131.
20 Tranchant, I., B. Thompson, C. Nicolazzi, N.
Mignet, and D. Scherman (2004) Physico-
chemical optimisation of plasmid delivery by
cationic lipids. J Gene Med 6 Suppl 1:S24S35.
21 Schmidt-Wolf, G. D. and I. G. Schmidt-Wolf
(2003) Non-viral and hybrid vectors in human
gene therapy: an update. Trends Mol Med 9:67
72.
22 Herweijer, H. and J. A. Wolff (2003) Progress
and prospects: naked DNA gene transfer and
therapy. Gene Ther 10:453458.
23 Tousignant, J. D., A. L. Gates, L. A. Ingram,
C. L. Johnson, J. B. Nietupski, S. H. Cheng,
S. J. Eastman, and R. K. Scheule (2000) Com-
prehensive analysis of the acute toxicities in-
duced by systemic administration of cationic
lipid:plasmid DNA complexes in mice. Hum
Gene Ther 11:24932513.
24 Leitner, W. W., H. Ying, D. A. Driver, T. W. Du-
bensky, and N. P. Restifo (2000) Enhancement
of tumor-specific immune response with plas-
mid DNA replicon vectors. Cancer Res 60:51
55.
25 Hodges, B. L., K. M. Taylor, M. F. Joseph, S. A.
Bourgeois, and R. K. Scheule (2004) Long-
term transgene expression from plasmid DNA
gene therapy vectors is negatively affected by
CpG dinucleotides. Mol Ther 10:269278.
26 Tousignant, J. D., H. Zhao, N. S. Yew, S. H.
Cheng, S. J. Eastman, and R. K. Scheule
(2003) DNA sequences in cationic lipid:pDNA-
References 207
mediated systemic toxicities. Hum Gene Ther
14:203214.
27 Schakowski, F., M. Gorschluter, C. Junghans,
M. Schroff, P. Buttgereit, C. Ziske, B. Schott-
ker, S. A. Konig-Merediz, T. Sauerbruch, B.
Wittig, and I. G. Schmidt-Wolf (2001) A novel
minimal-size vector (MIDGE) improves trans-
gene expression in colon carcinoma cells and
avoids transfection of undesired DNA. Mol
Ther 3:793800.
28 Minko, T., S. S. Dharap, R. I. Pakunlu, and Y.
Wang (2004) Molecular targeting of drug de-
livery systems to cancer. Curr Drug Targets
5:389406.
29 Moreno, S., L. Lopez-Fuertes, A. J. Vila-Coro,
F. Sack, C. A. Smith, S. A. Konig, B. Wittig,
M. Schroff, C. Juhls, C. Junghans, and M. Ti-
mon (2004) DNA immunisation with minima-
listic expression constructs. Vaccine 22:1709
1716.
30 Lopez-Fuertes, L., E. Perez-Jimenez, A.J . Vila-
Coro, F. Sack, S. Moreno, S. A. Konig, C. Jun-
ghans, B. Wittig, M. Timon, and M. Esteban
(2002) DNA vaccination with linear minima-
listic (MIDGE) vectors confers protection
against Leishmania major infection in mice.
Vaccine 21:247257.
31 Freund, J., J. Casals, and E. P. Hosmer (1937)
Sensitization and antibody formation after in-
jection of tubercle bacilli and paraffin oil. Proc
Soc Exp Biol Med 37:905913.
32 Persing, D. H., R. N. Coler, M. J. Lacy, D. A.
Johnson, J. R. Baldridge, R. M. Hershberg, and
S. G. Reed (2002) Taking toll: lipid A mimetics
as adjuvants and immunomodulators. Trends
Microbiol 10:S32S37.
33 Glenny, A. T., C. G. Pope, H. Waddington, and
V. Wallace (1926) The antigenic value of toxoid
precipitated by potassium alum. J Path Bact
29:3845.
34 Lindblad, E. B. (2004) Aluminium adjuvants
in retrospect and prospect. Vaccine 22:3658
3668.
35 Vasilakos, J. P., R. M. Smith, S. J. Gibson, J. M.
Lindh, L. K. Pederson, M. J. Reiter, M.H.
Smith, and M. A. Tomai (2000) Adjuvant activ-
ities of immune response modifier R-848:
comparison with CpG ODN. Cell Immunol
204:6474.
36 Minutello, M., F. Senatore, G. Cecchinelli, M.
Bianchi, T. Andreani, A. Podda, and P. Crovari
(1999) Safety and immunogenicity of an inac-
tivated subunit influenza virus vaccine com-
bined with MF59 adjuvant emulsion in elderly
subjects, immunized for three consecutive in-
fluenza seasons. Vaccine 17:99104.
37 Klinman, D. M. (2004) Immunotherapeutic
uses of CpG oligodeoxynucleotides. Nat Rev
Immunol 4:249258.
38 Latz, E., A. Schoenemeyer, A. Visintin, K. A.
Fitzgerald, B. G. Monks, C. F. Knetter, E. Lien,
N. J. Nilsen, T. Espevik, and D. T. Golenbock
(2004) TLR9 signals after translocating from
the ER to CpG DNA in the lysosome. Nat Im-
munol 5:190198.
39 Hemmi, H., O. Takeuchi, T. Kawai, T. Kaisho,
S. Sato, H. Sanjo, M. Matsumoto, K. Hoshino,
H. Wagner, K. Takeda, and S. Akira (2000) A
Toll-like receptor recognizes bacterial DNA.
Nature 408:740745.
40 Klinman, D. M. and D. Currie (2003) Hier-
archical recognition of CpG motifs expressed
by immunostimulatory oligodeoxynucleotides.
Clin Exp Immunol 133:227232.
41 Leifer, C. A., D. Verthelyi, and D. M. Klinman
(2003) Heterogeneity in the human response
to immunostimulatory CpG oligodeoxynucleo-
tides. J Immunother 26:313319.
42 Ballas, Z. K., A. M. Krieg, T. Warren, W. Ras-
mussen, H.L. Davis, M. Waldschmidt, and
G.J. Weiner (2001) Divergent therapeutic and
immunologic effects of oligodeoxynucleotides
with distinct CpG motifs. J Immunol
167:48784886.
43 Verthelyi, D., K. J. Ishii, M. Gursel, F. Takeshi-
ta, and D. M. Klinman (2001) Human periph-
eral blood cells differentially recognize and re-
spond to two distinct CPG motifs. J Immunol
166:23722377.
44 Krug, A., A. Towarowski, S. Britsch, S.
Rothenfusser, V. Hornung, R. Bals, T. Giese,
H. Engelmann, S. Endres, A. M. Krieg, and G.
Hartmann (2001) Toll-like receptor expression
reveals CpG DNA as a unique microbial stim-
ulus for plasmacytoid dendritic cells which
synergizes with CD40 ligand to induce high
amounts of IL-12. Eur J Immunol 31:3026
3037.
45 Rothenfusser, S., V. Hornung, A. Krug, A. To-
warowski, A. M. Krieg, S. Endres, and G.
Hartmann (2001) Distinct CpG oligonucleo-
tide sequences activate human gamma delta T
cells via interferon-alpha/-beta. Eur J Immunol
31:35253534.
46 Vollmer, J., R. Weeratna, P. Payette, M. Jurk,
C. Schetter, M. Laucht, T. Wader, S. Tluk, M.
Liu, H. L. Davis, and A. M. Krieg (2004) Char-
acterization of three CpG oligodeoxynucleotide
classes with distinct immunostimulatory activ-
ities. Eur J Immunol 34:251262.
47 Marshall, J. D., K. Fearon, C. Abbate, S. Subra-
manian, P. Yee, J. Gregorio, R. L. Coffman,
and G. Van Nest. (2003) Identification of a
novel CpG DNA class and motif that opti-
mally stimulate B cell and plasmacytoid den-
dritic cell functions. J Leukoc Biol 73:781792.
48 Hartmann, G., J. Battiany, H. Poeck, M.
Wagner, M. Kerkmann, N. Lubenow, S.
Rothenfusser, and S. Endres (2003) Rational
design of new CpG oligonucleotides that com-
bine B cell activation with high IFN-alpha in-
duction in plasmacytoid dendritic cells. Eur J
Immunol 33:16331641.
49 Cazenave, C., M. Chevrier, T. T. Nguyen, and
C. Helene (1987) Rate of degradation of [al-
pha]- and [beta]-oligodeoxynucleotides in Xeno-
pus oocytes. Implications for anti-messenger
strategies. Nucleic Acids Res 15:1050710521.
50 Sheehan, J. P. and H. C. Lan (1998) Phosphor-
othioate oligonucleotides inhibit the intrinsic
tenase complex. Blood 92:16171625.
51 Brown, D. A., S. H. Kang, S. M. Gryaznov, L.
DeDionisio, O. Heidenreich, S. Sullivan, X.
Xu, and M. I. Nerenberg (1994) Effect of phos-
phorothioate modification of oligodeoxynu-
cleotides on specific protein binding. J Biol
Chem 269:2680126805.
52 Levin, A. A. (1999) A review of the issues in
the pharmacokinetics and toxicology of phos-
phorothioate antisense oligonucleotides. Bio-
chim Biophys Acta 1489:6984.
53 Henry, S. P., G. Beattie, G. Yeh, A. Chappel,
P. Giclas, A. Mortari, M. A. Jagels, D. J. Korn-
brust, and A. A. Levin (2002) Complement ac-
tivation is responsible for acute toxicities in
rhesus monkeys treated with a phosphorothio-
ate oligodeoxynucleotide. Int Immunopharma-
col 2:16571666.
54 Heikenwalder, M., M. Polymenidou, T. Junt,
C. Sigurdson, H. Wagner, S. Akira, R. Zinker-
nagel, and A. Aguzzi (2004) Lymphoid follicle
destruction and immunosuppression after re-
peated CpG oligodeoxynucleotide administra-
tion. Nat Med 10:187192.
55 Erie, D., N. Sinha, W. Olson, R. Jones, and K.
Breslauer (1987) A dumbbell-shaped, double-
hairpin structure of DNA: a thermodynamic
investigation. Biochemistry 26:71507159.
56 Clusel, C., E. Ugarte, N. Enjolras, M. Vasseur,
and M. Blumenfeld (1993) Ex vivo regulation
of specific gene expression by nanomolar con-
centration of double-stranded dumbbell oligo-
nucleotides. Nucleic Acids Res 21:34053411.
57 Chu, B. C. and L. E. Orgel (1992) The stability
of different forms of double-stranded decoy
DNA in serum and nuclear extracts. Nucleic
Acids Res 20:58575858.
58 Kochling, J., S. A. Konig-Merediz, R. Stripecke,
D. Buchwald, A. Korte, H. G. Von Einsiedel,
F. Sack, G. Henze, K. Seeger, B. Wittig, and
M. Schmidt (2003) Protection of mice against
Philadelphia chromosome-positive acute lym-
phoblastic leukemia by cell-based vaccination
using nonviral, minimalistic expression vec-
tors and immunomodulatory oligonucleotides.
Clin Cancer Res 9:31423149.
59 Weihrauch, M. R., S. Ansen, E. Jurkiewicz, C.
Geisen, Z. Xia, K. S. Anderson, E. Gracien, M.
Schmidt, B. Wittig, V. Diehl, J. Wolf, H. Bohlen,
and L. M. Nadler (submitted) Phase I/II com-
bined immuno-/chemotherapy with CEA de-
rived HLA-A2 restricted CAP-1 peptide and
irinotecan, 5-fluorouracil and leucovorin in pa-
tients with primary metastatic colorectal cancer.
60 Wittig, B., A. Marten, T. Dorbic, S. Weineck,
H. Min, S. Niemitz, B. Trojaneck, D. Flieger,
S. Kruopis, A. Albers, J. Loffel, A. Neubauer,
P. Albers, S. Muller, T. Sauerbruch, T. Bieber,
D. Huhn, and I. G. Schmidt-Wolf (2001) Ther-
apeutic vaccination against metastatic carcino-
ma by expression-modulated and immunomo-
dified autologous tumor cells: a first clinical
phase I/II trial. Hum Gene Ther 12:267278.
61 Nagata, T., T. Aoshi, M. Uchijima, M. Suzuki,
and Y. Koide (2004) Cytotoxic T-lymphocyte-,
and helper T-lymphocyte-oriented DNA vacci-
nation. DNA Cell Biol 23:93106.
62 Giri, M., K. E. Ugen, and D. B. Weiner (2004)
DNA vaccines against human immunodefi-
ciency virus type 1 in the past decade. Clin
Microbiol Rev 17:370389.
63 Manoj, S., L. A. Babiuk, and S. van Drunen
Littel-van den Hurk (2004) Approaches to en-
hance the efficacy of DNA vaccines. Crit Rev
Clin Lab Sci 41:139.
64 Doria-Rose, N. A. and N. L. Haigwood (2003)
DNA vaccine strategies: candidates for im-
mune modulation and immunization regi-
mens. Methods 31:207216.
65 Sasaki, S., F. Takeshita, K. Q. Xin, N. Ishii,
and K. Okuda (2003) Adjuvant formulations
References 209
and delivery systems for DNA vaccines. Meth-
ods 31:243254.
66 Garmory, H. S., K. A. Brown, and R. W. Titball
(2003) DNA vaccines: improving expression of
antigens. Genet Vaccines Ther 1:2.
67 Jones, N. A., I. R. Hill, S. Stolnik, F. Bignotti,
S. S. Davis, and M. C. Garnett (2000) Polymer
chemical structure is a key determinant of
physicochemical and colloidal properties of
polymerDNA complexes for gene delivery.
Biochim Biophys Acta 1517:118.
68 Hirko, A., F. Tang, and J. A. Hughes (2003)
Cationic lipid vectors for plasmid DNA deliv-
ery. Curr Med Chem 10:11851193.
69 Hagstrom, J. E. (2003) Plasmid-based gene de-
livery to target tissues in vivo: the intravascu-
lar approach. Curr Opin Mol Ther 5:338344.
70 Hermanson, G., V. Whitlow, S. Parker, K.
Tonsky, D. Rusalov, M. Ferrari, P. Lalor, M.
Komai, R. Mere, M. Bell, K. Brenneman, A.
Mateczun, T. Evans, D. Kaslow, D. Galloway,
and P. Hobart (2004) A cationic lipid-formu-
lated plasmid DNA vaccine confers sustained
antibody-mediated protection against aeroso-
lized anthrax spores. Proc Natl Acad Sci USA
101:1360113606.
71 Ada, G. and I. Ramshaw (2003) DNA vaccina-
tion. Expert Opin Emerg Drugs 8:2735.
72 Rodriguez, E. G. (2004) Nonviral DNA vectors
for immunization and therapy: design and
methods for their obtention. J Mol Med
82:500509.
73 Schirmbeck, R., S. A. Konig-Merediz, P. Riedl,
M. Kwissa, F. Sack, M. Schroff, C. Junghans,
J. Reimann, and B. Wittig (2001) Priming of
immune responses to hepatitis B surface anti-
gen with minimal DNA expression constructs
modified with a nuclear localization signal
peptide. J Mol Med 79:343350.
74 Gonzalo, R. M., G. del Real, J. R. Rodriguez,
D. Rodriguez, R. Heljasvaara, P. Lucas, V. Lar-
raga, and M. Esteban (2002) A heterologous
prime-boost regime using DNA and recombi-
nant vaccinia virus expressing the Leishmania
infantum P36/LACK antigen protects BALB/c
mice from cutaneous leishmaniasis. Vaccine
20:12261231.
75 Maillard, I., P. Launois, H. Himmelrich, H.
Acha-Orbea, H. Diggelmann, R.M. Locksley,
and J.A. Louis (2001) Functional plasticity of
the LACK-reactive Vbeta4-Valpha8 CD4(+) T
cells normally producing the early IL-4 in-
structing Th2 cell development and suscepti-
bility to Leishmania major in BALB/c mice.
Eur J Immunol 31:12881296.
76 Julia, V., M. Rassoulzadegan, and N. Glaichen-
haus (1996) Resistance to Leishmania major
induced by tolerance to a single antigen.
Science 274:421423.
77 Leutenegger, C. M., F. S. Boretti, C. N. Mislin,
J. N. Flynn, M. Schroff, A. Habel, C. Jun-
ghans, S. A. Koenig-Merediz, B. Sigrist, A. Au-
bert, N. C. Pedersen, B. Wittig, and H. Lutz
(2000) Immunization of cats against feline
immunodeficiency virus (FIV) infection by
using minimalistic immunogenic defined
gene expression vector vaccines expressing
FIV gp140 alone or with feline interleukin-12
(IL-12), IL-16, or a CpG motif. J Virol
74:1044710457.
78 Boretti, F. S., C. M. Leutenegger, C. Mislin, R.
Hofmann-Lehmann, S. Konig, M. Schroff, C.
Junghans, D. Fehr, S. W. Huettner, A. Habel,
J. N. Flynn, A. Aubert, N. C. Pedersen, B. Wit-
tig, and H. Lutz (2000) Protection against FIV
challenge infection by genetic vaccination
using minimalistic DNA constructs for FIV
env gene and feline IL-12 expression. Aids
14:17491757.
79 Berzofsky, J. A., M. Terabe, S. Oh, I. M. Belya-
kov, J. D. Ahlers, J. E. Janik, and J. C. Morris
(2004) Progress on new vaccine strategies for
the immunotherapy and prevention of cancer.
J Clin Invest 113:15151525.
80 Parmiani, G., L. Pilla, C. Castelli, and L. Rivol-
tini (2003) Vaccination of patients with solid
tumours. Ann Oncol 14:817824.
81 Pecher, G. (2002) DNA-based tumor vaccines.
Onkologie 25:528532.
82 Wolchok, J. D., P. D. Gregor, L. T. Nordquist,
S. F. Slovin, and H. I. Scher. (2003) DNA vac-
cines: an active immunization strategy for
prostate cancer. Semin Oncol 30:659666.
83 Grimm, C. F., D. Ortmann, L. Mohr, S. Micha-
lak, T. U. Krohne, S. Meckel, S. Eisele, J. En-
cke, H. E. Blum, and M. Geissler (2000)
Mouse alpha-fetoprotein-specific DNA-based
immunotherapy of hepatocellular carcinoma
leads to tumor regression in mice. Gastroenter-
ology 119:11041112.
84 Han, R., N. M. Cladel, C. A. Reed, X. Peng,
L. R. Budgeon, M. Pickel, and N. D. Christen-
sen (2000) DNA vaccination prevents and/or
delays carcinoma development of papilloma-
virus-induced skin papillomas on rabbits. J
Virol 74:97129716.
85 Ren, J., L. Zheng, Q. Chen, H. Li, L. Zhang,
and H. Zhu (2004) Co-administration of a
DNA vaccine encoding the prostate specific
membrane antigen and CpG oligodeoxynu-
cleotides suppresses tumor growth. J Transl
Med 2:29.
86 Blattman, J. N. and P. D. Greenberg (2004)
Cancer immunotherapy: a treatment for the
masses. Science 305:200205.
87 Yamaguchi, Y., A. Ohshita, Y. Kawabuchi, K.
Ohta, K. Shimizu, K. Minami, J. Hihara, E.
Miyahara, and T. Toge (2003) Adoptive immu-
notherapy of cancer using activated autolo-
gous lymphocytes current status and new
strategies. Hum Cell 16:183189.
88 Morisaki, T., K. Matsumoto, H. Onishi, H.
Kuroki, E. Baba, A. Tasaki, M. Kubo, M. Naka-
mura, S. Inaba, K. Yamaguchi, M. Tanaka, and
M. Katano (2003) Dendritic cell-based combined
immunotherapy with autologous tumor-pulsed
dendritic cell vaccine and activated T cells for
cancer patients: rationale, current progress, and
perspectives. Hum Cell 16:175182.
89 Antonia, S. J. (1999) B7-1 gene-modified tu-
mor cell vaccines. Curr Opin Mol Ther 1:50
56.
90 Van Tendeloo, V. F., C. Van Broeckhoven, and
Z. N. Berneman (2001) Gene-based cancer vac-
cines: an ex vivo approach. Leukemia 15:545
558.
91 Brenner, M. K., H. E. Heslop, and C. M. Roo-
ney (1998) Gene and cell transfer for specific
immunotherapy. Vox Sang 74 Suppl 2:8790.
92 Meng, W. S. and L. H. Butterfield. (2002) Ra-
tional design of peptide-based tumor vaccines.
Pharm Res 19:926932.
93 Elsawa, S. F., D. A. Rodeberg, and E. Celis
(2004) T-cell epitope peptide vaccines. Expert
Rev Vaccines 3:563575.
94 Sarobe, P., E. Huarte, J. J. Lasarte, and F.
Borras-Cuesta (2004) Carcinoembryonic anti-
gen as a target to induce anti-tumor im-
mune responses. Curr Cancer Drug Targets
4:443454.
95 Brinkman, J. A., S. C. Fausch, J. S. Weber,
and W. M. Kast (2004) Peptide-based vac-
cines for cancer immunotherapy. Expert
Opin Biol Ther 4:181198.
96 Sundaram, R., N. K. Dakappagari, and P. T.
Kaumaya (2002) Synthetic peptides as cancer
vaccines. Biopolymers 66:200216.
97 Salit, R. B., W. M. Kast, and M. P. Velders
(2002) Ins and outs of clinical trials with
peptide-based vaccines. Front Biosci 7:e204
e213.
98 Zhang, E. P., J. Franke, M. Schroff, C. Jung-
hans, B. Wittig, and F. Hoffmann (2003)
Ballistic CTLA4 and IL-4 gene transfer into
the lower lid prolongs orthotopic corneal
graft survival in mice. Graefes Arch Clin Exp
Ophthalmol 241:921926.
99 Zhang, E. P., A. Muller, F. Schulte, M.S. Ko-
nig, F. Sack, C. Junghans, B. Wittig, and F.
Hoffmann (2002) Minimizing side effects of
ballistic gene transfer into the murine cor-
neal epithelium. Graefes Arch Clin Exp
Ophthalmol 240:114119.
100 Muller, A., E. P. Zhang, M. Schroff, B. Wit-
tig, and F. Hoffmann (2002) Influence of
ballistic gene transfer on antigen-presenting
cells in murine corneas. Graefes Arch Clin
Exp Ophthalmol 240:851859.
101 Konig Merediz, S. A., E. P. Zhang, B. Wittig,
and F. Hoffmann (2000) Ballistic transfer of
minimalistic immunologically defined ex-
pression constructs for IL4 and CTLA4 into
the corneal epithelium in mice after orthoto-
pic corneal allograft transplantation. Graefes
Arch Clin Exp Ophthalmol 238:701707.
References 211
Abstract
In the last decade, there were a growing
number of reports concerning novel genes
which produced transcripts without pro-
tein-coding capacity. Such RNAs, named
noncoding or nonprotein-coding RNAs
(npcRNAs), play important roles in many
regulatory processes in all organisms. In
eukaryotes, they are involved in cell differ-
entiation and development. Many of the
mammalian npcRNAs are localized within
chromosomal regions, which are linked to
certain diseases, including neurobehavioral
and developmental disorders, and cancer.
The understanding of npcRNA biology
may open new perspectives for molecular
diagnostics and modern biopharmaceuti-
cals.
Abbreviations
AS Angelman syndrome
BWS Beckwith-Wiedemann
syndrome
miRNA microRNA
npcRNA nonprotein-coding RNA
NSCLC nonsmall cell lung cancer
P-TEF positive transcription elongation
factor
Pol polymerase
PWS Prader-Willi syndrome
snoRNA small nucleolar RNA
RMS rhabdomyosarcoma
SRA steroid receptor activator
SRC steroid receptor co-activator
T-DMR tissue-dependent differentially
methylated region
UTR untranslated region
8.1
Introduction
It has been assumed for a long time that
the regulation of gene expression essen-
tially depends on the activity of specific
proteins transcription factors responsi-
ble for switching genes on and off. The ex-
pression of a particular gene can be regu-
lated on three different levels. The struc-
ture of chromatin and/or epigenetic fac-
tors (e.g., methylation) establishes the pre-
transcriptional level which determines if
the gene can be transcribed (active) or not
(inactive). The expression of active genes
is regulated on a transcriptional level by a
number of trans- and cis-acting factors
which govern the timing and efficiency of
transcription (see also Part III, Chapter 2).
The amount of the protein to be produced
is determined at the post-transcriptional
213
8
Nonprotein-coding RNAs and their Potential as Biopharmaceuticals
ISBN: 3-527-31184-X
level. The post-transcriptional regulatory
mechanisms can operate on different steps
of the pathway from the pre-mRNA spli-
cing to translation on the ribosome (see
also Part III, Chapter 3).
The last decade brought about an unpre-
cedented growth of biological data result-
ing mostly from nearly industrial-scale se-
quencing projects of genomes from a vari-
ety of organisms, including humans. The
results of these efforts demonstrated, be-
yond all doubt, that the views concerning
many aspects of molecular biology which
were obvious in the pre-genomic era need
to be revised. In particular, this applies to
the role of RNA in the cell. In the context
of recent discoveries in this field, it is evi-
dent that the functions of RNAs can no
longer be treated as accessory to proteins.
It seems that the correct gene expression
patterns are governed by the intricate net-
work of RNAs which control the flow of
information in the cell.
8.2
The Contents of the Genomes
The most striking finding revealed by the
analysis of mammalian genomes is a rela-
tively small number of protein genes. In
the initial analysis of the draft of the hu-
man genome, presented by the Human
Genome Sequencing Consortium, it was
estimated that the human genome con-
tains 2650029000 protein coding genes
[1]. Similar numbers (2700030500) were
revealed for the mouse genome [2]. More
recent analyses, using a gene prediction
program based on comparative analysis of
human and mouse genomes, yielded
44242 and 44770 protein genes, respec-
tively [3]. These figures are at least 23
times lower than was believed in the pre-
genomic era. Moreover, the protein-coding
part or the open reading frames (ORFs) of
these genes account for less than 2% of
the genomic DNA. ORFs together with the
untranslated regions (5'- and 3'-UTRs) and
introns represent around 2527% of the
genome. The rest of the nonprotein-coding
portions are composed primarily of repeti-
tive sequences which make up approxi-
mately 46% of the genome [1, 4]. Although
the functions of the remaining quarter of
the genomic DNA are largely unknown,
one can assume that at least some of its
portions are responsible for the spatial and
temporal coordination of gene expression.
These considerations presented above are
based on the assumption that any given
fragment of genomic DNA is transcribed
only in one direction and is part of one
transcription unit. This simplified view,
which is true for the organization of the
majority of bacterial genes, does not reflect
the situation found in more complex or-
ganisms. Certain parts of the yeast ge-
nome were found to be transcribed from
both strands, producing senseantisense
pairs of transcripts [5]. Bidirectional tran-
scription is also frequent in the Drosophila
genome as well as in mammalian ge-
nomes [68].
According to the prevailing opinion that
the repertoire of proteins encoded in the ge-
nome is solely responsible for the majority
of cellular functions and that its size deter-
mines the complexity of the system, the
nonprotein-coding DNA regions were once
regarded as nonfunctional junk. Interest-
ingly, from the comparisons of the contribu-
tion of protein-coding regions in the se-
quenced genomes, it turned out that the
more complex an organism is, the less of
its DNA fraction actually codes for proteins
[9]. In prokaryotes, where intergenic and
untranslated regions are short, and splicing
is an exception rather than the rule, the
ORFs of protein-coding genes account for
over 90% of genomic DNA. In simple eu-
karyotes, the nonprotein-coding DNA con-
stitutes 1040%. In invertebrates, the non-
coding part accounts for around 7090%
of the genome and in mammals as much
as 98% (Fig. 8.1). Therefore, these nonpro-
tein-coding regions of genomic DNA may
be responsible for the regulation of complex
mechanisms which underline development
and differentiation by means of controlling
the expression of proteins that play a role in
the cells hardware [10].
8.3
npcRNAs
The rationale for regarding the nonprotein-
coding DNA as an important functional
component of the genome became apparent
with the accumulation of data on transcrip-
tional activities of human chromosomes
and the availability of large collections of
full-length cDNA sequences. Based on the
data obtained from the analyses of single
human chromosomes it has been estimated
that approximately half of the human geno-
mic DNA is transcribed [11]. This means
that a significant fraction of the transcrip-
tional output from the genome does not
arise from protein-coding genes.
Over recent years there have been a
growing number of reports concerning the
discoveries of novel genes which produced
transcripts without protein-coding capacity,
collectively named noncoding or non-
protein-coding RNAs (npcRNA). The term
noncoding RNA is used in a broad sense
to describe any transcript or its fragment
that is not used as a template in ribosomal
protein synthesis [10]. In such a context
npcRNAs would also include all of the
housekeeping transcripts functioning in
the translation (tRNA, rRNA), splicing and
processing (small nuclear RNA, RNase P
RNA), RNA modifications [small nucleolar
RNA (snoRNA)], DNA replication (telo-
merase RNA) (see also Part I, Chapter 1)
or as components ribonucleoprotein parti-
cles (e.g., signal recognition particle RNA,
vault RNAs). Due to their involvement in
basic cellular activities, the housekeeping
RNAs are usually constitutively expressed
at the same level in all cell types, and their
transcription is carried out by RNA poly-
merases (Pol) I and III. In this respect,
the snoRNAs which are processed from in-
trons of RNA PolII-derived transcripts are
an exception.
There is also another group of npcRNAs
that consists of regulatory RNA molecules
(riboregulators) which seem to play a
crucial role in many mechanisms con-
trolling expression of genetic information
(Fig. 8.2). Unlike the housekeeping RNAs,
the riboregulators are usually not constitu-
tively expressed. Their production often
depends on tissue type or developmental
stage. In some cases, they are induced in
response to biotic and abiotic changes in
the environment [12] (see also Part IV,
Chapter 10). Regulatory mechanisms in-
8.3 npcRNAs 215
Fig. 8.1 The contribution of coding (ORF) (gray)
and noncoding (white) sequences in genomes of
Homo sapiens (Hs), Drosophila melanogaster (Dm),
Saccharomyces cerevisiae (Sc) and Escherichia coli
(Ec).
volving npcRNAs have been identified in
both prokaryotes and eukaryotes [13]. The
properties of RNAs and the fact that they
encode themselves in the genome make
them ideal candidates for intracellular sig-
naling molecules responsible for the mod-
ulation of gene expression. An expression
of a gene of which end-product is RNA re-
quires less energy. RNAs are also more
easily degraded than the proteins which in
some cases may be important for more
precise control of the duration of the
signal. Yet another issue is a high specific-
ity. Since a significant fraction of the
npcRNAs acts through complementary in-
teractions with target mRNAs, they pro-
vide a better alternative to specific proteins
with specialized RNA-binding domains
(see also Part I, Chapter 10 and Part II,
Chapter 8). It has to be noted that in a
role of key factors controlling the develop-
ment and differentiation, the npcRNAs of-
fer an unrivalled plasticity that can be a
driving force of evolution [10, 14]. It is
now clear that without detailed knowledge
about this group of RNAs, a full under-
standing of the genome will not be possi-
ble.
An interesting feature of many of
npcRNAs is that they show a strong re-
semblance to protein-coding mRNAs. The
majority of them are the products of RNA
PolII. They possess a modified cap struc-
ture protecting the 5'-end and a poly(A)
tail at the 3'-end. The primary transcripts
are often subject to alternative splicing,
giving rise to multiple mature variants.
One of the largest npcRNA genes found to
date is human BCMS (B cell neoplasia-as-
sociated gene with multiple splicing),
which spans a region of over 560 kb of the
chromosome 13q14 [15]. The lengths of
mature transcripts among the npcRNAs
vary. There are transcripts which do not
exceed 1 kb, but there are also very large
unspliced RNAs, like mouse Air RNA
which is over 100 kb long [16]. The simi-
larities between protein and npcRNA
genes go beyond structural attributes. A
detailed analysis of human chromosomes
21 and 22 clearly demonstrated that the
transcriptional regulation of the majority
of npcRNAs depends on the same tran-
scription factors that control the expres-
sion of protein-coding genes [17]. It has
also been found in a genomic mapping of
mouse putative npcRNAs obtained during
a large-scale cDNA sequencing project that
about a quarter of the respective genes are
associated with CpG islands [18]. These
findings strongly indicate that, despite the
lack of protein-coding potential, the prod-
ucts of npcRNA genes are functional and
do not result from spurious transcription.
Fig. 8.2 The activities of regulatory npcRNAs affect all stages
of transmission of genetic information from DNA to proteins.
Several mammalian regulatory RNAs
are transcribed by RNA PolIII. This rela-
tively small group consists of brain-specific
BC1 and BC200 RNAs [19, 20], 7SK RNA
[21], and a recently identified mouse B2
RNA [22]. These RNAs are smaller than
the mRNA-like transcripts (100200 nt)
and they are not spliced. They are also re-
ferred to as small cytoplasmic RNAs.
Although the number of regulatory PolIII
transcripts known to date is much lower
than that of mRNA-like npcRNAs, this
may be due to the fact that they are more
difficult to identify. Polyadenylated
npcRNAs transcribed by PolII can be
cloned alongside the mRNAs, which is not
the case of PolIII transcripts. However,
one cannot exclude the possibility that the
repertoire of PolIII-dependent genes pro-
ducing regulatory RNAs is much larger.
Until recently, most of the known
npcRNAs were identified as novel tran-
scripts, differentially expressed in various
cells or tissue types, or in response to
changing environmental conditions. These
accidental findings, however, did not show
the full scale of the problem and npcRNAs
were sometimes regarded as curiosities or
remnants of the RNA world. The growing
number of known npcRNAs suggested
that they may constitute a significant por-
tion of the transcriptome and play impor-
tant roles in the regulation of molecular
mechanisms underlying gene expression.
Large numbers of npcRNAs have been
identified recently in large-scale projects
aimed at sequencing of full-length mouse
and human cDNAs [18, 2, 24]. Within the
mouse FANTOM2 data set comprising
over 60000 full-length cDNAs, it was pos-
sible to distinguish 33409 transcriptional
units, 15815 of which were classified as
nonprotein coding [18]. Those results evi-
dently demonstrated that the mRNA-like
npcRNAs constitute a significant compo-
nent of mammalian transcriptomes. Sub-
sequent analyses also revealed that many
of the noncoding transcripts identified in
the mouse cDNA set have their counter-
parts in human and rat genomes [25]. Cur-
rently, more than 800 unique mammalian
npcRNAs are known, two-thirds of which
are micro-RNAs (miRNAs) and snoRNAs.
Also, almost 20000 putative npcRNAs
have been identified in high-quality cDNA
libraries [26, 27]. They are largely of un-
known function, but some are known to
be developmentally regulated, disease-asso-
ciated, imprinted, expressed pseudogenes
or antisense transcripts (see also Part II,
Chapter 8 and Part III, Chapter 3).
8.4
Functions of npcRNAs
In contrast to the large amount of se-
quence data obtained from the genomic
projects, there are a relatively small num-
ber of npcRNAs with clearly identified
functions. However, unlike amino acid se-
quences which can be used for more or
less reliable predictions of protein func-
tions, nucleotide sequences of RNAs do
not allow us to draw conclusions about
their activities. npcRNAs have been impli-
cated in various mechanisms which affect
expression of genes at virtually all steps of
transmission of genetic information from
DNA to proteins. Certain npcRNAs are in-
volved in remodeling of the chromatin,
thereby changing its transcriptional activ-
ity. This is accomplished by recruiting pro-
tein factors which alter the methylation
status of the chromosomal DNA and acety-
lation of histones. One of the best-studied
mammalian npcRNAs is XIST (X-inactive
specific transcript) expressed from the XIC
region (X-inactivation center) of the inac-
tive X chromosome [28]. An association of
8.4 Functions of npcRNAs 217
XIST with the chromatin of the X chromo-
some provides a signal for the proteins
which catalyze methylation of the CpG is-
lands and deacetylation of histone H4,
thus establishing the inactive state [29]. A
similar mechanism involving transcrip-
tional inactivation of a large chromosomal
region has been proposed for the mouse
Air RNA. That paternally expressed tran-
script is produced from the antisense
strand of the Igf2r (insulin-like growth fac-
tor type 2 receptor) gene. Air expression
leads to silencing of a cluster of genes
spanning approximately 400 kb [16]. There
is also a possibility that in some cases the
establishment and maintenance of an
open chromatin conformation may require
active transcription, producing probably
nonfunctional npcRNAs from the inter-
genic regions as demonstrated for the hu-
man b-globins locus [30].
A large fraction of the mammalian
npcRNAs is produced from the genes
which fully or partially overlap protein cod-
ing-genes in an antisense orientation [7,
25]. A perfect complementarity between
large regions of npcRNA and the corre-
sponding mRNA offers several ways in
which the antisense transcript can affect
the level of expression of a protein. The
most straightforward explanation is that
interactions between complementary re-
gions prevent translation. Alternatively,
npcRNA may affect mRNA transport or
splicing [31]. RNA duplexes between sense
and antisense transcripts can also be sub-
strates for Dicer, thus inducing the RNA
interference pathway [32]. In some in-
stances, however, the functions of anti-
sense transcripts seem to be independent
of the overlapping protein-coding gene. In
many cases, sense and antisense tran-
scripts are produced at the same time, and
there is no correlation between the levels
of npcRNA transcription and expression of
a protein [17]. Antisense npcRNAs are of-
ten produced from imprinted genes and
their expression is usually tissue specific
or limited to certain stages of the develop-
ment. The protein-coding (sense) and
RNA-coding (antisense) transcripts are re-
ciprocally imprinted, and it seems that the
npcRNAs may be required for silencing of
the protein-coding genes as demonstrated
in the case of Air RNA [16]. An involve-
ment of antisense transcript in an epige-
netic control of gene expression has been
recently demonstrated for the Sphk1
(sphingosine kinase-1) gene. The genera-
tion of alternative Sphk1 subtypes relies on
a tissue-dependent differentially methy-
lated region (T-DMR) located within the
Sphk1 CpG island. It has been demon-
strated that the expression of antisense
transcript Khps1, overlapping with the T-
DMR, induces demethylation of CG sites
and methylation of non-CG sites within
the T-DMR [33].
Certain npcRNAs have been shown to
be directly involved in the regulation of
gene expression by influencing the activity
of transcription factors. A steroid receptor
activator (SRA) RNA forms a ribonucleo-
protein complex with the steroid receptor
co-activator (SRC)-1 protein, which is a
strong co-activator of nuclear receptors for
steroid hormones including progestins, es-
trogens, androgens and glucocorticoids
[34, 35]. The interaction between SRC-1
and SRA RNA is probably mediated by a
subfamily of DEAD-box RNA-binding pro-
teins, p72/p68 [36]. Another protein inter-
acting with SRA RNA is a hormone-in-
duced transcriptional repressor, SHARP.
Thus, the competition between SRA RNA
and steroid receptors and SHARP may be
responsible for the modulation of expres-
sion of hormone-regulated genes [37]. In-
terestingly, there are variants of the SRA
transcripts with an extension at the 5'-end
which apparently encode the proteins de-
tectable in vivo. Thus, the SRA-encoding
gene is the first one to yield two distinct
end-products: npcRNA and a protein [38].
There are two PolIII transcripts, 7SK
and B2 RNA, which interact with and con-
trol the activity of components of PolII
transcription apparatus. 7SK RNA was
identified as a specific regulator of a posi-
tive transcription elongation factor (P-
TEFb) which is a cofactor of HIV-1 Tat
protein required for transcription from the
viral promoter (see also Part II, Chapters 7
and 8) [21, 39]. 7SK RNA acts as a P-TEFb
inhibitor by suppressing its cyclin-depen-
dent kinase (CDK9) activity. It has been
demonstrated that 7SK RNA-bound P-
TEFb cannot form complexes with the vir-
al promoter in HIV-1 transcription assay.
Another PolIII transcript involved in the
regulation of PolII-transcribed genes is
mouse B2 RNA encoded by short inter-
spersed elements. The level of B2 tran-
scripts increases after heat shock (see also
Part IV, Chapter 10) [40]. This is accompa-
nied by the repression of mRNA transcrip-
tion which can be reversed by the treat-
ment with antisense oligonucleotides to
B2 RNA [21] (see also Part I, Chapter 9).
The suppression of PolII activity results
from the interaction of the core polymer-
ase with B2 RNA within stable, yet tran-
scriptionally inactive, pre-initiation com-
plexes at promoters [41].
8.5
npcRNAs and Human Diseases
A significant number of npcRNAs identi-
fied to date are implicated in certain hu-
man diseases (Tab. 8.1). The changes of ex-
pression levels of certain npcRNAs which
accompany the malignant process strongly
support the functional role of RNA. It is
especially evident in cases when the chro-
mosomal aberrations disrupt the npcRNA
genes. The most severe developmental dis-
orders often associated with mental retar-
dation result from genetic or epigenetic
defects affecting imprinted genes. Geno-
mic imprinting is a process whereby the
expression of an allele depends upon
whether it comes from a mother or a
father [42]. A number of npcRNAs identi-
fied in humans are products of imprinted
genes. An abnormal pattern of expression
from imprinted gene clusters can result in
severe congenital disorders like Prader-
Willi (PWS), Beckwith-Wiedemann (BWS)
or Angelman (AS) syndromes [43]. Cur-
rently available data show that npcRNA
molecules probably play a pivotal role in
establishing and maintaining of the im-
printed status of certain chromosomal re-
gions.
AS and PWS arise from defects in im-
printed genes located within the chromo-
some 15q11q13 region. The imprinted
cluster consists of one maternally and 11
paternally expressed genes. The PWS re-
sults from genomic alterations suppress-
ing the expression of paternally expressed
genes as a result of maternal uniparental
disomy for chromosome 15, deletion of pa-
ternally inherited 15q11q13 region, pater-
nally inherited balanced translocations or
imprinting mutations. One of the pater-
nally expressed genes is IPW encoding
several alternatively spliced, polyadenylated
npcRNAs present at the same levels in all
tissues [44]. Interestingly, the mouse Ipw
shows tissue-specific variations with very
high expression in brain [45]. AS is a con-
sequence of disrupted expression of a sin-
gle maternally expressed gene, UBE3A. In-
terestingly, the imprinted maternal expres-
sion of UBE3A is restricted only to certain
brain cells, while in most tissues the gene
is expressed from both alleles. Both in hu-
man and mouse, the suppression of the
paternal alleles of UBE3A/Ube3a genes de-
pends on the expression of the large,
460 kb long, antisense paternally expressed
transcript, but the underlying mechanism
is not known [46, 47]. UBE3A-AS introns
encode a number of snoRNA. It has been
proposed that abnormalities in the im-
printed expression of these RNAs may be
responsible for the origin of PWS [48].
Another cluster of four paternally and
13 maternally expressed genes is located
on human chromosome 11p15 associated
with several human cancers and BWS.
Within this region, there are two npcRNA
genes with distinct features. H19 RNA
was the first identified npcRNA expressed
from an imprinted gene. Unlike the ma-
jority of imprinted npcRNA-coding genes,
H19 represents an independent transcrip-
tional unit and does not overlap with any
other gene. H19 RNA is expressed at high
levels, exclusively from the maternal allele,
in many embryonic tissues. Shortly after
birth, the transcriptional activity extin-
guishes except for skeletal muscles [49].
There is no evidence of a protein product,
but all H19 RNA sequences can be folded
Table 8.1 npcRNAs linked to human disorders
npcRNA Disorder
BCMS Contains regions frequently deleted in B cell neoplasia
OCC1 Overexpressed in colon carcinoma
MALAT-1 Metastasis associated in lung adenocarcinoma
TRNG10 Expressed in various cancer cells
CMPD Campomyelic displasia
HOST2 Expressed in ovarian cancer cells
NSCLC Expressed in nonsmall cell lung carcinoma
NCRMS Increased expression in alveolar rhabdomyosarcoma
DD3 Overexpressed in prostate cancer
PCGEM1 Overexpressed in prostate cancer
RAY1/ST7 Disrupted in autistic disorder
DGCR5 Disrupted in DiGeorge syndrome
22k48 HIRA intronic transcript deleted in DiGeorge syndrome
PSZA11q14 Reduced expression in brains of patients with schizophrenia
DISC2 Disrupted in schizophrenia
MEN1 Multiple endocrine neoplasia type 1 locus transcripts
HANC Expressed in CD4
+
T lymphocytes infected with HTLV-1
SRA-Del Steroid receptor activator RNA isoform expressed in breast cancer
C6orf37OS Antisense transcript from C6orf37 locus within diffuse panbronchiolitis
critical region
PEG8/IGF2AS Overexpressed in fetal tumors
SCA8
(KLHL1 antisense)
Spinocerebellar ataxia type 8
MESTIT1 Russel-Silver syndrome
COPG2IT1 Russel-Silver syndrome
IPW Prader-Willi syndrome
LIT1 BWS, and Romano-Ward, Jervell and Lange-Nielsen syndromes
UBE3A-AS AS
H19 Overexpressed in certain tumors
into a common secondary structure, which
is a strong argument for its functionality
[50]. H19 was shown to have the proper-
ties of a tumor suppressor gene, since its
expression reduced tumorigenicity and
growth of certain malignant cell lines [51,
52]. Elevated expression of H19 in some
cancer types and promotion of tumor pro-
gression by cells expressing a H19 trans-
gene suggest, however, that H19 is an on-
cogene [53, 54]. There is a possibility that
the effects of H19 RNA are different for
certain tissue-specific splice forms [55].
This variability can also be due to differ-
ences in the H19 RNA-binding protein re-
pertoire present in different cells [56].
The second npcRNA encoded within the
11p15 region is LIT1 (long QT intronic
transcript 1, KvLQT1-AS, KCNQ1OT1).
This RNA is transcribed from the paternal
allele and is responsible for maintaining
the silent state of several maternally ex-
pressed genes implicated in BWS [57, 58].
It is possible that the mechanism of LIT1-
dependent gene silencing is analogous to
that of mouse Air RNA and involves re-
modeling of the chromatin [16].
npcRNAs have been mapped to chromo-
somal regions associated with certain neu-
robehavioral disorders, including autism,
bipolar affective disorder and schizophre-
nia. Genetic studies revealed that a num-
ber of schizophrenia patients carry a bal-
anced translocation t(1: 11)(q43,q14). The
breakpoint region of chromosome 1q43 in
a balanced translocation segregating with
schizophrenia are the two genes: DISC1
and DISC2 (disrupted in schizophrenia 1
and 2). DISC1 is a protein-coding gene,
while DISC2 produces a range of tran-
scripts 2.59.5 kb long without protein-
coding potential. These genes overlap in
antisense orientation with their 3'-terminal
regions and it has been suggested that
DISC2 RNA may be involved in the regu-
lation of DISC1 expression [59]. The same
translocation affects another npcRNA gene
located on chromosome 11q14. The gene
called PSZA11q14 (putative schizophrenia
associated gene from 11q14) shows signifi-
cantly reduced expression for the patients
with schizophrenia when compared to nor-
mal individuals. PSZA11q14 is antisense
to the first intron of the DLG2 gene,
which suggests that its product may be a
cis-antisense regulator of DLG2 [60].
RAY1/ST7 is a complex locus located on
chromosome 7q31. It produces two major
transcripts ST7 and RAY1: two npcRNAs,
ST7OT4 and ST7OT3, form the sense
strand and two antisense RNAs, ST7OT1
and ST7OT2, from the antisense strand.
There are at least 18 alternatively spliced
variants of all transcripts [61]. The genetic
abnormalities affecting the long arm of
chromosome 7 have been reported for au-
tistic patients. In one case, it has been
demonstrated that the translocation dis-
rupts the RAY1/ST7 locus [62]. As in most
cases, the role of npcRNAs transcribed
from this locus is not known, but it has
been suggested that the antisense tran-
scripts may somehow regulate the expres-
sion of the sense protein-coding tran-
scripts [61].
Several human npcRNAs show altered
expression in tumors when compared with
normal cells. Wilms and other fetal tu-
mors show a markedly elevated expression
of a paternally expressed PEG8/IGF2AS
RNA antisense to IGF2 gene. In Wilms
tumors PEG8/IGF2AS overexpression is
observed only in tumor cells and not in
the normal kidney cells [63]. In colon car-
cinoma cells, a novel npcRNA gene OCC-1
(overexpressed in colon carcinoma 1)
shows significantly higher levels of expres-
sion than the cells of normal mucosa [64].
Two npcRNAs are significantly overex-
pressed in prostate cancers. The DD3 was
identified in a differential display as a
prostate-specific gene which shows ele-
vated expression in over 90% of the pros-
tate tumors analyzed [65]. Another pros-
tate-specific npcRNA-coding gene overex-
pressed in tumor tissue is androgen-re-
sponsive PCGEM1 [66]. Its upregulation is
correlated with increased proliferation and
colony formation [67].
B cell chronic lymphocytic leukemia and
mantle cell lymphoma are often associated
with deletions within the chromosomal re-
gion 13q14.3. These deletions affect the
largest npcRNA gene identified so far.
BCMS spans around 560 kb and is com-
posed of at least 50 exons. Alternative
splicing is tissue specific, but none of the
mature variants have a significant protein-
coding potential [15].
The differences in expression of
npcRNAs can distinguish between different
subtypes of cancer cells. In nonsmall cell
lung cancer (NSCLC), the expression of
MALAT-1 gene (metastasis associated in
lung adenocarcinoma transcript 1), encod-
ing an 8-kb npcRNA, is an indicator of a
metastasizing form of lung cancer, which
makes it a good candidate for the diagnosis
of lung cancer patients [68] (see also Part I,
Chapters 2 and 3, and Part V, Chapter 9).
The two histological subtypes of rhabdo-
myosarcoma (RMS) can also be distin-
guished based on the expression of the
npcRNA gene, NCRMS (noncoding RNA
in RMS). An increased expression of
NCRMS is observed in the alveolar, but
not in the embryonal, subtype of RMS.
The expression of NCRMS may be asso-
ciated with transcriptional deregulation
within the large chromosomal region, in-
cluding myogenic regulators Myf5 and
Myf6, and a growth factor Igf2, which leads
to cancer development as in neuroblastoma
and synovial sarcoma, which show similar
patterns of NCRMS expression [69].
8.6
miRNAs
The largest family of eukaryotic npcRNAs
includes small, 20- to 25-nt translational
regulators called miRNAs. These tiny
RNAs are the smallest functional RNA
molecules identified to date [70]. The first
two miRNAs, lin-4 and let-7, were recog-
nized in Caenorhabditis elegans as tempo-
rally regulated developmental switches
controlling the timing and sequence of
events in post-embryonic development [71,
72]. In recent years, hundreds of new miR-
NAs have been discovered both in animals
and in plants [70]. A common feature of
all miRNAs is their biogenesis. They are
processed by a Dicer ribonuclease from
70- to 80-nt hairpin precursors (see also
Part I, Chapter 10 and Part II, Chapter 8).
For a long time, there was a controversy
concerning transcription of miRNA-encod-
ing genes. Recently, it has been shown
that their primary transcripts are capped
and polyadenylated PolII products [73] (see
also Part III, Chapter 3). In the best-stud-
ied case of the lin-4 RNA from C. elegans,
it has been demonstrated that the RNA
acts as a translational repressor of at least
two genes, lin-14 and lin-28 [74, 75]. The
translational repression of lin-14 mRNA
depends on the presence of seven short se-
quence elements within the 3'-UTR, with
partial complementarity to lin-4 RNA [76].
The interaction of miRNA with the 3'-UTR
of mRNA causes suppression of transla-
tion. It seems that this mechanism is com-
mon for all animal miRNAs [77]. The
mechanisms by which the inhibition of
translation by miRNAs is achieved are not
fully understood. It seems that apart from
the complementary interactions between
miRNA and mRNA, it may require the
presence of additional factors like the pro-
teins of the Argonaute family [77, 78].
These highly specialized proteins are in-
dispensable for processing of precursors of
miRNAs and seem to determine the fate
of various Dicer products, directing them
to different regulatory pathways [78]. A dif-
ferent way of miRNAs downregulating the
expression of target mRNA was observed
in plants [79]. It involves a specific cleav-
age of mRNA at the site of complementary
interactions with miRNA. It has been
demonstrated in a case of HOXB8 mRNA
and miR-196 that such a mechanism is
also valid for certain mammalian miRNAs
[80].
miRNAs are probably the most impor-
tant factors regulating the expression of
genes during development. Many of them
show striking conservation during evolu-
tion. A systematic survey revealed that let-
7 RNA is present and almost totally con-
served in all bilaterally symmetrical ani-
mals, which together with specific expres-
sion patterns in humans and Drosophila
suggests that it may be involved in the
regulation of development and/or differen-
tiation [81, 82]. In the human leukemia
HL-60 cell line, the profiles of miRNA ex-
pression change during cell differentiation
into monocyte/macrophage-like cells in-
duced by 12-O-tetradecanoylphorbol-13-ace-
tate [83]. This demonstrates that the al-
tered patterns of miRNAs may be respon-
sible for the changes in the cells genetic
program, which in extreme cases results
in malignant growth. The differences in
miRNA expression patterns have been ob-
served in various cancer cell lines. In colo-
rectal cancers, the levels of miR-24-2 dif-
fered up to 50-fold between the samples
[84]. Moreover, many of the miRNA-coding
regions on human chromosomes are lo-
cated within the regions linked to the ori-
gin of certain forms of cancer [85]. It has
been also demonstrated using a microar-
ray analysis approach that the patterns of
miRNA expression can be used for distin-
guishing between the subtypes of human
B cell chronic lymphocytic leukemia [86].
8.7
Future Prospects
Our present knowledge of the nature of
npcRNAs and RNA-mediated regulation of
cellular processes is still very superficial.
Only a small fraction of npcRNAs identi-
fied to date have been characterized in
terms of function or expression patterns.
There is no doubt that npcRNAs hold the
key to understanding of functioning, devel-
opment and evolution of complex biologi-
cal systems. There are three crucial ques-
tions that have to be answered: (1) How
many npcRNAs are encoded in the ge-
nome? (2) What is their role in the cell?
(3) What is the mechanism of their action?
To answer these questions we need to de-
termine the expression profile for each
npcRNA. This includes tissue distribution,
developmental timing and possible induc-
tion under changing conditions (stress,
hormones, etc.). The most difficult task,
however, seems to be functional character-
ization. The large size of some of the
npcRNAs and possible interactions with
many cellular components make such
analyses unrealistic in vitro.
As demonstrated by several examples,
npcRNAs are potentially good markers for
human diseases including early detection
of certain forms of cancer. Systematic
studies on npcRNA expression profiles
may lead to the development of highly ac-
curate molecular diagnostic tools, which
could be used not only for detection, but
also for prognosis and the development of
modern biopharmaceuticals.
8.7 Future Prospects 223
Acknowledgments
This work was supported by the grants
from the Polish State Committee for Sci-
entific Research to J. B., and from the
Fonds der Chemischen Industrie e.V., the
Bundesministerium fr Wissenschaft, For-
schung und Technologie and the National
Foundation for Cancer Research to V. A. E.
References
1 Lander, E. S., Linton, L. M., Birren, B., et al.
2001. Initial sequencing and analysis of the
human genome. Nature 409, 860921.
2 Waterston, R. H., Lindblad-Toh, K., Birney, E.,
et al. 2002. Initial sequencing and compara-
tive analysis of the mouse genome. Nature
420, 520562.
3 Parra, G., Agarwal, P., Abril, J. F., et al. 2003.
Comparative gene prediction in human and
mouse. Genome Res. 13, 108117.
4 Venter, C. J., Adams, M. D., Myers, E. W., et al.
2001. The sequence of the human genome.
Science 291, 13041351.
5 Kumar A., Harrison, P. M., Cheung, K. H., et
al. 2002. An integrated approach for finding
overlooked genes in yeast. Nat. Biotechnol. 20,
5863.
6 Misra, S., Crosby, M. A., Mungall, C. J., et al.
2002. Annotation of the Drosophila melano-
gaster euchromatic genome: a systematic re-
view. Genome Biol. 3, research0083.
7 Yelin, R., Dahary, D., Sorek, R., et al. 2003.
Widespread occurrence of antisense transcrip-
tion in the human genome. Nat. Biotechnol.
21, 379386.
8 Veeramachaneni, V., Makalowski, W., Gald-
zicki, M., et al. 2004. Mammalian overlapping
genes: the comparative perspective. Genome
Res. 14, 280286.
9 Shabalina, S. A., Spiridonov, N. A. 2004. The
mammalian transcriptome and the function
of non-coding DNA sequences. Genome Biol.
5, 105.
10 Mattick, J. S. 2001. Non-coding RNAs: the ar-
chitects of eukaryotic complexity. EMBO Rep.
2, 986991.
11 Szymanski, M., Erdmann, V. A., Barciszewski,
J. 2003. Noncoding regulatory RNAs database.
Nucleic Acids Res. 31, 429431.
12 Scherer, S. W., Cheung, J., MacDonald, J. R., et
al. 2003. Human chromosome 7: DNA se-
quence and biology. Science 300, 767772.
13 Szymanski, M., Barciszewski, J. 2003. Regula-
tion by RNA. Int. Rev. Cytol. 231, 197258.
14 Herbert, A. 2004. The four Rs of RNA-directed
evolution. Nat. Genet. 36, 1925.
15 Wolf, S., Mertens, D., Schaffner, C., et al.
2001. B-cell neoplasia associated gene with
multiple splicing (BCMS): the candidate B-
CLL gene on 13q14 comprises more than
560 kb covering all critical regions. Hum. Mol.
Genet. 10, 12751285.
16 Sleutels, F., Zwart, R., Barlow, D. P. 2002. The
non-coding Air RNA is required for silencing
autosomal imprinted genes. Nature 415, 810
813.
17 Cawley, S., Bekiranov, S., Ng, H. H., et al.
2004. Unbiased mapping of transcription fac-
tor binding sites along human chromosomes
21 and 22 points to widespread regulation of
noncoding RNAs. Cell 116, 499509.
18 Numata, K., Kanai, A., Saito, R., et al. 2003.
Identification of putative noncoding RNAs
among the RIKEN mouse full-length cDNA
collection. Genome Res. 13, 13011306.
19 Martignetti, J. A., Brosius, J. 1993. BC200
RNA: a neural RNA polymerase III product
encoded by a monomeric Alu element. Proc.
Natl Acad. Sci. USA 90, 1156311567.
20 Martignetti, J. A., Brosius, J. 1995. BC1 RNA:
transcriptional analysis of a neural cell-specific
RNA polymerase III transcript. Mol. Cell. Biol.
15, 16421650.
21 Nguyen, V. T., Kiss, T., Michels, A. A., et al.
2001. 7SK small nuclear RNA binds to and in-
hibits the activity of CDK9/cyclin T com-
plexes. Nature 414, 322325.
22 Allen, T. A., Von Kaenel, S., Goodrich, J. A., et
al. 2004. The SINE-encoded mouse B2 RNA
represses mRNA transcription in response to
heat shock. Nat. Struct. Mol. Biol. 11, 816821.
23 Ota, T., Suzuki, Y., Nishikawa, T., et al. 2004.
Complete sequencing and characterization of
21,243 full length human cDNAs. Nat. Genet.
36, 4045.
24 Strausberg, R. L., Feingold, E. A., Grouse L. H.,
et al. 2002. Generation and initial analysis of
more than 15000 full-length human and
mouse cDNA sequences. Proc. Natl Acad. Sci.
USA 99, 1689916903.
25 Suzuki, M., Hayshizaki, Y. 2004. Mouse-cen-
tric comparative transcriptomics of protein
and non-coding RNAs. BioEssays 26, 833843.
26 Imanishi, T., Itoh, T., Suzuki, Y., et al. 2004.
Integrative annotation of 21,037 human genes
validated by full-length cDNA clones. PLoS
Biol. 2, E162.
27 Pang, K. C., Stuart S., Pr G., et al. 2005.
RNAdb a comprehensive mammalian non-
coding RNA database. Nucleic Acids Res. 33
(Database issue), D125D130.
28 Andersen A. A., Panning, B. 2003. Epigenetic
gene regulation by noncoding RNAs. Curr.
Opin. Cell Biol. 15, 281289.
29 Gilbert, S. L., Sharp, P. A. 1999. Promoter-spe-
cific hypoacetylation of X-inactivated genes.
Proc. Natl Acad. Sci. USA 96, 1382513830.
30 Gribnau, J, Diderich, K., Pruzina, S., et al.
2000. Intergenic transcription and develop-
mental remodeling of chromatin subdomains
in the human beta-globin locus. Mol. Cell. 5,
377386.
31 Lehner, B., Williams, G., Campbell, R. D., et
al. 2002. Antisense transcripts in the human
genome. Trends Genet. 18, 6365.
32 Novina, C. D., Sharp, P. A. 2004. The RNAi re-
volution. Nature 430, 161164.
33 Imamura, T., Yamamoto, S., Ohgane, J., et al.
2004. Non-coding RNA directed DNA de-
methylation of the Sphk1 CpG island. Bio-
chem. Biophys. Res. Commun. 322, 593600.
34 Lanz, R. B., McKenna, N. J., Onate, S. A., et al.
1999. A steroid receptor coactivator, SRA,
functions as an RNA and is present in an
SRC-1 complex. Cell 97, 727.
35 Lanz, R., Razani, B., Goldberg, A. D., et al.
2002. Distinct RNA motifs are important for
coactivation of steroid hormone receptors by
steroid receptor RNA activator (SRA). Proc.
36 Watanabe, M. Yanagisawa, J., Kitagawa, H., et
al. 2001. A subfamily of RNA-binding DEAD-
box proteins acts as an estrogen receptor a
coactivator through the N-terminal activation
domain (AF-1) with an RNA coactivator, SRA.
EMBO J. 20, 13411352.
37 Shi, Y., Downes, M., Xie, W., et al. 2001.
SHARP, an inducible cofactor that integrates
nuclear receptor repression and activation.
Genes Dev. 15, 11401151.
39 Emberley, E., Huang, G. J., Hamedani, M. K.,
et al. 2003. Identification of new human cod-
ing steroid receptor RNA activator isoforms.
Biochem. Biophys. Res. Commun. 301, 509515.
39 Li, T., Spearow, J., Rubin, C. M., et al. 1999.
Physiological stresses increases mouse short
interspersed element (SINE) RNA expression.
Gene 239, 367372.
40 Yang, Z., Zhu, Q., Luo, K., et al. 2001. The
7SK small nuclear RNA inhibits the CDK9/cy-
clin T1 kinase to control transcription. Nature
414, 317322.
41 Espinoza, C. A., Allen, T. A., Hieb, A. R., et al.
2004. B2 RNA binds directly to RNA polymer-
ase II to repress transcript synthesis. Nat.
Struct. Mol. Biol. 11, 822829.
42 Bartolomei, M. S., Tilghman, S. M. 1997.
Genomic imprinting in mammals. Annu. Rev.
Genet. 31, 493525.
43 Nicholls, R. D. 2000. The impact of genomic
imprinting for neurobehavioral and develop-
mental disorders. J. Clin. Invest. 105, 413418.
44 Wevrick, R., Kerns, J. A., Francke, U. 1994.
Identification of a novel paternally expressed
gene in the Prader-Willi syndrome region.
Hum. Mol. Genet. 3, 18771882.
45 Wevrick, R., Francke, U. 1997. An imprinted
mouse transcript homologous to the human
imprinted in Prader-Willi syndrome (IPW)
gene. Hum. Mol. Genet. 6, 325332.
46 Chamberlain, S. J., Brannan, C. I. 2001. The
Prader-Willi syndrome imprinting center acti-
vates the paternally expressed murine Ube3a
antisense transcript but represses paternal
Ube3a. Genomics 73, 316322.
47 Runte, M., Kroisel, P. M., Gillessen-Kaesbach,
G., et al. 2004. SNURF-SNRPN and UBE3A
transcript levels in patients with Angelman
syndrome. Hum. Genet. 114, 553561.
48 Runte, M., Huttenhofer, A., Gross, S., et al.
2001. The IC-SNURF-SNRPN transcript
serves as a host for multiple small nucleolar
RNA species and as an antisense RNA for
UBE3A. Hum. Mol. Genet. 10, 26872700.
49 Looijenga, L. H., Verkerk, A. J., De Groot, N.,
et al. 1997. H19 in normal development and
neoplasia. Mol. Reprod. Dev. 46, 419439.
50 Juan, V., Crain, C., Wilson, C. 2000. Evidence
for evolutionarily conserved secondary struc-
ture in the H19 tumor suppressor RNA. Nu-
cleic Acids Res. 28, 12211227.
References 225
51 Hao, Y., Crenshaw, T., Moulton, T., et al. 1993.
Tumour-suppressor activity of H19 RNA. Na-
ture 365, 764767.
52 Isfort, R. J., Cody, D. B., Kerckaert, G. A., et al.
1997. Role of the H19 gene in Syrian hamster
embryo cell tumorigenicity. Mol. Carcinogen.
20, 189193.
53 Ariel, I., Sughayer, M., Fellig, Y., et al. 2000.
The imprinted H19 gene is a marker of early
recurrence in human bladder carcinoma. Mol.
Pathol. 53, 320323.
54 Lottin, S., Adriaenssens, E., Dupressoir, T., et
al. 2002. Overexpression of an ectopic H19
gene enhances the tumorigenic properties of
breast cancer cells. Carcinogenesis 23, 1885
1895.
55 Matouk, I., Ayesh, B., Schneider, T., et al.
2004. Oncofetal splice-pattern of the human
H19 gene. Biochem. Biophys. Res. Commun.
318, 916919.
56 Ioannidis, P., Kottaridi, C., Dimitriadis, E., et
al. 2004. Expression of the RNA-binding pro-
tein CRD-BP in brain and non-small cell lung
tumors. Cancer Lett. 209, 245250.
57 Horike, S, Mitsuya, K., Meguro, M., et al.
2000. Targeted disruption of the human LIT1
locus defines a putative imprinting control
element playing an essential role in Beckwith-
Wiedemann syndrome. Hum. Mol. Genet. 9,
20752083.
58 Smilinich, N. J., Day, C. D., Fitzpatrick, G. V.,
et al. 1999. A maternally methylated CpG is-
land in KvLQT1 is associated with an anti-
sense paternal transcript and loss of imprint-
ing in Beckwith-Wiedemann syndrome. Proc.
59 Millar, J. K., Wilson-Annan, J. C., Anderson,
S., et al. 2000. Disruption of two novel genes
by a translocation co-segregating with schizo-
phrenia. Hum. Mol. Genet. 9, 14151423.
60 Polesskaya, O. O., Haroutunian, V., Davis,
K. L., et al. 2003. Novel putative nonprotein-
coding RNA gene from 11q14 displays de-
creased expression in brains of patients with
schizophrenia. J. Neurosci. Res. 74, 111122.
61 Vincent, J. B., Petek, E., Thevarkunnel, S., et
al. 2002. The RAY1/ST7 tumor-suppressor lo-
cus on a chromosome 7q31 represents a com-
plex multi-transcript system. Genomics 80,
283294.
62 Vincent, J. B., Herbrick, J. A., Gurling, H. M.,
et al. 2000. Identification of a novel gene on
chromosome 7q31 that is interrupted by a
translocation breakpoint in an autistic individ-
ual. Am. J. Hum. Genet. 67, 510514.
63 Okutsu, T., Kuroiwa, Y., Kagitani, F., et al.
2000. Expression and imprinting status of hu-
man PEG8/IGF2AS, a paternally expressed
antisense transcript from the IGF2 locus, in
Wilms tumors. J. Biochem. 127, 475483.
64 Pibouin, L., Villaudy, J., Ferbus, D., et al.
2002. Cloning of the mRNA of overexpression
in colon carcinoma-1: a sequence overex-
pressed in a subset of colon carcinomas. Can-
cer Genet. Cytogenet. 133, 5560.
65 Bussemakers, M. J., van Bokhoven A., Ver-
haegh G. W., et al. 1999. DD3: a new prostate-
specific gene, highly overexpressed in prostate
cancer. Cancer Res. 59, 59755979.
66 Srikantan, V., Zou, Z., Petrovics, G., et al.
2000. PCGEM1, a prostate-specific gene, is
overexpressed in prostate cancer. Proc. Natl
Acad. Sci. USA 97, 1221612221.
67 Petrovics, G., Zhang, W., Makarem, M., et al.
2004. Elevated expression of PCGEM1, a pros-
tate-specific gene with cell growth-promoting
function, is associated with high-risk prostate
cancer patients. Oncogene 23, 605611.
68 Ji, P., Diederichs, S., Wang, W., et al. 2003.
MALAT-1, a novel noncoding RNA, and thy-
mosin beta4 predict metastasis and survival in
early-stage non-small cell lung cancer. Onco-
gene 22, 80318041.
69 Chan, A. S., Thorner, P. S., Squire, J. A., et al.
2002. Identification of a novel gene NCRMS
on chromosome 12q21 with differential ex-
pression between Rhabdomyosarcoma types.
Oncogene 21, 30293037.
70 Moss, E. G. 2003. MicroRNAs. In Noncoding
RNAs: Molecular Biology and Molecular Medi-
cine. J. Barciszewski, V. A. Erdmann (eds.).
Landes, Georgetown, TX, pp. 98115.
71 Lee, R. C., Feinbaum, R. L., Ambros, V. 1993.
The C. elegans heterochromic gene lin-4 en-
codes small RNAs with antisense complemen-
tarity to lin-14. Cell 75, 843854.
72 Reinhart, B. J., Slack, F. J., Basson, M., et al.
2000. The 21-nucleotide let-7 RNA regulates
developmental timing in Caenorhabditis ele-
gans. Nature 403, 901906.
73 Lee, Y., Kim, M., Han, J., et al. 2004. Micro-
RNA genes are transcribed by RNA polymer-
ase II. EMBO J. 23, 40514060.
74 Wightman, B., Ha, I., Ruvkun, G. 1993. Post-
transcriptional regulation of the heterochro-
mic gene lin-14 by lin-4 mediates temporal
pattern formation in C. elegans. Cell 75, 855
862.
75 Moss, E. G., Lee, R. C., Ambros, V. 1997. The
cold shock domain protein LIN-28 controls de-
velopmental timing in C. elegans and is regu-
lated by the lin-4 RNA. Cell 88, 637646.
76 Ha, I., Wightman, B., Ruvkun, G. 1996. A
bulged lin-4/lin-14 RNA duplex is sufficient
for Caenorhabditis elegans lin-14 temporal gra-
dient formation. Genes Dev. 10, 30413050.
77 Bartel D. P., 2004. MicroRNAs: genomics, bio-
genesis, mechanism, and function. Cell 116,
281297.
78 Carmell, M. A., Xuan, Z., Zhang, M. Q., et al.
2002. The Argonaute family: tentacles that
reach into RNAi, developmental control, stem
cell maintenance, and tumorigenesis. Genes
Dev. 16, 27332742.
79 Dostie, J., Mourelatos, Z., Yang, M., et al.
2003. Numerous microRNPs in neuronal cells
containing novel microRNAs. RNA 9, 180
186.
80 Yekta, S., Shih, I. H., Bartel, D. P. 2004. Micro-
RNA-directed cleavage of HOXB8 mRNA.
Science 304, 594596.
81 Pasquinelli, A. E., Reinhart, B. J., Slack, F., et
al. 2000. Conservation of the sequence and
temporal expression of let-7 heterochromic
regulatory RNA. Nature 408, 8689.
82 Mansfield, J. H., Harfe, B. D., Nissen, R., et al.
2004. MicroRNA-responsive sensor trans-
gene uncover Hox-like and other developmen-
tally regulated patterns of vertebrate micro-
RNA expression. Nat. Genet. 10, 10791083.
83 Kasashima, K., Nakamura, Y., Kozu, T. 2004.
Altered expression profiles of microRNAs dur-
ing TPA-induced differentiation of HL-60
cells. Biochem. Biophys. Res. Commun. 322,
403410.
84 Schmittgen, T. D., Jiang, J., Liu, Q., et al.
2004. A high-throughput method to monitor
the expression of microRNA precursors. Nu-
cleic Acids Res. 32, e43.
85 Calin, G. A., Sevignani, C., Dumitru, C. D., et
al. 2004. Human microRNA genes are fre-
quently located at fragile sites and genomic
regions involved in cancers. Proc. Natl Acad.
Sci. USA 101, 29993004.
86 Calin, G. A., Liu, C. G., Sevignani, C., et al.
2004. MicroRNA profiling reveals distinct sig-
natures in B cell chronic lymphocytic leuke-
mias. Proc. Natl Acad. Sci. USA 101, 11755
11760.
References 227
Abstract
Double-stranded decoy oligodeoxynucleo-
tides (ODNs) represent a new class of po-
tential therapeutic drugs which can be de-
signed to target specifically transcription
factors involved in the pathogenesis of a giv-
en disease. There has been an explosion in
the use of transcription factor decoys as
tools for studying gene regulation and as ex-
perimental therapy to treat a variety of
pathological conditions. Ongoing preclini-
cal and clinical development programs at
various emerging biotech companies, as
well as academic research institutions, are
currently elucidating the potential of this
promising new class of biopharmaceuticals.
Abbreviations
AHR airway hyperresponsiveness
AP-1 activator protein-1
AV arteriovenous
C/EBP CCAAT/enhancer binding pro-
tein
CABG coronary artery bypass graft
CD Cluster of Differentiation
cdk2 cyclin-dependent kinase 2
CIA collagen-induced arthritis
c-myc c-myc oncogene
CRE cyclic AMP response element
DNA desoxy nucleic acid
dODN decoy oligodeoxynucleotide
E2F E2F transcription factor
EGF epidermal growth factor
ER estrogen receptor
ERE estrogen response element
ET-1 endothelin-1
FITC fluorescein isothiocyanate
HVJ hemagglutinating virus of Japan
IL interleukin
INF interferon
IRF-1 Interferon Regulatory Factor 1
MCP-1 monocyte chemotactic protein-1
mRNA messenger ribonucleic acid
NF-jB nuclear factor jB
ODN oligodeoxynucleotide
OVA ovalbumin
PCNA proliferating cell nuclear anti-
gen
PTCA percutaneous transluminal coro-
nary angioplasty
RNA ribonucleic acid
SCCHN squamous cell carcinoma of the
head and neck
siRNA short interfering RNA
STAT signal transducer and activator
of transcription
TAR transactivating response
Th T-helper
229
9
Double-stranded Decoy Oligonucleotides as new Biopharmaceuticals
ISBN: 3-527-31184-X
TNFa tumor necrosis factor-alpha
VCAM-1 vascular cell adhesion mole-
cule 1
9.1
Introduction
Transcription factors are DNA-binding pro-
teins which bind to the promoter regions
of one or several genes in the cell nucleus,
thereby controlling the expression of the
corresponding proteins [1]. These factors
are the master switches of gene expres-
sion, able to turn genes on and off with
an outstanding selectivity and sensitivity
in binding to promoter regulatory ele-
ments [2]. By activation and translocation
to the cellular nucleus, transcription fac-
tors serve as signal transduction mediators
in important biological processes including
embryonic development and cell differen-
tiation. Current results of the Human Ge-
nome Project suggest that about 50% of
all human genes are transcription factors.
The complexity and diversity of organisms
seems to be in part caused by the restric-
tion of expression of certain transcription
factors to specific cell types and/or to spe-
cific stages of development [3]. However,
transcription factors are not only impor-
tant control elements in mechanisms gov-
erning cellular differentiation and develop-
ment, but are also implicated in the patho-
genesis of many diseases. Because tran-
scriptional regulation is dependent on the
transcription factor activation by various
different stimuli, these regulatory proteins
became an attractive target for therapeutic
intervention. Particularly interesting ap-
pears the development of new pharmaceu-
tical interventions for the treatment of dis-
eases characterized by aberrant activation
and expression of genes whose products
are involved in the initiation and progres-
sion of a disease, including chronic in-
flammatory diseases and cancer [4].
Recent progress in molecular biology
has spurred the development of new tech-
niques for specifically inhibiting expres-
sion of a target gene. Antisense oligode-
oxynucleotides (ODN) are single-stranded
synthetic DNA molecules that after inter-
nalization hybridize with the mRNA of the
target gene, hence preventing its transla-
tion [5]. RNA interference (RNAi) is a cel-
lular mechanism to regulate gene expres-
sion targeting mRNA by double-stranded
small interfering RNA molecules (siRNA).
In contrast to these molecular interven-
tions, which are effective downstream at
the level of mRNA processing, it seems to
be of therapeutic advantage to interfere
with disease-mediating gene expression
more upstream at the level of transcription
factors. Thereby, pathological gene activa-
tion can be prevented early in the molecu-
lar events of a disease.
The specific interaction between enhan-
cer-containing molecules and cellular com-
ponents was first described by Schler and
Gruss in 1984 [1, 6]. In 1990, Bielinska et
al. [7] reported the application of cis-ele-
ment double-stranded decoy oligodeoxynu-
cleotides (decoy ODNs) as a powerful tool
in the study of transcriptional regulation
of genes and suggested a potential thera-
peutic application of this new DNA-based
pharmacological tool. The therapeutic po-
tential of the decoy ODN approach was
first demonstrated at the Falk cardiovascu-
lar research center of Stanford University
in work by Morishita et al. [8] in an in-vivo
cardiovascular animal model of restenosis.
Subsequently, decoy ODNs were employed
in several other animal models for the
treatment of cardiovascular diseases (for
reviews, see Refs. [911]). Short double-
stranded decoy ODN molecules imitate
the DNA binding promoter region of a
specific transcription factor. As a conse-
quence, the transcription factor is being
decoyed, thereby interfering with the
binding of the transcription factor to the
promoter region with subsequent preven-
tion of gene expression. Thus, the objec-
tive of this molecular intervention is to
cause a decrease of the interactions of
trans-factors with the target genomic cis-
elements, leading to alteration of transcrip-
tion.
9.1.1
Mechanism of Action
Known transcription factors are grouped
on the basis of shared DNA-binding mo-
tifs. Each transcription factor contains one
or more DNA binding domains which
bind to regulatory DNA promoter ele-
ments in a sequence-specific manner [3].
Decoy ODNs mimic specific consensus
binding sequences (cis-elements). These
consensus binding sequences are con-
served across several species with respect
to the length and the individual conserva-
tion of each nucleotide position within the
binding site. After entering the cells, decoy
ODNs compete for the binding of endoge-
nous trans-acting transcription factors,
thereby preventing the binding of tran-
scription factors to the endogenous cis-ele-
ments present within regulatory regions of
target genes. Eventually, the inhibition of
promoter activation results in suppression
of gene activation and subsequent RNA
processing (Fig. 9.1).
Since 1990, several researchers have
tried to influence transcription factor acti-
vation using molecular strategies. Initially,
overexpression of TAR-containing se-
quences (TAR RNA decoys) in a double-
copy murine retroviral vector was used to
render cells resistant to HIV replication
[12]. The big disadvantage of such RNA
decoys is the problematic continuous de-
coy expression in vivo, which may lead to
alterations of physiological cellular func-
tions. In contrast, the approach of employ-
ing synthetic double-stranded DNA decoy
ODN is particularly attractive for the fol-
lowing reasons:
The synthesis and hybridization of se-
quence-specific decoy ODNs is relatively
simple.
The potential drug targets (transcription
factors) are well conserved and easily
identifiable.
Fig. 9.1 Mechanism of action of
decoy oligodeoxynucleotides (ODN):
Decoy ODN therapeutics inhibit
transcription factors by imitating
their DNA binding sites
(dODN=decoy oligonucleotide,
TF=transcription factor).
Decoy ODN may be more effective than
antisense ODN in blocking constitutively
expressed transcription factors as well as
multiple transcription factors that bind
to the same cis element.
Direct cytosolic action leads to easier ac-
cess to target (transcription factor).
As mentioned earlier, the first experi-
mental evidence of in vivo efficacy of decoy
ODN was provided by Morishita et al. [8]
by employing a transfer of E2F-decoy
ODN in conjunction with Sendai virus li-
posomes into balloon-traumatized rat caro-
tid arteries, thereby inhibiting injury in-
duced smooth muscle cell proliferation
(neointimal lesion formation).
9.1.2
Delivery
Most nucleic acid-based drugs, such as
antisense or new RNAi-based drug candi-
dates, show weak cellular uptake without
the addition of uptake-enhancing agents
[9, 13] (see Part VI, Chapter 6).
Both mechanical means and transfection
reagents, among others, have been used to
facilitate the cellular uptake of oligonucleo-
tides. The application of intraluminal pres-
sure enhances the uptake of particular oli-
gonucleotides in vascular tissues such as
carotid arteries or venous bypass grafts
[14, 15]. Other approaches use chemical
modifications in order to secondarily mod-
ify the nucleic acid backbone [16, 17]. In
general, these modifications increase up-
take through the cell membrane based on
the classical receptor-mediated endocytosis
pathway. However, once inside the cell,
most nucleic acid compounds taken up by
endocytosis are ultimately trapped in the
lysosomal compartment.
In contrast to most other nucleic acid-
based therapeutic approaches, decoy ODN
generally do not require any auxiliary
means such as transfection reagents to
achieve efficient uptake by a variety of
cells. Recently, a cellular uptake mecha-
nism for short double-stranded DNA mole-
cules, which has also been shown to be
amenable to pharmacological modulation,
has been identified (Wagner et al., unpub-
lished observation). This carrier-dependent
transport route employs a membrane pro-
tein, which promotes the active uptake of
decoy ODNs. In human endothelial cells,
this occurs via a naturally occurring ion-
dependent transporter mechanism.
Decoy ODNs are very stable and readily
water-soluble, thereby enabling straightfor-
ward formulation development for intrave-
nous, pulmonary (aerosol for inhalation),
or topical (ointment) applications. Conse-
quently, local decoy ODN therapy is cur-
rently the predominant strategy for prod-
uct development by various biotech com-
panies worldwide (Corgentech Inc., South
San Francisco, USA; Avontec GmbH, Mu-
nich, Germany; AnGes MG, Osaka, Japan).
A clear advantage of local drug application
is the lack or significant reduction of
potential side effects potentially emerging
from systemic exposure to a drug.
9.2
Therapeutic Decoy ODN Application
Potential therapeutic benefits of decoy
ODNs have been described for various dis-
eases, as summarized in Table 9.1.
9.2.1
Inflammation (STAT-1, IRF-1, NF-jB)
A novel intracellular signaling pathway
was discovered in the early 1990s which,
after phosphorylation of cytokine receptors
by cytoplasmic Janus kinases, leads to acti-
vation of latent factors, later were termed
Signal Transducer and Activator of Tran-
scription (STAT) [1820]. Since then, rapid
scientific progress in this field has in-
creased the understanding that STAT mol-
ecules, after activation and translocation to
the nucleus, transactivate selective genes
and are involved in the regulation of a
multitude of physiological and pathophys-
iological conditions, including cellular and
humoral defense [20], growth [21, 22], lac-
tation, and normal mamma development
[2325].
STAT proteins have been shown to be
critically involved in inflammatory pro-
cesses of several immune and proliferative
disorders [26]. Among them, STAT-1 is ac-
tivated in response to many lymphocyte-ac-
tivating cytokines, mainly the interferons,
and is essential for cell-mediated immu-
nity. Interferon-c biases the immune sys-
tem towards a so-called Th1 response pro-
viding the rationale for the observation
that exaggerating Th1 responses predomi-
nantly underlie chronic inflammatory dis-
eases.
In 1999, Hecker and co-workers at the
University of Gttingen developed several
decoy ODNs targeting transcription factors
which play a key role in the expression of
several genes involved in inflammatory
processes [27, 28]. In rat and human cells,
STAT-1 induces expression of CD40 (part
of the important co-stimulatory CD40/
CD154 [CD40 ligand] signaling pathway)
in antigen-presenting cells and B cells, re-
spectively, after contact with pro-inflamma-
tory cytokines. STAT-1 is also responsible
for the de-novo synthesis of interferon reg-
ulatory factor-1 (IRF-1), which by itself in-
duces expression of the CD40 receptor
gene (Fig. 9.2). This is part of the co-stim-
ulatory CD40/CD154 receptor/ligand com-
plex, which is formed on antigen-present-
ing cells, such as dendritic cells and B
lymphocytes, during a so-called Th2-
mediated immune response.
Accordingly, STAT-1 is centrally involved
in the early regulation and manifestation
of immune responses following either the
Th1 or Th2 direction in establishing
chronic immune disease [29]. Inhibition of
STAT-1 is therefore expected generally to
attenuate an exaggerated inflammatory re-
action in chronic disease. A possible role
for STAT-1 in allergic airway inflammation
was suggested by Sampath and co-workers
[30], who demonstrated that STAT-1 is con-
Table 9.1 Potential therapeutic benefits of decoy oligodeoxynucleotides (ODNs).
Target of decoy ODN Potential clinical application Reference
AP-1 Restenosis
Diabetic nephropathy
44, 45
C/EBP Restenosis 46
E2F Bypass graft vasculopathy
Glomerular diseases
58, 65, 66
67
NF-jB Arthritis
Restenosis
Atopic dermatitis
35
50
37
STAT-1 Allergic asthma 68
STAT-3 Cancer 60
Estrogen response element (ERE) Breast cancer 62
Cyclic AMP response element (CRE) Cancer 61
stitutively expressed by bronchial epithelial
cells, thus inducing the up-regulation of
co-stimulatory molecules such as CD40
critical for the development of the disease.
Thus, potentially chronic inflammatory
diseases like allergic asthma or psoriasis
may greatly benefit from a local decoy
ODN therapy approach applied via stan-
dard inhalation or topical ointment, re-
spectively.
Hamelmann and co-workers at Hum-
boldt University Berlin applied a STAT-1
decoy ODN locally to the bronchial system
of BALB/c mice which were systemically
sensitized to ovalbumin (OVA) and chal-
lenged with OVA via the airways [38]. A
single application of decoy ODN markedly
and significantly reduced interleukin (IL)-5
production and numbers of eosinophils
and lymphocytes in bronchoalveolar lavage
fluid, and inhibited development of in vivo
airway hyperresponsiveness (AHR), com-
pared to sensitized, challenged controls
(Fig. 9.3). In association with decreased air-
way inflammation and AHR, expression
levels of CD40 in peribronchial infiltrates,
and of vascular cell adhesion molecule-1
(VCAM-1) on vascular endothelial cells, re-
spectively, were significantly reduced.
These data indicate that local application
of STAT-1 decoy ODN effectively inhibits
allergen-induced bronchial hyperreactivity
by potentially attenuating the up-regula-
Fig. 9.2 Effects of different decoy ODN on cyto-
kine-stimulated CD40 expression. Western blot de-
monstrating the effect of a STAT-1 or IRF-1 con-
sensus decoy ODN (4 h pre-incubation, 10 lM)
on TNFa (100 U mL
1
) plus INFc (1000 U ml
1
)
stimulated CD40 protein expression in human
monocytes (THP-1 cell line) after 14 h [27].
Fig. 9.3 Aerosolized STAT-1 decoy ODN decreases
development of airway hyperreactivity in allergen-
mediated bronchial asthma in mice. Airway re-
sponsiveness to aerosolized methacholine was
measured in unrestrained, conscious mice and
bronchial hyperresponsiveness (P
enh
) values were
determined [64]. Expressed is the fold increase in
P
enh
values of STAT-1 ODN-treated animals com-
pared with control ODN, vehicle, and negative
control (phosphate buffered saline; PBS) from
three independent experiments [68].
tion of immune response-mediating mole-
cules.
Initial clinical studies employing a
STAT-1 decoy ODN in allergic asthma as
well as psoriasis have been performed by
AVONTEC GmbH. No side effects were
seen in these early Phase I trials, and the
compound was well tolerated.
The eukaryotic transcription factor NF-
jB was identified as a protein that binds
specifically to a decameric DNA sequence
(ggg ACT TTC C), within the intronic en-
hancer of the immunoglobulin kappa light
chain in mature B and plasma cells, but
not pre-B cells [32, 33]. This transcription
factor consists of homo- or heterodimers
of different subunits which are members
of a family of structurally related proteins
(Rel/NF-jB proteins). NF-jB has been de-
tected in most cell types, and specific NF-
jB binding sites have been identified in
promoters and enhancers of numerous in-
ducible genes. Thus, it has been shown
that NF-jB plays a pivotal role in the coor-
dinated transactivation of many genes in-
volved in cytokine-mediated inflammation.
AnGes MG in Japan is developing a NF-
jB decoy ODN as therapeutic agent not
only for inflammatory diseases such as
atopic dermatitis and rheumatoid rheuma-
tism, but also as a prophylactic agent for
restenosis. In a rat model of collagen-in-
duced arthritis (CIA) and Sendai virus-
liposome (fusigenic liposomes) mediated-
intraarticular application of decoy ODN,
the presence of fluorescein isothiocyanate
(FITC)-labeled NF-jB decoy ODN was
found in the synovium until 28 days after
injection. The NF-jB decoy ODN de-
creased the severity of hind-paw swelling,
and both histologic and radiographic stud-
ies showed a marked suppression of joint
destruction. Furthermore, the production
of IL-1 and tumor necrosis factor-alpha
(TNFa) in the synovium of arthritic joints
was suppressed [34, 35]. Similar results
were found in a cynomolgus CIA-arthritis
model [36]. The efficacy of an ointment
containing NF-jB decoy ODN was investi-
gated regarding the development of atopic
dermatitis lesions in a mice model which
is characterized by the spontaneous onset
of atopic dermatitis under conventional
conditions [37]. Topical administration of
NF-jB decoy ODN twice monthly resulted
in a significant reduction in clinical skin
condition score, and a marked improve-
ment in histological findings. Improve-
ment of the atopic skin condition by NF-
jB decoy ODN was accompanied by a sig-
nificant decrease in the migration of mast
cells into the dermis, and an increase in
apoptotic cells. At the present time, AnGes
MG has completed Phase I studies in
rheumatoid arthritis and atopic dermatitis
(T. Tomita; personal communication).
Also, at Corgentech Inc., San Francisco,
US, a decoy ODN targeting NF-jB is cur-
rently under pre-clinical development for
rheumatoid arthritis and dermatitis (http://
www1.corgentech.com/cgt/inflammation).
9.2.2
Vascular Proliferative Diseases
(AP-1, C/EBP, NF-jB, E2F)
Restenosis occurring after balloon angio-
plasty and coronary artery stent implanta-
tion is characterized by an extensive neoin-
timal proliferation of vascular smooth
muscle cells and the formation of extracel-
lular matrix [38, 39]. This change in
smooth muscle cell phenotype from a con-
tractile to a synthetic state was shown to
be correlated with an increased synthesis
of the potent mitogen endothelin-1 (ET-1)
in the vessel wall [40]. Increased synthesis
of ET-1 has been demonstrated to occur
after pressure trauma to the vessel wall in
experimental animal models and in hu-
man vascular tissue samples [41, 42]. The
stretch-induced ET-1 synthesis by the vascu-
lar endothelial cells continues for one to two
days, suggesting in the case of restenosis
that it is important to block the excessive
formation of this growth factor within the
first days following balloon angioplasty.
Subsequent investigations into the molecu-
lar mechanism underlying the stretch-in-
duced increase in ET-1 synthesis have re-
vealed a crucial role for the activator pro-
tein-1 (AP-1) transcription factor. Namely,
the preproendothelin-1 gene is sensitive to
deformation stress in blood vessels and
uses AP-1 as an essential regulatory compo-
nent [43]. In cultured endothelial cells ex-
posed to mechanical deformation as well
as in isolated, intact blood vessels in re-
sponse to a non-physiological increase in
perfusion pressure, a decoy ODN directed
against AP-1 inhibited both preproendothe-
lin-1 gene expression and, as a conse-
quence, ET-1 protein synthesis [43].
In extending the above-mentioned in-vi-
tro data, the anti-AP-1 decoy ODN strategy
was modified in a way that the dODN-con-
taining solution was administered locally
through a Dispatch catheter (SCIMED
Lifesciences, USA) into the coronary ar-
teries of hypercholesterolemic minipigs at
the time of percutaneous transluminal cor-
onary angioplasty (PTCA) and subsequent
AVE-GFX stent-implantation [44]. AP-1 de-
coy ODN treatment significantly reduced
neointimal formation in the coronary ar-
teries after 4 weeks of follow-up (Fig. 9.4).
Noteworthy, this therapeutic effect was
maintained at 8 weeks after decoy ODN
application. Similar results have been de-
scribed by Ahn et al. using hemagglutinat-
ing virus of Japan (HVJ)-liposome based
transfer of a circular dumbbell AP-1 decoy
ODN in a model of balloon injury-induced
neointimal formation in the rat carotid ar-
tery [45].
In cooperation with Avontec GmbH,
Biotronik a Berlin-based medical device
manufacturer is currently developing a
second-generation drug-eluting stent com-
bining advanced vascular stent design with
AP-1 decoy ODN technology.
Many cytokine genes, including those
encoding acute-phase proteins and immu-
noglobulins, share binding sites for
CCAAT/enhancer binding protein (C/EBP)
in their 5'-flanking regions, and C/EBP-re-
lated transcription factors regulate cell pro-
liferation during terminal differentiation.
C/EBP therefore represents an attractive
target for inhibiting restenosis following
balloon angioplasty. In a rabbit model of
restenosis which combines balloon injury
of the carotid artery with cholesterol-
mediated chronic inflammation, a decoy
ODN capable of neutralizing C/EBP was
administered to the site of injury for
30 min [46]. Electrophoretic mobility shift
analysis confirmed that C/EBP activity in
decoy ODN-treated segments was virtually
absent after 2 days. Morphometric analysis
after 28 days revealed significant reduction
(up to 50%) of both neointima formation
and intravascular inflammation in decoy
ODN-treated segments as compared to
control ODN or vehicle-treated segments
(Fig. 9.5). In addition, de-novo synthesis of
ET-1, as well as the number of proliferat-
ing cell nuclear antigen-positive smooth
muscle cells in the vessel wall were mark-
edly attenuated at day 3. In extending
these findings, Ni et al. demonstrated that
a C/EBP decoy ODN blocked angiotensin
II-induced IL-6, TNF-a, and monocyte che-
motactic protein-1 (MCP-1) gene expres-
sion in rat aortic smooth muscle cells [47].
These findings suggest that decoy ODN-
based neutralization of C/EBP may be a
feasible and effective method to treat in-
flammatory cardiovascular disease and to
limit restenosis following angioplasty.
Expression of a constitutive NF-jB-like
activity plays an important role for pro-
liferation of cultured bovine vascular
smooth muscle cells brought about by the
coordinated transactivation of cytokine and
adhesion molecule genes [48]. Therefore,
it was a logical approach to study the
potential therapeutic benefit of the in-vivo
application of a NF-jB decoy ODN to treat
atherosclerosis and lesion formation after
vascular injury. Following vascular balloon
injury, NF-jB decoy ODN inhibited neo-
intimal lesion in rat carotid arteries [49]
and porcine coronary arteries (application
of decoy ODN via hydrogel balloon
catheter) for up to 4 weeks after a single
intravascular treatment [50]. The applica-
tion of NF-jB decoy ODN (combined with
Sendai virus liposomes) into rat hearts
resulted in a significant improvement in
tolerance against ischemiareperfusion in-
jury, together with inhibition of neutro-
phil adherence and tissue IL-8 production
[51], suggesting that NF-jB may play a sig-
nificant role in ischemiareperfusion in-
jury. It is even reported that myocardial
infarction may be prevented by in-vivo
transfection of a NF-jB decoy ODN
[52].
Mitogenic growth factors mediating vas-
cular cell proliferation in atherosclerosis
share a final common signaling pathway:
the cell cycle [53]. Within the cell cycle,
the E2F transcription factor family controls
expression of genes required in S phase
[54]. A single administration of an E2F de-
coy (containing the E2F cis-element) that
binds the transcription factor E2F inhib-
ited smooth muscle cell hyperplasia in a
rat carotid balloon injury model [8]. Bind-
ing of E2F to the decoy ODN prevents it
from transactivating the gene expression
of cell cycle regulatory proteins such as
proliferating cell nuclear antigen (PCNA),
c-myc, and cdk2, thereby inhibiting vascu-
lar smooth muscle cell proliferation and
subsequent neointima formation.
Early thrombosis, neointimal hyperpla-
sia, and atherosclerosis are hallmarks of
venous bypass graft disease resulting in
premature loss of bypass function. Local
ex-vivo treatment of bypasses with decoy
ODNs has been proposed as a potential ef-
ficacious preventive therapy to improve by-
pass longevity [10, 55, 56]. Experimental
application of an E2F decoy ODN resulted
in the long-term stabilization of vein graft
wall architecture and prolonged resistance
to experimental atherosclerosis [57]. In ex-
tending this work to the human situation,
Mann et al. at Harvard University [58] first
demonstrated therapeutic effects in a clini-
cal development program (PREVENT)
brought about by a pressure device-
mediated ex-vivo application of E2F decoy
ODN in vascular grafts of patients with
late-stage peripheral artery disease
(Fig. 9.6). The E2F decoy ODN, now under
development at Corgentech Inc., has been
shown to be effective in Phase I/II and IIb
trials, and is currently being evaluated in
two Phase III clinical trials. The peripheral
artery bypass study (PREVENT 3) is testing
edifoligide (E2F Decoy) in 1400 patients who
have undergone peripheral artery bypass
surgery at approximately 80 medical centers
throughout the United States. PREVENT 4
is evaluating edifoligide (E2F Decoy) in
2400 patients who have undergone coronary
artery bypass graft (CABG) surgery at more
than 100 United States medical centers. The
FDA has granted edifoligide (E2F Decoy)
fast track status for both coronary and pe-
ripheral indications due to the unmet med-
ical needs the product may address (bypass
atherosclerosis). Enrolment for both studies
has been completed, and data are expected
to be presented in late 2004 and early
2005, respectively (http://www1.corgen-
tech.com/cgt/edifoligide).
In May 2004, Corgentech initiated a
Phase I/II trial of E2F decoy ODN treat-
ment for the prevention of arteriovenous
(AV) graft failure in patients with end-
stage renal disease. This clinical trial is a
double-blind, randomized, placebo-con-
trolled study that will enroll 60 patients at
up to 20 research centers in the United
States. The companies expect to announce
initial data from this trial in the first half
of 2005.
9.2.3
Cancer (STAT-3, CRE, ERE)
Most head and neck cancers affecting the
mouth, nasal cavities, larynx and pharynx,
are squamous cell carcinomas (SCCHN).
Recently, it has been shown that increased
expression of epidermal growth factor
(EGF) receptor occurs early in squamous
cell carcinogenesis, and is critical for the
loss of growth control. Stimulation of the
Fig. 9.4 Proliferative vessel wall response 4 weeks
after stent angioplasty in three coronary arteries
from the same hypercholesterolemic minipig in-
fused with vehicle, activator protein-1 (AP-1) con-
sensus decoy ODN or control decoy ODN (Elasti-
ca van Giesons staining, original magnification
40) [44].
Fig. 9.5 Representative transverse sections of in-
jured carotid arteries showing Elastica van Gie-
sons staining, demonstrating the inhibitory effect
of the C/EBP consensus decoy ODN (b) com-
pared with the lack of effect of the mutant control
ODN (a) on neointimal lesion formation (arrows)
28 days after the intervention (original magnifica-
tion 25) [46].
Vehicle
EGF receptor in cultured tumor SCCHN
cells initiates signaling via persistent acti-
vation of STAT-3 transcription factor, an-
other family member of the STAT-proteins
[59]. To explore the possibility of targeting
STAT-3 for therapeutic benefit, Grandis
and co-workers at University of Pittsburgh
designed a transcription factor decoy ODN
approach that has been shown to inhibit
STAT-3-mediated gene expression and
SCCHN growth, but did not influence nor-
mal oral keratinocytes [60].
Another possibility to treat cancer and to
influence cellular regulatory processes
might be provided by interfering therapeu-
tically with the cyclic AMP response ele-
ment (CRE). The CRE-transcription factor
complex is a pleiotropic activator that par-
ticipates in the induction of a wide variety
of cellular and viral genes. It has been
shown that a CRE-palindromic decoy ODN
can penetrate into cells, compete with
CRE enhancers for binding transcription
factors, and interfere specifically with CRE
and AP-1-directed transcription in vivo
[61]. This decoy ODN restrained tumor cell
proliferation, without affecting the growth
of non-cancerous cells. Since there are
many cAMP-regulated genes distributed
ubiquitously in all cell types, preclinical
safety studies need to be performed to ex-
clude a potential harmful effect of CRE de-
coy ODN to cells and organisms.
Breast cancer, the most common malig-
nancy in women, has been shown to be
associated with the steroid hormone estro-
gen and its receptor (ER), a ligand-acti-
vated transcription factor. Wang et al., at
the National Cancer Institute in Frederick,
Maryland, USA [62], developed a phos-
phorothioate cis-element decoy ODN
against the estrogen response element
(ERE decoy) to target disruption of ER
DNA binding and transcriptional activity.
The ERE decoy ODN potently ablated the
17-b-estrogen-inducible cell proliferation
and induced apoptosis of human breast
carcinoma cells by functionally affecting
expression of the c-fos gene and AP-1 luci-
ferase gene reporter activity. Specificity of
the decoy ODN was demonstrated by its
ability to directly block ER binding to a cis-
element probe and transactivation. These
data may suggest that estrogen-mediated
cell growth of breast cancer cells can be
preferentially restricted via targeted disrup-
tion of ER at the level of DNA binding by
a decoy ODN strategy applied to steroid
nuclear receptors.
Fig. 9.6 E2F decoy ODN treatment of
human vascular bypass grafts (PREVENT
single-center, randomized, controlled
trial). Kaplan-Meier comparison of time
to graft failure between E2F-decoy and
untreated groups. (Reproduced with per-
mission from [58].)
In summary, decoy ODNs represent a
new class of potential biopharmaceuticals
which can be designed specifically to tar-
get transcription factors involved in the
pathogenesis of a given disease. There has
been an explosion in the use of transcrip-
tion factor decoys as tools for studying
gene regulation and as experimental thera-
py to treat a variety of pathological condi-
tions [63]. Ongoing preclinical and clinical
development programs at various emerg-
ing biotech companies as well as academic
research institutions will further elucidate
the potential of this promising new bio-
pharmaceutical drug class.
References
1 Scholer, H. R. and P. Gruss, Cell 1984, 36,
403411.
2 Morimoto, R. I., Curr Opin Cell Biol 1992, 4,
480487.
3 Brivanlou, A. H. and J. E. Darnell, Jr., Science
2002, 295, 81318.
4 Papavassiliou, A. G., Mol Med 1997, 3, 799
810.
5 Gyurko, R., D. Tran, and M. I. Phillips, Am J
Hypertens 1997, 10, 56S62S.
6 Scholer, H. R. and P. Gruss, EMBO J 1985, 4,
30053013.
7 Bielinska, A., R. A. Schivdasani, L. Zhang, and
G. J. Nabel, Science 1990, 250, 9971000.
8 Morishita, R., G. H. Gibbons, M. Horiuchi,
K. E. Ellison, M. Nakajima, L. Zhang, Y. Kane-
da, T. Ogihara, and V. J. Dzau, Proc Natl Acad
Sci USA 1995, 92, 58555859.
9 von der Leyen, H., M. J. Mann, and V. J. Dzau,
Gene therapy of cardiovascular disorders, in:
Hursts The Heart, V. Fuster, Editor. 1998,
McGraw-Hill Publishing Co: New York, NY.
p 213225.
10 Mann, M. J. and V. J. Dzau, J Clin Invest 2000,
106, 10711075.
11 Morishita, R., J. Higaki, N. Tomita, and T.
Ogihara, Circ Res 1998, 82, 10231028.
12 Sullenger, B. A., H. F. Gallardo, G. E. Ungers,
and E. Gilboa, Cell 1990, 63, 601608.
13 Brantl, S., Biochim Biophys Acta 2002, 1575,
1525.
14 von der Leyen, H. E., R. Braun-Dullaeus, M. J.
Mann, L. Zhang, and V. J. Dzau, Hum Gene
Ther 1999, 10, 23552364.
15 Mann, M. J., G. H. Gibbons, H. Hutchinson,
R. S. Poston, G. E. Hoyt, R. C. Robbins, and
V. J. Dzau, Proc Natl Acad Sci USA 1999, 96,
64116416.
16 Urban, E. and C. R. Noe, Il Farmaco 2003, 58,
243258.
17 Tomita, N., T. Tomita, K. Yuyama, T. Tougan,
T. Tajima, T. Ogihara, and R. Morishita, Curr
Opin Mol Ther 2003, 5, 107112.
18 Darnell, J. E., Jr., I. M. Kerr, and G. R. Stark,
Science 1994, 264, 14151421.
19 Ihle, J. N., T. Nosaka, W. Thierfelder, F. W.
Quelle, and K. Shimoda, Stem Cells 1997, 15
Suppl 1, 105111; discussion 112.
20 Durbin, J. E., R. Hackenmiller, M. C. Simon,
and D. E. Levy, Cell 1996, 84, 443450.
21 Takeda, K. and S. Akira, Cytokine Growth Fac-
tor Rev 2000, 11, 199207.
22 Horvath, C. M., Trends Biochem Sci 2000, 25,
496502.
23 Takeda, T., H. Kurachi, T. Yamamoto, H.
Homma, K. Morishige, A. Miyake, and Y.
Murata, J Endocrinol 1997, 153, R13.
24 Liu, X., G. W. Robinson, K. U. Wagner, L. Gar-
rett, A. Wynshaw-Boris, and L. Hennighausen,
Genes Dev 1997, 11, 179186.
25 Udy, G. B., R. P. Towers, R. G. Snell, R. J. Wil-
kins, S. H. Park, P. A. Ram, D. J. Waxman, and
H. W. Davey, Proc Natl Acad Sci USA 1997, 94,
7239744.
26 Schindler, C. W., J Clin Invest 2002, 109, 1133
1137.
27 Krzesz, R., A. H. Wagner, M. Cattaruzza, and
M. Hecker, FEBS Lett 1999, 453, 191196.
28 Wagner, A. H., M. Gebauer, B. Pollok-Kopp,
and M. Hecker, Blood 2002, 99, 520525.
29 Barnes, P. J. and I. M. Adcock, Eur Respir J
1998, 12, 221234.
30 Sampath, D., M. Castro, D. C. Look, and M. J.
Holtzman, J Clin Invest 1999, 103, 13531361.
31 Quarcoo, D., S. Weixler, R. A. Joachim, D.
Groneberg, B. Ahrens, A. H. Wagner, M.
Hecker, and E. Hamelmann, Immunology
2002, 206, 129.
32 Karin, M. and A. Lin, Nat Immunol 2002, 3,
221227.
33 Sen, R. and D. Baltimore, Cell 1986, 46, 705
716.
34 Tomita, T., E. Takeuchi, N. Tomita, R. Morishi-
ta, M. Kaneko, K. Yamamoto, T. Nakase, H.
Seki, K. Kato, Y. Kaneda, and T. Ochi, Arthritis
Rheum 1999, 42, 25322542.
35 Tomita, T., H. Takano, N. Tomita, R. Morishi-
ta, M. Kaneko, K. Shi, K. Takahi, T. Nakase, Y.
Kaneda, H. Yoshikawa, and T. Ochi, Rheuma-
tology (Oxford) 2000, 39, 749957.
36 Tomita, T., Y. Kunugiza, R. Morishita, H. Ha-
shimoto, K. Yoshida, A. Waki, S. Kuroda, and
H. Yoshikawa, Ann Rheum Dis 2004, 63
(Suppl I), 285.
37 Nakamura, H., M. Aoki, K. Tamai, M. Oishi,
T. Ogihara, Y. Kaneda, and R. Morishita, Gene
Ther 2002, 9, 12211229.
38 von der Leyen, H. E., M. J. Mann, and V. J.
Dzau, Semin Interv Cardiol 1996, 1, 20914.
39 Ross, R., Nature 1993, 362, 801809.
40 Serradeil-Le Gal, C., J. M. Herbert, C. Garcia,
M. Boutin, and J. P. Maffrand, Peptides 1991,
12, 575579.
41 Lauth, M., M. M. Berger, M. Cattaruzza, and
M. Hecker, Arterioscler Thromb Vasc Biol 2000,
20, 96103.
42 Schiffrin, E. L., Am J Hypertens 2001, 14, 83S
89S.
43 Lauth, M., A. H. Wagner, M. Cattaruzza, H. D.
Orzechowski, M. Paul, and M. Hecker, J Mol
Med 2000, 78, 441450.
44 Buchwald, A. B., A. H. Wagner, C. Webel, and
M. Hecker, J Am Coll Cardiol 2002, 39, 732
738.
45 Ahn, J. D., et al., Circ Res 2002, 90, 1325
1332.
46 Kelkenberg, U., A. H. Wagner, J. Sarhaddar,
M. Hecker, and H.E. von der Leyen, Arterios-
cler Thromb Vasc Biol 2002, 22, 949954.
47 Ni, W. H., D. L. Kong, M. Rezvani, A. A. Man-
gi, A. S. Pachori, M. Lopez-Ilasaca, C. C. Liew,
R. E. Pratt, and V. J. Dzau, Circulation 2002,
106 (suppl II), II13.
48 Bellas, R. E., J. S. Lee, and G. E. Sonenshein, J
Clin Invest 1995, 96, 25212527.
49 Yoshimura, S., R. Morishita, K. Hayashi, K.
Yamamoto, H. Nakagami, Y. Kaneda, N. Sakai,
and T. Ogihara, Gene Ther 2001, 8, 16351642.
50 Yamasaki, K., T. Asai, M. Shimizu, M. Aoki,
N. Hashiya, H. Sakonjo, H. Makino, Y. Kane-
da, T. Ogihara, and R. Morishita, Gene Ther
2003, 10, 356364.
51 Sawa, Y., R. Morishita, K. Suzuki, K. Kagisaki,
Y. Kaneda, K. Maeda, K. Kadoba, and H. Mat-
suda, Circulation 1997, 96, II-280284; discus-
sion II-285.
52 Morishita, R., T. Sugimoto, M. Aoki, I. Kida,
N. Tomita, A. Moriguchi, K. Maeda, Y. Sawa,
Y. Kaneda, J. Higaki, and T. Ogihara, Nat Med
1997, 3, 894899.
53 Elledge, S. J., Science 1996, 274, 166472.
54 Matsumura, I., H. Tanaka, and Y. Kanakura,
Cell Cycle 2003, 2, 333338.
55 Mann, M. J. and V. J. Dzau, Curr Opin Cardiol
1997, 12, 522527.
56 White, S. J. and A. C. Newby, J Card Surg
2002, 17, 549555.
57 Mann, M. J., E. Afshin, and V. J. Dzau, J Thor-
ac Cardiovasc Surg 2001, 121, 714722.
58 Mann, M. J., A. Whittemore, M. C. Donaldson,
M. D. Pollack, J. Orav, and V. J. Dzau, Lancet
1999, 354, 14931498.
59 Grandis, J. R., S. D. Drenning, Q. Zeng, S. C.
Watkins, M.F. Melhem, S. Endo, D. E. John-
son, L. Huang, Y. He, and J. D. Kim, Proc Natl
Acad Sci USA 2000, 97, 42274232.
60 Leong, P. L., G. A. Andrews, D. E. Johnson,
K. F. Dyer, S. Xi, J. C. Mai, P. D. Robbins, S.
Gadiparthi, N. A. Burke, S. F. Watkins, and
J. R. Grandis, Proc Natl Acad Sci USA 2003,
100, 41384143.
61 Park, Y. G., M. Nesterova, S. Agrawal, and Y. S.
Cho-Chung, J Biol Chem 1999, 274, 157380.
62 Wang, L. H., X. Y. Yang, X. Zhang, K. Mihalic,
W. Xiao, and W. L. Farrar, Cancer Res 2003, 63,
20462051.
63 Dzau, V. J., Circ Res 2002, 90, 12341236.
64 Hamelmann, E., K. Takeda, J. Schwarze, A. T.
Vella, C. G. Irvin, and E. W. Gelfand, Am J Re-
spir Cell Mol Biol 1999, 21, 480489.
65 Mann, M. J., G.H. Gibbons, R. S. Kernoff, F. P.
Diet, P. S. Tsao, J. P. Cooke, Y. Kaneda, and
V. J. Dzau, Proc Natl Acad Sci USA 1995, 92,
45024506.
66 Mann, M. J., G. H. Gibbons, P. S. Tsao, H. E.
von der Leyen, J. P. Cooke, R. Buitrago, R.
Kernoff, and V. J. Dzau, J Clin Invest 1997, 99,
12951301.
67 Tomita, N., J. Y. Kim, G. H. Gibbons, L.
Zhang, Y. Kaneda, R.A. Stahl, M. Ogborn, B.
Venderville, R. Morishita, D. Baran, and V.J.
Dzau, Int J Mol Med 2004, 13, 629636.
68 Quarcoo, D., S. Weixler, D. Groneberg, R. Joa-
chim, B. Ahrens, A. H. Wagner, M. Hecker,
and E. Hamelmann, J Allergy Clin Immunol
2004, 114, 288295.
References 241
Abstract
RNA interference (RNAi) reigns among
the most significant scientific discoveries
at the turn of the 21st century, both for its
impact on fundamental genetic research
and on biotechnology and the develop-
ment of biopharmaceuticals. This biologi-
cal phenomenon represents an evolutiona-
rily conserved mechanism that plays di-
verse roles, including protection from viral
infections or the products of aberrant tran-
scription and in the regulation of develop-
ment and differentiation. Elucidation of
the pathway through a series of indepen-
dent functional, bioinformatic, biochem-
ical, and genetic studies has revealed the
capability of unprecedented precision and
potent reduction of intended mRNA tar-
gets. Now well-documented in mammals,
RNAi results in gene suppression by cleav-
age or translational attenuation of target
mRNA using small interfering RNA (siR-
NA) or short hairpin RNA (shRNA), re-
spectively, as the functional intermediates
in a highly coordinated protein: RNA com-
plex known as RISC. Fortuitously, these
regulatory RNA molecules are readily
synthesized, and when artificially intro-
duced in vitro or in vivo, effect mRNA tar-
get-specific suppression. Coupled with the
ease of producing the siRNAs (and related
shRNAs), RNAi-mediated gene silencing
has now emerged as an extremely valuable
technology to reduce or knock down ex-
pression of specific genes and allow for as-
sessment of gene function. While applica-
tion of RNAi technology as an in vitro
functional genomics tool is now well es-
tablished, there are several challenges that
remain to be overcome before it may be
implemented as a viable therapeutic
approach. The challenges include: 1) en-
suring highly potent target inhibition; 2)
achieving appropriate target specificity; 3)
assuring stability of the active drug in bio-
logical fluids; 4) directing distribution to
the appropriate target organ; and 5) mini-
mizing target-based or chemical class-
based toxicity. Strategies to address some
of these issues include rational siRNA se-
quence selection (based on bioinformatics
and sophisticated design algorithms) and
use of chemical modifications of, and con-
jugations to, siRNA that enhance serum
stability and biodistribution. This chapter
provides a comprehensive overview of the
current landscape in the RNAi field, and
offers a glimpse of its potential in basic
and advanced research and its potential for
the development of biopharmaceuticals.
243
10
Rational siRNA Design for RNA Interference:
Optimizations for Therapeutic Use and Current Applications
ISBN: 3-527-31184-X
Abbreviations
dsRNA double-stranded RNA
FHV flock house virus
miRNA microRNA
nt nucleotide
RIP repeat induced point mutation
shRNA short hairpin RNA
SPECT single photon emission com-
puted tomography
10.1
RNAi: History and Mechanism
10.1.1
History of the Discovery of RNAi
RNA interference (RNAi) emerged onto
the biotechnology scene during the late
1990s as a previously recognized but rela-
tively uncharacterized phenomenon in
plants, fungi, and invertebrates. Now well-
documented in mammals, RNAi causes
gene suppression via cleavage of target
mRNA using a small interfering RNA
(siRNA) intermediate as part of an evolu-
tionarily conserved, ubiquitous multi-pro-
tein complex that is involved in an array
of regulatory functions in nature, includ-
ing protection against harmful mobile ge-
netic elements such as viruses or transpo-
sons, regulation of developmental events,
and elimination of unwanted run-on
mRNA transcripts. A critical advantage for
the pharmaceutical industry and basic re-
searchers alike is that these siRNA mole-
cules are readily synthesized and, when ar-
tificially introduced either in vitro or in
vivo, they effect mRNA target-specific
cleavage.
The elucidation of the RNAi mechanism
is an intriguing story, the denouement of
which may offer great hope in the ongoing
search for effective treatments for some of
the most intractable of human diseases.
The unfolding of our understanding of
RNAi began in 1990 with the report of an
unexpected co-suppression of homologous
genes in trans in petunias following the in-
troduction of a chimeric chalcone synthase
gene [1]. The mechanism of this co-sup-
pression was unknown at the time, and
was postulated to involve one or more of a
variety of processes, including DNA
methylation, repeat induced point muta-
tion (RIP), and transvection (in which one
allele influences the expression of a ho-
mologous allele). A similar homology-de-
pendent gene-silencing phenomenon was
reported in the fungus Neurospora crassa
in 1996 [2]. This suppression of gene ex-
pression was termed quelling, and re-
flected similar silencing effects, though
the cause of the mechanism remained un-
known.
The basis of this gene suppression in
plants and fungi became clearer in 1998
when Fire et al. published their seminal
work in Nature describing gene silencing
in Caenorhabditis elegans by the artificial
introduction of double-stranded RNA
(dsRNA) [3]. Previous studies by Guo and
Kemphues in 1995 [4] had shown that
sense-strand RNA was as effective as anti-
sense-strand RNA in suppressing gene ex-
pression in worms. Fires group found that
a senseantisense mixture injected into
the organisms caused gene-specific silenc-
ing, and that this silencing was not due
to single-stranded antisense mediation of
mRNA translation. Although they were
not at the time able to explain the mecha-
nism underlying the silencing, they deter-
10 Rational siRNA Design for RNA Interference: Optimizations for Therapeutic Use and Current Applications 244
mined several key aspects of the phenome-
non. dsRNA segments with identity to in-
trons or promoter regions did not effect si-
lencing; when silencing was attained by
use of dsRNA corresponding to the target
gene, a concomitant decrease in gene-spe-
cific mRNA transcripts was noted. Though
the overall silencing pathway was still
largely unknown, Fire et al. made a few
prescient observations presaging the explo-
sion of interest in and use of the RNAi
pathway. First, they remarked upon the
utility of this gene-silencing phenomenon
in probing the function of previously un-
characterized genes. Second, they postu-
lated that the silencing mechanism prob-
ably existed for a biological purpose. It is
the utilization of this underlying mecha-
nism for therapeutic purposes that gener-
ates widespread optimism from the discov-
ery of RNAi.
Subsequent to the results published by
Fire et al., the literature grew to include
reports of a similar gene suppression
mediated by dsRNA in Drosophila cell ly-
sates [5, 6] and in cultured Drosophila and
human cells [7]. Of special interest for the
pharmaceutical industry was the latter arti-
cle, in which Elbashir et al. reported use
of a variety of human cell lines, including
human embryonic kidney and HeLa cells,
and found that introduced 21-nucleotide
dsRNAs containing 2-nucleotide 3' over-
hangs could suppress endogenous and
heterologous genes in these cells. Since
that time, there have been numerous re-
ports of siRNA-induced RNAi in a wide
variety of human cells, including lung car-
cinoma [8], colorectal adenocarcinoma [9],
T cells [10], B lymphoblasts [11], cardiac
myocytes [12], metastatic prostate cells
[13], and glioma [14, 15].
10.1.2
Key Functional Components of the RNAi
Mechanism
The overall mechanism of siRNA-mediated
RNAi activity as currently understood is
outlined in Fig. 10.1 [5, 1623]. In nature,
the evolutionarily conserved RNAi mecha-
nism is activated by various forms of long
double-stranded precursor molecules
which are processed in the cytoplasm [18]
by the RNase III-type enzyme Dicer [24]
into the active 21- to 25-nucleotide (nt)
double-stranded siRNA intermediary con-
taining short 3' overhangs. In Drosophila,
the R2D2 protein was found to associate
with Dicer and bridge the transition be-
tween Dicer-mediated siRNA formation
and siRNA interaction with the RNA-in-
duced silencing complex (RISC) [25]. RISC
is a multi-protein complex found to be in-
volved in cleavage of the target mRNA
[26]. Different species-specific forms of an
RNA helicase protein have been shown to
be a critical component of the RISC mech-
anism [2731]. This helicase unwinds the
duplex intermediate (the mechanics of
which are described in Section 10.3), using
one of the strands to probe cytoplasmic
mRNA molecules for sequence comple-
mentarity to the RISC-associated func-
tional siRNA strand. When sufficient com-
plementarity is found, RISC effects cleav-
age of the mRNA molecule, thus causing
sequence-specific post-transcriptional gene
silencing.
Another component of the multi-protein
complex that comprises RISC is a member
of the Argonaute (Ago) family of proteins
[32, 33]. The Ago proteins are character-
ized by the PIWI domain, which is a
highly-conserved C-terminal motif of about
220 amino acids rich in basic residues.
The crystal structure of the Argonaute pro-
tein from the archaebacterium Pyrococcus
furiosus was solved to 2.25 [34] and re-
vealed an interesting feature of the PIWI
domain predictive of the mechanism for
siRNA-RISC-mediated cleavage. The PIWI
domain sits in a crescent-shaped fold
below another highly conserved region
known as the PAZ (PIWI/Argonaute/
Zwille) domain. The residues exposed by
this conformation contain a high number
of positive charges suitable for bond for-
mation with the negatively charged phos-
phate backbone of oligonucleotides and
the 2'-hydroxy moieties of the ribose su-
gars. Molecular modeling studies revealed
Fig. 10.1 Schematic representation of the RNAi
mechanism. 1) siRNA is bound by Dicer-R2D2-
pre-RISC complex. 2) siRNA duplex is unwound
and single-strand containing RISC is formed.
3) siRNA strand guides RISC identification of target
mRNA. 4) Ago2 member of the RISC cleaves the
target mRNA. 5) Release of cleavage products and
new mRNA target screening. Base pairs 28 (rela-
tive to the 5' AS end) play the primary role in
mRNA target site recognition.
that the PIWI domain could mediate cleav-
age of an mRNA target (bound to the com-
plementary strand of the siRNA) between
nucleotides 10 and 11 of the 5' end of the
antisense strand. This is exactly where in
vitro biochemical studies have shown that
RISC-mediated RNAi cleavage of target
mRNAs occurs [35]. Furthermore, the
PIWI domain, which has similarity to
RNase H, produces cleavage products with
5' phosphate and 3' OH groups. A sup-
porting view of this mechanism of cleav-
age is proposed by Martinez and Tuschl
[36]. Their biochemical studies showed
that RISC-associated mRNA cleavage gen-
erates products with 5' phosphate and 3'
OH groups. Meister et al. [37] confirmed
the involvement of Argonaute proteins in
RNAi, including the participation of Ago2
in mRNA cleavage. Additional insight into
the RNAi mechanism comes from the
NMR solution of the structure of the PAZ
domain, one of the functional domains
present in Dicer and in Argonaute pro-
teins. The PAZ domain has been shown
to interact with 2-nt 3' overhangs [38], em-
phasizing the importance of the end struc-
ture of siRNA.
In addition to siRNAs, the RNAi mecha-
nism utilizes other classes of short dsRNA
molecules that can be produced either exo-
genously or endogenously. Short hairpin
RNAs (shRNAs), which are *4050 nt
stem-loop molecules (or larger) may be
processed from *70 nt single-stranded
RNA transcribed from viral or plasmid
vectors. Dicer further processes these
stem-loop molecules into siRNAs that lead
to gene silencing. Another group of RNAi
mediators includes the microRNAs (mi-
RNAs), which are host-encoded transcripts
involved in the regulation of gene expres-
sion and organism development [39]. The
most current evidence suggests that, in
mammals, highly structured primary mi-
croRNAs (pri-miRNAs) are processed in
the nucleus by the RNase III endonuclease
Drosha, to *6070 nt stem-loop inter-
mediates known as precursor microRNAs
(pre-miRNAs) [39, 40]. The pre-miRNAs
are transported to the cytoplasm by Ran-
GTP and the export receptor Exportin-5,
where they undergo processing by the
RNase III-type molecule Dicer into the ac-
tive miRNA [39]. miRNAs have been asso-
ciated with RISC [41, 42], which has in
turn been associated with polyribosomes
in human cells [43]. Recent biochemical
studies in Drosophila suggest that RISC
may interact with ribosomes and interfere
with protein synthesis, resulting in transla-
tional attenuation [44]. This potential inter-
action with the translational apparatus
supports the notion that RNAi may be
closely linked with protein synthesis.
An additional feature of RNAi-mediated
post-transcriptional gene suppression
found in many organisms, but not in Dro-
sophila or humans, is the presence of an
RNA-dependent RNA polymerase (RdRP)
that is required for gene silencing [42].
RdRPs, present in C. elegans, D. discoideum
(see also Part III, Chapter 5), and plants
[45], are associated with amplification of
the RNAi response in these organisms,
leading to organism-wide gene silencing
effects following localized administration
of low concentrations of siRNAs. This am-
plification mechanism may function in a
similar role to that of the interferon re-
sponse in vertebrates, which is to rapidly
mount a global response to potential infec-
tion or other cellular challenges (e.g., tran-
sposons). Prior to the discovery of siRNAs
it was observed that long dsRNA caused
an interferon response in mammalian
cells mediated by the dsRNA-dependent
protein kinase PKR and RNase L [46]. This
non-specific response to long dsRNA pre-
cluded the use of these molecules as spe-
cific gene silencers in mammals. However,
the use of synthetic siRNAs introduced in-
tracellularly can circumvent this response,
allowing targeted, gene-specific suppres-
sion without induction of global, non-spe-
cific cellular toxicity. Thus, the specific tar-
geting of any gene transcript is possible by
employing an siRNA designed for efficient
entry into RISC to target native genes of
interest.
10.2
Early siRNA Design Parameters
10.2.1
Initial Considerations
A common expectation among many early
researchers in the field was that virtually
all siRNAs would be functional. However,
the RNAi mechanism involves the interac-
tion of siRNAs with several proteins,
pointing to potential sequence and struc-
tural characteristics that would determine
successful siRNA interaction with the
RNAi molecular machinery. Holen et al.
reported in 2002 that a panel of chemically
synthesized siRNAs targeting the human
Tissue Factor gene in cultured cells exhib-
ited a wide variation of silencing effective-
ness, with many of the siRNAs showing
limited ability to suppress gene expression
[47]. It was clear that not all siRNAs were
created equal; some were highly effective
gene silencers, some showed a moderate
effect, and a good number were ineffective
mediators of target gene suppression. This
and other similar observations provided
the impetus for further characterization of
the factors contributing to siRNA function-
ality.
Several groups explored the effect of tar-
get mRNA secondary and/or tertiary struc-
ture on the silencing efficiency of specific
siRNAs. RNAi studies performed by Yo-
kota et al. indicated a potential effect of
mRNA secondary structure on specific
siRNA functionality [48]. Vickers et al. [49]
supported these observations by demon-
strating for some siRNAs a correlation be-
tween higher order structure at the target
site and siRNA silencing efficiency, similar
to functional correlations observed for si-
lencing with antisense oligonucleotides.
Hohjohs studies similarly suggested that
observed differences in siRNA functional-
ity might have been due to distinct pre-
dicted secondary structure in the target
mRNAs tested [50]. A more recent study
characterized the specific inhibitory effects
of a well-characterized palindromic se-
quence known for its stable secondary
structure (the TAR: HIV-1 trans-activation
response region) on siRNA functionality
when positioned within the local context
of the target sequence [45, 51]. Thus, it ap-
pears that in certain cases the higher-order
structure of a target mRNA can affect
siRNA functionality; further exploration
and characterization of this relationship
will enhance our understanding of this
aspect of the RNAi mechanism.
Early guidelines for designing functional
siRNAs outlined by Elbashir et al. [52] did
not address mRNA or siRNA structural
features, and directed researchers to select
target regions from the open reading
frame (ORF) of the cDNA of interest, pre-
ferably 50100 nt downstream from the
start codon in order to avoid potential
blocking of the targeted region of the tran-
scribed mRNA by regulatory proteins.
Similar reasoning was applied for recom-
mendations to avoid the 5' and 3' untrans-
lated regions (UTRs). Additional sugges-
tions included selection of an mRNA tar-
get sequence with the general form of 5'-
AA(N19)UU-3' possessing approximately
50% G/C content, though 3279% G/C
was reported to work. High G-content se-
quences were to be avoided, as they form
very stable higher-order structures known
as G quartets. A BLASTn analysis was also
recommended in order to ensure targeting
of a unique gene. Furthermore, the use of
several unique siRNAs against one gene
was advised in order to confirm sequence-
specific gene silencing. Early discussions
of siRNA design highlighted the need for
2-nt 3' overhangs with a recommended
composition of U-U or dT-dT, the latter
contributing the added benefit of potential
nuclease resistance in mammalian cells
[35]. The secondary structure of target
mRNA was thought to have minimal effect
on siRNA-induced silencing [52].
10.2.2
Next-generation Design Parameters
While these early guidelines provided a
means for identifying siRNAs, silencing
performance could not be assured based
on these recommendations. In the absence
of more reliable strategies, researchers
were faced with time-consuming empirical
screens before identifying duplexes with
suitable potency. To address the challenge
of selecting siRNAs with a greater prob-
ability of potent silencing, systematic stud-
ies were undertaken to identify attributes
important for functionality and specificity.
Because the cellular processing of mi-
RNAs and siRNAs converges and achieves
similar functional outcomes (gene sup-
pression), the miRNA class of small regu-
latory RNAs serves as a potential source of
information regarding characteristics that
are essential for siRNA functionality. Mi-
croRNAs have apparently been part of the
cellular defense or regulatory mechanism
of plants and metazoans for much of their
evolutionary history, and therefore, as a re-
sult of selection pressure, miRNAs might
be expected to perform at a high level of
functionality. Analysis of the internal sta-
bility profiles (ISPs) of these naturally oc-
curring miRNAs from four different spe-
cies (human, mouse, Drosophila, and C.
elegans) determined a characteristic curve
that exhibits relative instability at the 5'
antisense end and, to a lesser extent, in-
ternally in the duplex at various locations
depending on the species analyzed. Func-
tional testing and analysis of the SPs of a
test panel of 360 randomly designed syn-
thetic siRNAs showed that highly func-
tional siRNAs (defined as F95, where F
refers to functionality as the denoted per-
cent of target gene suppression) exhibited
on average low thermodynamic stability at
the 5' antisense terminus and internally at
position 14 (Fig. 10.2). Conversely, non-
functional (F<50) siRNAs exhibited low 5'
sense stability and a stability profile that
negatively mirrors that of highly functional
siRNAs. Subsequent sequence analysis of
a collection of endogenous siRNAs origi-
nally isolated from Arabidopsis by Llave et
al. [53] determined that these siRNAs ex-
hibited a thermodynamic profile (Fig. 10.3)
similar to that shown in Fig. 10.2. Further-
more, comparison of the average ISPs of
highly functional (F95) exogenously pro-
duced (i.e., synthetic) siRNAs with the
ISPs of the recovered Arabidopsis siRNAs
showed a very high degree of similarity.
Two groups, one [22] taking an in vitro-
based biochemical approach, the other [54]
performing functional studies, investigated
the cellular mechanism of RISCsiRNA
interactions and found that the absolute
and relative instabilities of the base pairs
at the ends of the siRNA duplex determine
which strand (i.e., antisense or sense) in-
teracts with RISC and guides subsequent
cleavage of complementary mRNA se-
quences. The relatively low stability of the
5' antisense end of functional siRNA du-
10.2 Early siRNA Design Parameters 249
plexes suggests that RISC will preferen-
tially unwind the 5' antisense end (rather
than the 5' sense end) of functional siR-
NAs when probing siRNAs in the cell,
leading to antisense-mediated (rather than
sense-mediated) mRNA cleavage by the
functional siRNA. The combined observa-
tions of these two research groups strongly
support the hypothesis that a bias in the
internal stability profile governs the strand
preference or selectivity of the siRNA
RISC interaction.
Reynolds et al. [55] extended these re-
sults and identified other siRNA structure
and sequence-specific characteristics that
promote functionality. They found that the
Fig. 10.2 Calculated average internal stability profiles (AISPs)
for highly functional (F95) and non-functional (F<50) siRNAs
from a 360 siRNA data set. Gray squares indicate AISPs calculated
for the whole set.
Fig. 10.3 Calculated AISP for endogenously produced siRNAs
isolated by Llave et al. [50] from Arabidopsis (filled squares)
closely mimics AISP of functional siRNA subset (open squares).
presence of internal repeats conducive to
hairpin formation reduced the probability
of functionality of an siRNA. Additionally,
several site-specific sequence determinants
were found to promote siRNA functionality.
These include absence of a G at (sense) po-
sition 13, absence of a C or G at position 19,
and the presence of an A at position 3, a U
at position 10, and an A at position 19.
Using these additional guidelines for siR-
NA functionality, Reynolds et al. created
an algorithm for predicting functionality
of any siRNA (Fig. 10.4). These parameters
were later essentially confirmed, with a
few modifications, by Ui-Tei et al. [56] and
Amarzguioui and Prydz [57]. Ui-Tei et al. re-
ported that the presence of at least five A/U
pairs in the 5' antisense end promoted high
siRNA functionality. Amarzguioui and
Prydz determined that the three terminal
base pairs at the 5' antisense end were most
important positions for the presence of the
A/U pair. In their paper describing siDirect,
a new software program for siRNA design,
Naito et al. again essentially agreed with
these parameters, though they stipulated
the importance of A/U richness in the 5' ter-
minal third of the antisense strand [58]. Ka-
wasaki et al. [59] used a different approach;
they utilized a recombinant human Dicer to
randomly produce siRNAs that were effec-
tive and specific gene-silencing agents.
However, this method is of limited use
when targeting a specific gene while avoid-
ing the silencing of non-targeted genes. Al-
gorithms based on these and other thermo-
dynamic and sequence-specific parameters
improve the reliability of in silico prediction
of siRNA functionality compared with pre-
vious guidelines (e.g., those suggested by
Tuschl).
10.3
Current siRNA Design Considerations
10.3.1
An Enhanced Design Algorithm Broadens
the Scope of RNAi Application
The aforementioned strategies for rational
design of siRNAs were successful in that
Fig. 10.4 Correlation between an eight-component algorithm-derived score and ex-
perimentally determined siRNA functionality [52]. Significant correlation between
score value and functionality can be observed only at high score values.
the algorithms did provide some measure
of predictive ability for identifying func-
tional siRNAs. However, the differences
between the various algorithms and the
lack of high-level predictive capability
highlighted the need for a more robust al-
gorithm, one that was based on an exten-
sive analysis of a more comprehensive da-
taset. The analysis published by Reynolds
et al. on 180 siRNAs was extended to a
panel of almost 1000 siRNAs, and was
developed into an algorithm that takes into
account dozens of separate structural and
sequence-specific characteristics of siRNA
molecules. This algorithm generates a
scoring system for siRNA functionality
that significantly improves the reliability of
predictions of siRNA functionality (Fig.
10.5) [54, 55].
As mentioned previously, Schwarz et al.
[22] reported that a difference in thermo-
dynamic stability between the ends of an
siRNA or miRNA determines which of the
two strands is preferentially loaded into
the RISC. A lower stability profile for the
5' antisense end, for example, causes that
end of the double-stranded molecule to be
preferentially opened by the helicase asso-
ciated with RISC. Subsequent unwinding
of the duplex results, in this example, in
the antisense strand primarily being in-
cluded in the RISC assembly, though,
being an equilibrium-driven association, a
lower level of the sense strand will also in-
teract with RISC.
The importance of duplex asymmetry
for siRNA functionality was further inves-
tigated by analyzing a test panel of 340
siRNAs for thermodynamic-specific attri-
bute(s). This analysis resulted in the sepa-
ration of the panel into two populations:
1) siRNAs possessing antisense-biased
asymmetry; and 2) siRNAs possessing no
asymmetry or sense-biased asymmetry.
The population that exhibited 5' antisense
end instability was enriched in functional
siRNAs compared to the population
lacking this 5' antisense end instability
(Fig. 10.6). However, each population con-
tained a significant number of both func-
Fig. 10.5 Correlation between a 66-component algorithm-derived
score and experimentally determined siRNA functionality. A high
level of correlation is observed; most of the siRNAs with scores
above 90 are highly functional.
tional and non-functional siRNAs. Clearly,
a trait (or traits) other than 5' antisense
end instability determines functionality of
an siRNA. Our current understanding of
the RNAi mechanism lends insight into
the reason for this observation. The si-
RNA-induced mRNA cleavage process can
conceptually be divided into five different
steps (siRNA binding to RISC, preferential
unwinding of the siRNA duplex, mRNA
target recognition, target cleavage, and
product release; see Fig. 10.1). While the 5'
Fig. 10.6 (a) Separation of a panel of 340 siRNAs
into two groups based on sense/antisense ther-
modynamic asymmetry. While antisense-biased
siRNAs are substantially more functional at the F80
level than sense-biased or symmetric siRNAs (53%
versus 13%), approximately 16% of antisense-
biased siRNAs are non-functional (F<50), indi-
cating that there are other parameters in addition
to strand asymmetry required for overall siRNA
functionality. siRNAs were targeted against firefly
luciferase, DBI, or GAPDH. (b) Partitioning of
panel of 200 siRNAs into functional groups based
on whether antisense strand (RISC+) or sense
strand (RISC) is RISC-biased. Thermodynamic
asymmetry predominantly affects distribution of
highly functional and non-functional (NF) species.
antisense end instability promotes effective
RNAi function in step two, this character-
istic does not appear to be significant in
the other four stages.
With the development of the more ad-
vanced rational design algorithm described
above, RNAi researchers have a robust tool
for predicting siRNA functionality. This al-
gorithm was used to design a genome-
wide collection of siRNAs that targets all
known unique human, mouse, and rat
genes in the National Center for Biotech-
nology Information (NCBI) Reference Se-
quence database. Such an siRNA collection
can be used as a powerful tool in a variety
of basic research and biopharmaceutical
applications. For example, Li et al. used
siRNAs designed with this algorithm to ef-
ficiently identify the gene for vitamin K
epoxide reductase [60]. In another study,
high-throughput analysis of the effect of
siRNA-mediated silencing of a variety of
cell cycle-related genes was performed
using a subset of the genome-wide siRNA
library referred to above. This study,
coupled with novel intracellular imaging
techniques, allowed rapid determination of
both cell cycle and cell proliferation effects
resulting from knockdown of each of 110
cell-cycle related genes (Dharmacon and
Amersham, unpublished data). These and
similar studies illustrate the utility and
flexibility of rationally designed siRNAs. A
functional siRNA, however, is not neces-
sarily sufficient for performing unambigu-
ous and reproducible gene function anal-
yses. Specificity of the siRNA is equally
crucial, as unintended silencing of targets
other than the specified gene a phenom-
enon known as off-target effects can lead
to phenotypic changes that will interfere
with effective evaluation of experimental
results [6164].
10.3.2
Mitigation of Off-target Effects
An early hallmark of RNAi appeared to be
its ability to cause specific gene suppres-
sion by targeting and cleavage of a particu-
lar mRNA. However, several microarray
studies profiling the genome-wide effect of
siRNA-mediated silencing revealed more
extensive and non-specific off-target ef-
fects. One important study by Jackson et
al. [62] demonstrated that these off-target
effects could be linked to partial sequence
identity of the siRNA to unintended tar-
gets. With an siRNA concentration of
100 nM, dozens of off-target genes were
significantly down-regulated. In some
cases, these off-target signatures could be
minimized with lowered siRNA concentra-
tions; however, concomitant reduction of
targeted gene silencing was also observed.
Jackson et al. showed that for antisense-
biased RISC entry, the majority of the off-
target signature possesses sequence simi-
larity to the antisense strand; likewise,
when the sense strandRISC interactions
are favored, off-target effects are predomi-
nantly induced by the sense strand. These
off-target effects occur as a result of partial
complementarity to untargeted mRNA
strands. This partial sequence complemen-
tarity is not random and involves two sepa-
rate regions of the siRNA: 1) regions con-
taining a relatively long stretch of identity
(1114 contiguous nucleotides); and 2) 5'
regions of either strand containing eight
to nine nucleotides of identity. In another
study, Semizarov et al. [63] identified a
variety of strategies to minimize off-target
effects, including the selection of highly
potent siRNAs (thereby reducing siRNA
concentration needed to produce targeted
gene knockdown) and careful bioinfor-
matic and thermodynamic screening to
eliminate siRNAs that would interact with
non-target mRNA. The importance and
mechanics of bioinformatics for siRNA de-
sign will be expanded upon in the follow-
ing section.
10.3.2.1 Bioinformatics
Bioinformatics plays an important role in
the design and selection of functional
siRNAs. Current design strategies most
commonly rely on the basic local alignment
search tool, BLASTn, which is used specifi-
cally to compare a nucleotide query
sequence to a nucleotide database. The
BLAST family of programs was initially de-
veloped for quick identification of signifi-
cant sequence similarities between relative-
ly long stretches of nucleic acids and/or pro-
teins (www.ncbi.nlm.nih.gov/BLAST/) [65,
66]. The specificity of any given siRNA will
only be as good as the database of mRNA
sequences against which an analysis by
BLASTn is performed. Incomplete or poorly
annotated databases may result in the un-
intentional selection of siRNAs that have
high similarity to other genes. Recom-
mended sources of dependable informa-
tion are curated, organism-specific data-
bases of expressed sequences, many of
which are hosted by the National Center
for Biotechnology Information (NCBI) in
the United States, the Sanger Center, the
DNA Databank of Japan (DDBJ), or the
European Bioinformatics Institute (EBI), a
part of the European Molecular Biology In-
stitute. The Reference Sequence (RefSeq)
(www.ncbi.nlm.nih.gov/RefSeq/), Entrez
Gene (www.ncbi. nlm.nih.gov/entrez/
query.fcgi?db=gene), or UniGene (www.
ncbi.nlm.nih.gov/UniGene/) databases are
representative of such repositories of non-
redundant expressed sequence information
and are valuable resources for designing
siRNAs. However, the ability of bioinfor-
matics to predict off-target effects caused
by a specific siRNA is sometimes rather
limited. A high degree of identity (17
18 nt) between non-target mRNAs and the
targeting siRNA may lead to silencing of
the corresponding non-target genes. How-
ever, it is difficult to predict the effect of
shorter degrees of identity on the off-target
behavior of siRNAs. While BLASTn pos-
sesses the advantage of being able rapidly
to identify mRNA sequences with high
identity to potential siRNAs, it lacks the
ability to identify siRNAs that have a small
number of single mismatches to the target
mRNA. For example, an siRNA with iden-
tity to an mRNA sequence except for mis-
matches at positions 7, 13, and 14 may
prove functional but would be missed by a
BLASTn analysis. The Smith-Waterman dy-
namic programming sequence alignment
algorithm [67, 68] provides a more compre-
hensive means of identifying siRNAs that
have a low probability of causing off-target
silencing. However, this algorithm requires
extensive computational time (typically
hours to days) to perform the necessary cal-
culations, which in most cases exceeds the
practical limits of research and industry re-
quirements. Naito et al. [58] have developed
a new web-based bioinformatics program
(siDirect) designed to further optimize the
siRNA design process and minimize off-tar-
get effects based on sequence analysis.
siDirect can identify potential off-target
hybridization that BLAST analysis might
not be able to recognize. This approach of-
fers the advantages of rapid analysis and
the ability to identify sequences that have
a high probability of causing off-target ef-
fects. siDirect appears to provide a practical,
intermediate alternative between the
BLASTn and SmithWaterman bioinfor-
matics resources. However, with any of
these bioinformatics resources, the predic-
tion of off-target effects from sequence in-
formation alone is not always reliable. For
F
i
g
.
1
0
.
7
O
f
f
-
t
a
r
g
e
t
e
f
f
e
c
t
s
o
f
s
i
R
N
A
s
p
o
s
s
e
s
s
i
n
g
s
i
g
n
i
f
i
c
a
n
t
i
d
e
n
t
i
t
y
w
i
t
h
o
f
f
-
t
a
r
g
e
t
g
e
n
e
s
.
H
e
L
a
c
e
l
l
s
w
e
r
e
t
r
a
n
s
f
e
c
t
e
d
w
i
t
h
f
o
u
r
s
e
p
a
r
a
t
e
h
i
g
h
l
y
f
u
n
c
t
i
o
n
a
l
(
F
9
0
)
s
i
R
N
A
s
(
C
y
c
l
o
1
8
,
C
y
c
l
o
5
0
,
C
y
c
l
o
6
6
,
a
n
d
C
y
c
l
o
8
0
)
t
h
a
t
t
a
r
g
e
t
e
d
p
e
p
t
i
d
y
l
p
r
o
l
y
l
i
s
o
m
e
r
a
s
e
B
(
P
P
I
B
)
w
i
t
h
1
0
0
%
i
d
e
n
t
i
t
y
(
a
n
t
i
s
e
n
s
e
s
t
r
a
n
d
)
w
h
i
l
e
p
o
s
s
e
s
s
i
n
g
i
d
e
n
t
i
t
y
o
f
b
e
t
w
e
e
n
1
4
1
7
b
a
s
e
s
(
s
e
n
s
e
s
t
r
a
n
d
)
w
i
t
h
t
h
e
o
t
h
e
r
g
e
n
e
s
s
h
o
w
n
.
N
o
n
e
o
f
t
h
e
s
e
o
t
h
e
r
g
e
n
e
s
e
x
h
i
b
i
t
e
d
s
i
g
n
i
f
i
c
a
n
t
u
p
-
o
r
d
o
w
n
-
r
e
g
u
l
a
t
i
o
n
,
w
h
i
l
e
i
n
a
l
l
c
a
s
e
s
P
P
I
B
e
x
p
r
e
s
s
i
o
n
l
e
v
e
l
s
w
e
r
e
d
o
w
n
-
r
e
g
u
l
a
t
e
d
b
y
a
p
p
r
o
x
i
m
a
t
e
l
y
1
0
-
f
o
l
d
.
example, Anderson et al. performed experi-
ments to determine whether siRNA se-
quences could be intentionally designed to
produce off-target effects (unpublished re-
sults). They designed siRNAs containing
anywhere from 1417 bases of identity with
known genes. Although successful in silen-
cing the target gene (PPIB, cyclophilin B),
these siRNAs did not produce significant si-
lencing of the off-target genes to which they
had significant identity (Fig. 10.7). There-
fore, sequence information is not always a
reliable predictor of off-target effects; addi-
tional strategies for understanding and
minimizing these effects needed to be de-
veloped.
10.3.2.2 Pooling of Individual siRNAs
One approach for reducing off-target ef-
fects involves pooling a number of siRNAs
designed to target different regions of the
same gene. Individual siRNAs targeting
the same gene may produce a variety of
levels of silencing and variation in pheno-
type produced (Fig. 10.8a). When four of
these rationally designed individual si-
RNAs are pooled and used at the same to-
tal concentration as the individual siRNAs
in separate experiments, the pool produces
effective gene silencing and eliminates
misleading off-target false positive pheno-
types that may be produced by an individ-
ual siRNA (for example, the low cell viabil-
Fig. 10.8 (a) Pools of rationally designed siRNAs
retain potency while minimizing false-positive phe-
notypes. Four individual siRNAs (d1, d2, d3, and
d4, each at 100 nM) and one pool consisting of
the four individual siRNAs (total concentration
100 nM) were targeted against MEK1 and MEK2
genes. Light gray bars show level of cell survival
compared to controls. Black bars show level of
gene expression compared to controls. Pools of
targeting siRNAs exhibit effective gene knockdown
while minimizing false-positive phenotype of cell
death.
Fig. 10.8 (b) Microarray analysis of off-target gene
silencing by four individual siRNAs and by pool
consisting of the four individual siRNAs. A and B
are biological replicates of experiment. 1, 2, 3, and 4
are four individual siRNAs targeting PPIB. The
pool shows off-target silencing (green rows) of
genes at level comparable to the individual siRNA
producing the lowest level of off-target silencing.
Fig. 10.9 Chemical modification of siRNA may
substantially decrease levels of off-target activity.
Unmodified and modified siRNAs of identical se-
quences targeting mitogen-activated protein kinase
1 (MAPK1) were introduced into HeLa cells. Gene
expression changes were evaluated by Agilent
Human 1A arrays. Heat map of gene regulation:
upper row shows extensive off-target silencing by
unmodified duplex, lower row shows minimization
of off-target effects (without reduction in target
gene silencing) induced by duplex with both
strands modified.
ity resulting from use of the 1d siRNA tar-
geting MEK1 in Fig. 10.8a). Microarray
analysis of off-target signatures supports
the mitigating effects of the use of a pool
of siRNAs (Fig. 10.8b). The number of
genes down-regulated by a pool of four
siRNAs is comparable to the number
knocked down by the individual constitu-
ent siRNA that produced the lowest off-tar-
get signature, and is significantly fewer
than the individual siRNA that produced
the highest level of off-target gene silenc-
ing. The use of pools of rationally de-
signed siRNAs, then, can be an effective
tool for minimizing off-target silencing
and concomitant false positive phenotypes.
10.3.2.3 Chemical Modifications
Additional specificity of gene silencing can
be produced by the use of a variety of
commercially available, chemically modi-
fied siRNA duplexes. Modifications to the
sense strand have been shown to interfere
with this strands interaction with the
RISC, resulting in minimization of sense-
strand mediated off-target effects. Recent
investigations by Leake et al. (unpublished
results) have shown that specific modifica-
tions of both sense and antisense strands
can minimize off-target effects produced
by both strands (Fig. 10.9) while retaining
significant potency for targeted gene si-
lencing.
Clearly, a variety of integrated strategies
and technologies will be needed to con-
tinue to improve the specificity of siRNAs,
resulting in retention of effective targeted
gene silencing while minimizing unwant-
ed and potentially dangerous side effects.
10.4
Therapeutic Applications of RNAi
A number of companies are actively pur-
suing RNAi-based drug development pro-
grams. These include Acuity (Philadelphia,
PA, USA), Alnylam (Cambridge, MA,
USA), Benitec (St. Lucia, Queensland,
Australia), CytRx (Los Angeles, CA, USA),
Genta (Berkeley Heights, NJ, USA), Intra-
digm (Rockville, MD, USA), Isis (Carlsbad,
CA, USA), Mirus Bio (Madison, WI,
USA), Nucleonics (Malvern, PA, USA),
and Sirna Therapeutics (Boulder, CO,
USA) [69]. All attempts to bring siRNAs
into the marketplace as successful thera-
peutic agents will need to address several
potential obstacles, including potency,
specificity, delivery, stability, and bioavail-
ability and biodistribution.
10.4.1
Potency
Many reports have described the high po-
tency of siRNAs; effective gene silencing
has been observed in the sub-nanomolar
range (Dharmacon, unpublished results).
This potency is generally several orders of
magnitude greater than that found for
antisense or ribozymes, although at least
one study found very low levels of ribo-
zymes less than one copy per cell to
be effective [70].
10.4.2
Specificity
Specificity of any drug is paramount, as
the negative consequences of adverse side
effects can sometimes compete with the
benefits conferred by the drug. Considera-
tions and strategies for improving siRNA
specificity were discussed in Section
10.3.2. Briefly, specific chemical modifica-
tions to siRNAs, readily performed using
current RNA synthesis technologies, have
already been shown to increase specificity
of action and decrease off-target gene si-
lencing. Currently, commercial products
are available that enhance specificity with-
out sacrificing potency, and improvements
in this area are on the horizon. These
products take advantage of some of the in-
sights into the RNAi mechanism gained
through basic research and described pre-
viously, and they provide significant en-
hancements of specificity.
10.4.3
Delivery
Perhaps the key obstacle to overcome be-
fore siRNAs can be used as successful
therapeutic agents is effective delivery (see
also Part I, Chapter 7 and Part VI, Chap-
ters 1, 3, and 6). Current strategies for the
delivery of RNA and DNA molecules into
non-human animal models include elec-
troporation [71, 72], ultrasound [73, 74],
and rapid infusion of relatively large vol-
umes of solution [75]. In this last study,
Lewis et al. reported that transgene expres-
sion could be inhibited in postnatal mice
by rapid injection of a large volume of siR-
NA (along with the plasmid coding for the
transgenes) into the tail vein [75], echoing
the success of similar reports [76, 77].
More recent work has identified additional
delivery methods that may be more appli-
cable to strategies for use in pre-clinical
trials or in clinical settings. Ge et al. [78]
reported the inhibition of influenza virus
production in mice using two different de-
livery methods. First, intravenous delivery
of polyethyleneimine-complexed siRNAs
was shown to inhibit influenza virus pro-
duction. Second, intratracheal (as well as
intravenous) delivery of a DNA vector cod-
ing for shRNAs against the virus proved
effective in inhibiting virus production in
the lungs. In another study, Minakuchi et
al. [79] used Atelocollagen-complexed siR-
NA to inhibit targeted gene expression in
two separate xenograft tumors. They also
determined that these AtelocollagensiR-
NA complexes were resistant to nuclease
degradation in vitro, making them good
candidates for delivery of siRNAs in long-
er-term therapeutic regimens. Some of
these methods, shown to be effective in vivo
with animal models, may some day be
adapted for use in humans, though much
additional research needs to be performed
before this is possible.
10.4.4
Stability
One of the obstacles to overcome in the
use of siRNAs as therapeutic agents is the
low stability of unmodified RNA in human
serum. Endo- and exonucleases present in
human serum rapidly degrade RNA, ren-
dering unmodified RNA-based agents rela-
tively ineffective in therapeutic applications.
Fortunately, current RNA synthesis strate-
gies allow a variety of chemical modifica-
tions that can dramatically increase siRNA
stability in vivo. These modifications in-
clude the addition of a 2'-O-methyl group
at key positions on the sense or antisense
strand that confer nuclease resistance to
the siRNA molecule. Additional modifica-
tions include phosphorothioate internucleo-
tide linkages; Harborth et al. [80] found
that these modifications placed at various
positions in the siRNA did not signifi-
cantly affect siRNA-induced gene silenc-
ing, though in some cases duplexes with
greater than 50% phosphorothioate con-
tent led to cytotoxicity and reduced cell
growth and viability. Nucleotides modified
with 2'-fluoro moieties on the ribose sugar
also confer nuclease resistance, and Har-
borth et al. found that siRNAs with these
modifications silenced as efficiently as the
unmodified siRNAs, with no non-specific
toxicity. There are a host of additional
modifications that will increase siRNA sta-
bility, including 4'-thio-beta-d-oligoribonu-
cleotides [81] and methylated, 2'-O-amino-
propyl oligoribonucleotides [82]; these and
other modifications remain to be tested for
their effect on siRNA function and on gen-
eral cellular metabolism. Commercially
available siRNAs that are chemically modi-
fied to increase stability already exist. The
utility of these molecules is affected by
how they are tested and the biological sys-
tem in which they are employed; the abil-
ity to deliver these stabilized siRNAs to a
specific tissue in vivo will impact the func-
tionality of the siRNA.
10.4.5
Bioavailability and Biodistribution
siRNAs and their analogs (shRNAs and
miRNAs) occur naturally in vivo, making
them promising agents for therapeutic
use. In the study by Lewis et al. cited
above [75], the injection of siRNAs into a
tail vein resulted in targeted transgene
knockdown in the liver, spleen, lung,
kidney, and pancreas, indicating that the
siRNA was distributed widely throughout
the body while retaining its potency. The
in vivo use of oligonucleotide radiopharma-
ceuticals can be imaged using current
non-invasive methods such as single
photon emission computed tomography
(SPECT) and positron emission tomogra-
phy (PET) (see also Part V, Chapters 4 and
5). The latter technology is especially use-
ful for in vivo studies of RNAi. PET is the
molecular imaging technique with the
greatest sensitivity (of particular use for
siRNAs, which can be effective gene
silencers in the sub-nanomolar range) and
with the greatest level of quantitative abil-
ity [83]. However, reliable, targeted distri-
bution currently remains a challenge, and
efforts to develop effective conjugates or
delivery vehicles that enable delivery of
siRNA (and other molecules) to the appro-
priate tissue or organ for ready distribu-
tion and rapid uptake continue (see also
Part I, Chapter 7; Part V, Chapter 6; and
Part VI, Chapters 1, 3, and 6).
10.4.6
Target Validation
The drug development process historically
is a lengthy series of phases that includes
target identification, target validation, lead
identification, lead validation, and pre-clin-
ical and clinical testing prior to final FDA
approval (see also Part III, Chapter 3). Tar-
get validation is currently the primary bot-
tleneck; because of the linear nature of the
development process, it is often plagued
by costly late-stage failures. One of the key
advantages of the use of siRNA in drug
development is its capacity for streamlin-
ing the target validation stage of the pro-
cess. The broad utility of siRNA both as a
discovery tool and as a potential therapeu-
tic itself reduces the time to market with
fewer late stage setbacks.
Current drug discovery and development
programs are fed by fast-paced genome-se-
quencing projects. These projects define
the critical sets of genes that delineate nor-
mal biological function and lead to an un-
derstanding of how genetic mutations or
pathogens interfere with this normal func-
tion (see also Part I, Chapter 2). To this
end, cataloging whole genomes facilitates
the identification of potential gene candi-
dates against which small molecules or
therapeutic agents may be developed to
alleviate or abrogate disease-related syn-
dromes or specific pathologies (see also
Part I, Chapter 4). However, the drug de-
velopment process and more specifically
target validation is often hampered by
the plethora of genomic or mRNA se-
quence information, much of which re-
mains to be fully characterized. While
structure and function may be predicted
from this genomic data, validation of can-
didate genes as suitable targets requires re-
liable, practical approaches to performing
screens for functional analysis. Therefore,
even with complete sequence information
in hand, characterization of individual
genes can be an involved and daunting
task. The serendipitous discovery of RNAi
could not have been more opportune for
the pharmaceutical industry, as the rapid
output of functional information made
possible by RNAi-based strategies alleviates
the bottleneck of target validation.
Recent high-throughput analytical
approaches include combining siRNA-
mediated gene silencing with sophisticated
microarray assays, complex cell-based as-
says, and comprehensive bioinformatics
(see also Part I, Chapter 3 and Part V,
Chapter 8). For example, several microar-
ray studies (described previously in Section
10.3) of siRNA-treated cell populations re-
vealed both the occurrence of unintended
off-target effects and the actual genome-
wide expression profile of targeted gene si-
lencing. These studies led to the develop-
ment of modification strategies that en-
hance the specificity of siRNA-mediated si-
lencing. In another study, siRNA-mediated
silencing coupled with a sophisticated cell-
based assay that employs a reliable, robust
image analysis system permitted a com-
plete phenotypic assessment of the effects
of siRNA-induced knockdown of genes
that are involved in the cell cycle (Dharma-
con and Amersham, unpublished results).
Studies such as these illustrate the poten-
tial of new technologies to capture the
widespread cellular impact of modulating
gene function. The ultimate goal of inte-
grating RNAi biochemistry and biology,
high-throughput cell-based functional anal-
yses, and bioinformatics is to provide a
complete assessment of the biological im-
pact of small molecule therapies. The com-
bination of these methodologies promises
to accelerate the pace of drug discovery
and enhance the reliability of early target
identification and validation, maximizing
the investment in successful therapeutic
solutions.
10.4.7
RNAi in the Treatment of Viral Infection
and Cancer
The RNAi mechanism holds great promise
for development of antiviral agents and
therapeutic regimens (for a review, see
Ref. [84]). The antiviral action of RNAi is
naturally present in a wide range of organ-
isms. RNAi in plants appears to be an evo-
lutionarily conserved mechanism for pro-
tection against viruses. Several pieces of
data point to this virus-induced gene
silencing function of RNAi in plants. For
example, Lindbo et al. showed that natural
infection by plant viruses results in a
strong gene silencing response [85]. Reci-
procally, artificially induced RNAi in plants
can suppress viral infection [86]. Another
piece of evidence points to an evolutionary
link between plant viruses and RNAi:
plant viruses code for a variety of RNAi in-
hibitors [8790].
RNAi also appears to be involved in anti-
viral activity in invertebrates. Infection of
mosquito cells or whole organisms with
Sindbis virus carrying fragments of the
dengue virus genome inhibited replication
of the dengue virus [9193]. Of particular
interest is the finding that mosquito cells
transformed with a plasmid that tran-
scribed inverted-repeat RNA from the den-
gue virus were protected from viral infec-
tion and, notably, produced a collection of
RNA molecules 2125 nt long. These
small RNAs were complementary to sense
or antisense regions of the encoded virus
RNA genome [94]. Additional support for
the antiviral role of RNAi in invertebrates
was provided by a study in Drosophila cells
by Li et al. [95]. These authors showed that
flock house virus (FHV) infection of Droso-
phila S2 cells resulted in cellular produc-
tion of FHV-specific siRNAs.
Early consideration of whether RNAi
was involved in antiviral activity in mam-
mals raised several issues [84]. Unlike
plants and invertebrates, mammalian cells
possess a well-developed interferon re-
sponse to viruses and long dsRNA mole-
cules. It was possible that the development
of this antiviral mechanism resulted in the
loss of any RNAi-based antiviral defense
system. Even when it was discovered that
RNAi was active in mammalian cells,
there remained several potential obstacles
to implementing RNAi as a means to tar-
get viruses. RNA virus genomes are often
protected by a variety of structural protein
and nucleoprotein molecules, making
them less susceptible to siRNA-directed
degradation. A large portion of newly
synthesized viral-coded RNA is rapidly sur-
rounded by capsids, providing further pro-
tection against host-mediated degradation.
However, numerous studies have now
shown that siRNAs can block infection by
a variety of viruses; these include two ret-
roviruses, the human immunodeficiency
virus (HIV) [10, 14, 96] and Rous sarcoma
virus [97]; both positive-stranded (polio-
virus [98] and hepatitis C [99]) and nega-
tive-stranded (respiratory syncytial virus
[100] and influenza [101]) RNA viruses;
and the human papillomavirus, a DNA
virus [102]. This last report, involving sup-
pression of human papillomavirus expres-
sion, illustrated the specificity of RNAi.
Jiang and Milner found that levels of the
p53 protein (an important tumor suppres-
sor and cell cycle inhibitor) were stabilized
in Hdm2-deficient cell lines when these
cells were treated with siRNAs specific to
E6 (a viral-coded oncoprotein), which in
these cells is the sole regulator of p53. In
addition, Jiang and Milner showed that
the stabilization of p53 levels was a result
of specific siRNA-mediated effects and was
not a generalized stress response.
Recent investigations have shown the
ability of siRNA to inhibit viral replication
in vivo. As described previously, Ge et al.
[78] were able to inhibit influenza virus
production in mice using various methods
of delivery of polyethyleneimine (PEI)-
complexed siRNAs targeting the viral nu-
cleocapsid protein and components of the
viral RNA transcriptase. They found that
intravenous injection of the PEIsiRNA
complex inhibited virus production
whether given before or after influenza
infection. This group also used an shRNA-
expressing DNA vector and found that this
system could also effect inhibition of influ-
enza virus infection.
siRNA-mediated RNAi also holds prom-
ise for the treatment of cancer. The study
by Jiang and Milner [102] was performed
in human cervical carcinoma cells; greater
than 90% of human cervical carcinoma
cells contain papillomavirus. As described
above, Jiang and Milner were able selec-
tively to silence expression of genes from
this virus. In another study, Filleur et al.
[103] used siRNAs to inhibit expression of
the angiogenesis-stimulating vascular en-
dothelial growth factor (VEGF) in mice.
This gene suppression allowed administra-
tion of the anti-angiogenic molecule
thrombospondin-1 (TSP1) to significantly
inhibit tumor growth. These and other in
vivo studies illustrate the potential of siR-
NA-directed RNAi to effectively treat viral
diseases and cancers in man.
10.4.8
Optimization of the Drug Development
Process
10.4.8.1 High-throughput, High-content
Analysis
The high potency and high specificity of
siRNAs makes this technology well-suited
for high-throughput screening strategies.
For example, a collaboration between No-
vartis, Kalypsys, and the Scripps Research
Institute [104] produced a genome-wide
functional profiling of the mammalian ac-
tivator protein-1 (AP-1) signaling pathway.
This study investigated approximately
20 000 cDNAs for their ability to modulate
AP-1, which is involved in cell growth and
mitogenesis. After identification of *129
cDNAs that increased AP-1 reporter activ-
ities, a series of additional analytic tests
were performed, including use of intro-
duced siRNAs and plasmid-expressed
shRNAs to further characterize the AP-1
pathway. This study demonstrated the
practicability of using RNAi in the rapid,
large-scale, high-content functional charac-
terization of important mammalian cellu-
lar pathways. A high-throughput func-
tional genomics study that made even
more extensive use of RNAi was per-
formed by Aza-Blanc et al. [105] of the No-
vartis Foundation. This study investigated
the biology and mechanism of TRAIL-in-
duced apoptosis. TRAIL is a widely ex-
pressed member of the tumor necrosis fac-
tor (TNF) superfamily that induces selec-
tive cytotoxicity of tumor cells after bind-
ing to its cognate receptors. Aza-Blanc et
al. used a library of 510 siRNAs to test the
effect of their targeted gene silencing on
apoptosis and TRAIL-mediated signaling
pathways. Their study resulted in the
further delineation of the TRAIL-induced
apoptotic response. In addition, they were
able to determine the function of two pre-
viously unknown genes and characterize
the role of known genes in the apoptotic
pathway under study.
10.5
Summary: The Future of RNAi
in Biopharmaceutical Development
Gene silencing by siRNAs has emerged as
an extremely useful technology to knock
down expression of specific genes and al-
low for assessment of gene function. In
addition, RNAi technology possesses a lev-
el of potency and specificity that makes its
implementation as a therapeutic interven-
tion strategy very appealing (see Part I,
Chapter 1). While the application of the
technology as an in vitro functional geno-
mics tool has been well established, there
are several challenges that need to be over-
come before it can be considered a viable
biopharmaceutical approach. The chal-
lenges facing siRNA are similar to those
that any potential drug candidate would
face: 1) ensuring highly potent target inhi-
bition; 2) achieving appropriate target
specificity; 3) assuring stability of the ac-
tive drug in biological fluids; 4) directing
distribution to the appropriate target or-
gan; and 5) minimizing target-based or
chemical class-based toxicity. Strategies to
address some of these issues include ra-
tional siRNA sequence selection (based on
bioinformatics and sophisticated design al-
gorithms) and the use of chemical modifi-
cations of, and conjugation to, siRNA that
enhance serum stability, pharmacokinetics,
and biodistribution. The primary challenge
to the therapeutic use of siRNA continues
to be efficient delivery of the molecules to
target tissues and organs. Several groups
are attempting to address this issue via lo-
calized delivery to target organs (e.g., intra-
vitreal injection), though long-term strate-
gies must address effective systemic deliv-
ery to target tissues. Conjugation strategies
using lipid and polymer formulations that
allow targeted and measured drug release
or optimized viral delivery methods may
prove useful in this regard.
The successful development and imple-
mentation of RNAi-based therapeutic strat-
egies will rely on an interdisciplinary
approach requiring contributions from
bioinformatics, molecular and cell biology,
chemistry, and pharmacology. The discov-
ery of RNAi has, in a very short time, ini-
tiated a revolution in molecular biology
and the study of gene expression. The up-
side of RNAi is tremendous, as it pos-
sesses characteristics that may make it a
standard of hope for the effort to develop
biopharmaceuticals to previously intract-
able human diseases.
Acknowledgments
The authors thank Cindy McElhiney and
Julia Kendall for help with the manuscript
preparation, and Mike Sportiello and Jon
Karpilow for invaluable contributions and
discussions regarding this chapter.
References
1 C. Napoli, C. Lemieux and R. Jorgensen, Plant
Cell 1990, 2(4), 279289.
2 C. Cogoni, N. Romano and G. Macino, A. Van
Leeuwenhoek 1994, 65(3), 205209.
3 A. Fire, S. Xu, M. K. Montgomery, S. A. Kos-
tas, S. E. Driver and C. C. Mello, Nature 1998,
391(6669), 806811.
4 S. Guo and K. J. Kemphues, Cell 1995, 81(4),
611620.
5 T. Tuschl, P. D. Zamore, R. Lehmann, D. P.
Bartel and P. A. Sharp, Genes Dev. 1999,
13(24), 31913197.
6 S. M. Elbashir, W. Lendeckel and T. Tuschl,
Genes Dev. 2001, 15(2), 188200.
7 S. M. Elbashir, J. Harborth, W. Lendeckel, A.
Yalcin, K. Weber and T. Tuschl, Nature 2001,
411(6836), 494498.
8 B. Spankuch-Schmitt, J. Bereiter-Hahn, M.
Kaufmann and K. Strebhardt, J. Natl. Cancer
Inst. 2002, 94(24), 18631877.
9 S. Moskalenko, D. O. Henry, C. Rosse, G. Mi-
rey, J. H. Camonis and M. A. White, Nat. Cell
Biol. 2002, 4(1), 6672.
10 C. D. Novina, M. F. Murray, D. M. Dykxhoorn,
P. J. Beresford, J. Riess, S. K. Lee, R. G. Coll-
man, J. Lieberman, P. Shankar and P. A.
Sharp, Nat Med. 2002, 8(7), 681688.
11 S. F. Soriano, P. Hernandez-Falcon, J. M.
Rodriguez-Frade, A. M. de Ana, R. Garzon,
C. Carvalho-Pinto, A. J. Vila-Coro, A. Zaballos,
D. Balomenos, C. Martinez-A. and
M. Mellado, J. Exp. Med. 2002, 196(3), 311
321.
12 F. Wu, W. Yan, J. Pan, J. Morser and Q. Wu,
J. Biol. Chem. 2002, 277(19), 1690016905.
13 J. D. Debes, L. J. Schmidt, H. Huang and D. J.
Tindall, Cancer Res. 2002, 62(20), 56325636.
14 J. Capodici, K. Kariko and D. Weissman, J.
Immunol. 2002, 169(9), 51965201.
15 M. Leirdal and M. Sioud, Biochem. Biophys.
Res. Commun. 2002, 295(3), 744748.
16 P. D. Zamore, T. Tuschl, P. A. Sharp and D. P.
Bartel, Cell 2000, 101(1), 2533.
17 A. Nykanen, B. Haley and P. D. Zamore, Cell
2001, 107(3), 309321.
18 E. Billy, V. Brondani, H. Zhang, U. Muller
and W. Filipowicz, Proc. Natl. Acad. Sci. USA
2001, 98(25), 1442814433.
19 J. Martinez, A. Patkaniowska, H. Urlaub, R.
Luhrmann and T. Tuschl, Cell 2002, 110(5),
563574.
20 D. S. Schwarz, G. Hutvagner, B. Haley and
P. D. Zamore, Cell 2002, 10(3), 537548.
21 H. Zhang, F. A. Kolb, V. Brondani, E. Billy
and W. Filipowicz, EMBO J. 2002, 21(21),
58755885.
22 D. S. Schwarz, G. Hutvagner, T. Du, Z. Xu, N.
Aronin and P. D. Zamore, Cell 2003, 115(1),
199208.
23 H. Zhang, F. A. Kolb, L. Jaskiewicz, E. West-
hof and W. Filipowicz, Cell 2004, 118, 5768.
References 265
24 E. Bernstein, A. A. Caudy, S. M. Hammond
and G. J. Hannon, Nature 2001, 409(6818),
363366.
25 Q. Liu, T. A. Rand, S. Kalidas, F. Du, H.-E.
Kim, D. P. Smith and X. Wang, Science 2003,
301(5641), 19211925.
26 S. M. Hammond, E. Bernstein, D. Beach and
G. J. Hannon, Nature 2000, 404(6775), 293
296.
27 H. Tabara, E. Yigit, H. Siomi and C. C. Mello,
Cell 2002, 109(7), 861871.
28 M. Tijsterman, R. F. Ketting, K. L. Okihara,
T. Sijen and R. H. Plasterk, Science 2002,
295(5555), 694697.
29 A. Ishizuka, M. C. Siomi and H. Siomi, Genes
Dev. 2002, 16(19), 24972508.
30 D. Wu-Scharf, B.-R. Jeong, C. Zhang and H.
Cerutti, Science 2000, 290, 11591162.
31 W. Stapleton, S. Das and B. McKee, Chromoso-
ma 2001, 110(3), 228240.
32 M. Fagard, S. Boutet, J. B. Morel, C. Bellini
and H. Vaucheret, Proc. Natl. Acad. Sci. USA
2000, 97(21), 1165011654.
33 S. M. Hammond, S. Boettcher, A. A. Caudy,
R. Kobayashi and G. J. Hannon, Science 2001,
293(5532), 11461150.
34 J.-J. Song, S. K. Smith, G. J. Hannon and
L. Joshua-Tor, Science 2004, 305(5689), 1434
1437.
35 S. M. Elbashir, J. Martinez, A. Patkaniowska,
W. Lendeckel and T. Tuschl, EMBO J. 2001,
20(23), 68776888.
36 J. Martinez and T. Tuschl, Genes Dev. 2004, 18,
975980.
37 G. Meister, M. Landthaler, A. Patkaniowska,
Y. Dorsett, G. Teng and T. Tuschl, Cell 2004,
15(2), 185197.
38 A. Lingel, B. Simon, E. Izaurralde and M. Sat-
tler, Nature 2003, 426, 465469.
39 D. Bartel, Cell 2004, 116(2), 281297.
40 Y. Lee, K. Jeon, J. T. Lee, S. Kim and V. N.
Kim, EMBO J. 2002, 21(17), 46634670.
41 J. Pham, J. Pellino, Y. Lee, R. Carthew and E.
Sontheimer, Cell 2004, 117(1), 8394.
42 G. Hutvagner and P. D. Zamore, Science 2002,
297(5589), 20562060.
43 J. Kim, A. Krichevsky, Y. Grad, G. D. Hayes,
K. S. Kosik, G. M. Church and G. Ruvkun,
Proc. Natl. Acad. Sci. USA 2004, 101, 360365.
44 J. C. Carrington and V. Ambros, Science 2003,
301(5631), 336338.
45 R. H. Plasterk, Science 2002, 296(5571), 1263
1265.
46 G. Stark, I. Kerr, B. Williams, R. Silverman
and R. Schreiber, Annu. Rev. Biochem. 1998,
67, 227264.
47 T. Holen, M. Amarzguioui, M. T. Wiiger, E.
Babaie and H. Prydz, Nucleic Acids Res. 2002,
30(8), 17571766.
48 T. Yokota, N. Sakamoto, N. Enomoto, Y. Ta-
nabe, M. Miyagishi, S. Maekawa, L. Yi, M.
Kurosaki, K. Taira, M. Watanabe and H. Mizu-
sawa, EMBO Rep. 2003, 4, 602608.
49 T. A. Vickers, S. Koo, C. F. Bennett, S. T.
Crooke, N. M. Dean and B. F. Baker, J. Biol.
Chem. 2003, 278(9), 71087118.
50 H. Hohjoh, FEBS Lett. 2002, 521(13), 195
199.
51 K. Yoshinari, M. Miyagishi and K. Taira,
Nucleic Acids Res. 2004, 32(2), 691699.
52 S. M. Elbashir, J. Harborth, K. Weber and
T. Tuschl, Methods 2002, 26(2), 199213.
53 C. Llave, K. D. Kasschau, M. A. Rector and
J. C. Carrington, Plant Cell 2002, 14(7), 1605
1619.
54 A. Khvorova, A. Reynolds and S. Jayasena, Cell
2003, 115(1), 209216.
55 A. Reynolds, D. Leake, S. Scaringe, W. S. Mar-
shall, Q. Boese and A. Khvorova, Nature Bio-
technol. 2004, 22(3), 326330.
56 K. Ui-Tei, Y. Naito, F. Takahashi, T. Haragu-
chi, H. Ohki-Hamazaki, A. Juni, R. Ueda and
K. Saigo, Nucleic Acids Res. 2004, 32, 936948.
57 M. Amarzguioui and H. Prydz, Biochem. Bio-
phys. Res. Commun. 2004, 316(4), 10501058.
58 Y. Naito, T. Yamada, K. Ui-Tei, S. Morishita
and K. Saigo, Nucleic Acids Res. 2004,
32(DOI:10.1093/nar/gkh442).
59 H. Kawasaki, E. Suyama, M. Iyo and K. Taira,
Nucleic Acids Res. 2003, 31(3), 981987.
60 T. Li, C.-Y. Chang, D.-Y. Jin, P.-J. Lin, A.
Khvorova and D. W. Stafford, Nature 2004,
427(6974), 541544.
61 P. C. Scacheri, O. Rozenblatt-Rosen, N. J. Ca-
plen, T. G. Wolfsberg, L. Umayam, J. C. Lee,
C. M. Hughes, K. S. Shanmugam, A. Bhatta-
charjee, M. Meyerson and F. S. Collins, Proc.
Natl. Acad. Sci. USA 2004, 101(7), 18921897.
62 A. L. Jackson, S. R. Bartz, J. Schelter, S. V. Ko-
bayashi, J. Burchard, M. Mao, B. Li, G. Cavet
and P. S. Linsley, Nature Biotechnol. 2003,
21(6), 635637.
63 D. Semizarov, L. Frost, A. Sarthy, P. Kroeger,
D. N. Halbert and S. W. Fesik, Proc. Natl. Acad.
Sci. USA 2003, 100(11), 63476352.
64 S. P. Persengiev, X. Zhu and M. R. Green,
RNA 2004, 10(1), 1218.
65 S. F. Altschul, W. Gish, W. Miller, E. W. Myers
and D. J. Lipman, J. Mol. Biol. 1990, 215(3),
403410.
66 T. L. Madden, R. L. Tatusov and J. Zhang,
Methods Enzymol. 1996, 266, 131141.
67 T. F. Smith and M. S. Waterman, J. Mol. Biol.
1981, 147(1), 195197.
68 O. Gotoh, J. Mol. Biol. 1982, 162(3), 705708.
69 J. Minkel, Drug Disc. Dev. 2004, 7(8), 2834.
70 M. Andng, J. Hinkula, G. Hotchkiss, S. Lars-
son, S. Britton, F. Wong-Staal, B. Wahren and
L. hrlund-Richter, Proc. Natl. Acad. Sci. USA
1999, 96, 1274912753.
71 D. Dean, D. Machado-Aranda, K. Blair-Parks,
A. Yeldandi and J. Young, Gene Ther. 2003,
10(18), 16081615.
72 L. Zhang, E. Nolan, S. Kreitschitz and D. Ra-
bussay, Biochim. Biophys. Acta 2002, 1572(1),
19.
73 A. Lawrie, A. Brisken, S. Francis, D. Cumber-
land, D. Crossman and C. Newman, Gene
Ther. 2000, 7(23), 20232027.
74 Y. Taniyama, K. Tachibana, K. Hiraoka, M.
Aoki, S. Yamamoto, K. Matsumoto, T. Naka-
mura, T. Ogihara, Y. Kaneda and R. Morishita,
Gene Ther. 2002, 9(6), 372380.
75 D. L. Lewis, J. E. Hagstrom, A. G. Loomis,
J. A. Wolff and H. Herweijer, Nat Genet. 2002,
32(107).
76 A. P. McCaffrey, K. Ohashi, L. Meuse, S. Shen,
A. M. Lancaster, P. J. Lukavsky, P. Sarnow and
M. A. Kay, Mol. Ther. 2002, 5(6), 676684.
77 A. P. McCaffrey, L. Meuse, T. T. Pham, D. S.
Conklin, G. J. Hannon and M. A. Kay, Nature
2002, 418(6893), 3839.
78 Q. Ge, L. Filip, A. Bai, T. Nguyen, H. N. Eisen
and J. Chen, Proc. Natl. Acad. Sci. USA 2004,
101(23), 86768681.
79 Y. Minakuchi, F. Takeshita, N. Kosaka, H. Sa-
saki, Y. Yamamoto, M. Kouno, K. Honma, S.
Nagahara, K. Hanai, A. Sano, T. Kato, M. Ter-
ada and T. Ochiya, Nucleic Acids Res. 2004, 32,
109.
80 J. Harborth, S. Elbashir, K. Vandenburgh,
H. Manninga, S. Scaringe, K. Weber and
T. Tuschl, Nucleic Acid Drug Dev. 2003, 13(2),
83105.
81 C. Leydier, L. Bellon, J. Barascut, F. Morvan,
B. Rayner and J. Imbach, Antisense Res Dev.
1995, 5(3), 167174.
82 M. Teplova, S. T. Wallace, V. Tereshko, G. Mi-
nasov, A. M. Symons, P. D. Cook, M. Mano-
haran and M. Egli, Proc. Natl. Acad. Sci.
USA 1999, 96, 1424014245.
83 B. Tavitian, Gut 2003, 52(Suppl IV), 4047.
84 L. Gitlin and R. Andino, J. Virol. 2003,
77(13), 71597165.
85 J. Lindbo, L. Silva-Rosales, W. Proebsting
and W. Dougherty, Plant Cell 1993, 5(12),
17491759.
86 P. Waterhouse, M. Graham and M. Wang,
Proc. Natl. Acad. Sci. USA 1998, 95(23),
1395913964.
87 R. Anandalakshmi, G. Pruss, X. Ge, R. Ma-
rathe, A. Mallory, T. Smith and V. Vance,
1307913084.
88 G. Brigneti, O. Voinnet, W. Li, L. Ji, S. Ding
and D. Baulcombe, EMBO J. 1998, 17(22),
67396746.
89 K. Kasschau and J. Carrington, Cell 1998,
95(4), 461470.
90 O. Voinnet, Y. Pinto and D. Baulcombe,
1414714152.
91 Z. Adelman, C. Blair, J. Carlson, B. Beaty
and K. Olson, Insect Mol. Biol. 2001, 10(3),
265273.
92 P. Gaines, K. Olson, S. Higgs, A. Powers,
B. Beaty and C. Blair, J. Virol. 1996, 70,
21322137.
93 K. E. Olson, S. Higgs, P. J. Gaines, A. M.
Powers, B. S. Davis, K. I. Kamrud, J. O. Carl-
son, C. D. Blair and B. J. Beaty, Science 1996,
272(5263), 884886.
94 Z. N. Adelman, I. Sanchez-Vargas, E. A. Tra-
vanty, J. O. Carlson, B. J. Beaty, C. D. Blair and
K. E. Olson, J. Virol. 2002, 76, 1292512933.
95 H. Li, W. X. Li and S. W. Ding, Science 2002,
296(5571), 13191321.
96 G. A. Coburn and B. R. Cullen, J. Virol. 2002,
76(18), 92259231.
97 W. Hu, C. Myers, J. Kilzer, S. Pfaff and F.
Bushman, Curr. Biol. 2002, 12(15), 13011311.
98 L. Gitlin, S. Karelsky and R. Andino, Nature
2002, 418, 430434.
99 J. A. Wilson, S. Jayasena, A. Khvorova, S. Sa-
batinos, I. G. Rodriguez-Gervais, S. Arya, F.
Sarangi, M. Harris-Brandts, S. Beaulieu and
C. D. Richardson, Proc. Natl. Acad. Sci. USA
2003, 100(5), 27832788.
100 V. Bitko and S. Barik, BMC Microbiol. 2001,
1(1), 34.
References 267
101 Q. Ge, M. T. McManus, T. Nguyen, C. H.
Shen, P. A. Sharp, H. N. Eisen and J. Chen,
27182723.
102 M. Jiang and J. Milner, Oncogene 2002,
21(39), 60416048.
103 S. Filleur, A. Courtin, S. Ait-Si-Ali, J. Gu-
glielmi, C. Merle, A. Harel-Bellan, P. Clzar-
din and F. Cabon, Cancer Res. 2003, 63,
39193922.
104 S. K. Chanda, S. White, A. P. Orth, R. Reis-
dorph, L. Miraglia, R. S. Thomas, P. DeJe-
sus, D. E. Mason, Q. Huang, R. Vega, D.-H.
Yu, C. G. Nelson, B. M. Smith, R. Terry, A. S.
Linford, Y. Yu, G.-W. Chirn, C. Song, M. A.
Labow, D. Cohen, F. J. King, E. C. Peters,
P. G. Schultz, P. K. Vogt, J. B. Hogenesch
and J. S. Caldwell, Proc. Natl. Acad. Sci. USA
2003, 100(21), 1215312158.
105 P. Aza-Blanc, C. Cooper, K. Wagner, S. Bata-
lov, Q. Deveraux and M. Cooke, Cell 2003,
12(3), 627637.
Abstract
One approach to overcome the transplant
rejection of human embryonic stem (ES)
cells is to derive them by nuclear transfer
of the patients own cells. In the absence
of an efficient protocol for human somatic
cell nuclear transfer (SCNT), several critical
steps must be optimized, namely repro-
gramming time, activation method, and in
vitro culture conditions. Reprogramming
time was defined as the time between cell
fusion and oocyte activation to permit prop-
er embryonic development. A 2 h repro-
gramming time led to *25% of the recon-
structed embryos developing to blastocysts.
In SCNT, in the absence of sperm-mediated
activation, an artificial stimulus is needed to
initiate embryo development. Addition of
10 lM ionophore for 5 min, and incubation
with 2.0 mM 6-dimethyl aminopurine for
4 h, was the most efficient chemical activa-
tion protocol for human SCNT embryos.
Encouragingly, inefficiencies in embryo cul-
ture have been overcome by supplementing
culture medium with different energy sub-
strates and macromolecules, or implement-
ing a sequential culture system tailored to
different stages of embryo development.
We prepared human modified synthetic ovi-
ductal fluid with amino acids (mSOFaa) by
supplementing hmSOFaa with human se-
rum albumin and fructose instead of bovine
serum albumin (BSA) and glucose, respec-
tively. The culture of human SCNT-derived
embryos in G1.2 medium for 48 h, followed
by hmSOFaa medium, produced more blas-
tocysts than using G1.2 medium for 48 h
followed by culture in G1.2 medium, or in
continuous hmSOFaa medium. The present
protocol led to the production of cloned blas-
tocysts at rates of 19 to 29%, compared to
rates with established SCNT methods of
*25% in cattle and *26% in pigs. A total
of 30 SCNT-derived blastocysts was cul-
tured, 20 inner cell masses (ICMs) were iso-
lated by immunosurgical removal of the tro-
phoblast, and one human cloned ES cell line
(SCNT-hES1) with typical ES cell morphol-
ogy and pluripotency was derived.
269
11
The First Cloned Human Embryo: An Unlimited Source of Stem Cells
for Therapeutic Cloning
ISBN: 3-527-31184-X
Mobilis in Mobile Human Embryonic Stem Cells
and Other Sources for Cell Therapy
Abbreviations
bFGF basic fibroblast growth factor
BSA bovine serum albumin
COC cumulusoocyte complex
DMAP dimethyl aminopurine
EG embryonic germ
ES embryonic stem cells
hmSOFaa human modified synthetic
oviductal fluid with amino acids
HSA human serum albumin
ICM inner cell mass
IVC in vitro culture
IVF in vitro fertilization
LIF leukemia inhibitory factor
MII metaphase II
PVA poly-vinyl alcohol
SCNT somatic cell nuclear transfer
SOF synthetic oviductal fluid
STR short tandem repeat
11.1
Introduction
The isolation of pluripotent human em-
bryonic stem (ES) cells [1], combined with
breakthroughs in somatic cell nuclear
transfer (SCNT) in mammals [2], have
raised the possibility of performing human
SCNT to generate virtually unlimited sup-
plies of undifferentiated cells for research,
with potential applications for tissue repair
and transplantation. This concept known
as therapeutic cloning refers to transfer
of the nucleus of a somatic cell into an
enucleated donor oocyte [3]. In theory, the
oocytes cytoplasm would reprogram the
transferred nucleus by silencing all the
somatic cell genes and activating the em-
bryonic genes. A certain reprogramming
time is needed to return the gene expres-
sion pattern of the somatic cell to one that
is appropriate and necessary for embryo
development. This period plays a critical
role on chromatin remodeling, and is
known to determine the developmental
competence, both in vivo and in vitro, of
SCNT embryos. ES cells would be isolated
from the inner cell masses (ICMs) of the
cloned preimplantation embryo. In a ther-
apeutic setting, these cells would carry the
nuclear genome of the patient; thus it is
proposed that, following directed cell dif-
ferentiation, the cells could be trans-
planted without immune rejection to
treat degenerative disorders such as dia-
betes, osteoarthritis, and Parkinsons dis-
ease. Previous reports in animals have
identified the possibility of therapeutic
cloning by generating bovine ES-like cells
[4] and mouse ES cells from ICMs of
cloned blastocysts [57], and development
of the cloned embryos to the 8- and 10-cell
stages [8]. The possibility that these find-
ings could be reproduced in humans was
demonstrated only recently [9]. Here, we
describe the successful derivation and
characterization of human cloned ES cells
after SCNT.
11.2
Human Somatic Cell Nuclear Transfer
(SCNT)
Before commencing human SCNT experi-
ments, approval for the study was obtained
from the Institutional Review Board on
Human Subjects Research and Ethics
Committees of Hanyang University Hospi-
tal, Seoul, Korea.
11.2.1
Donor Cells and Oocytes
The type of donor cell influences develop-
ment of the cloned embryos. Zakhartchen-
ko et al. [10] reported that nuclei from bo-
vine ear skin fibroblasts supported better
11 The First Cloned Human Embryo: An Unlimited Source of Stem Cells for Therapeutic Cloning 270
development of cloned embryos to the
blastocyst stage than did mammary gland
cells. Cho et al. [11] evaluated four types of
bovine cell (cumulus, ear fibroblasts, ovi-
duct, and uterine), and showed that cumu-
lus cells or ear fibroblasts yielded higher
rates of fusion and blastocyst formation
than did the other two cell types. Uhm et
al. [12] showed that porcine fetal fibroblast
cells could direct the development of re-
constructed oocytes to morula or blastocyst
stage at higher rates than could cumulus
cells, while Lee et al. [13] compared the de-
velopmental competence of porcine SCNT
embryos using four different donor cell
types (adult fibroblasts, fetal fibroblasts,
cumulus cells or oviductal cells) and
showed more blastocysts to be derived
from SCNT of fetal fibroblasts than from
other donor cell types. In mice, it has been
suggested that cumulus cells could be par-
ticularly suitable as nuclear donor cells
[14, 15].
In a preliminary study to determine the
optimal donor cell type for human SCNT,
we evaluated three cell types (adult fibro-
blasts from abdominal skin, cultured cu-
mulus cells, or freshly isolated cumulus
cells) using oocytes collected from surgi-
cally removed ovaries. The use of freshly
isolated cumulus cells was seen to direct
better embryo development than the other
cell types. A total of 242 oocytes was ob-
tained from 16 volunteers after ovarian
stimulation: 176 metaphase II (MII) oo-
cytes were used directly for SCNT after in-
cubation in vitro for 30 min, while the re-
maining 66 oocytes were allowed to ma-
ture to MII before use in SCNT. On the
basis of preliminary results, we performed
autologous SCNT, whereby the donors
own cumulus cell, freshly isolated from
the cumulusoocyte complex (COC) and
after treatment with 0.1% hyaluronidase,
was transferred back into the donors own
enucleated oocyte. Enucleation and injec-
tion of donor cells were performed as pre-
viously described [16]. In order to confirm
removal of the oocytes DNA during enu-
cleation, the extruded DNAMII spindle
complex from each oocyte was imaged
using Hoechst 33342 fluorescent DNA dye
(Fig. 11.1A and B).
11.2.2
Fusion and Activation
Exit from MII arrest of the ovulated oocyte
is accomplished by fertilization with
sperm, and is commonly referred to as
oocyte activation. This activation is trig-
gered by intracellular calcium oscillations
induced by fertilization [17]. Due to an ab-
sence of fertilization by sperm in SCNT
embryos, reconstructed oocytes must be
stimulated artificially to fuse between do-
nor cells and oocytes, and to initiate em-
bryo development. A variety of chemical,
physical and mechanical agents have been
shown to induce the activation of recon-
structed oocytes, with different efficiencies.
The in vitro development of bovine cloned
embryos was significantly improved by de-
layed activation with an electrical pulse for
46 h after fusion [18]. Other materials, in-
cluding a calcium inducer (calcium iono-
phore or ionomycin), a protein kinase in-
hibitor [6-dimethyl aminopurine (DMAP)]
or a protein synthesis inhibitor (cyclohexa-
mide or puromycin) were known oocyte
activators. In cattle, incubation with cyclo-
hexamide following exposure of the em-
bryo to ionomycin or electrical pulses re-
sulted in embryo development to term [19,
20]. Shin et al. [21] reported that an im-
proved development of bovine oocytes re-
constituted with ear fibroblasts was
achieved by applying a separate procedure
of electric fusion and chemical activation
(calcium ionomycin) 4 h apart. In pigs,
Betthauser et al. [22] used 6-DMAP and
calcium ionomycin for chemical activation
after electric fusion, while Lai et al. [23]
used only simultaneous electric fusion/ac-
tivation. Hyun et al. [24] reported that an
additional chemical activation with 6-
DMAP after electric stimulus did not im-
prove the developmental competence of
porcine SCNT embryos compared to si-
multaneous electric fusion/activation, indi-
cating that no further chemical stimula-
tion with ionomycin or 6-DMAP is neces-
sary for post-activation of porcine SCNT
embryos. Nakagawa et al. [25] induced
parthenogenetic activation of human oo-
cytes using the calcium ionophore A23187
for 5 min, followed by treatment with pu-
romycin, and observed a 91% activation
rate without blastocyst formation. Cibelli
et al. [8] activated human oocytes by treat-
ment with ionomycin, followed by 6-
DMAP, and reported the cleavage rate to
be 90% and the blastocyst rate as 27%.
Likewise, Cibelli et al. [8] used the same
activation protocol to activate human re-
constructed oocytes, but failed to obtain
cloned human blastocysts.
Since there is an absence of reports detail-
ing the activation of human SCNT oocytes
and the successful production of cloned
blastocysts, it was necessary to identify sev-
eral parameters, including the reprogram-
ming time (the time between cell fusion
and egg activation, returning the gene ex-
pression profile of the somatic cell to that
needed for appropriate embryonic develop-
ment) and activation methods. Based on re-
sults from animal SCNT oocytes and the
parthenogenetic activation of human oo-
cytes, we initially employed a porcine activa-
tion protocol (simultaneous fusion and acti-
vation with electrical pulse) which used hu-
Fig. 11.1 Confirmation of enucleation, photographs
of human SCNT ES cells and their undifferentiated
progeny. Images (200) of extruded DNAMII
spindle complexes (arrows) from oocyte before (A)
and after enucleation (B). The morphology of
cloned human blastocysts (C) and isolated inner
cell masses (D). The contrast (E, 100) micro-
graphs, and higher magnification (F, 200) of a
colony of SCNT-hES-1 cells. Scale bars=100 lm
(A, B, E and F) and 50 lm (C and D).
man oocytes collected from surgically re-
moved ovary, mainly because the porcine
activation protocol was simple and did not
require any reprogramming time. However,
as we observed low fusion and cleavage
rates, and no blastocyst development, we
adapted a bovine SCNT protocol of waiting
for a few hours between fusion and activa-
tion, and of using combination of electrical
pulse and a chemical, as with bovine and
human oocytes. After electrical fusion, the
reconstructed oocytes were incubated for
various times of reprogramming (2, 4, 6 or
20 h) before chemical activation. Any oo-
cytes with donor somatic cells remaining
in the perivitelline space were re-fused by
electric stimulation. Fused donor oocytes
and somatic cells were then activated in
either a calcium ionophore A123187 (5 or
10 lM) or ionomycin (5 or 10 lM) for
5 min, followed by activation with 2 mM 6-
DMAP for 4 h (Table 11.1). It was found
Table 11.1 Conditions for human somatic cell nuclear transfer
Experi- Activation condition
a)
Repro- 1
st
step 2
nd
step No. of No. (%) of cloned embryos
ment gram- medi- medium oocytes
ming
time
[h]
um
b)
2-cell Com-
pacted
morula
Blasto-
cyst
1
st
set 10 lM
Ionophore
6-DMAP 2 G 1.2 hmSOFaa 16 16 (100) 4 (25) 4 (25)
10 lM
Ionophore
6-DMAP 4 G 1.2 hmSOFaa 16 15 (94) 1 (6) 0
10 lM
Ionophore
6-DMAP 6 G 1.2 hmSOFaa 16 15 (94) 1 (6) 1 (6)
10 lM
Ionophore
6-DMAP 20 G 1.2 hmSOFaa 16 9 (56) 1 (6) 0
2
nd
set 10 lM
Ionophore
6-DMAP 2 G 1.2 hmSOFaa 16 16 (100) 5 (31) 3 (19)
5 lM
Ionophore
6-DMAP 2 G 1.2 hmSOFaa 16 11 (69) 0 0
10 lM
Ionomycin
6-DMAP 2 G 1.2 hmSOFaa 16 12 (75) 0 0
5 lM
Ionomycin
6-DMAP 2 G 1.2 hmSOFaa 16 9 (56) 0 0
3
rd
set 10 lM
Ionophore
6-DMAP 2 G 1.2 hmSOFaa 16 16 (100) 4 (25) 3 (19)
10 lM
Ionophore
6-DMAP 2 G 1.2 G 2.2 16 16 (100) 0 0
10 lM
Ionophore
6-DMAP 2 Continuous
hmSOFaa
16 16 (100) 0 0
4
th
set 10 lM
Ionophore
6-DMAP 2 G 1.2 hmSOFaa 66 62 (93) 24 (36) 19 (29)
a) Fused donor oocytes and somatic cells were activated in either
calcium ionophore A23187 (5 or 10 lM) or ionomycin (5 or
10 lM) for 5 min, followed by 2 mM 6-dimethylaminopurine
(6-DMAP) treatment for 4 h.
b) Oocytes were incubated in first medium for 48 h.
that incubation in 10 lM A23187 for 5 min,
followed by incubation with 2.0 mM 6-
DMAP for 4 h, provided an efficient chemi-
cal activation for human SCNT oocytes.
11.2.3
In-vitro Culture (IVC) of the Reconstructed
Oocytes
Although in vitro fertilization (IVF) and
embryo production using in vitro-matured
animal and human oocytes have been suc-
cessful, differential developmental compe-
tence in response to various culture media
has been demonstrated in IVF and SCNT
embryos [26], indicating the importance of
optimizing IVC conditions for preimplan-
tation development of SCNT embryos.
Many attempts have been made to over-
come the inadequacies of IVC systems by
supplementing the culture medium with
energy substrates or proteins, and the re-
sults have been encouraging (see Sections
11.2.3.1 and 11.2.3.2). Furthermore, the
implementation of sequential culture sys-
tems tailored to different stages of embryo
development has significantly improved
the percentage of embryos developing to
the blastocyst stage in vitro. The recent de-
velopment of serum-free sequential media,
formulated according to the carbohydrate
composition of the oviduct and adjusted
for the changing physiology and metabolic
requirements of the human embryo, has
led to considerable improvements in the
rate of pregnancies generated using as-
sisted reproductive technologies [27]. For
human IVF embryos, G1.2/G2.2 media are
most commonly used. Langendonckt et al.
[27] compared the developmental compe-
tency of G1.2/G2.2 sequential media with
Sydney IVF cleavage media/Sydney IVF
blastocyst media, and showed an increased
rate of blastocyst formation in the former
system compared to the latter. Macklon et
al. [28] also compared developmental com-
petency between G1.2/G2.2 sequential cul-
ture media and Rotterdam medium and,
again, identified a higher rate of blastocyst
formation in the G1.2/G2.2 media. CR2
medium [29] and synthetic oviductal fluid
(SOF) [30] are widely used to culture bo-
vine IVF and SCNT embryos. However, it
has been shown recently that the use of a
modified SOF with amino acid (mSOFaa)
improved the developmental competence
of bovine cloned embryos with regard to
cleavage rate and morula and blastocyst
formation compared to modified CR2 with
amino acid (mCR2aa). The formula of
mSOFaa is basically the same as that of
SOF, except for the glucose concentration
and the addition of essential and non-es-
sential amino acids, insulin, transferrin,
selenium and BSA. Among the media
used for IVC of porcine embryos, North
Carolina State University (NCSU)-23 medi-
um is known to be one of the most suc-
cessful [31, 32].
11.2.3.1 Role of Energy Substrates
and Protein Supplementation
for In-vitro Culture
The energy substrate is an important in-
gredient for optimum preimplantation em-
bryo development in IVC medium. Glu-
cose is widely supplemented as the major
energy substrate, and as such is known to
be important for blastocyst formation in
the post-compaction period of bovine em-
bryos [33]. However, exposure to high con-
centrations of glucose during early em-
bryonic stages caused developmental retar-
dation in many species including hamsters
[34], mice [35], rats [36], cattle [37], sheep
[38], and human [39]. Replacement of glu-
cose with fructose in the medium signifi-
cantly improved the quality of blastocysts
by increasing the number of total cells in
mice [40] or total and TE (Trophectoderm)
cells in hamsters. Recently, Kwun et al.
[16] showed that the use of 1.5 mM fruc-
tose in mSOFaa medium significantly en-
hanced blastocyst formation in both SCNT
and IVF embryos compared to 1.5 mM
glucose. As in bovine embryos, NCSU-23
containing glucose as an energy substrate
has been widely accepted to produce por-
cine embryos. This detrimental effect of
glucose during the early IVC period may
occur because early porcine embryos can-
not metabolize it readily before the 8-cell
stage. Replacing glucose with pyruvate/lac-
tate in the culture medium proved benefi-
cial for the development of porcine IVF
embryos to blastocysts [41, 42]. In porcine
SCNT embryos, culturing reconstructed
embryos in NCSU-23 medium supplemen-
ted with lactate (5.0 mM)/pyruvate
(0.5 mM) improved the in vitro develop-
ment of porcine SCNT embryos in terms
of cleavage rate and blastocyst formation
[43]. Unlike bovine embryos, the replace-
ment of glucose with fructose did not im-
prove the in vitro development of porcine
SCNT embryos (our unpublished results).
11.2.3.2 Role of Protein Supplementation
for IVC
Protein is widely used as a supplement in
culture media, and is known to improve
the developmental competence of embryos
[4446]. Serum and/or serum albumin are
commonly used as such protein sources,
though serum may negatively affect em-
bryo development [47]. If a serum-free me-
dium is required, the serum is replaced
with BSA or a synthetic macromolecule
(e.g. poly-vinyl alcohol; PVA). Supplement-
ing BSA in the culture media significantly
improved the blastocyst formation rate in
bovine IVF embryos compared to PVA
supplementation [44, 48]. In porcine IVF
embryos, BSA increased both blastocyst
formation and the number of total cells in
blastocysts [49, 50]. This beneficial effect
of BSA may vary between suppliers, and
even between lots, however [51]. Likewise,
BSA is considered a semi-defined compo-
nent that may be contaminated with fatty
acids and citrate, and also be a possible
source of disease agents [47]. When a com-
pletely defined culture medium is re-
quired, without reducing the rate of em-
bryo development, BSA can be replaced
with recombinant human serum albumin
(HSA), which has equal developmental po-
tential to BSA and is safe for culturing bo-
vine IVF embryos [52].
11.2.3.3 In-vitro Culture of Human SCNT
Oocytes
Based on results from bovine SCNT oo-
cytes, human modified SOF with amino
acids (hmSOFaa) was prepared by supple-
menting mSOFaa with HSA (10 mg mL
1
)
and fructose (1.5 mM), instead of BSA
(8 mg mL
1
) and glucose (1.5 mM), respec-
tively. For culturing human SCNT em-
bryos, sequential or continuous culture
systems were evaluated: after activation,
oocytes were washed with fresh G1.2 me-
dium and cultured in G1.2 medium for
48 h. On the third day of culture, cleaved
embryos were transferred to the second
medium (hmSOFaa or G2.2) and cultured
for another 6 days. One group of activated
oocytes was cultured in hmSOFaa
throughout the IVC period. As a result,
the reconstructed oocytes were developed
to 2-, 4-, 8- to 16-cell stages, morulae and
blastocysts. Culturing human SCNT em-
bryos in G1.2 medium for the first 48 h,
followed by hmSOFaa medium, produced
more blastocysts compared to G1.2 medi-
um for the first 48 h followed by culture
in G2.2 medium or continuous hmSOFaa
medium (see Table 11.1). Oocyte limita-
tions precluded full optimization of all the
parameters for human SCNT; nonetheless,
the protocol for oocyte activation and cul-
ture produced cloned blastocysts at rates
of 19 to 29% (as a percentage of recon-
structed eggs), these being comparable
with rates from established SCNT methods
in cattle (*25%) [16] and pigs (*26%)
[24, 53].
11.3
Establishment and Characterization
of Human SCNT ES Cells
A total of 30 SCNT-derived blastocysts was
cultured after removal of the zona pelluci-
da (ZP) with 0.1% pronase treatment
(Fig. 11.1C). In comparison, 20 ICMs were
isolated by immunosurgical removal of the
trophoblast (Fig. 11.1D), first incubating
them with 100% anti-human serum anti-
body for 20 min, followed by an additional
30 min exposure to guinea pig comple-
ment. Isolated ICMs were cultured on mi-
tomycin C mitotically inactivated primary
mouse embryonic fibroblast feeder layers
in gelatin-coated, 4-well tissue culture
dishes. The culture medium was Dulbec-
cos modified Eagles medium (DMEM)/
DMEM F12 (1: 1) supplemented with 20%
Knockout Serum Replacement, 0.1 mM b-
mercaptoethanol, 1% non-essential amino
acids, 100 units mL
1
penicillin, 100 lg
mL
1
streptomycin, and 4 ng mL
1
basic fi-
broblast growth factor (bFGF). During the
early stage of SCNT embryonic stem cell
culture, the medium was supplemented
with 2000 units mL
1
human leukemia in-
hibitory factor (LIF). As a result, one ES
cell line (SCNT-hES-1) was derived. The
cell colonies displayed similar morphology
to that reported previously for hES cells
derived from IVF (Fig. 11.1E). The SCNT-
hES-1 cells had a high nucleus to cyto-
plasm ratio, and prominent nucleoli
(Fig. 11.1F). When cultured in the defined
medium conditioned for neural cell differ-
entiation [54], SCNT-hES-1 cells differen-
tiated into nestin-positive cells, an indica-
tion of primitive neuroectoderm differen-
tiation. The SCNT-hES-1 cell line was me-
chanically passaged every 57 days using a
hooked needle, and successfully main-
tained its undifferentiated morphology
after continuous proliferation for more
than 130 passages, while still maintaining
a normal female (XX) karyotype. When
characterized for cell-surface markers,
SCNT-hES-1 cells express ES cell markers
such as alkaline phosphatase, SSEA-3,
SSEA-4, TRA-1-60, TRA-1-81, and Oct-4,
but not SSEA-1 (Fig. 11.2). As previously
described in monkey [55] and human ES
cells [1, 56, 57], and mouse SCNT-ES cells
[6], SCNT-hES-1 cells do not respond to ex-
ogenous LIF, suggesting that a pluripotent
state is maintained by a gp130-indepen-
dent pathway. Pluripotency of SCNT-hES-1
cells was tested in vitro and in vivo. For
embryoid body formation, clumps of the
cells were cultured in vitro for 14 days in
suspension (plastic Petri dishes) in
DMEM/DMEM F12 without hLIF and
bFGF. The resulting embryoid bodies were
stained with three dermal markers, and
found to differentiate into a variety of cell
types including derivatives of endoderm,
mesoderm, and ectoderm. When undiffer-
entiated SCNT-hES-1 cells (clumps consist-
ing of *100 cells) were injected into the
testes of 6- to 8-week-old SCID mice, tera-
tomas were obtained at 67 weeks after in-
jection. These teratomas contained tissue
which was representative of all three germ
layers, including neuroepithelial rosset,
pigmented retinal epithelium, smooth
muscle, bone, cartilage, connective tissues,
and glandular epithelium. Confirmation
that the cells were of SCNT origin, and
not due to the parthenogenetic activation
of oocytes, was made by performing a
DNA fingerprinting analysis with human
short tandem repeat (STR). The statistical
probability that the cells may have derived
from an unrelated donor was 8.8 10
16
.
Furthermore, RT-PCR amplification of pa-
ternally expressed (hSNRPN and ARH1)
and maternally expressed (UBE3A and
H19) genes demonstrated biparental, and
not unimaternal, expression of imprinted
genes. Further confirmation of the com-
plete removal of oocyte DNA, DNA finger-
print assay and imprinted gene analysis
provided three lines of evidence support-
ing the SCNT origin of SCNT-hES-1 cells.
11.4
Reprogramming Adult Cells
into an Embryonic State
Although normal cell development appears
to involve a progressive restriction in the
developmental potential of cells, recent evi-
dence has suggested that such restriction
is not irreversible, and may be altered to
reveal novel phenotypic potentials of stem
cells, progenitor, and even differentiated
cells. These reversible events are explained
by a dedifferentiated or transdifferentiated
process. Although dedifferentiation and
transdifferentiation have the same ulti-
mate end-point, a distinction should be
made between the two [58]. In general,
dedifferentiation requires the cytokine and
sequential markers of an earlier precursor
cell that can be identified during the nor-
mal pathway of differentiation. However, if
11.4 Reprogramming Adult Cells into an Embryonic State 277
Fig. 11.2 Expression of characteristic cell-surface
markers in human SCNT ES cells. SCNT-hES-1
cells expressed cell-surface markers including alka-
line phosphatase (A), SSEA-3 (C), SSEA-4 (D),
TRA-1-60 (E), TRA-1-81 (F), and Oct-4 (G), but
not SSEA-1 (B). Magnification (A to G: 40). Scale
bars=100 lm.
the transition is rapid, does not follow a
normal sequence, or cannot easily be ex-
plained by our understanding of normal
sequences, then this process would be con-
sidered a transdifferentiation. Cellular de-
differentiation in salamanders supports
the reversible events in vertebrates. In sal-
amanders, new stem cells or progenitor
cells are created through a process of cel-
lular dedifferentiation in which differen-
tiated cells reverse the normal develop-
mental processes and become precursor
cells [5962]. Furthermore, when triploid
muscle tissue labeled with
3
H-thymidine
was transplanted into a regenerating di-
ploid salamander limb, the labeled cells
were found in all mesodermally derived
tissue of limb, suggesting that transdiffer-
entiation had occurred [63, 64]. Fully dif-
ferentiated myotubes also transdifferenti-
ate into chondrocytes through dedifferen-
tiation and redifferentiation when they are
implanted into the regenerating blastem
[62, 65].
As with the regenerating systems in sal-
amanders, several methods are available
for the reprogramming of human adult
cells into an embryonic state. One proven
method is nuclear transfer of an adult cell
into an enucleated oocyte to produce a
cloned embryo to produce cloned ES cells
(therapeutic cloning), though the limited
supply of donated human eggs, as well as
ethical concerns, are clear limitations to
this approach. Consequently, alternative
routes are being taken in an attempt to en-
hance the regenerative capacity of mam-
malian cells. It has been shown that adult
cells may be differentiated into other cell
types by fusing them with ES cells, or with
the embryonic germ (EG) cells that give
rise to sperm and oocytes. Terada et al.
demonstrated that bone marrow cells, by
fusing with ES cells, were transformed
into several cell lineages including myo-
cytes, hepatocytes, and neurons [66]. Tada
et al. also reported that EG cells induced
epigenetic reprogramming of the somatic
nucleus in EG-thymic lymphocyte hybrid
cells [67]. Tada et al. also showed that,
after cell fusion, these ES-thymocyte hy-
brids had pluripotency, including reactiva-
tion of the Oct-4 gene, and contributed to
all three germ layers [68]. In these experi-
ments, the ES or EG cells were small and
difficult to work with, and could not easily
be stripped from their DNA; this resulted
in hybrid cells which contained chromo-
somes from the ES cells, as well as the
cells to be reprogrammed. To overcome
these problems, Collas et al. [69] carried
out the functional reprogramming of fibro-
blast cells into T-cell lineage using a nucle-
ar and cytoplasmic extract derived from
transformed T-cell lines. These results sug-
gest that T-cell-specific factors diffuse into
the permeabilized fibroblasts, actively
taken up by the nuclei and induce the re-
programming process in fibroblast cells.
In identifying the nuclear factors required
for the reprogramming of cellular events,
Kilkyo and Gonda found a group of pro-
teins which were released in the remodel-
ing of local chromatin and disassembling
of nucleoli in Xenopus egg cytoplasm, and
termed these nucleosomal ATPase ISWI
and Xenopus germ cell proteins FRGY2a
and FRGY2b [70, 71]. In addition to these
cellular factors, Chen et al. also sought
small synthetic molecules which could in-
duce cellular dedifferentiation. By screen-
ing libraries of heterocyclic compounds,
these authors identified reversine, a com-
pound which differentiated myogenic line-
age-committed cells to multipotent me-
senchymal progenitor cells that can prolif-
erate and redifferentiate into bone and fat
cells [72].
Although these in vitro cell-based experi-
ments for the reprogramming of differen-
tiated adult cells into embryonic state cells
or other cell lineage are in their early
stages, the approaches employed will un-
doubtedly prove functional on a large scale
for cell replacement and other therapeutic
applications. Before such claims may be
met, however, a better understanding is re-
quired of the molecular mechanisms
which regulate cellular differentiation.
11.5
Discussion and Conclusion
Success in the production of human
SCNT-ES-1 cell lines has been attributed
to the optimization of several factors, in-
cluding donor cell type, reprogramming
time, activation protocol and use of a se-
quential culture system with newly devel-
oped IVC medium. One factor of utmost
importance appears to be the use of a less-
invasive enucleation method the squeez-
ing method (for a description of this tech-
nique, see the video animation on the
supplement CD-ROM). In this method the
MII oocytes are squeezed using a glass
pipette so that the DNAspindle complex
is extruded through a small hole in the
ZP, rather than being aspirated with a
glass pipette, as described elsewhere [73].
Using an aspiration method, Simerly et al.
[74] recently reported defective mitotic
spindles after SCNT in non-human pri-
mate embryos, this perhaps being the re-
sult of depletion of the microtubule motor
and centrosome proteins lost to the meio-
tic spindle after enucleation.
In the present study, we were success-
fully able to develop preimplantation em-
bryos after SCNT, the fused SCNT embryo
being developed into 2-cell, 4-cell, 8-cell
stage, morula and blastocysts (Fig. 11.3).
In order successfully to derive immuno-
compatible human ES cells from a living
donor, a reliable and efficient method for
producing cloned embryos and ES isola-
tion must be developed. Thomson et al.
11.5 Discussion and Conclusion 279
Fig. 11.3 Preimplantation development of em-
bryos after somatic cell nuclear transfer. The fused
SCNT embryo (A) was developed into 2-cell (B),
4-cell (C), 8-cell (D), morula (E) and blastocyst
(F). Magnification=200 (A to E) and 100 (F).
Scale bar =100 lm (A to E) and 50 lm (F).
[1], Reubinoff et al. [56], and Lanzendorf et
al. [75] each produced human ES cell lines
at high efficiency. Briefly, five ES cell lines
were derived from a total of 14 ICMs, two
ES cell lines from four ICMs, and three
ES cell lines from 18 ICMS, respectively.
In the present study, one SCNT-hES cell
line was derived from 20 ICMs. It remains
to be determined if this low efficiency is
due to faulty reprogramming of the so-
matic cells, or to subtle variations in the
experimental procedures utilized. The pos-
sibility cannot be ruled out that the genetic
background of the cell donor had an im-
pact on the overall efficiency of the proce-
dure. Further improvements in IVC sys-
tems for ES cells are needed before con-
templating the use of this technique for
cell therapy. In addition, those mecha-
nisms governing the differentiation of hu-
man tissues must be elucidated in order to
produce tissue-specific cell populations
from undifferentiated ES cells.
In conclusion, our study describes the
first establishment of pluripotent ES cells
from SCNT of a human adult repro-
grammed cell and provides the feasibility
of using autologous cells in transplant
medicine.
One such example is described by Lior
Gepstein in the next chapter (Part I, Chap-
ter 12).
References
1 J. A. Thomson et al., Science 1998, 282, 1145
1147.
2 D. Solter, Nat Rev Genet 2000, 1, 199207.
3 R. P. Lanza et al., Nat Med 1999, 5, 975977.
4 J. B. Cibelli et al., Nat Biotechnol 1998, 16,
642646.
5 M. J. Munsie et al., Curr Biol 2000, 10, 989
992.
6 E. Kawase et al., Genesis 2000, 28, 156163.
7 T. Wakayama et al., Science 2001, 292, 740
743.
8 J. B. Cibelli et al., J Regen Med 2001, 26, 25
31.
9 W. S. Hwang et al., Science 2004, 303, 1669
1674.
10 V. Zakharchenko et al., Mol Reprod Dev 1999,
54, 264272.
11 C. K. Cho et al., Theriogenology 2002, 57,
18191828.
12 S. J. Uhm et al., Mol Reprod Dev 2000, 57,
331337.
13 G. S. Lee et al., Theriogenology 2003, 59,
19491957.
14 W. M. Rideout et al., Science 2001, 293, 1093
1098.
15 F. Xue et al., Nat Genet 2002, 31, 216220.
16 J. Kuwn et al., Mol Reprod Dev 2003, 65, 167
174.
17 M. J. Whitaker and R. Patel, Development
1990, 108, 525542.
18 D. N. Wells et al., Reprod Fertil Dev 1998, 10,
369378.
19 C. Kubota et al., Proc Natl Acad Sci USA
2000, 97, 990995.
20 P. Kasinathan et al., Biol Reprod 2001, 64,
14871493.
21 S. J. Shin et al., Theriogenology 2001, 55,
16971704.
22 J. Betthauser et al., Nat Biotechnol 2000, 18,
10551059.
23 L. Lai et al., Science 2002, 295, 10891092.
24 S. H. Hyun et al., Biol Reprod 2003, 69, 1060
1068.
25 K. Nakagawa et al., Zygote 2001, 9, 8388.
26 Y. G. Chung et al., Biol Reprod 2002, 66,
11781184.
27 A. Langendonckt et al., Fertil Steril 2001, 76,
10231031.
28 N. S. Macklon et al., Human Reprod 2002, 17,
27002705.
29 C. F. Rosenkrans Jr. et al., Biol Reprod 1993,
49, 459462.
30 H. T. Tervit et al., J Reprod Fertil 1972, 30,
493497.
31 R. M. Petters et al., J Reprod Fertil 1990, 89,
269275.
32 Z. Machaty et al., Biol Reprod 1998, 59, 451
455.
33 D. Rieger et al., Reprod Fertil Dev 1992, 4,
547557.
34 D. K. Barnett and B. D. Bavister, Mol Reprod
Dev 1996, 43, 105133.
35 L. Scott and D. C. Whittingham, Mol Reprod
Dev 1996, 4, 336346.
36 K. Miyoshi et al., J Reprod Fertil 1994, 100,
2126.
37 J. H. Kim et al., Biol Reprod 1993, 48, 1320
1325.
38 J. G. Thompson et al., Mol Reprod Dev 1992,
31, 253257.
39 J. Conaghan et al., J Reprod Fertil 1993, 99,
8795.
40 D. Sakkas et al., Biol Reprod 1993, 49, 1288
1292.
41 J. Kikuchi et al., J Reprod Fertil 1995, 105,
325330.
42 K. Yoshioka et al., Biol Reprod 2002, 66, 112
119.
43 G. S. Lee et al., Mol Reprod Dev 2003, 661,
1723.
44 R. L. Krisher et al., Biol Reprod 1999, 60,
13451352.
45 C. Wrenzycki et al., Mol Reprod Dev 1999, 53,
818.
46 Y. H. Choi et al., Theriogenology 2002, 58,
11871197.
47 B. D. Bavister, Hum Reprod Update 1995, 1,
91148.
48 C. Wrenzycki et al., Hum Reprod 2001, 16,
893901.
49 M. Ott et al., Anat Histol Embryol 2002, 31,
151157.
50 H. S. Kim et al. Theriogenology 2004, 61,
13811393.
51 S. H. McKiernan et al., In Vitro Cell Dev Biol
1992, 28A, 154156.
52 M. Lane et al., Mol Reprod Dev 2003, 64, 70
78.
53 B. Kuhholzer et al., Biol Reprod 2004, 64,
16351698.
54 S. H. Lee et al., Nat Biotechnol 2000, 18, 675
679.
55 J. A. Thomson et al., Proc Natl Acad Sci USA
1995, 92, 78447848.
56 B. E. Reubinoff et al., Nat Biotechnol 2000, 18,
399404.
57 S. R. John, Trends Biotechnol 2002, 20, 417
419.
58 Y. Liu and M. S. Rao, J Cell Biochem 2003, 88,
2940.
59 C. W. Bodemer and N. B. Everett, Dev Biol
1959, 1, 327342.
60 E. D. Hay and D. A. Fischman, Dev Biol 1961,
3, 2659.
61 C. S. Thornton, J. Morphol. 1983, 62, 1747.
62 A. Kumar et al., Dev Biol 2000, 218, 125136.
63 T. P. Steen, J Exp Zool 1968, 167, 4978.
64 M. Namenwirth, Dev Biol 1974, 41, 4256.
65 D. C. Lo, F. Allen, J. P. Brockes, Proc Natl
Acad Sci USA 1993, 90, 72307234.
66 N. Terada et al., Nature 2002, 416, 542545.
67 M. Tada, EMBO J 1997, 16, 65106520.
68 M. Tada, Curr Biol 2001, 11, 15531558.
69 A. M. Hakelien et al., Nat Biotechnol 2002, 20,
460466.
70 N. Kikyo, Science 2000, 289, 23602362.
71 K. Gonda, Nature Cell Biol 2003, 5, 205210.
72 C. Shuibing et al., J Am Chem Soc 2004, 126,
410411.
73 I. Wilmut et al., Nature 1997, 385, 810813.
74 C. Simerly et al., Science 2003, 300, 297.
75 S. E. Lanzendorf et al., Fertil Steril 2001, 76,
132137.
References 281
Supplement 1 (Video 1). Enucleation, injection of donor cell and fusion donor cell and oocyte. Under
the DIC microscope equipped with a micromanipulation system, an oocyte is secured in place using a
holding pipette. A slit was cut in the zona pellucida (ZP) adjacent to the polar body using a fine glass
needle by rubbing the perforated ZP against the holding tip. The oocyte is released from the holding
pipette and squeezed between the cutting pipette and the holding pipette. A fraction of the eggs cyto-
plasm, presumably containing the metaphase II chromosomes (about 10% of the total cytoplasm),
along with the polar body, was expelled through the slit made in the ZP. Using an injection needle, a
cumulus cell was aspirated and deposited into the perivitelline space of oocyte using the same opening
in the ZP made during enucleation. After transferring donor cells, a reconstructed oocyte was placed in
a fusion chamber containing two stainless steel electrodes 3.4 mm apart, and is subjected to electric
stimuli to fuse a donor cell and oocyte.
Supplement 2 (Video 2). Subculture of SCNT-hES-1 cells. Subculture was performed mechanically every
57 days. Under an inverted microscope, feeder cells were detached from a ES cell colony using a
hooked Pasteur pipette, and the colony dissected into 200- to 300-cell clumps.
Abstract
Adult cardiomyocytes have limited regen-
erative capacity, and therefore any signifi-
cant cell loss, such as occurs during myo-
cardial infarction, may result in the devel-
opment of progressive heart failure. Simi-
larly, the same processes of tissue loss or
dysfunction, occurring at critical sites in
the cardiac electrical conduction system,
may result in inefficient rhythm initiation
or impulse conduction, requiring the im-
plantation of a permanent electronic pace-
maker. Cell replacement therapy is a pro-
mising new approach for myocardial re-
pair, but has been hampered by the pau-
city of cell sources for functional human
cardiomyocytes and by the lack of direct
evidence for functional integration be-
tween host and donor cardiomyocytes. The
recent establishment of the pluripotent hu-
man embryonic stem (hES) cell lines may
present a novel solution for this cell-sour-
cing problem. The hES lines were derived
from human blastocysts, and shown cap-
able of continuous undifferentiated prolif-
eration in vitro, while retaining the capabil-
ity to form derivatives of all three germ
layers. More recently, we were able to gen-
erate a reproducible cardiomyocyte differ-
entiation system from these unique cells.
This chapter describes the derivation and
properties of hES cells and the molecular,
ultrastructural, and functional characteris-
tics of the cardiomyocyte tissue derived
using this unique differentiating system.
Evidence is also provided for the ability of
hES cell-derived cardiomyocytes to prolifer-
ate following differentiation and to inte-
grate structurally and functionally with
host cardiomyocytes in both in vitro and in
vivo models. Possible applications of this
unique cardiomyocyte-differentiating sys-
tem in several research and clinical areas
will be discussed, as will be the steps re-
quired to fully harness the potential of this
new technology in the fields of myocardial
cell replacement and tissue engineering.
Finally, the many obstacles and possible
solutions that need to be overcome on the
way to successful clinical utilization of
these cells will be presented.
Abbreviations
ANP atrial natriuretic peptide
AV atrioventricular
BIO 6-bromoindirubin-3'-oxime
BMPs bone morphogenetic proteins
CMV cytomegalovirus
DMSO dimethylsulfoxide
EBs embryoid bodies
ES embryonic stem (cells)
283
ISBN: 3-527-31184-X
12
Myocardial Regeneration Strategies using Human Embryonic
Stem Cells
FACS fluorescence-activated cell sort-
ing
GFP green fluorescent protein
GSK-3 glycogen synthase kinase 3
hES human embryonic stem
ICM inner cell mass
IVF in vitro fertilization
MEA micro-electrode array
MEF mouse embryonic fibroblasts
MHC major histocompatibility com-
plex
Mef myocyte enhancer factor Mef
RA retinoic acid
RT-PCR reverse transcriptase polymerase
chain reaction
TGF-b transforming growth factor-b
VE visceral endoderm
12.1
Introduction
One of the most exciting areas in basic re-
search today involves the use of stem cells.
These unique cells have the capability to
transform and replenish the different tis-
sue types that make up the body, and also
represent the fundamental building blocks
of human development. Recent advances
in the areas of stem cell biology and tissue
engineering, coupled with parallel achieve-
ments in molecular and cell biology, have
paved the way to the development of a
new field in biomedicine regenerative
medicine.
This approach seeks to develop new bio-
logical solutions to replace or modify the
function of diseased, absent, or malfunc-
tioning tissue.
The adult heart represents an attractive
candidate for these emerging technologies.
This is because adult cardiomyocytes have
limited regenerative capacity, and hence
any significant loss of heart cells is mostly
irreversible and may lead to progressive
and irretrievable loss of ventricular func-
tion and finally to the development of
heart failure. Congestive heart failure is a
growing epidemic that affects more than 5
million Americans [1], and is associated
with significant morbidity and mortality.
Therefore, it is not surprising throughout
the years that much effort has been spent
on the development of different therapeu-
tic modalities. Yet, despite advances in the
pharmacological, interventional, and surgi-
cal therapeutic measures, the prognosis
for heart failure patients remains poor.
Non-pharmacological treatments such as
heart transplantation (see Part I, Chapter
15) for end-stage patients are of limited
impact, as the chronic lack of donors lim-
its the number of patients that could bene-
fit from heart transplantation. Given these
circumstances, the development of new
therapeutic strategies for the treatment of
heart failure has become imperative.
A possible novel therapeutic strategy for
heart failure following myocardial infarc-
tion may be to increase the number of
functional myocytes within the diseased
area by the implantation of exogenous
myogenic cells. Early studies used neona-
tal rat cardiomyocytes for transplantation,
as these cells have cardiac phenotype and
still retain some proliferation capacity [2
4]. Fetal cardiomyocyte cell grafts showed
the formation of cell-to-cell contacts, com-
plete with gap junction proteins [4]. More-
over, cultured human fetal cardiomyocytes
were shown to survive, and fetal rat cardio-
myocytes were shown to be present in the
infarcted rats hearts for up to 6 months
after transplantation [5]. Further studies in
animal models of myocardial infarction
showed that grafting of cardiomyocytes
from fetal and neonatal sources was asso-
ciated with smaller infarcts, prevented car-
diac dilatation and remodeling, and also
improved ventricular function [6, 7]. The
mechanisms underlying these functional
improvements may be multifactorial, and
may include a direct contribution to con-
tractility by the transplanted cells, attenua-
tion of the remodeling process by chang-
ing the architectural and structural proper-
ties of the scar, and improvement in the
function of viable tissue within the border
zone by induction of angiogenesis.
Since human fetal cardiomyocytes can-
not be obtained in sufficient quantities for
clinical use, a search for alternate cell
sources for transplantation has begun. A
series of studies conducted during the past
few years showed that dispersed prepara-
tions of myogenic cells such as skeletal
myoblasts could survive and even differ-
entiate when engrafted onto recipient
hearts [8, 9]. Skeletal myoblasts strongly
resist ischemia, thereby allowing for in-
creased survival and engraftment in areas
of poor coronary perfusion, which is often
the case in patients with coronary artery
disease (see Part I, Chapter 6). Using skel-
etal myoblasts implanted into a cryoinfarct
rabbit model, Taylor et al. demonstrated
improvement in myocardial performance
[9]. Initial clinical studies followed these
encouraging pre-clinical results, and Me-
nasch and colleagues [10] reported on
possible improvement in local myocardial
contractility and viability in the grafted
scar on echocardiography and positron
emission tomography (PET; see Part V,
Chapter 5) after autologous skeletal myo-
blasts were injected into the post-infarction
scars of 10 patients during coronary artery
bypass grafting [10]. However, although
the initial reports suggested that these
cells may have the ability to adopt a cardi-
ac-like phenotype following cardiac trans-
plantation, it is now clear that they do not
possess such a capability. Moreover, the
relatively high rate of ventricular arrhyth-
mias observed in the initial Phase I clini-
cal trials, which probably resulted from
differences in the electrophysiological
properties between host cardiomyocytes
and the engrafted myotubes, may further
limit this approach.
Other groups have turned toward bone
marrow-derived cells for this purpose. En-
dothelial progenitor cells are bone mar-
row-resident cells that can be released into
circulation after an acute myocardial in-
farction, and can enhance neovasculariza-
tion [1113]. Endothelial progenitor cells
are excellent donor cells because they al-
low autologous harvesting, thereby obviat-
ing the need for immunosuppression.
Transplanted bone marrow-derived hema-
topoietic stem cells were initially shown to
differentiate into myocytes and form con-
nections with the host tissue [16]. How-
ever, two recent studies called the previous
report into question and, using a genetic
marking technique, showed that very few
if any of the transplanted cells actually
transdifferentiated to cardiomyocytes or
even survived for long in the infarcted
mouse heart [17, 18]. Despite the contro-
versy regarding bone marrow progenitors,
several groups have performed small, ran-
domized clinical trials and have reported
some success in the improvement of ejec-
tion fraction following transplantation of
endothelial progenitors in the setting of
acute myocardial infarction [19, 20].
Mesenchymal stem cells are another
group of adult stem cells that have been
suggested as potential donor cells (see Part
I, Chapter 13). These cells are accessible
from the bone marrow and peripheral
blood, allow autologous transplantation,
may be multipotent, and can also differ-
entiate into specialized tissues, including
possibly cardiomyocytes, endothelial cells,
and smooth-muscle cells [21, 22]. Initial
studies in a swine model showed encour-
aging results following the implantation of
autologous or allogenic swine and human
mesenchymal stem cells after myocardial
infarction. These studies claimed sus-
tained engraftment in host myocardium,
differentiation into cardiomyocytes, and
possibly improved cardiac function [23].
Most of the aforementioned donor cells
are derived from different types of stem
cells. All stem cells whether from adult
or embryonic sources share a number of
properties [24]. First, they are capable of
self-renewal, which means that they can
generate stem cells with similar properties.
Second, the stem cells are clonogenic,
which means that each cell can form a col-
ony in which all the cells are derived from
this single cell and have identical genetic
constitution. Third, they are capable of dif-
ferentiation into one or more mature cell
types. The different stem cells can be cate-
gorized anatomically, functionally, or by
cell surface markers, transcription factors,
and the proteins they express. One clear
division of the stem cell family is between
those in adult somatic tissue (known as
adult stem cells) and those isolated from
the embryo (known as embryonic stem
cells) (see Part I, Chapter 11).
12.2
Derivation of Human Embryonic Stem Cells
Although adult stem cells have been found
to be more versatile than originally be-
lieved, they typically can differentiate to a
relatively limited number of cell types. In
contrast, cells in the early preimplantation
mammalian embryo have the potential to
contribute to all adult tissues. At the blas-
tocyst stage, a group of cells begins to sep-
arate from the outer cells and forms the
inner cell mass (ICM). While the outer
cells become the trophoectoderm, the ICM
cells will ultimately give rise, through spe-
cialized progenitor cells, to all the tissues
in the body and are therefore truly pluripo-
tent. In 1981, the ICM cells, isolated from
mouse blastocysts, were used to generate
pluripotent stem cell lines that were
termed embryonic stem (ES) cells [25, 26].
The mouse ES cells were shown to be
capable of prolonged in vitro proliferation
and self-renewal, but also retained the abil-
ity to differentiate into derivatives of all
three germ layers, both in vitro and in vivo.
Following cultivation in suspension, the
murine ES cells tend spontaneously to cre-
ate three-dimensional aggregates of differ-
entiating tissue known as embryoid bodies
(EBs). Among other cell types, cardiomyo-
cyte tissue appears within this multicellu-
lar arrangement, as spontaneously con-
tracting areas that can be studied as a clus-
ter or as dispersed cells [27].
Given the outstanding potential demon-
strated by the mouse ES cells, it was not
surprising that much effort was spent on
the development of similar human ES cell
(hES) lines. This quest has ended recently
when two groups described the generation
of hES cell lines [28, 29]. The origin of the
hES cell lines, similar to that in the mouse
and rhesus models, is from the preimplan-
tation embryo produced by in vitro fertili-
zation (IVF) for clinical purposes and do-
nated by individuals after informed con-
sent. The hES cell lines were established
by isolating the ICM cells after removal of
the trophoectoderm with specific antibod-
ies (immunosurgery). The cells were then
plated on a feeder layer of mitotically in-
activated mouse embryonic fibroblasts
(MEF). The resulting colonies were se-
lected, passaged, and expanded for the
creation of the hES cell lines (Fig. 12.1).
The hES cells were shown to fulfill all
the criteria defining embryonic stem cells
[28, 29], namely: derivation from the pre-
or peri-implantation embryo, prolonged
undifferentiated proliferation under special
conditions, and the capacity to form deri-
vatives of all three germ layers. Thus,
when cultured on top of the MEF feeder
layer, the hES cells could be maintained in
the undifferentiated state for prolonged
periods. When removed from the MEF
feeder layer, and allowed spontaneously to
12.2 Derivation of Human Embryonic Stem Cells 287
Fig. 12.1 Early embryonic development (A), deri-
vation of the hES cell lines (B) and establishment
of an in vitro cardiomyocyte differentiation system.
The hES cells were generated from the early-stage
embryo at the blastocyst stage. At this stage, the
embryo is composed of the trophoectoderm and
the inner cell mass (ICM), which eventually will
give rise to all tissue types in the embryo (A).
ICM cells isolated by immunosurgery and plated
on the MEF feeder layer were used to generate
the ES lines (B). The resulting colonies were pro-
pagated and expanded. Following establishment
of the hES lines, they can be propagated continu-
ously in the undifferentiated state when grown on
top of the MEF feeder layer (B, top). When re-
moved from these conditions and grown in sus-
pension, they form 3-D cell aggregates that are
termed embryoid bodies (EBs) (B, middle). This
in vitro differentiating system can be used to gen-
erate a plurality of tissue types, including cardio-
myocytes (B, bottom). (C) Photomicrographs
showing the different stages in the in vitro cardio-
myocyte differentiation of the hESCs. Top: Initially,
hESC colonies are propagated in the undifferen-
tiated state on top of the MEF feeder layer. Mid-
dle: To induce differentiation, hESC are removed
from the MEFs and grown in suspension where
they form EBs. Bottom: The EBs are then plated
and observed for the appearance of spontaneously
contracting areas (arrows).
differentiate, the hES cells could form EBs
containing cell derivatives of all three
germ layers [30] (Fig. 12.1). A subsequent
study described the generation of clonally
derived hES cell lines [31], and demon-
strated the pluripotency of single hES
cells, the maintenance of pluripotency dur-
ing an extended period of culture, and the
long-term self-renewing properties of cul-
tured hES cells. The undifferentiated hES
cell lines and their clonal derivatives were
also shown to express high levels of telo-
merase, and to retain normal karyotype for
prolonged culture periods.
Several differences distinguish human
from mouse ES cells [32]. The hES cells
have a slightly different morphology and
form flatter colonies. The stage-specific
embryonic antigen-3 and -4 as well as
TRA-1-60 and TRA-1-81 are expressed by
the human, but not by mouse, ES cells
[32]. Moreover, hES cells grow more slowly
than mouse ES cells; the population dou-
bling time of mouse ES cells is ~12 hours,
whereas that of hES cells is about
36 hours.
The most important difference between
the two cell lines, however, is probably in
the mechanisms involved in their self-re-
newal, as reviewed by Rao et al. [33]. The
maintenance of the undifferentiated state
of the ES cells is mediated both by differ-
entiation inhibiting signals as well as by
the lack of expression of differentiation in-
ducing genes. Among the differentiation-
inhibiting signals, the well-known, leuke-
mia inhibitory factor (LIF) is considered to
be a sufficient trigger for the maintenance
of undifferentiated proliferation in the
mouse ES system in the presence of se-
rum. In addition, recent reports have sug-
gested that induction of the expression of
inhibitor of differentiation genes by bone
morphogenic proteins (BMPs) in concert
with LIF can maintain the in vitro self-re-
newal capabilities of mouse ES cells in se-
rum-free conditions [34]. In contrast, in
the hES system LIF is insufficient or even
not required for this purpose, and the
presence of the MEF feeder layer itself, its
conditioned medium or other cell support
system such as human fetal fibroblast,
adult epithelial cells or foreskin cells, is re-
quired [35, 36].
Except for the difference in LIF-gp130
signaling between human and mouse ES
cells, several key regulators of stem cell
self-renewal have been shown to be con-
served between the human and mouse sys-
tems. Sato et al. [37] elucidated the role of
the canonical Wnt pathway in the mainte-
nance of embryonic stem cell self-renewal.
This study showed that 6-bromoindirubin-
3'-oxime (BIO) a specific glycogen
synthase kinase 3 (GSK-3) inhibitor is
sufficient for the maintenance of stem-
ness propagation and pluripotency of
both human and mouse ES cells. The
homeobox domain-containing protein Na-
nog was recently shown to act in parallel
to the LIF pathway in maintaining ES cell
self-renewal both in mouse and human ES
cells [38]. Further studies elucidating the
factors participating in ES cell self-renewal
are crucial for establishing a reproducible,
well-defined, animal- and serum-free sup-
porting system that may be up-scaled and
will not only facilitate research practices
but also provide a safer alternative for fu-
ture clinical applications of hES cells.
The pluripotency of ES cells can be es-
tablished traditionally using three different
approaches. Mouse ES cells can be re-
transferred into early mouse embryos
where they eventually give rise to all so-
matic cells of the chimeric embryo, includ-
ing the germ cells [39]. Such a test cannot
be applied to hES for obvious ethical rea-
sons. The second approach relates to the
demonstration that ES cells can differenti-
ate to generate derivatives of all three
germ layers in vivo. When hES cells were
injected into immunodeficient mice, they
formed benign teratomas containing ad-
vanced differentiated tissue types repre-
senting all three germ layers [28, 29].
The third and most exciting approach
establishes ES pluripotency during in vitro
differentiation. Both mouse and human
ES cells, when removed from the MEF
feeder layer and allowed to differentiate,
could form three-dimensional cell aggre-
gates, termed embryoid bodies (EBs), that
contain tissue derivatives of endodermal,
ectodermal, and mesodermal origin [30,
32]. The ability of the hES cells to generate
a variety of mature somatic cell types was
demonstrated using both spontaneous and
directed in vitro differentiation systems.
Hence, since the initial report of the deri-
vation of the hES cell they were shown to
be able to differentiate into cardiac tissue
[40], neuronal tissue [41] including dopa-
minergic cells [42], b-islet pancreatic cells
[43], hematopoietic progenitors [44], kera-
tinocytes [45], bone tissue [46] and endo-
thelial cells [47].
12.3
Cardiomyocyte Differentiation of ES Cells
As described above, the most common
method used to induce differentiation of
the ES cells requires an initial aggregation
step to form EBs (see Fig. 12.1). This is
performed in the mouse ES model by re-
moving the cells from the MEF feeder
layer, or by discontinuing LIF and cultivat-
ing them in suspension. Different proto-
cols have been used in the murine ES cells
for such cultivation, including the mass
culture technique, cultivation in methyl-
cellulose, or the hanging drops tech-
nique [48]. Among other differentiating
cell types within the mouse EBs, cardio-
myocyte tissue can be identified by the ap-
pearance of spontaneously contracting
areas.
The formation of cardiomyocytes within
the murine EB provided investigators with
a unique in vitro tool for the investigation
of early cardiomyogenesis. Detailed ultra-
structural, immunohistochemical, molecu-
lar, and electrophysiological studies
showed that the developmental stages of
the ES-derived cardiomyocytes in vitro par-
allel those of the in vivo murine heart [49
53].
The advent of the murine ES model has
also provided descriptive and mechanistic
information regarding the development of
excitability and electromechanical coupling
in early cardiac tissue, including patterns
of gene expression, myofibrillogenesis, ion
channel development and function, cal-
cium handling, receptor development, and
the signal machinery involved in these
processes [4953]. The murine ES cell-de-
rived cardiomyocytes displayed diverse ac-
tion-potential morphologies including ven-
tricular, atrial, and sinus nodal types, as
demonstrated by intracellular recordings
[51]. During development, the percentage
of the different cell types transformed
from a pacemaker-like cell predominance
to atrial or ventricular cell predominance
in older EBs.
Detailed electrophysiological analyses of
the developing murine EBs have also re-
vealed a developmental cascade of ion
channel expression and modulation [47,
52]. The non-contracting precursor cells
display voltage-dependent L-type Ca
2+
channels at very low densities. Cardiomyo-
cytes of an early differentiation stage ex-
hibit a primitive pacemaker action poten-
tial generated by voltage-dependent L-type
Ca
2+
channels and transient outward K
+
channels. Terminally differentiated cardio-
myocytes express various additional ion
channels according to their various pheno-
types. Ventricle-like cells express voltage-
dependent Na
+
channels, delayed outward
rectifying K
+
channels, and inward rectify-
ing K
+
channels. Additional ion channels,
such as muscarinic acetylcholine-activated
K
+
channels and the hyperpolarization-acti-
vated pacemaker channels were demon-
strated in atrial-like cells and sinus node-
like cells, respectively.
12.3.1
Human ES System
Recently, we used a slightly different dif-
ferentiating scheme to that reported in the
mouse model to generate a reproducible
spontaneous cardiomyocyte differentiating
also in the hES system (Fig. 12.1) [40]. Un-
differentiated hES of the single-cell clone
H9.2 were propagated on top of the MEF
feeder layer. The hES cells were then re-
moved from the feeder layer, dissociated
into small clumps of between three and 20
cells, and grown in suspension for 710
days, where they formed EBs. The EBs
were then plated on gelatin-coated culture
dishes and observed microscopically for
the appearance of spontaneous contraction
(Fig. 12.1). Rhythmically contracting areas
appeared at 422 days after plating in
about 10% of the EBs.
Several lines of evidence confirmed the
cardiomyocyte nature of the cells within
the beating EBs (Fig. 12.2) [40]. Reversed-
phase polymerase chain reaction (RT-PCR)
studies demonstrated the expression of
cardiac-specific transcription factors (e.g.,
GATA4 and Nkx2.5) and cardiac-specific
structural genes [cTnI, cTnT, atrial natri-
uretic peptide (ANP), MLC-2V, MLC-2a].
Initial analysis of gene expression pattern
during in vitro cardiomyocyte differentia-
tion of the hES revealed a reproducible
developmental temporal pattern. This was
manifested initially by a gradual decrease
during differentiation in the expression of
undifferentiated stem cell markers, such
as OCT-4.
The first event that may be related to
cardiomyogenesis was an early increase,
during the suspension phase, in the ex-
pression of growth factors known to be in-
volved in cardiac differentiation, such as
Wnt11 and BMP-2. This was followed by
expression of cardiac-specific transcription
factors (Nkx2.5, Mef2c, and GATA4) to-
wards the end of the suspension phase
and the immediate post-plating period.
These events were consequentially fol-
lowed by the expression of cardiac-specific
structural genes such as ANP and myosin
heavy chains.
Immunostaining studies of cells isolated
from the contracting areas within the EBs
confirmed the presence of cardiac-specific
proteins (MHC, sarcomeric a-actinin, des-
min, cTnI, ANP). These studies also dem-
onstrated the presence of early-cardiac
morphology with a typical early-striated
staining pattern. The cells, however, did
not exhibit immunoreactivity with anti-
nebulin monoclonal antibodies (mAbs), a
specific skeletal muscle sarcomeric protein
shown to be expressed early in skeletal
myoblast differentiation.
Ultrastructural analysis of the differen-
tiating cardiomyocytes showed that these
cells were mainly mononuclear, contained
varying degrees of myofibrillar bundle or-
ganization, and exhibited nascent interca-
lated discs. Transmission electron micro-
scopy of EBs at varying developmental
stages showed the progressive ultrastruc-
tural maturation from an irregular myofi-
lament distribution to a more mature sar-
comeric organization in late-stage EBs
[54]. These results are consistent with ul-
trastructural properties of early-stage cardi-
omyocytes, and with the developmental
process of myofibrillar assembly.
Interestingly, during the process of ultra-
structural maturation of the hES-derived
cardiomyocytes they gradually start to
withdraw from the cell cycle. Using
[
3
H]thymidine incorporation or Ki-67 im-
munolabeling, the hES cell-derived cardio-
myocytes demonstrated a gradual with-
drawal from cell cycle with cessation of
DNA synthesis after 7 weeks. Hence, our
results demonstrate a reproducible tempor-
al pattern of early cardiomyocyte cell pro-
liferation, cell-cycle withdrawal, and cellu-
lar hypertrophy and maturation [54].
In addition to the molecular and struc-
tural studies described above, several func-
tional assays including extracellular and
intracellular electrophysiological record-
ings, calcium imaging, and pharmacologi-
Fig. 12.2 The contracting areas within the embry-
oid bodies (EBs) displayed molecular, structural,
and functional properties of early-stage human
cardiomyocytes. These properties include expres-
sion of cardiac-specific genes and transcription
factors and the positive immunocytochemical
staining for cardiac-specific proteins (e.g., ANP
and cTnI). Transmission electron microscopy stud-
ies demonstrated the presence of an early sarco-
meric ultrastructural pattern and intercalated
discs typical of cardiomyocytes. Finally, the cells
were also shown to display cardiac-specific action
potentials and ionic transients at the cellular level
during patch-clamp recordings, and spontaneous
pacemaker activity and electrical conduction at
the tissue level using multi-electrode recordings
(see color-coded activation map).
cal studies clearly demonstrated that the
contracting areas within the EBs also dis-
play physiological properties consistent
with an early-stage human cardiac pheno-
type [40, 55, 56]. Hence, all the compo-
nents of normal cardiac excitationcontrac-
tion coupling were shown to be present
within this tissue, including the typical
electrical activation, increase in [Ca
2+
]
i
, and
the resulting contraction.
Extracellular electrophysiological record-
ings using microelectrodes demonstrated a
sharp and a slow component, consistent
with a relatively long action potential dura-
tion characteristic of cardiomyocytes [40,
56]. The positive and negative chronotropic
responses to isoproterenol and carbamyl-
choline demonstrated the presence of
functional adrenergic and cholinergic re-
ceptors respectively in these cells [40]. A
major pathway of the b-adrenoreceptor-de-
pendent chronotropic response is the acti-
vation of adenylate cyclase and the conse-
quent rise in cytosolic cAMP and stimula-
tion of protein kinase. The positive chro-
notropic effect exerted by forskolin (a
direct activator of adenylate cyclase) and by
IBMX (a phosphodiesterase inhibitor) sug-
gests that this signaling pathway is already
present early in human cardiomyocytic dif-
ferentiation [40].
Similar to the mouse ES model, whole-
cell patch-clamp studies showed that the
hES cell-derived cardiomyocytes also dis-
play cardiac-specific action potential
morphologies and ion currents (Fig. 12.2)
[55]. Additional studies conducted in our
laboratory revealed the basis for the spon-
taneous automaticity in these cells, at least
at the mid-differentiation stages. These
studies revealed that during this stage, the
spontaneous electrical activity is mediated
by the absence of significant inward recti-
fier K
+
current and a prominent Na
+
cur-
rent sensitive to TTX coupled with the
presence of the HCN pacemaker current
(If) [55]. The paucity of the inward current
creates a high-input resistance state that
allows a small inward current to bring the
membrane potential to threshold.
The next step was to determine whether
the hES differentiating system is limited
to the creation of isolated cardiomyocytes,
or whether a functional cardiac tissue is
generated. In order to answer this ques-
tion, the spontaneously contracting areas
within the EBs were microdissected and
plated on top of a micro-electrode array
(MEA) mapping technique. This allowed
long-term, high-resolution electrophys-
iological recordings from the EBs. These
measurements demonstrated the presence
of a functional syncytium with stable
spontaneous pacemaking activity and syn-
chronous action-potential propagation [56].
Both the site of earliest focal activation
and the conduction properties within each
EB were relatively reproducible during
both short-term (3 h) and long-term (105
days) recordings.
An attempt was also made to define the
tissues structural properties. This analysis
identified an isotropic tissue with the car-
diomyocytes arranged in various orienta-
tions [56]. The cells were relatively small
and round-, triangular-, or rod-shaped.
Next, we determined the presence and
properties of gap junctions within the con-
tracting areas because their number, size,
and distribution are important determi-
nants of conduction during physiological
and pathological conditions. The gap junc-
tions were relatively small and distributed
homogeneously along the cell circumfer-
ence, with no preferential polar orienta-
tion. This pattern is similar to the one ob-
served in human fetal and neonatal tissue.
We also identified a predominance of con-
nexin45 (Cx45) in the gap junction con-
necting the hES-CM. The significance of
Cx45 in this model is not surprising.
Although almost absent in adult ventricu-
lar myocardium, Cx45 has been shown to
play a major role in early cardiac embryo-
nic development.
Following our initial studies, other
groups have reproduced our results, and
generated cardiomyocytes from different
hES cell lines [5759]. The beating cells they
described expressed markers characteristic
of cardiomyocytes, such as cardiac a-myosin
heavy chain, cardiac troponin I and T, atrial
natriuretic factor, and cardiac transcription
factors GATA-4, Nkx2.5, and MEF-2. The
human cardiomyocytes they describe dis-
played an immature sarcomeric pattern,
and also possessed functional adrenergic
and cholinergic receptors.
12.4
Possible Research and Clinical Applications
of the hES-derived Cardiomyocytes
The absence of in vitro sources for human
cardiac tissue imposes significant limita-
tions for cardiovascular research. Conse-
quentially, the generation of a reproducible
cardiomyocyte differentiating system from
the hES lines provides an indispensable ex
vivo source for human cardiac tissue. This
ability may provide researchers with a
unique tool for the investigation of the
mechanisms involved in early human car-
diac lineage commitment, differentiation,
and maturation. In addition, the genera-
tion of a long-term in vitro model to study
human cardiac tissue may be used for sev-
eral pathophysiological studies, for func-
tional genomics, drug and growth factor
discovery, drug testing, and reproductive
toxicology. Finally, the ability to generate
ex vivo human cardiac tissue may bring a
unique value to the developing field of car-
diovascular regenerative medicine.
12.5
Early Cardiac Lineage Differentiation
In contrast to the fairly well-characterized
process of the morphogenic transforma-
tion of the primitive heart into the four-
chambered structure, the inductive clues
that lead to the specification and terminal
differentiation of cardiomyocytes are some-
what less well-known. Although organo-
genesis or significant tissue organization
does not occur within the EB model, valu-
able information can be gathered regard-
ing the process involved in lineage com-
mitment and differentiation. In fact, in vi-
tro differentiation within the EB model
system may provide a number of advan-
tages over comparable approaches in the
whole embryo. First, it provides access to
population of early precursor cells that are
difficult if not impossible to identify in
vivo. Second, it could allow the study of
targeted mutations of genes that may be
lethal in vivo but can be studied in vitro.
These advantages are even more important
for human embryology, due to the limited
access to early-stage human tissue.
The currently available cardiomyocytes
differentiation system of the hES is essen-
tially spontaneous, and is characterized by
relatively low efficacy. Understanding the
mechanisms that drive early-cardiac differ-
entiation of the hES cells may also have
important clinical implications. A major
obstacle for the use of these cells in future
myocardial regeneration strategies is the
insufficient number of cardiomyocytes
achieved by the currently available differ-
entiation scheme. Therefore, further ef-
forts aimed at developing a more produc-
tive cardiomyocyte differentiation system
are essential for the generation of the large
quantities of cells needed to achieve the
successful application of this strategy.
12.5 Early Cardiac Lineage Differentiation 293
The development of a directed differen-
tiation system is hampered by the relative
lack of data regarding the inductive cues
that lead to commitment and terminal dif-
ferentiation of human cardiomyocytes.
Thus, strategies for directed differentiation
should undoubtedly follow the research
conducted in a number of model organ-
isms (see Part III, Chapter 4), most notab-
ly the chick, amphibians, zebrafish, and
mouse.
Embryologically, the heart arises from
cells in the anterior lateral plate mesoderm
of the early embryo. The cells of the car-
diogenic mesoderm adopt a crescent-like
morphology and are therefore termed the
cardiac crescent [60]. The endoderm that
is in direct contact with the cardiac cres-
cent is considered to have an obligatory
role in induction of the cardiac fate [61].
Various genetic and biochemical perturba-
tions in several organisms have shown a
key role for bone morphogenetic proteins
(BMPs) members of the transforming
growth factor-b (TGF-b) superfamily ex-
pressed in endoderm, as well as in adja-
cent ectoderm and extra-embryonic tissues,
in specifying and/or maintaining the myo-
cardial lineage [62].
Studies conducted in Xenopus and chick
models suggested that cardiogenesis is in-
hibited by Wnt-mediated signals from the
underlying neural tube activating the cano-
nical Wnt pathway. Based on the same ani-
mal models, cardiac differentiation was in-
duced by Wnt-binding proteins (Crescent
and Dkk-1) secreted from the anterior en-
doderm. However, two recent articles have
suggested that the role of the Wnt family
of proteins in cardiomyogenesis is much
more complex. Pandur et al. [63] showed
that Wnt-11, an activator of the non-cano-
nical Wnt/JNK pathway, is required for
cardiogenesis using the Xenopus model
and the pluripotent mouse embryonic car-
cinoma stem cell line P19. Nakamura et
al. [64], also using the P19 cell line, re-
vealed that the canonical b-catenin path-
way of Wnt signaling is actually activated
very early during mammalian cardioge-
nesis.
In response to the inductive signal, the
cardiac crescent activates several transcrip-
tional regulators of the cardiac pro-
gramme, including Gata4/Gata5/Gata6,
Nkx 2-5, Myocyte enhancer factor (Mef2b/
Mef2c) and T-Box 5/20 and a positive
cardiac cross regulatory network is estab-
lished [65]. A powerful transcription factor
termed myocardin was recently identified
and shown to coactivate transcription of
several cardiac-specific gene promoters in
conjunction with serum response factor.
Possible strategies for increasing the car-
diomyocyte yield during hES differentia-
tion may thus include the use of different
growth factors, overexpression of cardiac-
specific transcription factors, co-culturing
with feeder layers, and mechanical factors.
Directed differentiation of ES cells to the
cardiac lineage in the murine model was
achieved using a variety of soluble factors
including dimethylsulfoxide (DMSO), reti-
noic acid (RA) and, more recently, BMP-2
and TGF-b, and ascorbic acid. Xu et al.
[57] showed that cardiac differentiation in
the human ES model was enhanced by 5-
aza-2'-deoxycytidine, but surprisingly not
by DMSO or RA.
There is also evidence to suggest that les-
sons learned from early cardiac differentia-
tion in the model systems (as described
above) may also apply to the hES cells.
The cardiogenic inductive role of the primi-
tive visceral endoderm (VE) was also shown
to play a role in cardiomyocyte differentia-
tion of the hES line in an elegant study con-
ducted by Mummery et al. [59]. Co-cultur-
ing of a human ES cell line (hES2) that does
not regularly differentiate spontaneously to
cardiomyocytes, with END-2 cells (a VE-like
cell line) provided the missing trigger for
cardiac differentiation.
Another important property of the hES
differentiating system is the ability to pro-
vide, reproducibly, differentiated, non-
transformed, cardiomyocytes for the long-
term in vitro assessment of cardiac tissue.
Although the heart has been thoroughly
investigated in its intact form, only a small
number of in vitro models are currently
available for the study of its structural and
functional properties during normal physi-
ological and pathological states. These
models include a number of primary cul-
tures, which may be limited by their rela-
tively short-term availability and by the
lack of a similar human model. The ability
of hES cells to provide in vitro cardiomyo-
cyte tissue for long-term assessment may
also prove invaluable for drug discovery,
drug screening, and toxicity testing.
Furthermore, by using the differentiation
of the murine ES cells to cardiomyocytes,
a standardized in vitro model (the so-called
embryonic stem cell test) has already been
derived to analyze the embryotoxic effects
of chemical compounds.
12.6
Myocardial Regeneration Strategies
using hES-derived Cardiomyocytes
Cell replacement therapy is emerging as
an innovative therapeutic approach for the
treatment of degenerative diseases (see
Part I, Chapter 14). This therapeutic
approach for degenerative heart diseases is
based on the assumption that myocardial
function may be improved by repopulating
diseased areas with a new pool of func-
tional cells. Although a number of cell
types have been suggested for tissue graft-
ing (see Section 12.1), the ideal donor cell
should probably exhibit the electrophysio-
logical, structural and contractile proper-
ties of cardiomyocytes and should be able
to integrate both structurally and function-
ally with host tissue. In addition, it has to
retain an initial high proliferative potential
that may enable improved colonization of
the scar tissue. The ability to undergo ge-
netic manipulation ex vivo in order to pro-
mote desirable characteristics, such as re-
sistance to ischemia and apoptosis and im-
proved contractile functions, may be an-
other advantage of such an ideal cell type.
Finally, the optimal candidate cell should
have an autologous origin or retain mini-
mal immunogenicity and should be readily
available in large quantities for transplan-
tation.
Unfortunately, none of the currently
available candidate cell sources exhibits all
of the aforementioned properties. The de-
rivation of the hES cell lines offers a num-
ber of potential advantages over the cur-
rently available candidate donor cells. The
hES cells are currently the only cell source
that potentially can provide, ex vivo, an un-
limited number of human cardiac cells for
transplantation. Because of their inherent
cardiac phenotype, hES-derived cells are
more likely to achieve a functional connec-
tion with host myocardium than other
non-cardiomyocyte cell grafts. Although
several of the aforementioned studies,
using a variety of adult stem cells, have
shown an improvement in cardiac ejection
fraction, the mechanisms involved are not
clear. Such an improvement may result
from changes in cardiac architecture fol-
lowing transplantation, from changes in
the passive diastolic properties of heart, or
from the prevention of remodeling. True
systolic augmentation following cell trans-
plantation, however, would depend on
functional integration between graft and
host cardiomyocytes.
12.6 Myocardial Regeneration Strategies using hES-derived Cardiomyocytes 295
Another possible advantage of the hES
cells is their ability to differentiate into a
plurality of cell lineages. This capability
may be utilized for transplantation of differ-
ent cell types such as endothelial progenitor
cells for induction of angiogenesis, and
even specialized cardiomyocytes subtypes
(pacemaking cells, atrial, ventricular, etc.)
tailored for specific applications. In addi-
tion, due to their clonal origin, the hES-de-
rived cardiomyocytes could lend themselves
to extensive characterization and genetic
manipulation to promote desirable charac-
teristics such as resistance to ischemia and
apoptosis, improved contractile function,
and specific electrophysiological properties.
Furthermore, the hES-derived cells could
also serve as a platform and a cellular vehi-
cle for different gene therapy procedures
aiming to manipulate the local myocardial
environment by local secretion of growth-
promoting factors, various drugs, and an-
giogenic growth factors. Finally, the ability
to generate potentially unlimited numbers
of cardiomyocytes ex vivo from the hES cells
may also bring a unique value to tissue en-
gineering approaches.
Although hES cell-derived cardiomyo-
cytes could, in theory, have the potential to
fulfill most of the properties of the ideal do-
nor cell, a number of critical obstacles must
be overcome prior to clinical application:
1. Studies assessing the ability of the cells
to survive and integrate upon transplan-
tation to the normal and diseased myo-
cardial host tissue should be conducted.
2. Strategies need to be developed for di-
recting hES cell differentiation into the
cardiac lineage (as discussed above).
3. Purification of the cardiomyocyte popu-
lation should be achieved using selection
protocols.
4. Up-scaling of the culturing techniques is
needed to yield clinically relevant num-
ber of cells for transplantation.
5. A transplantation technique should be
developed to enable proper alignment of
the graft tissue, high seeding rate of the
transplanted cells, and minimal damage
to the host tissue.
6. Strategies aimed at preventing immuno-
logical rejection of the cells should be
developed.
12.7
Functional Integration of the Cell Grafts
Optimal functional improvement following
cell grafting would require structural, elec-
trophysiological, and mechanical coupling
of donor cells to the existing network of host
cardiomyocytes. For example, although
transplantation of skeletal myoblasts was
shown to improve myocardial performance,
gap junctions were not observed between
graft and host tissues [66]. Yet even the pres-
ence of such gap junctions between host
and donor cardiomyocyte tissues, as ob-
served in some studies, does not guarantee
functional integration. For such integration
to occur, currents generated in one cell
passing through gap junctions must be suf-
ficient to depolarize neighboring cells.
In a recent study, we tested the ability of
the hES cell-derived cardiomyocytes to in-
tegrate structurally and functionally with
host cardiac tissue both in vitro and in vivo
[67]. Initially, the ability of the hES cardio-
myocytes to form electromechanical con-
nections with primary cardiac cultures was
assessed using a high-resolution, in-vitro
coculturing system (Fig. 12.3a). The con-
tracting areas within the EBs were dis-
sected and added to primary neonatal rat
cardiomyocyte cultures. Within 24 hours of
grafting it was possible already to detect
microscopically, in all 22 cocultures stud-
ied, synchronous mechanical activity (as
impressively shown in a video on the sup-
12.7 Functional Integration of the Cell Grafts 297
Fig. 12.3 (A) Multi-electrode recordings showing
in vitro electrical integration between the hES cell-
derived cardiomyocytes (white cluster of cells in
the phase-contrast micrograph) and primary rat
cardiomyocyte cultures (black area in the micro-
graph). The activation map (top, right) generated
demonstrated propagation of the electrical activity
from the rat tissue (red) in this example to the
rest of the co-culture, activating also the human
tissue. Simultaneous recordings from both human
and rat tissues (red and green electrodes) de-
picted long-term synchronous activity. (BD) Gen-
eration of a biological pacemaker using the hES
cell-derived cardiomyocytes in the swine slow
heart rate (complete AV block) model. (B) Electro-
cardiographic recordings following the creation of
AV block demonstrated complete dissociation be-
tween the atrial and ventricular activities with a
slow ventricular rate (top). Following cell trans-
plantation, it was possible to detect episodes of
an ectopic ventricular rhythm that had a signifi-
cantly different morphology and faster rate than
the initial rhythm. (C) Three-dimensional electroa-
natomical activation map of the normal activation
pattern of the ventricle prior to cell grafting (left)
and of the new ectopic rhythm shown from a left
lateral view. Note that the origin of the new activ-
ity (red) is in the posterolateral wall, at the site of
cell grafting. (D) During mapping, a focal ablation
(brown label) was performed 2 cm away from the
earliest activation site (red). Excellent spatial cor-
relation was noted in pathology, with the ablation
site being exactly 2 cm away from the cell injec-
tion site. Histological examination of the earliest
activation site revealed the presence of the grafted
cells (positive immunostaining using anti-human
mitochondrial antibodies, red) and their cardio-
myocyte phenotype (green staining using anti-cTnI
antibodies).
plement CD-ROM). We then mapped the
electrical activity of the co-cultures with
the high-resolution MEA mapping tech-
nique and documented synchronous activ-
ity and tight electrophysiological coupling
between the two tissue types. We also
showed electromechanical connections and
structural integration, as identified by the
presence of gap junctions between the hu-
man and rat cardiomyocytes. The high de-
gree of coupling was evident by the lack of
local conduction delay at the tissues junc-
tion, by the continuous long-term cou-
pling, and by the persistent coupling dur-
ing altered pacemaker position, adrenergic
stimulation and partial (but not total) gap-
junction uncoupling.
In order to demonstrate the ability of
the hES cell-derived cardiomyocytes to sur-
vive, function, and integrate in the in vivo
heart, we assessed their ability to pace the
heart and to function as a biological pace-
maker in an animal model of slow heart
rate (Fig. 12.3bd). An animal model of
complete atrioventricular (AV) block was
generated in pigs by ablating their AV
node, the major electrical conduction path-
way between the atria and the ventricles.
This resulted in complete dissociation be-
tween the atrial and ventricular electrical
activities, and the generation of a slow
ventricular rate, mimicking the clinical
scenario of patients suffering from com-
plete AV block, requiring the implantation
of an electronic pacemaker (Fig. 12.3c).
Following the creation of AV block in
these animals, the spontaneously contract-
ing EBs were injected into the posterolat-
eral left ventricular wall and their electro-
cardiogram monitored.
Following cell grafting, a new ectopic ven-
tricular rhythm was detected in 11 out of 13
animals studies, in six of which it was char-
acterized by sustained and long-term activ-
ity. Three-dimensional electrophysiological
mapping showed that this ectopic ventricu-
lar rhythm originated from the area of cell
transplantation (Fig. 12.3c, d). Pathological
studies validated the presence and integra-
tion of the grafted hES cell-derived cardio-
myocytes at the site of cell transplantation
(Fig. 12.3d).
12.8
Cardiomyocyte Enrichment, Purification,
and Up-scaling Strategies
Although cardiomyocyte tissue can be re-
producibly generated from ES cells using
the EB differentiating system in both the
murine and human models, the differen-
tiating cardiomyocytes typically account for
only a minority of the cells within the
EBs. Similarly, spontaneously contracting
areas are not observed in all EBs, even less
so in the human model. Since the number
of cardiomyocytes generated may have an
important effect on the ultimate success of
cell grafting procedures, cardiomyocyte en-
richment of the EB differentiating system
may be of crucial importance.
Although cardiomyocyte differentiation
may be enhanced by one of the possible
directed differentiation approaches de-
scribed above, it is unlikely that the degree
of purity achieved would be sufficient for
clinical purposes. Given the heterogeneous
mixture of the differentiating cells within
the EB, the task of obtaining a relatively
pure culture of cardiomyocytes would
probably require some form of selection
strategy. Such a strategy is required to in-
crease the number of cardiomyocytes and
to avoid the presence of other cell deriva-
tives or remaining pluripotent stem cells
in the graft.
A relatively simple and elegant strategy
for cardiomyocyte selection during ES cell
differentiation was reported in a mouse
model [68]. In this approach, a cardiac-re-
strictive promoter is used to drive a selec-
tion marker such as an antibiotic resis-
tance gene (Neo
). Once a clone that sta-

bly expresses the vector is isolated, undif-
ferentiated genetically modified ES cells
could be propagated and expanded. The
ES cells are then allowed to differentiate
in vitro and subjected to selection with the
appropriate antibiotic (neomycin, G418).
Using this selection process during in vitro
differentiation, Klug et al. [68] showed that
>99% pure cardiomyocyte cultures could
be generated in the murine model. The se-
lected cardiomyocytes were further shown
to form stable grafts following transplanta-
tion into adult dystrophic mice hearts.
Using a slightly different approach, re-
searchers have transfected murine ES cells
with a construct encoding a cytomegalo-
virus (CMV) enhancer and a ventricular-
specific (MLC-2V) promoter, driving the
green fluorescent protein (GFP) product
[69] (see Part I, Chapter 6). The use of Per-
coll gradient centrifugation and subse-
quent fluorescence-activated cell sorting
(FACS) yielded 97% pure cardiomyocyte
fractions. Approximately 80% of these car-
diomyocytes displayed a typical ventricular
action potential.
It is estimated that, typically, hundred of
millions of cardiomyocytes are lost in a
large myocardial infarction that results in
the development of heart failure. More-
over, the transplantation of an even a
greater number of cells may be required to
replace this cell loss because of the signifi-
cant number of cardiomyocytes that die
following cell grafting. Therefore, a major
barrier for the possible use of hES cells in
cell transplantation strategies is the gen-
eration of sufficient numbers of cardio-
myocytes.
Strategies to increase numbers of cardio-
myocytes generated during hES cell differ-
entiation may, in theory, be employed at
several levels:
By increasing the initial number of un-
differentiated hES cells used for differ-
entiation.
By increasing the percentage of hES
cells differentiating to the cardiac line-
age using the possible directed differen-
tiating systems described above.
By increasing the ability of the cells to
proliferate following cardiomyocyte dif-
ferentiation.
By up-scaling the entire process using
bioreactors and related technologies (see
Part IV, Chapter 1).
12.9
Prevention of Immunological Rejection
A major obstacle for the utilization of hES
cell derivatives in the regeneration of dif-
ferent tissue types is to prevent their im-
mune rejection. Although this issue is be-
yond the scope of this chapter, and is the
subject of an excellent review [70], we will
briefly discuss some possible strategies to
overcome this problem. The first question
is precisely how immunogenic are tissues
derived from ES cells. An initial character-
ization of the immunogenicity of the hES
cells was conducted recently [71], and the
hES cells were shown to express relatively
low level of human leukocyte antigen
(HLA) class I molecules. This expression
was increased only moderately after differ-
entiation in vitro (to EBs) and in vivo (to
teratoma cells), but was significantly aug-
mented following interferon-c treatment.
No expression of HLA class II molecules
and the ligands for NK cell receptors was
detected on the ES cells or their differen-
tiated products. However, this study did
not examine in detail the possible effects
of the degree of EB maturation and differ-
12.9 Prevention of Immunological Rejection 299
entiation and the heterogeneity of the cells
within the EB on the immunogenicity of
the cells. In addition, possible expression
of HLA class II molecules should also be
assessed in vivo, especially in the setting of
an inflammatory response such as occurs
during graft rejection.
The absence of major histocompatibility
complex (MHC)-II antigens on hES cells
and their cardiomyocyte derivatives may
be important, since cells expressing these
proteins (e.g., B cells, macrophages, den-
dritic cells, endothelial cells) are believed
to be highly immunogenic. This may pro-
vide an inherent immune advantage to
hES-derived grafts, and possibly require
milder immunosuppressive regimens. In
addition, strategies aimed at reducing the
mass of alloreactive T cells are being devel-
oped, and these and other novel therapies
with particular relevance to the anticipated
immune response mounted against ES-de-
rived cell transplants will probably be em-
ployed.
Other approaches for reducing graft
rejection may be to establish banks of
MHC antigen-typed hES cells. An alterna-
tive solution to prevent immune rejection
may be to generate a universal donor ES
cell line. This could be achieved by silenc-
ing genes associated with the assembly or
transcriptional regulation of the MHCs, or
by inserting or deleting other genes that
can modulate the immune response. An-
other attractive strategy for inducing toler-
ance is hematopoietic chimerism which,
in theory, may be achieved by transplant-
ing hematopoietic stem cells derived from
hES cells. Following cell engraftment, the
host will obtain tolerance due to the nega-
tive selection of alloreactive T cells in the
thymus. Hence, various differentiated deri-
vatives of the specific ES cell line could be
then safely transplanted, without the risk
of immune rejection.
One of the most promising strategies is
based on the generation of isogenic ES cell
lines tailored specifically for each patient.
Recently, Hwang et al. [72] impressively
demonstrated that, by using somatic nu-
clear transfer technology, this strategy may
become technically possible. These authors
were able to derive a hES cell line from an
enucleated oocyte following somatic cell
nuclear transfer (SCNT) (see Part I, Chap-
ter 11).
12.10
Conclusions
The development of the hES lines and
their ability to differentiate into cardiomyo-
cyte tissue holds great promise for several
areas of cardiovascular research and clini-
cal investigation. Research based on these
cells may help to elucidate the mecha-
nisms involved in early human cardiac
lineage commitment, differentiation and
maturation. Moreover, this research may
promote the discovery of novel growth and
transcriptional factors by using gene trap-
ping techniques, functional genomics and
proteomics, as well as providing a novel in
vitro model for drug development and test-
ing. Finally, the ability to generate human
cardiac tissue in vitro provides an exciting
and promising cell source for the emer-
ging discipline of regenerative medicine,
tissue engineering, and myocardial repair.
Despite these advances, important knowl-
edge in these areas is still lacking, many
more basic investigations are required, sev-
eral methodologies aspects must be re-
solved, and several milestones achieved in
order to harness fully the enormous re-
search and clinical potential of this unique
biopharmaceutical technology.
References
1 Cohn, J. N., et al. Report of the National
Heart, Lung, and Blood Institute Special Em-
phasis Panel on Heart Failure Research. Circu-
lation 95, 766770 (1997).
2 Koh, G. Y., et al. Stable fetal cardiomyocyte
grafts in the hearts of dystrophic mice and
dogs. J Clin Invest 96, 20342042 (1995).
3 Leor, J., et al. Bioengineered cardiac grafts:
A new approach to repair the infarcted myo-
cardium? Circulation 102, III56III61 (2000).
4 Soonpaa, M. H., Koh, G. Y., Klug, M. G. &
Field, L. J. Formation of nascent intercalated
disks between grafted fetal cardiomyocytes
and host myocardium. Science 264, 98101
(1994).
5 Muller-Ehmsen, J., et al. Rebuilding a dam-
aged heart: long-term survival of transplanted
neonatal rat cardiomyocytes after myocardial
infarction and effect on cardiac function. Cir-
culation 105, 17201726 (2002).
6 Etzion, S., Battler, A., Barbash, I. M., Cagnano,
E., Zarin, P., Granot, Y., et al. Influence of
embryonic cardiomyocyte transplantation on
the progression of heart failure in a rat model
of extensive myocardial infarction. J Mol Cell
Cardiol 33, 13211330 (2001).
7 Scorsin, M., Hagege, A. A., Marotte, F., Mi-
rochnik, N., Copin, H., Barnoux, M., et al.
Does transplantation of cardiomyocytes im-
prove function of infarcted myocardium?
Circulation 96, 188193 (1997).
8 Murry, C. E., Wiseman, R. W., Schwartz, S. M.
& Hauschka, S. D. Skeletal myoblast trans-
plantation for repair of myocardial necrosis.
J Clin Invest 98, 25122523 (1996).
9 Taylor, D. A., et al. Regenerating functional
myocardium: improved performance after
skeletal myoblast transplantation. Nat Med 4,
929933 (1998).
10 Menasche, P., et al. Autologous skeletal myo-
blast transplantation for severe postinfarction
left ventricular dysfunction. J Am Coll Cardiol
41, 10781083 (2003).
11 Shintani, S., et al. Mobilization of endothelial
progenitor cells in patients with acute myocar-
dial infarction. Circulation 103, 27762779
(2001).
14 Takahashi, T., et al. Ischemia- and cytokine-in-
duced mobilization of bone marrow-derived
endothelial progenitor cells for neovasculariza-
tion. Nat Med 5, 434438 (1999).
15 Asahara, T., et al. Bone marrow origin of en-
dothelial progenitor cells responsible for post-
natal vasculogenesis in physiological and
pathological neovascularization. Circ Res 85,
221228 (1999).
16 Orlic, D., et al. Bone marrow cells regenerate
infarcted myocardium. Nature 410, 701705
(2001).
17 Murry, C. E., et al. Haematopoietic stem cells
do not transdifferentiate into cardiac myocytes
in myocardial infarcts. Nature 428, 664668
(2004).
18 Balsam, L. B., et al. Haematopoietic stem cells
adopt mature haematopoietic fates in isch-
aemic myocardium. Nature 428, 668673
(2004).
19 Assmus, B., et al. Transplantation of Progeni-
tor Cells and Regeneration Enhancement in
Acute Myocardial Infarction (TOPCARE-AMI).
Circulation 106, 30093017 (2002).
20 Wollert, K. C., et al. Intracoronary autologous
bone-marrow cell transfer after myocardial in-
farction: the BOOST randomised controlled
clinical trial. Lancet 364, 141148 (2004).
21 Toma, C., Pittenger, M. F., Cahill, K. S., Byrne,
B. J. & Kessler, P. D. Human mesenchymal
stem cells differentiate to a cardiomyocyte
phenotype in the adult murine heart. Circula-
tion 105, 9398 (2002).
22 Makino, S., et al. Cardiomyocytes can be gen-
erated from marrow stromal cells in vitro.
J Clin Invest 103, 697705 (1999).
23 Shake, J. G., et al. Mesenchymal stem cell im-
plantation in a swine myocardial infarct mod-
el: engraftment and functional effects. Ann
Thorac Surg 73, 19191925; discussion 1926
(2002).
24 Weissman, I. L., Anderson, D. J. & Gage, F.
Stem and progenitor cells: origins, pheno-
types, lineage commitments, and transdiffer-
entiations. Annu Rev Cell Dev Biol 17, 387403
(2001).
25 Evans, M. & Kaufman, M. Establishment in
culture of pluripotent cells from mouse em-
bryos. Nature 292, 154156 (1981).
26 Martin, G. Isolation of a pluripotent cell line
from early mouse embryos cultured in medi-
um conditioned by teratocarcinoma stem cells.
Proc Natl Acad Sci USA 78, 7635 (1981).
27 Doetschman, T. C., Eistetter, H., Katz, M.,
Schmidt, W. & Kemmler, R. In vitro develop-
ment of blastocyst derived embryonic stem
cells lines: Formation of visceral yolk sac,
References 301
blood islands and myocardium. J Embryol
Morphol 87, 2745 (1985).
28 Thomson, J. A., et al. Embryonic stem cell
lines derived from human blastocysts. Science
282, 11451147 (1998).
29 Reubinoff, B. E., Pera, M. F., Fong, C. Y.,
Trounson, A. & Bongso, A. Embryonic stem
cell lines from human blastocysts: somatic dif-
ferentiation in vitro. Nat Biotechnol 18, 399
404 (2000).
30 Itskovitz-Eldor, J., et al. Differentiation of hu-
man embryonic stem cells into embryoid
bodies compromising the three embryonic
germ layers. Mol Med 6, 8895 (2000).
31 Amit, M., et al. Clonally derived human em-
bryonic stem cell lines maintain pluripotency
and proliferative potential for prolonged peri-
ods of culture. Dev Biol 227, 271278 (2000).
32 Odorico, J. S., Kaufman, D. S. & Thomson,
J. A. Multilineage differentiation from human
embryonic stem cell lines. Stem Cells 19, 193
204 (2001).
33 Rao, M. Conserved and divergent paths that
regulate self-renewal in mouse and human
embryonic stem cells. Dev Biol 275, 269286
(2004).
34 Ying, Q. L., Nichols, J., Chambers, I. & Smith,
A. BMP induction of Id proteins suppresses
differentiation and sustains embryonic stem
cell self-renewal in collaboration with STAT3.
Cell 115, 281292 (2003).
35 Xu, C., Inokuma, M. S., Denham, J., Golds,
K., Kundu, P., Gold, J. D., et al. Feeder-free
growth of undifferentiated human embryonic
stem cells. Nat Biotechnol 19, 971974 (2001).
36 Richards, M., Fong, C. Y., Chan, W. K., Wong,
P. C. & Bongso, A. Human feeders support
prolonged undifferentiated growth of human
inner cell masses and embryonic stem cells.
Nat Biotechnol 20, 933936 (2002).
37 Sato, N., Meijer, L., Skaltsounis, L., Green-
gard, P. & Brivanlou, A. H. Maintenance of
pluripotency in human and mouse embryonic
stem cells through activation of Wnt signaling
by a pharmacological GSK-3-specific inhibitor.
Nat Med 10, 5563 (2004).
38 Bhattacharya, B., et al. Gene expression in hu-
man embryonic stem cell lines: unique molec-
ular signature. Blood 103, 29562964 (2004).
39 Nagy, A., Rossant, J., Nagy, R., Abramow-New-
erly, W. & Roder, J. C. Derivation of completely
cell culture-derived mice from early-passage
embryonic stem cells. Proc Natl Acad Sci USA
90, 84248428 (1993).
40 Kehat, I., et al. Human embryonic stem cells
can differentiate into myocytes with structural
and functional properties of cardiomyocytes. J
Clin Invest 108, 407414 (2001).
41 Reubinoff, B. E., et al. Neural progenitors
from human embryonic stem cells. Nat Bio-
technol 19, 11341140 (2001).
42 Zeng, X., et al. Dopaminergic differentiation
of human embryonic stem cells. Stem Cells 22,
925940 (2004).
43 Assady, S., Maor, G., Amit, M., Itskovitz-Eldor,
J., Skorecki, K. L. & Tzukerman, M. Insulin
production by human embryonic stem cells.
Diabetes 50, 16911697 (2001).
44 Kaufman, D. S. & Thomson, J. A. Human ES
cells: haematopoiesis and transplantation
strategies. J Anat 200, 243248 (2002).
45 Green, H., Easley, K. & Iuchi, S. Marker suc-
cession during the development of keratino-
cytes from cultured human embryonic stem
cells. Proc Natl Acad Sci USA 100, 15625
15630 (2003).
46 Sottile, V., Thomson, A. & McWhir, J. In vitro
osteogenic differentiation of human ES cells.
Cloning Stem Cells 5, 149155 (2003).
47 Levenberg, S., Golub, J. S., Amit, M., Itskovitz-
Eldor, J. & Langer, R. Endothelial cells derived
from human embryonic stem cells. Proc Natl
Acad Sci USA 99, 43914396 (2002).
48 Wobus, A. M., Wallukat, G. & Hescheler, J.
Pluripotent mouse embryonic stem cells are
able to differentiate into cardiomyocytes ex-
pressing chronotropic responses to adrenergic
and cholinergic agents and Ca
2+
channel
blockers. Differentiation 48, 173182 (1991).
49 Hescheler, J., et al. Embryonic stem cells: a
model to study structural and functional prop-
erties in cardiomyogenesis. Cardiovasc Res 36,
149162 (1997).
50 Metzger, J. M., Lin, W. I., Johnston, R. A.,
Westfall, M. V. & Samuelson, L. C. Myosin
heavy chain expression in contracting myo-
cytes isolated during embryonic stem cell car-
diogenesis. Circ Res 76, 710719 (1995).
51 Klug, M. G., Soonpaa, M. H. & Field, L. J.
DNA synthesis and multinucleation in em-
bryonic stem cell-derived cardiomyocytes. Am
J Physiol 269, H1913H1921 (1995).
51 Maltsev, V. A., Rohwedel, J., Hescheler, J. &
Wobus, A. M. Embryonic stem cells differenti-
ate in vitro into cardiomyocytes representing
sinus nodal, atrial and ventricular cell types.
Mech Dev 44, 4150 (1993).
52 Maltsev, V. A., Wobus, A. M., Rohwedel, J., Ba-
der, M. & Hescheler, J. Cardiomyocytes differ-
entiated in vitro from embryonic stem cells
developmentally express cardiac-specific genes
and ionic currents. Circ Res 75, 233244
(1994).
53 Maltsev, V. A., Ji, J. G., Wobus, A. M., Fleisch-
mann, B. K. & Hescheler, J. Establishment of
beta adrenergic modulation of L-type Ca
2+
cur-
rent in the early stages of cardiomyocytes
development. Circ Res 84, 136145 (1999).
54 Snir, M., Kehat, I., Gepstein, A., Coleman, R.,
Itskovitz-Eldor, J., Livne, E. & Gepstein, L. As-
sessment of the ultrastructural and prolifera-
tive properties of human embryonic stem cell-
derived cardiomyocytes. Am J Physiol Heart
Circ Physiol 285, H2355H2363 (2003).
55 Satin, J., et al. Mechanism of spontaneous ex-
citability in human embryonic stem cell de-
rived cardiomyocytes. J Physiol 559, 479496
(2004).
56 Kehat, I., Gepstein, A., Spira, A., Itskovitz-El-
dor, J. & Gepstein, L. High-resolution electro-
physiological assessment of human embryonic
stem cell-derived cardiomyocytes: a novel in
vitro model for the study of conduction. Circ
Res 91, 659661 (2002).
57 Xu, C., Police, S., Rao, N. & Carpenter, M. K.
Characterization and enrichment of cardio-
myocytes derived from human embryonic
stem cells. Circ Res 91, 501508 (2002).
58 He, J. Q., Ma, Y., Lee, Y., Thomson, J. A. &
Kamp, T. J. Human embryonic stem cells de-
velop into multiple types of cardiac myocytes:
action potential characterization. Circ Res 93,
3239 (2003).
59 Mummery, C., et al. Differentiation of human
embryonic stem cells to cardiomyocytes: role
of coculture with visceral endoderm-like cells.
Circulation 107, 27332740 (2003).
60 Zaffran, S. & Frasch, M. Early signals in cardi-
ac development. Circ Res 91, 457469 (2002).
61 Nascone, N. & Mercola, M. An inductive role
for the endoderm in Xenopus cardiogenesis.
Development 121, 515523 (1995).
62 Monzen, K., Nagai, R. & Komuro, I. A role
for bone morphogenetic protein signaling in
cardiomyocyte differentiation. Trends Cardio-
vasc Med 12, 263269 (2002).
63 Pandur, P., Lasche, M., Eisenberg, L. M. &
Kuhl, M. Wnt-11 activation of a non-canonical
Wnt signalling pathway is required for cardio-
genesis. Nature 418, 636641 (2002).
64 Nakamura, T., Sano, M., Songyang, Z. &
Schneider, M. D. A Wnt- and beta-catenin-de-
pendent pathway for mammalian cardiac
myogenesis. Proc Natl Acad Sci USA 100,
58345839 (2003).
65 Harvey, R. P. Patterning the vertebrate heart.
Nat Rev Genet 3, 544556 (2002).
66 Reinecke, H., MacDonald, G. H., Hauschka,
S. D. & Murry, C. E. Electromechanical cou-
pling between skeletal and cardiac muscle.
Implications for infarct repair. J Cell Biol 149,
731740 (2000).
67 Kehat, I., et al. Electromechanical integration
of human embryonic stem cell derived cardio-
myocytes. Nat Biotechnol 22, 12821289
(2004).
68 Klug, M. G., Soonpaa, M. H., Koh, G. Y. &
Field, L. J. Genetically selected cardiomyocytes
from differentiating embryonic stem cells
form stable intracardiac grafts. J Clin Invest
98, 216224 (1996).
69 Muller, M., et al. Selection of ventricular-like
cardiomyocytes from ES cells in vitro. FASEB
J 14, 25402548 (2000).
70 Bradly, J. A., Bolton, E. M. & Pederson, R. A.
Stem cell medicine encounters the immune
system. Nat Rev Immun 2, 859871 (2002).
71 Drukker, M., et al. Characterization of the ex-
pression of MHC proteins in human embryo-
nic stem cells. Proc Natl Acad Sci USA 99,
98649869 (2002).
72 Hwang, W. S., Ryu, Y. J., Park, J. H., Park,
E. S., Lee, E. G., Koo, J. M., et al. Evidence of a
pluripotent human embryonic stem cell line
derived from a cloned blastocyst. Science 303,
16691674 (2004).
References 303
Movie A video showing the hybrid culture grown on top of the micro-electrode array plate. The image
is shown at a high magnification (40). Note the synchronous contractions of the hES cell-derived
cardiomyocyte tissue (cell cluster on the right side of the image) and the neonatal rat ventricular myo-
cyte monolayer.
Abstract
The incidence of cardiovascular diseases
continues to increase worldwide, despite
recent advances in therapeutic strategies.
As a result, growing numbers of patients
are or will be failing, current state-of-the-
art therapies, thereby generating an in-
creased demand for innovative therapies.
Gene and cell-based strategies have
evolved into powerful therapeutic plat-
forms capable of influencing the patho-
physiology of complex, acquired, polyge-
netic diseases. The simultaneous elucida-
tion of the molecular mechanisms in-
volved in atherosclerosis, ischemic heart
disease and myocardial failure, and the de-
velopment of sophisticated surgical and
catheter-based systems capable of deliver-
ing a new generation of safe, effective and
stable vectors, makes gene and cell-based
therapeutics for cardiovascular disease a
strategy whose time has come. In this
broad-based review, we examine the role of
gene and cell-based approaches to the
management of lower-extremity and myo-
cardial ischemia, bypass graft failure, arte-
rial restenosis after intervention, myocar-
dial protection and myocardial failure.
Abbreviations
bARK-1 bAR Kinase-1
b
1
ARs b
1
-adrenergic receptors
AAV adeno-associated viral vectors
ACE angiotensin-converting enzyme
cAMP cyclic adenosine monophosphate
CHF congestive heart failure
ecSOD extracellular superoxide dismutase
eNOS endothelial nitric oxide synthase
EPCs endothelial progenitor cells
GCSF granulocyte colony stimulating factor
HGF hepatocyte growth factor
HIF hypoxia inducible factor
HO-1 hemo-oxygenase-1
HVJ hemagglutinating-virus of Japan
I/R ischemiareperfusion
iNOS inducible nitric oxide synthase
LAD left anterior descending
NOS nitric oxide synthase
ODN 1-oligodeoxynucleotide
PCNA proliferating cell nuclear antigen
PKA protein kinase A
PTI percutaneous transluminal
intervention
ROS reactive oxygen species
SPECT single photon emission computed
tomogram
VSMC vascular smooth muscle cell
305
13
Gene and Cell-based Therapies for Cardiovascular Disease
Abeel A. Mangi
ISBN: 3-527-31184-X
13.1
Introduction
Cardiovascular disease is the leading cause
of lost productivity, morbidity and mortal-
ity in industrialized nations, and it is pro-
jected that by 2020 it will be the leading
contributor to the worldwide burden of
disease. The past several decades have wit-
nessed significant advances in cardiovascu-
lar therapeutics the invention of new in-
terventional and electrophysiological de-
vices, the development of minimally inva-
sive surgical techniques, and the discovery
of new and effective drugs that have al-
tered the natural history of cardiovascular
disease. Ironically, the increased survival
resulting from these treatment modalities
has produced a growing population with
chronic cardiovascular diseases who are, or
who will be failing, current state-of-the-
art therapies. Accordingly, there is an
increasing need for the development of
advanced and innovative therapeutics.
Recent advances in gene and cell-based
approaches provide unprecedented oppor-
tunities for the discovery of novel therapies
to address this pressing demand.
A confluence of scientific, technical, and
medical advancements has made genetic
therapeutics for cardiovascular disease a
promising and exciting field. The molecu-
lar mechanisms of major cardiovascular
disorders such as atherosclerosis, ischemic
heart disease, and myocardial failure have
been well characterized. Sophisticated sur-
gical and catheter-based systems that can
enable the delivery of therapeutic genes or
cells in vivo are in clinical use, and clinical
therapeutic end-points for the evaluation
of treatment efficacy have been clearly de-
fined. Simultaneously, gene therapy has
evolved from a modality restricted to the
potential cure of monogenetic diseases to
a therapeutic platform that enables cus-
tomized local delivery of genes, the prod-
ucts of which can act as drugs, thereby
exerting therapeutic actions even on com-
plex, acquired polygenetic diseases. This
paradigm shift has been facilitated by the
identification of gene(s), the alteration of
which is linked to pathophysiologic de-
rangements, and the manipulation of
which (by over-expression or deletion)
disrupts pathophysiologic processes. Ad-
vances in vector biology have resulted in a
new generation of vectors that are safe, ef-
ficient and stable. These can be further en-
hanced by the use of endogenous, induci-
ble promoter systems engineered to en-
sure tissue-specific delivery and tissue-re-
stricted expression. Recent developments
in stem cell biology and ex vivo genetic
manipulation makes regenerative medi-
cine for blood vessels and the myocardium
a distinct possibility. The aim of this chap-
ter is to highlight the current status of
gene and cell-based therapy for cardiovas-
cular disease, and to provide a preview
into the future of genetic biopharmaceuti-
cals for these diseases.
13.2
Gene Therapy as Novel Drug Delivery
The local delivery of high levels of thera-
peutic genes (through direct transfer or via
transplanted cells), the products of which
disrupt disease processes, while avoiding
high (and potentially toxic) systemic levels,
constitutes the basis for the paradigm of
genetic therapeutics. We will discuss appli-
cations of this strategy with respect to ther-
apeutic angiogenesis, the inhibition of by-
pass graft failure, the prevention of re-ste-
nosis after vascular intervention, myocar-
dial protection, and myocardial failure.
13.2.1
Gene Therapy for Therapeutic Angiogenesis
The application of vascular endothelial
growth factor (VEGF) for therapeutic an-
giogenesis provides an excellent illustra-
tion of this paradigm. VEGF is a mitogen
produced by endothelial cells, and is in-
duced under ischemic conditions by hyp-
oxia inducible factor (HIF), a property that
makes HIF an attractive candidate for an-
giogenesis gene therapy itself. VEGF then
acts in an autocrine fashion to induce rap-
id replication of endothelial cells the crit-
ical cellular elements in the formation of
new functional blood vessels. These prop-
erties are shared by angiogenic factors he-
patocyte growth factor (HGF) and fibro-
blast growth factor (FGF) [1].
During the mid-1990s it was postulated
that the over-expression of angiogenic
growth factors in ischemic environments
might augment endogenous reserves and
serve as a potent stimulus for angiogen-
esis. Using a rabbit model of acute hind-
limb ischemia, several groups demon-
strated that intramuscular or intra-arterial
delivery of naked plasmid DNA coding for
FGF, HGF, VEGF or HIF-1a/VP16; or re-
combinant FGF or HGF resulted in an ap-
proximate doubling of capillary density by
histology, in two- to three-fold increases in
collateralization on angiography, and in a
comparable magnitude of improvement in
clinical criteria such as calf muscle atro-
phy, limb necrosis, transcutaneous oxime-
try, calf blood flow and pressure ratios [2].
In experimental models, adenoviral deliv-
ery of FGF and VEGF to ischemic lower
extremities resulted in similar improve-
ments, with no evidence of systemic toxici-
ty [3]. This class of gene products also in-
duced therapeutic angiogenesis in experi-
mental models of myocardial ischemia.
Both intramyocardial and intracoronary in-
jection of recombinant protein, or plasmid
DNA coding for HGF, FGF-1 or 2, and
VEGF, resulted in 1.5- to three-fold in-
creases in the numerical density of distri-
bution vessels, with comparable increases
in blood flow ratios, significant reduction
in left ventricular infarct size, and im-
provements in fractional shortening and
left ventricular end-diastolic pressure.
Fusigenic-hemagglutinating-virus of Japan
(HVJ)liposome complex-mediated VEGF
gene transfer (with and without laser
transmyocardial revascularization) was
shown to reverse ischemia-induced abnor-
malities in myocardial contractility in pigs,
while adenoviral delivery of VEGF and
FGF resulted in similar improvements in
myocardial perfusion, collateralization and
function [4].
These preclinical studies have led to the
initiation of over 20 clinical trials investi-
gating the potential efficacy of angiogenic
gene therapy on critical limb and myocar-
dial ischemia (Table 13.1) [5]. These Phase
I and Phase II studies have reported re-
ductions in ischemic symptoms, improve-
ments in exercise time, and improvements
in quality of life subjective measures that
have been supported by improvements in
objective parameters such as perfusion
(SPECT) scanning and angiography.
As outlined in Table 13.1, three Phase I
trials for intramuscular and intra-arterial
administration of naked plasmid VEGF
165
in human lower-extremity ischemia and
Buergers disease have been completed,
having achieved their end-points within
one year. Three Phase II clinical trials are
currently recruiting patients to investigate
the role of intramuscular VEGF plasmid
delivery in lower-extremity ischemia. A
Phase I trial for intra-arterial administra-
tion of recombinant FGF-2 has been com-
pleted (Table 13.1), and the Phase II
TRAFFIC trial further evaluating the effect
F
i
g
.
1
3
.
1
E
x
p
e
r
i
m
e
n
t
a
l
a
n
d
c
l
i
n
i
c
a
l
*
g
e
n
e
t
i
c
t
h
e
r
a
p
e
u
t
i
c
s
f
o
r
c
a
r
d
i
o
v
a
s
c
u
l
a
r
d
i
s
e
a
s
e
s
.
T
a
b
l
e
1
3
.
1
H
u
m
a
n
c
l
i
n
i
c
a
l
t
r
i
a
l
s
f
o
r
t
h
e
r
a
p
e
u
t
i
c
a
n
g
i
o
g
e
n
e
s
i
s
N
a
m
e
Y
e
a
r
I
n
d
i
c
a
t
i
o
n
I
n
c
l
u
s
i
o
n
c
r
i
t
e
r
i
a
P
r
o
t
e
i
n
e
/
G
e
n
e
R
o
u
t
e
o
f
a
d
m
i
n
i
s
t
r
a
t
i
o
n
T
y
p
e
o
f
t
r
i
a
l
F
o
l
l
o
w
-
u
p
O
u
t
c
o
m
e
B
a
u
m
g
a
r
t
n
e
r
1
9
9
8
C
r
i
t
i
c
a
l
l
o
w
e
r
e
x
t
r
e
m
i
t
y
i
s
c
h
e
m
i
a
C
r
i
t
i
c
a
l
i
s
c
h
e
m
i
a
,
r
e
s
i
s
t
a
n
t
t
o
m
a
x
i
m
a
l
m
e
d
i
c
a
l
t
h
e
r
a
p
y
,
n
o
t
s
u
r
g
i
c
a
l
c
a
n
d
i
d
a
t
e
s
N
a
k
e
d
p
l
a
s
m
i
d
V
E
G
F
1
6
5
I
n
t
r
a
m
u
s
c
u
l
a
r
i
n
j
e
c
t
i
o
n
P
h
a
s
e
I
2
1
1
m
o
n
t
h
s
I
m
p
r
o
v
e
d
A
n
k
l
e
-
B
r
a
c
h
i
a
l
i
n
d
e
x
,
e
x
e
r
c
i
s
e
t
i
m
e
,
n
e
w
v
e
s
s
e
l
s
o
n
a
n
g
i
o
-
g
r
a
m
,
l
i
m
b
s
a
l
v
a
g
e
,
i
m
-
p
r
o
v
e
d
t
i
s
s
u
e
i
n
t
e
g
r
i
t
y
I
s
n
e
r
1
9
9
8
B
u
e
r
g
e
r
s
d
i
s
e
a
s
e
C
r
i
t
i
c
a
l
i
s
c
h
e
m
i
a
,
r
e
s
i
s
t
a
n
t
t
o
m
a
x
i
m
a
l
m
e
d
i
c
a
l
t
h
e
r
a
p
y
,
n
o
t
s
u
r
g
i
c
a
l
c
a
n
d
i
d
a
t
e
s
N
a
k
e
d
p
l
a
s
m
i
d
V
E
G
F
1
6
5
I
n
t
r
a
m
u
s
c
u
l
a
r
i
n
j
e
c
t
i
o
n
P
h
a
s
e
I
4
m
o
n
t
h
s
I
m
p
r
o
v
e
m
e
n
t
s
i
n
w
o
u
n
d
h
e
a
l
i
n
g
,
r
e
s
t
p
a
i
n
,
a
n
k
l
e
b
r
a
c
h
i
a
l
i
n
d
i
c
e
s
,
n
e
w
v
e
s
s
e
l
s
o
n
M
R
I
a
n
d
a
n
g
i
o
g
r
a
m
I
s
n
e
r
1
9
9
6
C
r
i
t
i
c
a
l
l
o
w
e
r
e
x
t
r
e
m
i
t
y
i
s
c
h
e
m
i
a
C
r
i
t
i
c
a
l
i
s
c
h
e
m
i
a
,
r
e
s
i
s
t
a
n
t
t
o
m
a
x
i
m
a
l
m
e
d
i
c
a
l
t
h
e
r
a
p
y
,
n
o
t
s
u
r
g
i
c
a
l
c
a
n
d
i
d
a
t
e
s
N
a
k
e
d
p
l
a
s
m
i
d
V
E
G
F
1
6
5
I
n
t
r
a
-
a
r
t
e
r
i
a
l
a
d
m
i
n
i
s
t
r
a
t
i
o
n
P
h
a
s
e
I
1
2
w
e
e
k
s
I
m
p
r
o
v
e
d
a
n
g
i
o
g
r
a
p
h
i
c
a
n
d
D
o
p
p
l
e
r
e
v
a
l
u
a
t
i
o
n
s
h
e
m
a
n
g
i
o
m
a
s
f
r
o
m
a
n
d
r
e
g
r
e
s
s
L
a
z
a
r
o
u
s
2
0
0
0
C
r
i
t
i
c
a
l
l
o
w
e
r
e
x
t
r
e
m
i
t
y
i
s
c
h
e
m
i
a
C
r
i
t
i
c
a
l
i
s
c
h
e
m
i
a
,
e
s
i
s
t
a
n
t
t
o
m
a
x
i
m
a
l
m
e
d
i
c
a
l
t
h
e
r
a
p
y
,
n
o
t
s
u
r
g
i
c
a
l
c
a
n
d
i
d
a
t
e
s
F
G
F
-
2
p
r
o
t
e
i
n
I
n
t
r
a
-
a
r
t
e
r
i
a
l
a
d
m
i
n
i
s
t
r
a
t
i
o
n
P
h
a
s
e
I
7
m
o
n
t
h
s
I
m
p
r
o
v
e
d
b
l
o
o
d
f
l
o
w
L
e
d
e
r
m
a
n
n
/
T
r
a
f
f
i
c
2
0
0
1
C
r
i
t
i
c
a
l
l
o
w
e
r
e
x
t
r
e
m
i
t
y
i
s
c
h
e
m
i
a
R
e
c
r
u
i
t
i
n
g
F
G
F
-
2
p
r
o
t
e
i
n
I
n
t
r
a
-
a
r
t
e
r
i
a
l
a
d
m
i
n
i
s
t
r
a
t
i
o
n
P
h
a
s
e
I
/
I
I
R
e
c
r
u
i
t
i
n
g
G
e
n
z
y
m
e
2
0
0
2
L
o
w
e
r
e
x
t
r
e
m
i
t
y
i
s
c
h
e
m
i
a
C
r
i
t
i
c
a
l
i
s
c
h
e
m
i
a
,
r
e
s
i
s
t
a
n
t
t
o
m
a
x
i
m
a
l
m
e
d
i
c
a
l
t
h
e
r
a
p
y
,
n
o
t
s
u
r
g
i
c
a
l
c
o
n
d
i
d
a
t
e
s
N
a
k
e
d
p
l
a
s
m
i
d
H
I
F
1
a
/
V
P
1
6
I
n
t
r
a
m
u
s
c
u
l
a
r
P
h
a
s
e
I
R
e
c
r
u
i
t
i
n
g
S
c
h
u
m
a
c
h
e
r
1
9
9
8
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
T
h
r
e
e
-
v
e
s
s
e
l
d
i
s
e
a
s
e
a
n
d
d
i
s
t
a
l
L
A
D
d
i
s
e
s
e
a
f
t
e
r
L
I
M
A
i
n
s
e
r
t
i
o
n
F
G
F
-
1
p
r
o
t
e
i
n
I
n
t
r
a
m
y
o
c
a
r
d
i
a
l
i
n
j
e
c
t
i
o
n
P
h
a
s
e
I
1
2
w
k
s
t
o
3
y
r
s
N
e
w
v
e
s
s
e
l
s
d
i
s
t
a
l
t
o
L
A
D
d
e
c
r
e
a
s
e
d
a
n
g
i
n
a
a
n
d
u
s
e
o
f
d
r
u
g
s
T
a
b
l
e
1
3
.
1
(
c
o
n
t
i
n
u
e
d
)
N
a
m
e
Y
e
a
r
I
n
d
i
c
a
t
i
o
n
I
n
c
l
u
s
i
o
n
c
r
i
t
e
r
i
a
P
r
o
t
e
i
n
e
/
G
e
n
e
R
o
u
t
e
o
f
a
d
m
i
n
i
s
t
r
a
t
i
o
n
T
y
p
e
o
f
t
r
i
a
l
F
o
l
l
o
w
-
u
p
O
u
t
c
o
m
e
S
t
e
g
m
a
n
2
0
0
0
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
S
e
v
e
r
e
a
n
g
i
n
a
n
o
t
a
m
e
n
a
b
l
e
t
o
r
e
v
a
s
c
u
l
a
r
i
z
a
t
i
o
n
F
G
F
-
1
p
r
o
t
e
i
n
I
n
t
r
a
m
y
o
c
a
r
d
i
a
l
i
n
j
e
c
t
i
o
n
P
h
a
s
e
I
6
a
n
d
1
2
w
k
s
I
m
p
r
o
v
e
d
S
P
E
C
T
p
e
r
f
u
s
i
o
n
i
n
c
r
e
a
s
e
i
n
e
x
e
r
c
i
s
e
t
o
l
e
r
a
n
c
e
S
e
l
l
k
e
1
9
9
8
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
V
e
s
s
e
l
n
o
t
g
r
a
f
t
a
b
l
e
,
u
n
d
e
r
g
o
i
n
g
C
A
B
G
f
o
r
v
a
r
i
a
b
l
e
m
y
o
c
a
r
d
i
u
m
F
G
F
-
2
p
r
o
t
e
i
n
P
e
r
i
a
d
v
e
n
t
i
t
i
a
l
i
m
p
l
a
n
t
a
t
i
o
n
P
h
a
s
e
I
3
m
o
V
a
r
i
a
b
l
e
S
P
E
C
T
s
c
a
r
s
,
a
r
e
p
e
r
i
o
p
e
r
a
t
i
v
e
i
n
f
a
r
c
t
i
o
n
,
a
l
l
p
a
t
i
e
n
t
s
f
r
e
e
o
f
a
n
g
i
n
a
L
a
h
a
m
2
0
0
0
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
C
A
D
,
n
o
t
a
m
e
n
a
b
l
e
t
o
r
e
v
a
s
c
u
l
a
r
i
z
a
t
i
o
n
F
G
F
-
2
p
r
o
t
e
i
n
I
n
t
r
a
c
o
r
o
n
a
r
y
a
d
m
i
n
i
s
t
r
a
t
i
o
n
P
h
a
s
e
I
1
6
m
o
I
m
p
r
o
v
e
d
n
u
c
l
e
a
r
a
n
d
M
R
I
p
e
r
f
u
s
i
o
n
,
d
e
c
r
e
a
s
e
d
a
n
g
i
n
a
,
i
m
p
r
o
v
e
d
e
x
e
r
c
i
s
e
t
o
l
e
r
a
n
c
e
,
a
n
d
q
u
a
l
i
t
y
o
f
l
i
f
e
U
n
g
e
r
2
0
0
0
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
C
A
D
,
n
o
t
a
m
e
n
a
b
l
e
t
o
r
e
v
a
s
c
u
l
a
r
i
z
a
t
i
o
n
F
G
F
-
2
p
r
o
t
e
i
n
I
n
t
r
a
c
o
r
o
n
a
r
y
a
d
m
i
n
i
s
t
r
a
t
i
o
n
P
h
a
s
e
I
1
m
o
T
r
a
n
s
i
e
n
t
i
n
c
r
e
a
s
e
i
n
L
e
f
t
M
a
i
n
(
L
M
)
d
i
a
m
e
t
e
r
N
o
c
h
a
n
g
e
i
s
e
x
e
r
c
i
s
e
t
o
l
e
r
a
n
c
e
K
l
e
i
m
a
n
/
F
I
R
S
T
2
0
0
0
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
C
A
D
,
n
o
t
a
m
e
n
a
b
l
e
t
o
r
e
v
a
s
c
u
l
a
r
i
z
a
t
i
o
n
F
G
F
-
2
p
r
o
t
e
i
n
I
n
t
r
a
c
o
r
o
n
a
r
y
a
d
m
i
n
i
s
t
r
a
t
i
o
n
P
h
a
s
e
I
6
m
o
N
o
d
i
f
f
e
r
e
n
c
e
b
e
t
w
e
e
n
p
l
a
c
e
b
o
g
r
o
u
p
H
e
n
r
y
2
0
0
0
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
C
A
D
,
n
o
t
a
m
e
n
a
b
l
e
t
o
r
e
v
a
s
c
u
l
a
r
i
z
a
t
i
o
n
V
E
G
F
1
6
5
p
r
o
t
e
i
n
I
n
t
r
a
c
o
r
o
n
a
r
y
a
d
m
i
n
i
s
t
r
a
t
i
o
n
P
h
a
s
e
I
2
m
o
I
m
p
r
o
v
e
d
S
P
E
C
T
s
c
a
n
r
e
s
u
l
t
s
,
d
e
c
r
e
a
s
e
d
a
n
g
i
n
a
i
n
1
3
/
1
5
p
a
t
i
e
n
t
s
H
e
n
d
e
l
2
0
0
0
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
C
A
D
,
n
o
t
a
m
e
n
a
b
l
e
t
o
r
e
v
a
s
c
u
l
a
r
i
z
a
t
i
o
n
V
E
G
F
1
6
5
p
r
o
t
e
i
n
I
n
t
r
a
c
o
r
o
n
a
r
y
a
d
m
i
n
i
s
t
r
a
t
i
o
n
P
h
a
s
e
I
2
m
o
I
m
p
r
o
v
e
d
d
o
s
e
-
d
e
p
e
n
-
d
e
n
t
r
e
s
t
a
n
d
s
t
r
e
s
s
S
P
E
C
T
i
m
a
g
e
s
H
e
n
r
y
/
V
I
V
A
T
r
i
a
l
2
0
0
1
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
C
A
D
,
n
o
t
a
m
e
n
a
b
l
e
t
o
r
e
v
a
s
c
u
l
a
r
i
z
a
t
i
o
n
V
E
G
F
1
6
5
p
r
o
t
e
i
n
I
n
t
r
a
c
o
r
o
n
a
r
y
a
d
m
i
n
i
s
t
r
a
t
i
o
n
P
h
a
s
e
I
I
2
m
o
t
o
1
y
r
T
r
e
n
d
t
o
w
a
r
d
d
e
c
r
e
a
s
e
d
a
n
g
i
n
a
b
u
t
e
q
u
i
v
a
l
e
n
t
e
x
e
r
c
i
s
e
t
o
l
e
r
a
n
c
e
i
n
e
x
p
e
r
i
m
e
n
t
a
l
a
n
d
p
l
a
c
e
b
o
g
r
o
u
p
s
T
a
b
l
e
1
3
.
1
(
c
o
n
t
i
n
u
e
d
)
N
a
m
e
Y
e
a
r
I
n
d
i
c
a
t
i
o
n
I
n
c
l
u
s
i
o
n
c
r
i
t
e
r
i
a
P
r
o
t
e
i
n
e
/
G
e
n
e
R
o
u
t
e
o
f
a
d
m
i
n
i
s
t
r
a
t
i
o
n
T
y
p
e
o
f
t
r
i
a
l
F
o
l
l
o
w
-
u
p
O
u
t
c
o
m
e
S
y
m
e
s
/
V
a
l
e
1
9
9
9
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
S
y
m
p
t
o
m
a
t
i
c
C
A
D
,
n
o
t
a
m
e
n
a
b
l
e
t
o
r
e
v
a
s
c
u
l
a
r
i
z
a
t
i
o
n
N
a
k
e
d
p
l
a
s
m
i
d
V
E
G
F
1
6
5
I
n
t
r
a
m
y
o
c
a
r
d
i
a
l
i
n
j
e
c
t
i
o
n
a
t
t
h
o
r
a
c
o
t
o
m
y
P
h
a
s
e
I
1
y
r
I
m
p
r
o
v
e
d
S
P
E
C
T
,
n
o
r
e
s
t
p
a
i
n
V
a
l
e
1
9
9
9
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
S
y
m
p
t
o
m
a
t
i
c
C
A
D
,
n
o
t
a
m
e
n
a
b
l
e
t
o
r
e
v
a
s
c
u
l
a
r
i
z
a
t
i
o
n
N
a
k
e
d
p
l
a
s
m
i
d
V
E
G
F
1
6
5
I
n
t
r
a
m
y
o
c
a
r
d
i
a
l
i
n
j
e
c
t
i
o
n
v
i
a
e
l
e
c
t
r
o
m
e
c
h
a
n
i
-
c
a
l
c
a
t
h
e
t
e
r
P
h
a
s
e
I
9
0
d
a
y
s
S
i
g
n
i
f
i
c
a
n
t
r
e
d
u
c
t
i
o
n
i
n
a
n
g
i
n
a
a
n
d
N
T
G
u
s
e
i
m
p
r
o
v
e
d
S
P
E
C
T
s
c
a
n
V
a
l
e
2
0
0
1
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
S
y
m
p
t
o
m
a
t
i
c
C
A
D
,
n
o
t
a
m
e
n
a
b
l
e
t
o
r
e
v
a
s
c
u
l
a
r
i
z
a
t
i
o
n
N
a
k
e
d
p
l
a
s
m
i
d
V
E
G
F
1
6
5
I
n
t
r
a
m
y
o
c
a
r
d
i
a
l
i
n
j
e
c
t
i
o
n
v
i
a
e
l
e
c
t
r
o
m
e
c
h
a
n
i
-
c
a
l
c
a
t
h
e
t
e
r
P
h
a
s
e
I
/
I
I
m
u
l
t
i
c
e
n
t
e
r
9
0
d
a
y
s
S
i
g
n
i
f
i
c
a
n
t
r
e
d
u
c
t
i
o
n
i
n
a
n
g
i
n
a
a
n
d
N
T
G
u
s
e
I
m
p
r
o
v
e
d
S
P
E
C
T
s
c
a
n
R
o
s
e
n
g
a
r
t
1
9
9
9
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
S
y
m
p
t
o
m
a
t
i
c
C
A
D
,
n
o
t
a
m
e
n
a
b
l
e
t
o
r
e
v
a
s
c
u
l
a
r
i
z
a
t
i
o
n
A
d
e
n
o
v
i
r
a
l
V
E
G
F
1
6
5
I
n
t
r
a
m
y
o
c
a
r
d
i
a
l
i
n
j
e
c
t
i
o
n
a
t
t
h
o
r
a
c
o
t
o
m
y
P
h
a
s
e
I
3
0
d
I
n
c
r
e
a
s
e
d
R
e
n
t
r
o
p
s
c
o
r
e
D
e
c
r
e
a
s
e
i
n
a
n
g
i
n
a
R
o
s
e
n
g
a
r
t
1
9
9
9
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
S
y
m
p
t
o
m
a
t
i
c
C
A
D
,
n
o
t
a
m
e
n
a
b
l
e
t
o
r
e
v
a
s
c
u
l
a
r
i
z
a
t
i
o
n
A
d
e
n
o
v
i
r
a
l
V
E
G
F
1
6
5
V
A
T
S
i
n
t
r
a
m
y
o
-
c
a
r
d
i
a
l
i
n
j
e
c
t
i
o
n
P
h
a
s
e
I
3
0
d
N
o
t
r
e
p
o
r
t
e
d
G
r
i
n
e
s
/
A
G
E
N
T
2
0
0
1
M
y
o
c
a
r
d
i
a
l
i
s
c
h
e
m
i
a
C
l
a
s
s
I
I
o
r
I
I
I
s
t
a
b
l
e
a
n
g
i
n
a
,
>
1
v
e
s
s
e
l
o
p
e
n
w
i
t
h
o
t
h
e
r
t
r
e
a
t
m
e
n
t
o
p
t
i
o
n
s
A
d
e
n
o
v
i
r
a
l
F
G
F
-
4
I
n
t
r
a
c
o
r
o
n
a
r
y
P
h
a
s
e
I
/
I
I
1
2
w
k
s
N
o
n
-
s
i
g
n
i
f
i
c
a
n
t
i
n
c
r
e
a
s
e
i
n
e
x
e
r
c
i
s
e
t
o
l
e
r
a
n
c
e
of FGF-2 in limb ischemia is ongoing. Six
Phase I trials exploring the use of recom-
binant FGF for myocardial ischemia via in-
tramyocardial injection (in two trials), peri-
adventitial implantation (in one trial) and
intracoronary administration (in three
trials) have been concluded, and there are
currently two Phase II trials for intracoro-
nary recombinant FGF administration in
progress (Table 13.1). Two Phase I trials,
and one Phase II clinical trial have studied
intracoronary injection of recombinant
VEGF protein for myocardial ischemia (Ta-
ble 13.1). Whilst two studies have shown
significant improvement, the results in
one study (Henry et al.) revealed no signif-
icant difference in treadmill walking time
in both experimental and placebo groups
at 2 months of follow-up. Long-term fol-
low-up data are awaited. Adenoviral VEGF
delivery to the myocardium by intramyo-
cardial injection with or without concomi-
tant coronary artery bypass grafting
(CABG) has been successfully studied in
two Phase I clinical trials (Table 13.1) and
Schering AG is currently enrolling patients
for a Phase II trial in Europe. Finally, Gen-
zyme is enrolling patients for a Phase I
trial of naked plasmid HIF1a/VP16 intra-
muscular injection for lower-extremity
ischemia (Table 13.1).
Enthusiasm for the results from these
preliminary studies must be tempered by
their limitations, which include small sam-
ple size, lack of controls, open-label and
non-randomized design. In addition, be-
cause angiogenic therapy is sometimes ad-
ministered in conjunction with revasculari-
zation procedures, it is difficult to discern
the relative contributions of each. Safety
concerns surrounding the use of angio-
genic genes for humans have yet to be
adequately addressed. These include po-
tential formation of hemangiomas, retino-
pathy, edema, and tumor progression. He-
mangiomas have been noted in immuno-
compromised animals undergoing trans-
plantation of myoblasts retrovirally
engineered to express VEGF; in animals
receiving high doses of VEGF plasmid;
and in those receiving adeno-associated
virus vectors which express the gene for
upwards of 9 months. Hemangiomas have
also been noted in one human trial, but
they resolved spontaneously. Angiogenic
factors such as FGF and VEGF have been
implicated in the pathophysiology of prolif-
erative diabetic retinopathy. The same con-
dition has been noted in transgenic mice
engineered to over-express VEGF, and
after sub-retinal injection of adenoviral
VEGF. The effects of injection at remote
sites on proliferative retinopathy are un-
known. VEGF is known to augment vascu-
lar permeability, and life or limb-threaten-
ing edema has been noted in transgenic
mice. In humans, transient (but treatable)
peripheral edema has been documented
after lower-extremity intramuscular VEGF
injection. The concern for tumor progres-
sion is based on the studies of Folkman et
al. in identifying angiogenesis as a critical
stimulus for tumor growth. While there is
little evidence in preclinical trials that
would support the notion that administra-
tion of angiogenic growth factors to the
heart or limb stimulates the growth of tu-
mors, this issue clearly merits further
scrutiny [6].
In summary, the findings of early Phase
I and II clinical trials suggest subjective
and objective improvements in patients
with ischemic lower-extremity or myocar-
dial disease, and merit further investiga-
tion in additional Phase II, and pivotal
Phase III and IV clinical studies. Issues
that need to be addressed in these studies
include limitations and safety profiles,
definition of the populations that would
benefit from this therapy, determination of
whether angiogenic gene therapy is suit-
able as a primary or adjunctive therapeutic
modality, and whether its use is most ef-
fective early or late in the course of isch-
emic cardiovascular disease.
13.2.2
Gene Therapy for Bypass Graft Failure
and Arterial Re-stenosis
Atherosclerosis is the most common cause
of occlusive arterial disease and associated
tissue ischemia. The two most commonly
employed revascularization procedures in-
clude surgical bypass and percutaneous
transluminal intervention (PTI), involving
angioplasty with or without stenting. Due
to graft occlusion and arterial re-stenosis,
almost half of the CABGs fail after 10
years, while 20% of infra-inguinal bypass
grafts fail within 1 year, and one-third of
vessels treated by PTI are re-occluded with-
in 6 months. A common pathologic fea-
ture for both graft occlusion and re-steno-
sis is neointimal hyperplasia. Injury to the
graft or vessel wall initiates vascular
smooth muscle cell (VSMC) proliferation
and migration, inflammation, endothelial
cell dysfunction, and matrix expansion [7].
Therapeutic strategies designed to inhibit
neointimal hyperplasia attempt to restore
endothelial cell function, block cell cycle
progression, and prevent extracellular ma-
trix remodeling. Approaches include local
drug delivery, irradiation, the use of ribo-
zymes, transcription factor decoy and anti-
sense oligodeoxynucleotides (ODNs), and
gene transfer.
Cytotoxic strategies such as radioactive
stents designed to emit either a-particles
or c-radiation, while demonstrating im-
pressive short-term results, are susceptible
to late failure in as many as 50% of hu-
man interventions, particularly at the
stent-to-artery transition, resulting in the
so-called edge effect [7]. Gene transfer of
cytosine deaminase or thymidine kinase
plus gancyclovir (Table 12.2) [7] induces
the death of large numbers of cells, with
resultant inflammation that weakens the
vessel wall [7]. To date, this strategy has
not found clinical application.
On the other hand, cytostatic strategies
have yielded encouraging results in experi-
mental and early clinical studies (Table
13.3) [813]. Examples include coronary
Table 13.2 Cytotoxic strategies to reduce neointimal hyperplasia
Technologies/
route of delivery
Specific targets Animal model/human trial Author/
sponsor
Lesion inhibi-
tion [%]
Cytotoxic gene therapy
Direct delivery,
adenoviral
Cytosine deaminase Rat carotid artery Harrell et al. 45
Direct delivery,
adenoviral
Thymidine kinase:
Gancyclovir
Rabbit carotid artery Steg et al. 42
Direct delivery,
adenoviral
Fas ligand Rat carotid artery Luo et al. 5860
Radiation
b-particle Non-specific G
1
/S
blockade
Phase I, II human coronary
artery
Albiero et al. 6974
c-radiation Non-specific G
1
/S
blockade
Phase I, II human coronary
artery
Teirstein et al. 74
Table 13.3 Cytostatic strategies and gene transfer to reduce neoin-
timal hyperplasia in animal models and humans
Route of delivery and vectors Target genes/
pathways
Animal model/
human trial
Author
Pharmacologic
Paclitaxel (Taxol) Microtubules Phase I, II human
coronary artery
Heldman et al.
Rapamycin (Sirolimus) Non-specific G
1
/S
blockade via mTOR
protein
Phase I, II human
coronary artery
Sousa et al.
Ribozymes Cdk-1 PCNA Rat carotid artery Dev et al.
Antisense ODN c-myb Rat carotid artery Simons et al.
c-myc Phase I human coronary/
ITALICS trial
Kutryk et al.
Cdk-2 Rat carotid artery Morishita et al.
PCNA Rat carotid artery Simons et al.
Cyclin B Rat carotid artery Morishita et al.
Decoy ODN E2F Phase I, II, III
human GSV bypass grafts
Mann et al.
Gene transfer targeting G
1
/S phase
Catheter-based, adenoviral Rb (non-phosphorylatable) Porcine femoral artery Chang et al.
Direct delivery, adenoviral RB2/p130 Rat carotid artery Claudio et al.
Catheter-based, adenoviral p21 Rat carotid artery Chang et al.
Direct delivery, adenoviral p27 Rat carotid artery Chen et al.
Direct, plasmid with fusigenic
liposome
p53 Rabbit carotid artery Yonomitso et al.
Catheter-based, adenoviral GAX Rabbit carotid artery Maillard et al.
Direct, plasmid ras (transdominant nega-
tive)
Rat carotid artery Indolfi et al.
Direct delivery, adenoviral Fas ligand and p35 Rabbit femoral artery Luo et al.
Nitric oxide gene transfer
Catheter-based, plasmid with
fusigenic liposome
Bovine eNOS III Rat carotid artery von der Leyen
et al.
Direct delivery, cationic
liposome
Human iNOS II Porcine femoral artery
stent model
von der Leyen
et al.
Direct delivery, adenovirus Human eNOS III Pig coronary artery Vorenne et al.
Peri-adventitial delivery,
adenovirus
Bovine eNOS III Rat carotid and pig
coronary artery
Kullo et al.
Catheter-based, adenoviral Human iNOS II Pig coronary artery Theng et al.
Catheter-based, adenoviral Human iNOS II Rat carotid artery Shears et al.
Catheter-based, adenoviral Human eNOS III Rat carotid artery Janssens et al.
Catheter-based, adenoviral Human iNOS II Rat aorta allograft model Shears et al.
Intravascular seeding of
transduced SMCs
Human eNOS III Rat carotid artery Chen et al.
Catheter-based, plasmid with
fusigenic liposome
Human iNOS Phase I, human coronary
artery
REGENT-I
Gene transfer for rapid re-endothelialization
Direct, plasmid Human HGF Rat carotid artery Hayashi et al.
stents impregnated with the anti-prolifera-
tive drugs paclitaxel and rapamycin, and
E2F decoy ODN for bypass grafting. Pacli-
taxel alters the dynamic equilibrium be-
tween microtubules and a- and b-tubulin
by favoring the formation of abnormally
stable microtubules. This leads to the inhi-
bition of cell division and migration,
which relies on the rapid and efficient
depolymerization of microtubules. Two
Phase II prospective, randomized and dou-
ble-blinded (ELUTES and TAXUS) clinical
trials in patients undergoing PTI with
stenting reported encouraging results.
Phase III and IV studies are currently in
progress. Rapamycin acts by binding to a
cytosolic protein [8]. This complex binds
to a specific cell-cycle regulatory protein
(mTOR) and inhibits its activation, which
in turn induces cell-cycle arrest in late G
1
.
Preliminary results in humans are strik-
ing, with two Phase II trials using rapamy-
cin eluting stents after PTI (RAVEL and
Sousa et al.) reporting almost complete
suppression of in-stent stenosis, and a
97% event-free survival rate at 9 months
of follow-up (Table 13.3). Phase I trials
with stents designed to elute actinomy-
cin-D and estradiol are currently ongoing
[8].
A particularly promising approach is the
use of decoy ODNs to the E2F family of
transcription factors that regulate cell-cycle
progression at the G
1
/S checkpoint. When
molar excesses of double-stranded DNA
bearing the consensus binding sequence
for E2F-1 were delivered into the blood
vessel, they specifically bound and seques-
tered the target transcription factor, render-
ing it incapable of binding to the promoter
region of specific cell-cycle regulatory
genes, thereby inhibiting their expression
and blocking the cells from progressing
beyond the G
1
/S checkpoint. Using a nov-
el pressure-mediated transfection system,
we have demonstrated that the E2F decoy
inhibited neointimal hyperplasia in bal-
loon-injured rat carotid arteries and rabbit
carotid artery interposed vein grafts. These
studies led to a human Phase IIB study
the PREVENT-I trial that studied a cohort
of patients at high risk for lower-extremity
graft failure in prospective, double-blinded,
randomized and controlled fashion. Intra-
operative ex vivo pressure-mediated trans-
fection of the saphenous vein graft with
E2F decoy was performed prior to transpo-
sition into the arterial position. The pri-
mary end-points were safety, feasibility
and biological. Indeed, the procedure was
Route of delivery and vectors Target genes/
pathways
Animal model/
human trial
Author
Anti-thrombogenic gene transfer
Intravenous, adenoviral Human kallikrein Mouse carotid artery Emanueli et al.,
Murakami et al.
Catheter-based, adenoviral Human plasminogen
activator inhibitor type 1
Rat carotid artery DeYoung et al.
Direct delivery, adenoviral uPA-BPTI Rat carotid artery Lamfers et al.
Gene transfer to prevent matrix remodeling
Catheter-based, adenoviral Human TGF-beta-1 Porcine coronary artery Kingston et al.
Direct delivery, adenoviral Human tissue inhibitor
of metalloproteinase-1
Ex-vivo human
saphenous vein grafts
George et al.
safe and feasible. Furthermore, E2F decoy
inhibited the target cell cycle gene (c-myc
and proliferating cell nuclear antigen;
PCNA) expression, and consequently sup-
pressed cellular proliferation in the treated
human grafts. At 1-year follow-up, there
was a statistically significant 50% reduc-
tion in graft failure in the treated group as
compared to the control group. A second
Phase II-B clinical trial studied CABG
grafts in a randomized, blinded and con-
trolled manner. E2F decoy resulted in sig-
nificant (3040%) decreases in CABG graft
failure associated with a reduction in
three-dimensional neointimal volume as
assessed by intracoronary ultrasound. Giv-
en these early successes, two Phase III
clinical trials evaluating the efficacy of this
genetically modified graft in CABG and
peripheral arterial bypass are now under-
way (Table 13.3) [9]. Although several anti-
sense ODNs to PCNA, c-myb, c-myc, non-
muscle myosin heavy chain and cdc2-ki-
nase have been reported to be effective in
reducing experimental re-stenosis and
graft failure [9], clinical results in human
have generally been unimpressive. For ex-
ample, a recent study of intracoronary lo-
cal delivery of c-myc antisense ODN after
human coronary stenting yielded disap-
pointing results (Table 13.3) [9].
Local gene therapy is also progressing
towards human application in PTI. In ex-
perimental models, nitric oxide synthase
(NOS) gene transfer has been shown to
provide therapeutic benefit for balloon-in-
jured vessels or bypass grafts by inducing
vasorelaxation, by exerting cytoprotective
and anti-inflammatory effects, and by in-
hibiting VSMC proliferation by G
1
/S
phase blockade [10] via tyrosine phosphor-
ylation of paxillin and focal adhesion ki-
nase [11]. Together, these actions result in
approximately 70% reduction in neointi-
mal hyperplasia after direct or catheter-
based delivery of the endothelial NOS
(eNOS) gene to the balloon-injured arterial
wall (Table 13.3) [10]. Inducible NOS
(iNOS) is a powerful feedback regulator of
vascular inflammation. It reduces mono-
cyte and platelet adhesion, aggregation,
and activation, similar to the actions of
anti-thrombotic genes kallikrein, and hu-
man plasminogen activator inhibitor type
1. iNOS also protects endothelial cells
from superoxide radical and lipopolysac-
charide-induced apoptosis. iNOS gene
transfer has also resulted in a significant
reduction in neointimal hyperplasia in ex-
perimental models of PTI (Table 13.3) [12].
NOS gene transfer has a clear advantage
over NO adducts or NO donors, in that it
achieves high concentrations of NO locally
in the target tissue, without the potential
adverse effects of excessively high systemic
levels. Both eNOS and iNOS appear to be
effective and safe, and iNOS gene transfer
is not associated with the cytotoxicity nor-
mally observed with activation of the en-
dogenous gene (Table 13.3) [10, 13]. The
success of NOS gene therapy in animal
models constitutes the basis of the RE-
GENT-I clinical trial on coronary re-steno-
sis. This is a Phase I safety, feasibility and
dose-finding trial, which has completed pa-
tient recruitment to investigate catheter-
based administration of human iNOS
(using the infiltrator catheter plus lipoplex
delivery system) to prevent re-stenosis of
coronary arteries treated by PTI. In order
for arterial gene therapy to be successfully
and routinely deployed in humans, it must
overcome several hurdles, including safety
and efficacy of delivering the vector into
the vessel wall. Furthermore, it is unclear
whether a transient surge of therapeutic
gene expression of sufficient duration is
adequate to inhibit the waves of mitogen
activation and VSMC proliferation; or
whether stable long-term expression with
chromosomal integration of the transgene
is needed. These issues have to be ad-
dressed in more detailed experimental
studies and in larger clinical trials.
13.2.3
Gene Therapy for Myocardial Protection
An unmet need in cardiovascular therapy
is effective myocardial protection and pre-
servation during ischemia and/or reperfu-
sion. When myocardial oxygen demand ex-
ceeds oxygen supply, cardiomyocytes are
deprived of oxygen and other nutrients.
After cessation of blood flow, concentra-
tions of high-energy phosphate com-
pounds such as ATP and creatine phos-
phate fall. As the cells shift from oxidative
metabolism to anaerobic glycolysis, the in-
tracellular pH falls, impairing the function
of membranous energy-dependent ion
pumps, and impairing the contractile
forces generated by actinmyosin cross-
bridge formation. Without timely removal
of the ischemic insult, cytosolic levels of
calcium, free fatty acids, modified lipids
and phospholipid intermediates increase,
which, in turn, compromises the integrity
of cellular membranes, resulting in cell
death. The injurious processes initiated by
coronary ischemia may, paradoxically, be
exacerbated by reperfusion a phenome-
non known as ischemiareperfusion (I/R)
injury. Reperfusion of the ischemic myo-
cardium results in the formation of free
radical reactive oxygen species (ROS),
which damage proteins and membrane
structures, and can activate signal trans-
duction pathways that lead to apoptosis.
Leukocytes that adhere to injured endothe-
lial cells release inflammatory mediators,
which in turn, worsen myocyte and endo-
thelial cell injury. The accumulation of
ROS during reperfusion depletes the buf-
fering capabilities of endogenous anti-oxi-
dant reserves, thereby exacerbating the
deleterious effects of ROS. With time,
repeated I/R injury leads to progressive
impairment of contractile function, culmi-
nating in hemodynamic failure [14].
An understanding of the mechanisms of
the I/R cascade has made myocardial pro-
tection against I/R-induced injury an op-
portunity for genetic therapeutics. It has
been hypothesized that an increase in pro-
oxidant scavenging activity imparted by
constitutive over-expression of anti-oxidant
enzymes such as hemo-oxygenase-1 (HO-
1) or extracellular superoxide dismutase
(ecSOD) can confer cytoprotection against
future I/R episodes. Indeed, our group has
used adeno-associated viral vectors (AAV)
to deliver HO-1 and ecSOD by intramyo-
cardial injection several months prior to
induction of I/R by ligation and release of
the left anterior descending (LAD) coro-
nary artery. AAV delivery resulted in long-
term expression of HO-1 and ecSOD in
the myocardium, yielding near-complete
prevention of myocardial infarction from
I/R, and thereby providing proof of con-
cept of a preventive gene therapy strategy
for long-term myocardial protection
against future repeated episodes of isch-
emia [15]. Other groups have had similar
success by over-expressing a repertoire of
genes that are induced by oxidative stress,
such as heat shock protein 70, anti-
apoptotic gene Bcl-2, protein kinase B
(Akt) and other immunosuppressive cyto-
kines. However, these therapies were ad-
ministered during the acute event of I/R
using adenoviral vectors that provided only
short-term transgene expression. Similar
results have been achieved using a ODN
strategy to block the pro-inflammatory
transcription factor nuclear factor kappa B
(NF-jB); and an antisense-ODN strategy
directed at angiotensin-converting enzyme
(ACE) mRNA in myocardial ischemic in-
jury has been shown to ameliorate myocar-
dial dysfunction and injury after I/R [16].
Taken together, gene therapy holds pro-
mise for acute and long-term myocardial
protection, and may eventually be used in
the management of human coronary ar-
tery disease.
13.2.4
Gene Therapy for Myocardial Failure
When prevention proves inadequate, gene
therapy can also be used to rescue contrac-
tile function in the failing myocardium.
The failing heart is characterized by altera-
tions in calcium handling, decreased myo-
filament sensitivity and adrenergic recep-
tor down-regulation and desensitization.
Specifically, cardiac b
1
-adrenergic receptors
(b
1
ARs) mediate the myocardial contractile
response to sympathetic stimulation.
b
1
ARs are coupled to the stimulatory gua-
nine nucleotide binding protein G
s
. Stimu-
lation of the b
1
AR by agonists results in
dissociation of the G
a
subunit from the
G
2a
subunit. The stimulatory subunit G
s `
binds to, and activates adenylate cyclase,
causing production of cyclic adenosine
monophosphate (cAMP) and activation of
protein kinase A (PKA). In the myocar-
dium, PKA phosphorylates and activates L-
type Ca
2+
channels, the sarcoplasmic retic-
ulum Ca
2+
ATPase (SERCA2a) inhibitor
phospholamban, and myofibrillar protein
troponin I. The net result is an increase in
cytosolic Ca
2+
transience, and an increase
in cardiac contractility. Accordingly, strate-
gies for gene transfer to increase cardiac
contractility have included adenoviral deliv-
ery of SERCA2a, and cardiac overexpres-
sion of adenylate cyclase, both of which
have been shown to increase contractility
in aortic-banded rats and cardiomyopathic
mice, respectively. Conversely, antisense
ODN inhibition of phospholamban has
achieved similar results in isolated rat ven-
tricle myocytes [17].
A negative feedback loop which results
in desensitization and down-regulation of
activated b
1
ARs is mediated by bAR Ki-
nase-1 (bARK-1). This kinase has affinity
for the membrane-bound G
2a
subunits of
the activated b
1
AR, is activated by binding
this subunit, and uncouples the b
1
AR by
phosphorylation. bARK-1 is itself negative-
ly regulated by the peptide bARK
ct
, which
inhibits bARK-1 activity by competitively
binding the G
2a
subunit. Congestive heart
failure (CHF) is known to impair this sig-
naling and regulatory mechanism. A 50%
reduction in the number of b
1
ARs, with a
concomitant increase in bARK-1 level and
activity results in decrease of the basal and
b-agonist-stimulated contractility in pa-
tients with CHF. b
1
ARs are not down-
regulated, and their sensitivity is thought
to be increased. This class of bARs is
thought to function independently of the
cAMP-PKA axis. It was therefore hypothe-
sized that myocardial over-expression of
the b
1
AR would improve cardiac function.
This hypothesis has subsequently been va-
lidated in a transgenic model of b
1
AR
over-expression, as well as by adenoviral
delivery of the b
1
AR to rabbit myocardium.
The dissection of this pathway has pro-
vided additional targets for gene therapy to
increase cardiac contractility such as block-
ade of bARK-1 via adenoviral over-expres-
sion of bARK
ct
which resulted in marked
reversal of ischemia-induced left ventricu-
lar dysfunction in an experimental model
[17]. Given the increasing prevalence of
CHF and the limited repertoire of thera-
peutic options (including devices and
transplantation), human gene therapy may
have a useful place in the future treatment
of this disorder.
13.3
Cell-based Gene Therapy and Regenerative
Cardiovascular Medicine
Recent discoveries of nests of replicating
and self-renewing cells that have the ability
to differentiate into highly specialized and
functional post-mitotic cells (such as neu-
rons and cardiomyocytes) has introduced
yet another paradigm shift in the lexicon
of genetic therapeutics. The emerging field
of regenerative medicine investigates the
possibilities of transplanting regeneration
competent cells into injured or damaged
organs as a means of tissue regeneration
and repair. The appeal of such an
approach is further heightened by the dis-
covery of autologous stem cells in adult
animal and human tissue. When com-
bined with customized genetic manipula-
tion of these cells prior to transplantation,
and without the need for immunosuppres-
sion, such a strategy has the potential to
revolutionize clinical medicine. This excit-
ing field is finding application in cardio-
vascular therapeutics in the areas of vascu-
logenesis, angiogenesis and myogenesis.
13.3.1
Endothelial Progenitor Cells
for Vascular Re-endothelialization
Adult bone marrow contains endothelial
progenitor cells (EPCs) that are derived
from hemangioblasts. EPCs can be mobi-
lized from the bone marrow in response
to systemic administration of angiogenic
growth factors and cytokines, making it
possible to isolate EPCs from the mono-
nuclear fraction of rat, rabbit, and human
peripheral blood. Embryonic stem cells
can also be manipulated to differentiate
into functional endothelial cells [18]. It has
been hypothesized that one inciting event
in the pathogenesis of neointimal hyper-
plasia endothelial loss or dysfunction
can be arrested by therapeutic re-endothe-
lialization using EPCs engineered to over-
express cytoprotective genes eNOS and/or
HO-1. Our group has demonstrated that
EPCs re-endothelialize the denuded arteri-
al bed with 85% coverage at day 7 and 70
80% at day 14 [18]. In addition, Kaushal et
al. have shown that EPCs can be seeded
onto decellularized porcine iliac artery
grafts, which then gain the ability to pro-
duce NO, and to relax and contract both in
vitro and in vivo [18]. If vessels in treated
this way reduce graft stenosis and throm-
bosis, they may find application in clinical
settings.
13.3.2
Endothelial Progenitor Cells and Angioblasts
for Angiogenesis
Human EPCs are mobilized from the bone
marrow during ischemic episodes, and are
thought to participate in angiogenesis (the
proliferation of pre-existing vasculature)
and vasculogenesis (the formation of new
blood vessels) in the ischemic areas. While
endogenous reserves may not always pro-
vide enough EPCs to rescue blood supply
in an ischemic area and prevent loss of tis-
sue, augmenting local levels by local injec-
tion prove to be a successful therapeutic
strategy. Systemic intravenous injection of
human angioblasts into nude athymic rat
hearts resulted in an increase in capillary
density which, in turn, protected against
cardiomyocyte apoptosis, decreased local
collagen deposition, and improved cardiac
function [19]. In separate experimental
studies, EPCs engineered to over-express
VEGF demonstrated improved proliferative
and adhesive capabilities, both in vitro and
in vivo. Animals treated with these VEGF-
EPCs demonstrated improved blood flow
and less limb loss when compared to ani-
13.3 Cell-based Gene Therapy and Regenerative Cardiovascular Medicine 319
mals treated with Lac-Z-transduced EPCs
[19].
13.3.3
Myocardial Regeneration using Stem Cells
The heart, when damaged by ischemic in-
jury, compensates for the loss of functional
tissue by undergoing remodeling. This in-
volves the replacement of infarcted tissue
by fibrous scar and compensatory hyper-
trophy of surviving myocytes. Even though
cardiomyocytes may be capable of divid-
ing, their replicative potential is limited,
and is overwhelmed by rapidly proliferat-
ing cardiac fibroblasts. Cardiomyocytes are
unable to more than double in size before
they succumb to eventual exhaustion [20].
Increased wall stress exerted upon the thin
and non-functional scar results in patho-
logic alteration of ventricular geometry
and perpetuates progressive loss of cardio-
myocytes, eventually leading to cardiac fail-
ure. It has been hypothesized that repopu-
lating the damaged zone with contractile
cells coupled with appropriate matrix mod-
ulation may normalize the hemodynamic
load on the surviving cardiomyocytes,
thereby avoiding the deleterious conse-
quences of ventricular remodeling.
Skeletal muscle is able to repair itself
after injury because of the presence of res-
ident satellite cells (myoblasts) that pro-
liferate in response to injury, and fuse
with damaged muscle fibers to regenerate
functional skeletal muscle. The injection
of skeletal myoblasts has been reported to
improve myocardial stroke work, end-dia-
stolic segment length, contractile function,
and diastolic relaxation. The first report of
skeletal myoblast transplantation for hu-
man heart failure required the injection of
810
8
cells into the myocardium of a sin-
gle patient with revascularizable New York
Heart Association class III heart failure at
the time of CABG. After 5 months, the
grafted area was viable and contractile.
Since this initial report, an additional four
patients have undergone myoblast trans-
plantation in open-label, uncontrolled fash-
ion, with an average 13% increase in ejec-
tion fraction.
As a prerequisite for fine motor control,
skeletal muscle fibers are electrically iso-
lated from one another, and, accordingly,
do not express either connexin-43 (the
major gap junction protein) or N-cadherin
(the major adherens protein in cardiac in-
tercalated discs). Asynchronous islands of
intramyocardial skeletal muscle can result
in lethal arrhythmias in mice [21].
Although cell types such as fetal cardio-
myocytes [22] and cardiomyocytes derived
from murine or human embryonic stem
cells [23] are capable of electromechanical
coupling, their clinical use has unfortu-
nately been hampered by technical, ethical,
moral, social and legal hurdles.
Over the past three years, several groups
have reported the existence of cardiac myo-
cyte precursor cells in the bone marrow of
adult animals. The potential of harnessing
this population for an autologous thera-
peutic strategy involving cardiac regenera-
tion has great appeal. When bone marrow
cells are treated with 5-azacytidine in vitro,
spontaneously and synchronously beating
cells with phenotypic characteristics of dif-
ferentiated cardiac myocytes have been re-
ported to develop. When transplanted into
the heart, these cells augment ventricular
function [24]. Anversa et al. have demon-
strated that a pool of human cardiomyo-
cytes in the myocardium is capable of en-
tering the cell cycle, and of undergoing cy-
tokinesis. Evidence of cardiomyocyte prolif-
eration was provided by staining human
heart sections for Ki67, an essential ele-
ment of the outer dense fibrillar compo-
nent of the nucleolus, where it facilitates
the rapid production of ribosomes for the
increased metabolic requirements of cells
that are actively dividing, and accordingly
is expressed in all phases of the cell cycle
except G
0
. These investigators also demon-
strated that shortly after myocardial infarc-
tion, the mitotic index of cardiomyocytes
in the border zone increased almost three-
fold. In a separate set of studies, this
group isolated a population of resident
primitive undifferentiated c-kit
+
cells from
the hearts of senescent rats. These cells
were capable of proliferating in culture
conditions in an undifferentiated state,
and differentiated into cardiomyocytes
when transplanted into the infarcted rat
heart. The origin of these cells is unclear,
but one possibility is that they originate
from the bone marrow. Arguments sup-
porting this hypothesis include a report
from Jackson et al., who have demon-
strated that the poorly characterized SP
population (putative hematopoietic pro-
genitors isolated using the Hoechst epi-flu-
orescence technique) is capable of partici-
pating in cardiac regeneration. In addition,
Orlic et al. demonstrated that c-kit
+
cells
could be isolated from the bone marrow
using flow cytometry, and participated in
cardiac regeneration when injected into
the ischemic murine heart. Furthermore,
this group demonstrated that these cells
are mobilized from the bone marrow after
systemic administration of granulocyte col-
ony stimulating factor (GCSF) and stem
cell factor and home to the myocardium,
where they induce myocardial repair after
infarction, and reduce mortality from in-
farction. The bone marrow origin of this
cell, and its ability to migrate into the
heart was verified by the demonstration of
Y-chromosome-labeled cardiomyocytes and
resident c-kit
+
cells in hearts transplanted
from female donors to male recipients
[25]. Our group has characterized a highly
purified population of mesenchymal stem
cells harvested from the bone marrow of
adult animals, that is easily expandable
and scalable, is amenable to ex vivo genetic
manipulation (to increase cell survival, for
example), and induces recovery of cardiac
function after myocardial infarction by dif-
ferentiating into cardiomyocytes in vivo
[26].
Several unknown issues need to be in-
vestigated before human cardiac stem cell
transplantation can be safely started. These
questions include: What percentage of
transplanted stem cells are viable after
transplantation? What is the proliferative
and regenerative capacity of the trans-
planted cells? What are the local signals
and mediators of homing, trafficking, pro-
liferation and differentiation of these cells?
What is the optimal timing of transplanta-
tion that is, is it better to transplant dur-
ing the acute ischemic event, or several
weeks thereafter? What are the relative
contributions of CD34
+
and CD34
cells to
angiogenesis and myogenesis? Do angio-
genesis and myogenesis complement one
another? What are the most appropriate
criteria to assess myocardial function after
transplantation? Do regenerated cardio-
myocytes induce primarily systolic or dia-
stolic improvement, or both?
13.4
Future Directions and Challenges
The adaptability of the powerful technique
of genetic therapeutics has allowed re-
searchers to pursue lines of investigation
not thought possible as recently as a de-
cade ago. Genetic manipulation and nucle-
ar transfer cloning have been combined to
generate transgenic swine that are immu-
nocompatible with humans, potentially
making xenotransplantation a possibility.
13.4 Future Directions and Challenges 321
Cell-based therapy is being combined with
developments in artificial organs to gener-
ate artificial heart valves which can, in the-
ory, be lined with cells harvested from the
patients own body and engineered to ex-
press anti-thrombotic compounds. Genes
implicated in lethal congenital cardiac mal-
formations are being identified, and rapid
developments in in utero gene therapy
promise to cure these defects before they
can exert a deleterious effect [26].
These astonishing strides demand great
responsibility on the part of scientists and
physicians. Many issues will need to be ad-
dressed as genetic therapeutics finds it
way into the clinics: Who should be
treated? How much better is genetic ma-
nipulation than tried and tested tradi-
tional therapies? What are the risks and
benefits, and how does the physician use
this information to make clinically relevant
decision? In this age of healthcare cost
constraints, who will pay for this therapy?
Will patients want genetic information re-
leased to insurance companies? Can gene
therapy for non-lethal conditions be ratio-
nalized? Can gene therapy as a means of
prevention, or as a means toward en-
hanced health be justified? Will society
permit manipulation of the germ line for
human therapeutic cloning (see also Part
I, Chapter 11). These important issues
must be satisfactorily addressed as cell-
and gene-based therapies are being intro-
duced for human therapy. Innovative ad-
vances in basic science have allowed the
rapid translation of genetic information to
manipulation for clinical therapy, espe-
cially in cardiovascular medicine, making
results that were once thought unachieva-
ble within the realm of possibilities.
Acknowledgments
These studies were supported by Grants
HL 35610, HL 58516, HL 59316 and HL
54527 from the National Heart, Lung and
Blood Institute. A.A.M. is the recipient of
a National Research Service Award (1 F32
NHL 10503-01) from the National Insti-
tutes of Health, Bethesda, MD; and the
Linton Research Fellowship from the De-
partment of Surgery, Massachusetts Gener-
al Hospital, Boston, MA. V.J.D. is the re-
cipient of a National Heart, Lung and
Blood Institute MERIT Award.
References and Notes
1 D. W. Leung, et al., Science, 246, 1306 (1989);
P. J. Keck, et al., Science, 246, 1309 (1989); S.
Takeshita, et al., Circulation, 98, 1261 (1998);
Alfranca, A., et al., Mol Cell Biol, 22,12
(2002); Hattori, K., et al., J Exp Med, 193,
1005 (2001); Altavilla, D., et al., Diabetes, 50,
667 (2001); Schweigerer, L., et al., Nature, 325,
257 (1987); Lee, S. H., et al., N Engl J Med,
342, 626 (2000); Namiki, A., et al., J Biol
Chem, 270, 31189 (1995).
2 Taniyama, Y., et al., Circulation, 104, 2344
(2001); Baffour, R., et al., J Vasc Surg, 16, 181
(1992); Pu, L. Q., et al., Circulation, 88, 208
(1993); Taniyama, Y., et al., Gene Ther, 8, 181
(2001); Takeshita, S., et al., Biochem Biophys
Res Commun, 227, 628 (1996); Takeshita, S.,
et al., Circulation, 90, II228 (1994); Tsurumi,
Y., et al., Circulation, 96, II-382 (1997); Vin-
cent, K. A., et al., Circulation, 102, 2255
(2000).
3 Ohara, N., et al., Gene Ther, 8, 837 (2001);
Gowdak, L. H., et al., Circulation, 102, 565
(2000); Mack, C. A., et al., J Vasc Surg, 27, 699
(1998).
4 Aoki, M., et al., Gene Ther, 7, 417 (2000); Ya-
nagisawa-Miwa, A., et al., Science, 257, 1401
(1992); Unger, E. F., et al., Am J Physiol, 266,
H1588 (1994); Harada, K., et al., J. Clin In-
vest, 94, 632 (1994); Mack, C. A., et al., J Thor-
ac Cardiovasc Surg, 115, 168 (1998); Lee, L. Y.,
et al., Ann Thorac Surg, 69, 14 (2000); Safi J,
Jr., et al., Microvasc Res, 58, 238 (1999); U.
Sayeed-Shah, et al., J Thorac Cardiovasc Surg,
116, 763 (1998).
5 Baumgartner, I., et al., Circulation, 97, 1114
(1998); Isner, J. M., et al., J. Vasc Surg, 28, 964
(1998); Isner, J. M., et al., Lancet, 348, 370
(1996); Lazarous D. F., et al., J Am Coll Cardi-
ol, 36, 1239 (2000); Lederman, R. J., et al., Am
J Cardiol, 88, 192 (2001); Schumacher, B., et
al., Circulation, 97, 645 (1998); Schumacher,
B., et al., J Cardiovasc Surg (Torino), 39, 739
(1998); Stegmann, T. J., et al., Cardiac and Vas-
cular Regeneration, 1, 5 (2000); Sellke, F. W.,
et al., Ann Thorac Surg, 65, 1540 (1998); La-
ham, R. J., et al., J Am Coll Cardiol, 35, 73A
(2000); Laham, R. J., et al., J. Am Coll Cardiol,
36, 2132 (2000); Under, E. F., et al., Am J Car-
diol, 85, 1414 (2000); Kleiman, N. S., et al., J
Am Coll Cardiol, 36, 310 (2000); Henry, T. D.,
et al., Am Heart J, 142, 872 (2001); Henry,
T. D., et al. Circulation, 102, II-309 (2000);
Hendel, R. C., et al., Circulation, 101, 118
(2000); Vale, P. R., et al., Circulation, 100,
I-477 (1999); Symes, J. F., et al., Ann Thorac
Surg, 68, 830 (1999); Vale, P. R., et al., Circula-
tion, 102, II-563 (2000); Vale, P. R., et al., Cir-
culation, 102, II-563 (2001); Rosengart, T. K.,
et al., Circulation, 100, 468 (1999); Rosengart,
T. K., et al., Circulation, 100, I-770; Grines,
C. L., Late breaking clinical trial, American
College of Cardiology, Mar 1821, 2001.
www.acc.org/media/session_info/late/
ACC2001/index.htm.
6 Springer, M. L., et al., Mol Cell, 2, 549 (1998);
Schwarz, E. R., et al., J Am Cell Cardiol, 35,
1323 (2000); Frank, R. N., et al., Am J
Ophthalmol, 122, 393 (1996); Okamoto, N.,
et al., Am J Pathol, 151, 281 (1997); Baffi, J.,
et al., Invest Ophthalmol Vis Sci, 41, 3582
(2000); Thurstaon, G., et al., Science, 286,
2511 (1999); Folkman, J., N Engl J Med, 285,
1182 (1971); Su, H., et al., Proc Natl Acad Sci
USA, 97, 13801 (2000); Su, H., et al., Proc
Natl Acad Sci USA, 97, 13801 (2000).
7 M. Grossman, et al., Nat Med, 1, 1148 (1995);
G. H. Gibbons and V. J. Dzau, Science, 272,
689 (1996); Albiero, R., et al., Circulation, 101,
18 (2000); Teirstein, P.S., et al., N Engl J Med,
336, 1697 (1997); van Der Giessen, W. J., et al.,
Circulation, 104, 2236 (2001); Luo, Z., et al.,
Circulation 99, 1776 (1999).
8 Heldman, A. W., et al., Circulation, 103, 2289
(2001); Sousa, J. E., et al., Circulation, 104,
2007 (2001); Kay, I. P., et al., Circulation, 102,
1434 (2000); Harrell, R.L., et al., Circulation,
96, 621 (1997); Steg, P.G., et al., Circulation,
96, 408 (1997); Dev, V., et al., Circulation, 92,
I-634 (1995); Shy, Y., et al., Circulation, 90,
944 (1994); Gregory, C. R., et al., Transplant
Res, 25, 770 (1993); Poon, M., et al., J Clin In-
vest, 98, 2277 (1996).
9 Morishita, R., et al., J Clin Invest, 93, 1458
(1993); Morishita, R., et al., Gene, 149, 13
(1994); Mann, M.J., et al., Lancet, 354, 1493
(1999); Grube, E., Circulation, 104, suppl II,
A2730; Simons, M., et al., J Clin Invest, 93,
2351 (1994); Kutryk, M. J., et al., Am J Coll
Cardiol, 39, 281 (2002); Simons, M., et al.,
Nature, 359, 67 (1992); Morishita, R., et al.,
Arterioscler Thromb Vasc Biol, 20, 915 (2000);
Morishita, R., et al., Proc Natl Acad Sci USA,
90, 8474 (1993).
10 Chang, M.W., et al., Science, 267, 518 (1995);
Claudio, P.P., et al., Circ Res, 85, 1032 (1999);
Chang, M.W., et al., J Clin Invest, 96, 2260
(1995); Chen, D., et al., J Clin Invest, 99, 2334
(1997); Yonomitso, Y., et al., Circ Res, 82, 147
(1998); Maillard, L., et al., Circ Res, 94, I-637
(1996); Indolfi, C., et al., Nat Med, 1, 541
(1996); Luo, Z., et al., Hum Gene Ther, 12,
2191 (2001); von der Leyen, H.E., et al., Proc
Natl Acad Sci USA, 92, 1137 (1995); von der
Leyen, H.R., et al., J Am Coll Cardiol, 35,
281A (2000); Vorenne, O., et al., Circulation,
98, 2254 (1998); Kullo, I.J., et al., Circulation,
96, 2254 (1997); Tzeng, E., et al., Mol Med, 2,
211 (1996); Shears, L.L., et al., J Am Coll
Surg, 187, 295 (1995); Janssens, S., et al., Cir-
culation, 97, 1274 (1998); Shears, L.L., et al., J
Clin Invest, 100, 2035 (1997); Chen, L., et al.,
Circ Res, 82, 862 (1998); Hayashi, K. et al.,
Gene Ther, 7, 1664 (2000); Emanueli, C., et
al., Arterioscler Thromb Vasc Biol, 20, 2500
(2000); DeYoung, M.B., et al., Circulation,
104, 1972 (2001); Lamfers, M.L., et al., Gene
Ther, 8, 534 (2001); Kingston, P.A., et al., Cir-
culation, 104, 2595 (2001); George, S.J., et al.,
Hum Gene Ther, 9, 867 (1998); Murakami,
H., et al., Hypertension, 34, 164 (1999).
11 Fang, S., et al., Biochem Biophys Res Com-
mun, 236, 706 (1997); von der Leyen, H. E., et
al., Circulation, 103, 2760 (2001).
12 Peng, H. B., et al., J Immunol, 161, 1970
(1998); Tzeng, E., et al., Surgery, 122, 255
(1997); Tsao, P., et al., Circulation, 96, 934
(1997).
References and Notes 323
13 Nathan C., J Clin Invest, 199, 2417 (1997); De
Meyer, G. R. Y., et al., Arterioscler Thromb
Vasc Biol, 20, 1896 (2000); Tzeng, E., et al.,
Proc Natl Acad Sci USA, 92, 11771 (1995).
14 D. L. Carden and D. N. Granger, Am J Pathol,
190, 255 (2000); T. L. Vanden Hoek, et al., Am
J Physiol, 270, H1334 (1996); P. K. Singal, et
al., Cardiovasc Res, 40, 426 (1998); Williams,
R. S., et al., J Clin Invest, 106, 813 (2000).
15 Melo, L. G., et al., Circulation, 105, 602 (2002);
Agrawal, R. S., et al., Mol Ther, 3, A782
(2001); Li, Q., et al., Circulation, 103, 1893
(2001).
16 N. Li, et al., FASEB J, 13, 1137 (1999); R. Mor-
ishita, et al., Nat Med, 3, 894 (1997); K. Suzu-
ki, et al., J Clin Invest, 99, 1645 (1997); S.
Okudu, et al., Circulation, 103, 877 (2001); Z.
Chen, et al., Am J Physiol, 280, H2313 (2001);
W. Miao, et al., J Mol Cell Cardiol, 32, 2397
(2000); L. Qin, et al., J Immunol, 156, 2316
(1996); Chen, H., et al., Gene Ther, 8, 804
(2001).
17 M. I. Miyamoto, et al., Proc Natl Acad Sci
USA, 97, 793 (2000); K. Eizema, et al., Circu-
lation, 101, 2193 (2000); J. P. Maurice, et al., J
Clin Invest, 104, 21 (1999); D. M. Roth, et al.,
Circulation, 99, 3099 (1999); England, P. J., et
al., Biochem J, 246, 687 (1987); Freedman,
L. J., et al., J Biol Chem, 270, 17953 (1995);
Pitcher, J. A., et al., J Biol Chem, 270, 11707
(1995); Pitcher, J. A., et al., Science, 257, 1264
(1992); Xiao, R.-P., et al., J Biol Chem, 269,
19151 (1994); Altschuld, R. A., et al., Circula-
tion, 92, 1612 (1995); Ungerer, M., et al., Cir-
culation, 87, 454 (1993); Milano, C. A., et al.,
Science, 264, 582 (1994); Cross, H. R., et al.,
Circ Res, 85, 1077 (1999); Akhter, S., et al.,
Proc Natl Acad Sci USA, 94, 12100 (1997);
Shah, A. S., et al., Circulation, 103, 1311
(2001); Tevaearai, H. T., et al., Circulation, 104,
2069 (2001).
18 T. Asahara, et al., Science, 275, 964 (1997); G.
Balconi, et al., Arterioscler Thromb Vasc Biol,
20, 1443 (2000); Kaushal, S., et al., Nat Med,
7, 1035 (2001); Reyes, M., et al., J Clin Invest,
109, 337 (2002).
19 Kawamoto, A., et al., Circulation, 103, 634
(2001); C. Kalka, et al., Proc Natl Acad Sci
USA, 97, 3422 (2000); A. A. Kocher, et al., Nat
Med, 7, 412 (2001); S. Shintani, et al., Circula-
tion, 103, 2776 (2001); Iwaguro, H., et al., Cir-
culation, 105, 732 (2002).
20 J. Kajstura, et al., Proc Natl Acad Sci USA, 95,
8801 (1998); A. Beltrami, et al., N Engl J Med,
334, 1750 (2001); P. Anversa and J. Kajstura,
Circ Res, 83, 1 (1998).
21 R. C.-J. Chiu, A. Zibaitis, R. L., Ann Thorac
Surg, 60, 12 (1995); G. Ferrari, et al., Science,
279, 1528 (1998); T. A. Rando, H. M. Blau, J
Clin Biol, 125, 1275 (1994); D. A. Taylor, et al.,
Nat Med, 4, 929 (1998); C. E. Murry, et al.,
J Clin Invest, 98, 2512 (1996); B. Z. Atkins, et
al., J Surg Res, 85, 234 (1999); Menasch, P.,
et al., Lancet, 357, 279 (2001); Menasch, P.,
et al., Circulation, 104, suppl II, A2830; H.
Reinecke, et al., J Clin Biol, 149, 731 (2000);
J. Gutstein, et al., Circ Res, 88, 333 (2001).
22 M. Scorsin, et al., J Thorac Cardiovasc Surg,
119, 1169 (2000); H. Soonpaa, et al., Science,
264, 696 (1994).
23 M. G. Klug, et al., Am J Physiol, 269, H1913
(1995); M. G. Klug, et al., J Clin Invest, 98,
216 (1996); M. Muller, et al., FASEB J, 14,
2540 (2000); M. Kehat, et al., J Clin Invest,
108, 407 (2001).
24 S. Makino, et al., J Clin Invest, 103, 697
(1999); S. Tomita, et al., Circulation, 100, II-
247 (1999); J.-S. Wang, et al., J Thorac Cardio-
vasc Surg, 120, 999 (2000).
25 Beltrami, A. P., et al., N Engl J Med, 344, 1750
(2001); Beltrami, A. P., et al., Circulation, 104,
suppl II, A1559; Quaini, F., et al., N Engl J
Med, 346, 5 (2002); Orlic, D., et al., Proc Natl
Acad Sci USA, 98, 10344 (2001); Orlic, D., et
al., Nature, 410, 701 (2001).
26 K. Jackson, et al., J Clin Invest, 107, 1395
(2001); A. A. Mangi, et al., Nat Med, 2003;
9(9): 11951201. http://dx.doi.org/10.1038/
nm912.; D. H. Adams, et al., Ann Thorac
Surg, 70, 320 (2000); U. Galili, et al., Proc
Natl Acad Sci USA, 84, 1369 (1987); J. L. Platt,
et al., Transplant Proc, 22, 1066 (1990); B. J. W.
van Denderen, et al., Transplantation, 64, 882
(1997); G. W. Byrne, et al., Transplantation, 63,
149 (1997); M. L. Schwarze, et al., Ann Thorac
Surg, 71, 131 (2000); R. Siduan, et al., Ann
Thorac Surg, 70, 140 (2000); L. E. Nikalson, et
al., Science, 284, 289 (1999); P. Zilla, et al., J
Vasc Surg, 19, 540 (2000); J. M. Wilson, et al.,
Science, 244, 1344 (1989); C. A. Mason, et al.,
Nat Med, 5; 176 (1999); D. Srivastava and
E. N. Olsen, Nature, 407, 221 (2000); Lai, L.,
et al., Science, 295, 1089 (2002).
Abstract
This paper deals with epidemiological,
clinical and social features of Parkinsons
disease (PD), one of the most common
neurodegenerative disorders. An overview
of recent neuroscientific work for etiologi-
cal and pathological aspects of the disease
is given, and mechanisms of disease on a
molecular basis are discussed together
with potential risk and protection factors.
Drug therapy of PD is described by sum-
marizing conventional oral medications
aiming at L-DOPA substitution or admin-
istration of dopaminergic drugs. Other
substances active for PD, their mechanism
of action and their therapeutic restrictions
are also described. The reasons for failure
of drug therapy after years of progressing
disease are discussed and the medical
need for innovative therapies, particularly
in advanced disease, is explained. Surgical
therapeutic approaches for PD including
pallidotomy and deep brain stimulation
are described. The anatomical and electro-
physical rationale for these therapies is
presented, as is the benefit and the draw-
backs of these therapeutic innovations.
The history of transplantation of fetal me-
sencephalic cells for PD is set out, to-
gether with the reasons why fetal trans-
plantation programs are presently not
further pursued. The expectations the sci-
entific community puts into future poten-
tial stem cell therapies are analyzed and
recent research on growth factor treatment
in PD is also reported. The last part of the
chapter is a description of biochemical,
pharmacological and immunological as-
pects of human retinal pigment epithelial
(hRPE) cells used in an experimental ther-
apeutic approach requiring neurosurgical
cell implantation into the brain. The pro-
duction of Spheramine
, a preparation of
hRPE cells on microcarriers for implanta-
tion into the brain of PD patients, is de-
scribed, and preclinical development in-
cluding work with animal models and effi-
cacy studies in animals are set forth. Clini-
cal aspects and questions regarding the
clinical development of Spheramine are
described. Ethical questions of sham sur-
gery, double-blind placebo-controlled clini-
cal trials in a complex setting of physicians
who need to know treatment assessment
and others who must not, and the design
of clinical studies involving stereotactic
neurosurgery in general are discussed.
Abbreviations
AD Alzheimers disease
ADL activity of daily living
325
14
Spheramine
: A Cell Therapeutic Approach to Parkinsons Disease

ISBN: 3-527-31184-X
DHCA dihydroconiferyl alcohole
DHI 5,6-dihydroxyindole
GDNGF glial-derived nerve growth factor
h human
HBSS Hanks balanced salt solution
MPTP 1-methyl-4-phenyl-1,2,3,6-tetra-
hydropyridine
MRI magnetic resonance imaging
PD Parkinsons disease
RPE retinal pigment epithelial
UPDRS unified Parkinsons disease
rating scale
14.1
Introduction
Neurodegenerative disease can be regarded
as the greatest challenge in neurology.
Although the decade of the brain (1990
2000) has produced a tremendous quantity
of research results elucidating mechanisms
of neuronal decay, and cell degeneration in
general, no realistic chance of a cure is in
sight. However, various promising new
therapies are beginning to emerge.
Alzheimers disease (AD), Parkinsons dis-
ease (PD) and multiple sclerosis (MS), the
most frequent diseases affecting the human
brain, are neurodegenerative disorders, and
individuals suffering from one of these con-
ditions make up the vast majority of neurol-
ogy patients both in clinics/doctors offices
and as in-patients.
Parts of the pathophysiology in AD and
PD are well understood. In MS, an impor-
tant portion of the pathophysiology is due
to immunological processes, but the roots
of the disease seem to extend much earlier
than onset of inflammation and neurode-
generation, either hereditary or acquired,
is thought to underpin the disease as a
primary cause [1].
For PD, a number of very effective
symptomatic treatment approaches exist.
However, two fundamental questions re-
main without answer:
Is there a way to stop this progressive
disorder?
Can we replace neurons that have died
in the course of disease is brain repair
a possibility?
14.2
PD
14.2.1
Terminology
As with most aspects of PD, the terminol-
ogy of the disease itself is presently under-
going revision. So far, the belief that PD is
one disease, that Lewy bodies in the sub-
stantia nigra are the hallmark of PD and
that Lewy bodies are responsible for the
death of neurons in PD has been unani-
mously accepted. However, all three of
these hypotheses may require correction [2].
14.2.2
Epidemiology
After AD, PD is the second most common
neurological disease. It is a disease of the el-
derly. All cases with early onset of parkin-
sonism under 40 years of age are accounted
for by familial, genetically determined diag-
noses and the incidence of true idiopathic
PD cases under 50 years of age represents
only 3.8% of cases. Monozygotic twins with
idiopathic PD have the same concordance
as dizygotic twins or other siblings, while
monozygotic twins with early onset are
100% concordant for their disease [3].
Age-adjusted prevalence studies in Cana-
da, Scotland, New Zealand, Sicily and var-
ious parts of the USA resulted in prevalence
figures between 76 and 257 per 100000.
Studies in the UK, Norway and the Faroe Is-
14 Spheramine

lands resulted in crude prevalence values
between 111 and 187 per 100000. Broken
down into age groups, the high prevalence
value of 329 per 100000 found in one epide-
miologic study in Nebraska indicates preva-
lences of 406/298 (men/women) for the 60
70 years age group, 1794/991 for 7080
years and 4248/2069 for the population over
80 [4]. Men seem to have a higher risk for
PD than women [5].
While it is highly probable that national
or ethnic differences in prevalence found
in former years (e.g., low prevalence in Ja-
pan) [6] are due to imperfect counting
methods, underdiagnosis and other sys-
tematic epidemiologic errors, it cannot al-
together be excluded that prevalences dif-
fer in rural and urban environments. The
hypothesis resulting from this distribution,
i.e., that pesticides may induce PD, will
hopefully be clarified by a large agricultur-
al health study in 52000 pesticide sprayers
that is funded by the National Institute of
Neurologic Disease and Stroke and is pres-
ently ongoing.
14.2.3
Risk and Protection Factors
The pattern and time course of neurodegen-
eration in PD strongly suggests environ-
mental causation [7] of the disease; however,
the actual toxic agent or agents remain un-
known. High iron intake in combination
with high manganese intake have been ac-
cused [8]. There is very clear evidence that
1-methyl-4-phenyl-1,2,3,6-tetrahydropyri-
dine (MPTP), an impurity contained in il-
legally produced and distributed heroin in
the late 1960s in California, caused parkin-
sonism in consumers and that rotenone, a
natural toxin used for pest control and
commercial fishing in East Asia, cause par-
kinsonism in animals [9]. The incidental
findings in humans led to the development
of animal models of PD [10].
Little is also known regarding protecting
factors: the favorable effect of smoking
could be confirmed in multiple studies
and coenzyme Q10 supplementation [11]
has a mild beneficial effect. Nonsteroidal
anti-inflammatory drugs may reduce the
risk, but only animal experiments and ret-
rospective studies in humans support this;
no prospective trial is available [12].
14.2.4
Pathology
Generations of neurologists were taught
that PD starts in the substantia nigra with
the degeneration of dopaminergic neurons,
and that the destruction of neurons in this
area of the brain has its most palpable effect
in the dysfunction of the basal ganglia,
where those nerve cells project. More recent
pathological findings, however, suggest that
PD is a pathologic condition with an onset
far more peripheral in the nervous system,
in the olfactory bulb, the peripheral vegeta-
tive nerves as the Auerbach plexus of the in-
testine, and that in fact it may even start
from the peripheral nerve system and as-
cend to the brain stem and the mesence-
phalon until it finally spreads over the
entire brain, including phylogenetically
younger parts like the forebrain.
PD pathology is characterized by thread-
like proteinaceous neurite inclusions and
intracellular Lewy bodies (Fig. 14.1) occur-
ring exclusively in unmyelinated fibers. In
pathological stage I and II of the disease,
first Lewy neurites, then Lewy bodies ap-
pear almost simultaneously in the olfac-
tory bulb and the dorsal visceromotor nu-
cleus of the vagal nerve. Degeneration in
the olfactory bulb seems to be very slow.
Early signs of Lewy neurites and Lewy
bodies can also be found in Auerbachs
14.2 PD 327
plexus in the peripheral vegetative nerve
system. The nucleus ceruleus is affected
in stage II, as are the raphe nuclei of the
lower brain stem and the gigantocellular
reticular nucleus.
Areas above the tegmentum and pons
(the substantia nigra and amygdala, the ba-
sal forebrain and the hypothalamus) are not
affected before stage III. Stage IV describes
involvement of the temporal mesocortex, an
area exhibiting little myelination, and stage
V is defined as the ascension to the neocor-
tex, while the final stage VI is related to in-
volvement of the pre-motor and primary
motor and sensory cortices [13].
14.2.5
Clinical Symptoms
Early clinical symptoms of PD correlate well
with these stages: olfactory disturbances and
obstipation, but also cognitive decline other
than dementia, traditionally referred to as
bradyphrenia (deficits in attention, visual
and spacial cognition, and REM sleep disor-
der), are the earliest, although non-specific,
signs of PD that become apparent long be-
fore the typical PD symptoms summarized
under the TRAP acronym (tremor, rigidity,
akinesia and postural instability) appear.
Asymmetry of symptoms is an important
feature of sporadic PD, differentiating it
from other forms of parkinsonism. As the
disease progresses, symptoms become bilat-
eral. Postural instability and nonmotor signs
including dementia, symptoms that are dif-
ficult to treat, become the main cause of lim-
itation in the patients quality of life [14].
14.2.6
Etiopathophysiology
Although some single steps in PD patho-
physiology are well understood, the entire
pathway leading from one or more puta-
tive toxic agents to progressive disease is
far from being elucidated. Genetic disor-
ders with a clinical picture resembling
sporadic PD and intoxications causing par-
kinsonism have helped to obtain insight in
a number of mechanisms involved in the
pathophysiology of PD.
A basic molecular step on the pathway
of neuronal damage is dysfunction of the
mitochondrial complex I. MPTP, rotenone
and paraquat are toxins inhibiting complex
I. Inhibition leads to the formation of free
radicals and, as a consequence, oxidative
stress, which makes the cells vulnerable to
glutamate excitotoxicity. Complex I dys-
function may also mediate cell death via
caspase-dependent and caspase-indepen-
dent apoptosis, necrosis, and inflamma-
tion-induced injury [15, 16].
An interesting recent finding in trans-
genic mice overexpressing L-3-hydroxyacyl-
CoA dehydrogenase II supports this model
of MPTP toxicity. Overexpression of this
enzyme, which is a mitochondrial oxire-
ductase system involved in neuronal sur-
vival, protects the animals from the conse-
quences of MPTP intoxication [17].
Polymorphisms in a number of genes
can cause parkinsonism or create suscepti-
14 Spheramine

Fig. 14.1 Lewy body.
bility for PD, the best investigated ones
being a-synuclein gene, Parkin gene, DJ1
gene and PINK1 gene [18] polymorphisms
(see also Part I, Chapter 2). Mutations with
a questionable risk increase are found in
cytochrome P-450 [19] (see also Part VII,
Chapters 2 and 3) and the dopamine
transporter (DAT) gene [20].
a-Synuclein is a protein of unknown
function that is prone to form insoluble
oligomers in vitro and proteins carrying a
missense mutation are even more prone
to do so. The importance of a-synuclein
was greatly enhanced by the discovery that
Lewy bodies and Lewy neurites in PD in
general contain a-synuclein aggregates.
However, it remains unknown if these ag-
gregates have a causative effect or are a
mere epiphenomenon in the pathophysiol-
ogy of the disease [21, 22].
Interestingly, in patients with autosomal
recessive juvenile parkinsonism, no Lewy
bodies or Lewy neurites are found at au-
topsy [23].
Mutations in the Parkin gene are seen
as another cause of familial parkinsonism.
Parkin is a nervous system protein attach-
ing short ubiquitin peptide chains to pro-
teins a process tagging these proteins to
mark them as candidates for degradation
by the proteasome complex. This Parkin-
controlled process may be important for
the normal turnover of a-synuclein and
impaired function could lead to an accu-
mulation of a-synuclein [24].
Proteolytic stress as a result of protea-
some dysfunction itself is also believed to
be a potential cause of protein accumula-
tion, independent of ubiquination. The as-
sembly of the proteasome complex is ATP
dependent a potential link in the causal
chain of PD etiology. Furthermore, striatal
dopaminergic cells have a high burden of
oxidized proteins to be degraded due to
enzymatic oxidation and auto-oxidation of
dopamine, but relatively low proteosomal
activity. They may therefore be particularly
susceptible to any molecular mechanism
inhibiting or injuring their proteasomal
system [25].
Systemic exposure of adult rats to epoxo-
micin, a naturally occurring proteasome in-
hibitor, causes progressive parkinsonism in
these animals. This supports both the pos-
tulated pathomechanism of proteasome
dysfunction in PD and the hypothesis of en-
vironmental causation of the disease [26].
Inflammatory processes are almost cer-
tain to play a role in PD. There is one
striking historic event depicting the inter-
relationship of neuroinflammation and
PD: v. Economos encephalitis, a viral pan-
demic at the beginning of the 20th cen-
tury, was associated with parkinsonism.
Activated microglia was found in the
brains of drug addicts who suffered invo-
luntary intoxication with MPTP contained
in impure heroin. Increased levels of cyto-
kines interleukin (IL)-1b, IL-2, IL-4, IL-6
and tumor necrosis factor-a in the cerebro-
spinal fluid as well as increased density of
glial cells expressing pro-inflammatory
cytokines in the substantia nigra of PD
patients are further evidence for the in-
volvement of neuroinflammatory processes
in the pathological cascade of neurodegen-
eration in PD. Viral agents have not been
found in PD patients, and the mechanism,
time of onset and impact of neuroinflam-
mation in the overall balance of PD
pathology is unclear [2729].
14.2.7
Therapy
14.2.7.1 Drug Therapy
The greatest and to date unequalled break-
through in PD therapy occurred in 1961
when, based on research by Carlsson and
Hornykiewicz, Birkmayer started giving
14.2 PD 329
PD patients high doses of L-DOPA [30].
Since then, L-DOPA has been the gold
standard in PD therapy. As dopamine is
metabolically sensitive and is too polar to
easily cross the bloodbrain barrier, its pre-
cursor L-DOPA must instead be used. De-
carboxylase inhibitors administered simul-
taneously, almost always in combinatory
formulations, prevent immediate transfor-
mation into dopamine in the periphery
and allow for therapeutic concentrations in
the brain to be built up. Despite this trick
and the development of slow-release for-
mulations, the half-life of L-DOPA is still
very short and multiple doses over the day
are required, at least in advanced PD pa-
tients.
Amantadin, originally an antiviral agent,
exhibits weaker effects than L-DOPA, but
still maintains its position in the inventory
of anti-PD drugs as it is available in i.v. for-
mulations and can be given to patients with
temporary swallowing difficulties [31]. The
same is true for apomorphin, a drug with
good efficacy, but unpleasant side-effects,
which can be administered via a pump to
provide even drug levels over the day.
A variety of dopaminergic agents mi-
micking L-DOPA effects are available, all
with a smaller efficacy than L-DOPA, but
with varying beneficial features like long
half-life, better tolerability or modified, but
not principally different, side-effect profiles
[32].
The paradigm of L-DOPA toxicity was
the driving force behind the development
of those drugs. In a scientific controversy
lasting for decades, the question has been
discussed whether L-DOPA or the disease
itself accounts for complications that arise
as the disease progresses [33].
In the course of disease, L-DOPA loses
efficacy in treating PD symptoms. The
therapeutic window becomes narrower
from both sides, i.e., higher doses are re-
quired to achieve an effect, while at the
same time the maximum tolerable dose
decreases. The most prominent overdosing
signs in L-DOPA therapy are dyskinesia
and hallucinations. In advanced disease,
there is the on/off phenomenon, sudden
loss of efficacy despite normal drug levels
and freezing (a sudden inability to move).
Maintaining the balance between appro-
priate symptom control and overdosing
may become impossible in advanced PD
patients. It was postulated and vigorously
debated that the duration of L-DOPA treat-
ment and the accumulated life dose were
associated with the onset and severity of
dyskinesia and other advanced-stage com-
plications of PD [34, 35].
Although the final assessment of studies
supporting the L-DOPA toxicity model
may still be controversial, there seems to
be clear evidence for the fact that pulsatory
stimulation of dopaminergic neurons in
the PD brain contributes to the develop-
ment of dyskinesias, and that continuous
stimulation mimicking the physiological
conditions of dopamine titers in the brain,
e.g., by infusion, can prevent or delay dys-
kinesias and can even reverse them to a
certain degree [36].
Adjunct therapies include inhibitors of
enzymes supporting the degradation of do-
pamine, such as catechyl-O-methyltransfer-
ase and monoaminooxidase B.
Ergotamine preparations have lost some
of their popularity due to the availability of
stronger and more tolerable dopaminer-
gics, and acetylcholinesterase inhibitors
are regarded as being obsolete due to their
negative effects on cognition.
In summary, a rich choice of effective
treatments exists for symptomatic treat-
ment of PD. Although the introduction of
these therapies, in particular the pioneer-
ing work of L-DOPA implementation, has
achieved dramatic improvement in PD
14 Spheramine

mortality, none of these drugs can stop or
slow disease progression. No adequate
alternatives exist for advanced patients
who have a reduced L-DOPA tolerance and
at the same time require elevated doses of
L-DOPA for achieving a symptomatic
effect.
Ambitious research striving to overcome
these problems is being performed. How-
ever, it must be acknowledged that many
efforts have not provided the desired suc-
cess. A summary of recent failures in PD
drug development lists many dopaminer-
gic, serotoninergic, noradreneric, GABAer-
gic, adenosine-A2a-receptor antagonist-re-
lated, glutamate antagonist-related, canna-
binoid agonist- and antagonist-related, and
opioid receptor agonist- and antagonist-re-
lated approaches. All these missed their
goal of improving PD drug therapy [37].
14.2.7.2 Surgical Therapies
Pallidotomy Complications and limita-
tions of L-DOPA therapy became apparent
several years after its implementation.
Based on the accidental finding that stroke
in certain brain areas can sometimes alle-
viate PD symptoms, ablative surgical
therapies were investigated or, rather, re-
discovered and further developed after ac-
tivity in this field had ceased due to the
overwhelming success of L-DOPA therapy.
Pallidotomy has been an efficient surgi-
cal tool for treating patients who no longer
responded adequately to L-DOPA therapy,
and it is still indicated and performed in
certain complex cases. Its efficacy is due to
the fact that no electrophysiological stimu-
lation in the pallidum is less deleterious
than the overstimulation caused by the
loss of dopaminergic projections from the
substantia nigra.
Pallidotomy can only be performed uni-
laterally as bilateral destruction of the ba-
sal ganglia would prompt the risk of ser-
ious cognitive side-effects [38, 39].
Deep brain stimulation
A very elegant way to achieve the same
effects as with pallidotomy, but with no
destruction of brain tissue and by a revers-
ible method, was the development of deep
brain stimulation. Electrical overstimula-
tion of the nucleus subthalamicus or the
internal part of the globus pallidus
achieves the same effect as destruction of
nerve cells of the pallidum. The stimula-
tor can be removed in the case of compli-
cations or lack of efficiency and bilateral
treatment can be performed. Deep brain
stimulation is now an approved treatment
for advanced PD, and is largely accepted
by neurologists and patients. The obvious
disadvantage is an artifact remaining in
the brain, and the necessity of numerous
and time-consuming fine-tuning sessions
to adapt the device to the patients needs,
as well as repeated surgeries for chang-
ing batteries, etc. [40, 41]. Negative effects
on cognition, more pronounced in pa-
tients over 69 years, may also limit the use
of deep brain stimulation in PD patients
[42].
Other experimental surgical therapies
Transplantation of human fetal mesencepha-
lic cells Transplantation of dopaminergic
neurons obtained from human embryos
has a long and controversial history. The
pioneer work was performed by Scandina-
vian groups who transplanted fresh mes-
encephalic tissue from four to eight em-
bryonic donors per patient into the stria-
tum of advanced PD patients. Spanish and
Mexican researchers followed them, and at
least 600 patients have been treated in
open studies.
14.2 PD 331
A controlled trial was initiated by Curt
Freed and Stanley Fahn [43]. Of the 40 par-
ticipants in the trial, 20 received trans-
plants and 20 underwent sham surgery
with no neuronal tissue being trans-
planted. Very surprisingly, only a subgroup
of patients under 60 years of age was
found to benefit from the transplantation
after a 1-year follow-up.
Even more disturbing, after 24 months,
33% of these patients had developed side-ef-
fects that were described as OFF dyskinesia
or late-running dyskinesias. These patients
suffered violent involuntary movements
even when they were off their anti-PD med-
ications and they could no longer tolerate
even the lowest possible doses of L-DOPA.
Similar results were reported from a sec-
ond randomized and placebo-controlled
study by Olanow et al. [44], where only
limited and transient effects were found,
and late-occurring or OFF dyskinesias in
56% of the patients.
To date, it is entirely unclear why these
side-effects occur. They could be due to
certain properties in parts of the patient
population, they could be related to immu-
nosuppression or to the fact that a cell
mixture rather than one uniform cell line
was transplanted, or they could simply re-
flect the fact that transplanted neurons
can grow into the host brain and form sy-
napses, but that they do not always form
the correct connections to the surrounding
brain tissue. There is little chance for the
embryonic transplantation programs to re-
sume before these questions have been an-
swered.
Transplantation of porcine fetal mesencepha-
lic cells In order to circumvent ethical
problems encountered in the collection of
donors for human fetal mesencephalic
cells, dopaminergic neurons from an ani-
mal source were studied as transplants for
PD. The clinical program advanced as far
as phase II (proof-of-concept) trials, but
the placebo-controlled trial failed to show
efficacy [45, 46].
Stem cells Culturing human stem cells
and differentiating them into dopaminer-
gic neurons for cell transplantation in PD
is a dream that inspires many academic
researchers and the biotechnology indus-
try. The idea of creating a reliable and re-
newable source of uncontaminated, well-
defined and characterized cells that could
replace degenerated neurons and create
the possibility of functional brain repair is
indeed intriguing (see also Part I, Chapters
11, 12 and 15).
Differentiating stem cells down the
dopaminergic pathway seems to be one
possibility. The crucial breakthrough was
the observation that for survival and differ-
entiation into dopaminergic neurons, stem
cells need an environment with a low oxy-
gen content.
Several cell sources have been used for
experimental cell therapy approaches:
pluripotent stem cells like embryonic stem
cells from the blastocyst and embryonic
germ cells [47, 48], human cord-blood de-
rived cells [49], but also cultured cell lines.
Neural stem cells (stem cells forming neu-
ronal as well as glial cell lineages) isolated
from different parts of the brain known to
harbor multipotent neural stem cells [50
52] are also an option. In a non-neurode-
generative indication, a cultured cell line
was used to treat patients after stroke in a
pilot clinical trial. The cells were trans-
planted in the paramedian plane and treat-
ment was reported to be safe, with signs
of efficacy [53].
However, there remain problems to be
solved before these approaches can be re-
garded mature enough to proceed to clini-
cal experiments. First, it is not clear how
14 Spheramine

long stem-cell-derived dopaminergic neu-
rons can maintain their phenotype in vivo.
They do so for a limited time in culture,
but there are no data on the survival of
transplanted dopaminergic neurons in the
human brain. For a clinically acceptable
therapy it should be years rather than
months. Second, the tumorigenicity of
stem-cell-derived neurons is not suffi-
ciently excluded, although no tumor has
been found after adult stem cell transplan-
tation in short- to mid-term experimental
models. Third, one or several animal mod-
els for testing the transplantation para-
digm would be desirable before human ex-
periments are started. Fourth, for PD,
there should be conclusive data from ani-
mal models showing that stem cell trans-
plantation does not cause similar late-oc-
curring dyskinesias or OFF dyskinesias as
were seen with the human fetal trans-
plants discussed earlier.
In summary, use of stem-cell-derived
neurons as transplants in PD appears to
be a very attractive approach. It will not be
realized, however, before additional scientif-
ic work and combined efforts in cell culture,
preclinical (animal model) and genetic re-
search have been accomplished [54, 55].
Transplantation of human non-neuronal do-
pamine-producing cells Three human cell
types have undergone evaluation for their
suitability to serve as dopamine sources
after transplantation into the brain of PD
patients: autologous adrenal cells, autolo-
gous carotid body cell aggregates and allo-
geneic human retinal pigment epithelial
(hRPE) cells. Adrenal cell transplantation
has undergone thorough riskbenefit as-
sessment and there is now agreement that
this treatment, particularly the harvesting
of adrenal cells from the patient, bears too
much risk to be recommended [56]. In the
future, carotid body aggregates may share
this fate of being regarded as too risky,
although no general recommendation or
warning has been issued by any sounding
board so far [57, 58]. hRPE cells are the
cellular component of Spheramine and
their development is described in more de-
tail below.
Growth factors The first attempt to pro-
mote a very interesting approach towards
a new PD treatment has recently failed.
Successful animal experiments including
nonhuman primates had nourished hope
that infusion of glial-derived nerve growth
factor (GDNF) could alleviate PD symp-
toms and at the same time act as a neuro-
protective agent that could stop or at least
delay disease progression [59]. Numerous
ambitious experimental programs were de-
signed based on the efficacy of GDNF in
animal models there were plans to use
genetically modified cells designed for
transplantation into the brain that could
directly release growth factors into the de-
generating neuronal tissue or co-transplan-
tation of nerve cells and growth factor-pro-
ducing cells to enhance transplant func-
tion and survival [60]. At this point, it is
unlikely that these projects will succeed
after the failure of direct GDNF infusion
[61].
14.2.8
Medical Need for the Development
of Alternative Therapies for PD
Of the many attempts and endeavors to
improve PD therapy, modify the disease or
even accomplish brain repair to help pa-
tients suffering from this devastating and
debilitating condition, none has so far
been able to provide a cure. Effective
symptomatic therapies, both as conven-
tional oral medications and surgical
approaches, are available. The effectiveness
14.2 PD 333
of oral medications is limited to treatment
periods of 57 years, a timespan after
which treatment complications will inevita-
bly occur. Of the many surgical therapies
that have been investigated to date, only
deep brain stimulation exhibits an accept-
able riskbenefit balance. It is indicated
for a subgroup of patients only those
who are young enough to tolerate both the
surgical risk and the extensive fine-tuning
procedures required to maintain the func-
tion of the device. Drawbacks such as the
continued and cumulative risk of infection
of an indwelling artifact in the brain, re-
peated re-surgery for changing batteries,
etc., are obvious.
14.3
Spheramine
Spheramine is a biological product com-
posed of hRPE cells placed on microcarriers
for implantation into the human brain. The
pharmacologically active part of Sphera-
mine is composed of the hRPE cells which
produce L-DOPA. They are placed on mi-
crocarriers of crosslinked porcine gelatin
to enhance their survival and function.
14.3.1
Pharmacological Rationale
Spheramine is thought to provide an
approach for dopamine substitution in the
brain at a constant and even level, with no
or insignificant fluctuations. As is known
from experimental work with continuous
dopaminergic stimulation [62], this repre-
sents a more physiologic substitution of
the neurotransmitter lacking in PD than
achievable by oral therapy. It may be ap-
propriate to reduce fluctuations of the do-
pamine response such as seen in the on/
off phenomena and dyskinesias.
Spheramine is not believed to provide
any brain repair, nor even modulation or
slowing of disease progression per se,
although it is conceivable that should pre-
liminary results from a pilot study be con-
firmed, Spheramine may delay the onset
of treatment complications if administered
in time.
14.3.2
Cell Source
The cellular source for Spheramine is the
RPE obtained from non-embryonic neona-
tal eye donor (premature infants) tissue. It
is collected by approved organ transplanta-
tion banks.
The RPE of the eye is a monolayer
membrane located between the photore-
ceptors and the chorioid (Fig. 14.2), and it
is thought to function as a support tissue
for the photoreceptors. hRPE cells are in-
volved in retinoid transport, esterification
and isomerization processes, and the pha-
gocytosis and degradation of shed outer
segments. It participates in the regulation
of ion channels and the extravasation of
metabolites, and depending on the devel-
opmental state of the individual synthe-
sis of melanin.
hRPE pigment is a blackish-brown sub-
stance called eumelanin. It differs from
the melanin contained in melanocytes.
While melanocytes are derived from the
mesoderm, RPE cells are derived from the
neuroectoderm. In mammals, about 160
genes are involved in the biosynthesis of
melanin pigments. Both in cutaneous mel-
anocytes and RPE cells, tyrosinases synthe-
size DOPA and subsequently dopaqui-
none, starting from the substrate L-tyro-
sine. However, DHCA and DHI oligomers
are formed via dopachrome in hRPE cells,
whereas cutaneous melanocytes form
pheomalins via cysDOPA.
14 Spheramine

14.3 Spheramine 335
F
i
g
.
1
4
.
2
L
o
c
a
t
i
o
n
a
n
d
a
n
a
t
o
m
y
o
f
t
h
e
R
P
E
.
The eumelanin pigment in the hRPE cell
layer is thought to provide a light and UV
radiation shelter for the inner parts of the
eye [63]. Dopamine produced from L-DOPA
by a selective decarboxylase is involved in
pigment migration via D1 receptors, and
potassium and chloride channel activity
via D2 receptors. The same receptor is in-
volved in the control of disk shedding and
other rhythmic processes. Rhythmic synthe-
sis of melatonin in the photoreceptors con-
trols dopamine release from the retinal in-
terplexiform and the amacrine cells. Dopa-
mine is thus an important regulator in the
function and protection of the photorecep-
tor of the eye. Cherksey et al. [63] concluded
that hRPE cells may therefore be suitable
for the treatment of PD.
14.3.3
Expansion, Characterization
and Immunosuppression
Harvesting the hRPE membrane from do-
nated eyes of premature newborns (gesta-
tional age approximately 20 weeks or more)
is followed by dissociation and growth in
culture until confluence of a monolayer cell
sheet is reached in the culture flask.
After implantation into the brain of 6-
OHDA-lesioned rats (an animal model of
PD), loose hRPE cells exhibit a positive
effect on the rotational behavior of these
lesioned and amphetamine-challenged ani-
mals, but the effect is only transient. It is
only after placing the hRPE cells on a gela-
tin substrate that they survive after im-
plantation and exhibit continuous function
(see Fig. 14.3).
Gelatin microcarriers are manufactured
from denatured porcine collagen (gelatin)
obtained from processing pig skin from
certified slaughterhouses and from stock
certified to be free of transmittal disease.
After denaturation under harsh thermal
conditions, the gelatin is sieved to isolate
spherules meeting very narrow specifica-
tions. It should be noted that Spheramine
microcarrier beads for use in rats have dif-
ferent specifications (smaller diameter)
than those used in human and nonhuman
primates (see Fig. 14.4).
hRPE cells may be deep-frozen after har-
vesting, then thawed and attached to the
microcarriers in a precisely defined and
patent-protected procedure, to provide
Spheramine. After the combination of
cells and carriers, the substance is ready
14 Spheramine

Fig. 14.3 A single hRPE cell.
Fig. 14.4 Gelatin microcarriers.
for transport and then implantation within
a short period of time. A set of rigid speci-
fications for morphology, L-DOPA content
and a number of other parameters is in
place to make sure that cells with a com-
parable profile are used in all preclinical
and clinical experiments.
After implantation into the human
brain, no immunosuppression is required
for inhibiting graft rejection and graft ver-
sus host disease.
14.3.4
Animal Models
The three most recognized and widely
used animal models for PD are the 6-
OHDA and MPTP rat models, and the
MPTP monkey model. Parkinson-like
symptoms in rats can be produced by ex-
posing one brain side to 6-OHDA, a toxic
dopamine metabolite. After receiving a
challenging dose of amphetamine, these
animals exhibit rotational behavior which
can be modified by anti-PD drugs.
In the MPTP monkey model, like in hu-
mans, MPTP produces Parkinson-like
symptoms that can be assessed by a modi-
fied Unified PD Rating Scale (UPDRS),
the most commonly used tool for the eval-
uation of PD symptoms in humans.
Both 6-OHDA rat and MPTP monkey
models were used for the preclinical devel-
opment of Spheramine.
14.3.4.1 Animal Safety Studies
A Salmonella typhimurium reverse muta-
tion test for determination of potential
mutagenic activity of the crosslinked gela-
tin microcarrier component was per-
formed with Spheramine.
Additional safety studies included the
short-term exposure of nonhuman pri-
mates to crosslinked gelatin implanted in-
tracranially. Microcarriers suspended in
Hanks buffered salt solution (HBSS), or
HBSS alone, were injected into different
coordinate sites of the forebrain in two
male and two female Macaca fascicularis
monkeys, which were sacrificed after 1
and 4 weeks. No evidence of granuloma-
tous or immune-mediated reactions was
found. Weight loss of approximately 10%
was observed in one monkey. Hemor-
rhage, necrosis and neutrophilic cell infil-
tration 1 week after surgery, mild to mod-
erate glial fiber proliferation, and mono-
nuclear cell infiltration of mild to moder-
ate severity continuing until week 4 were
nonspecific findings in these animals.
Thirty-two Cynomologous monkeys as-
signed to five treatment groups (Group 1:
55000 cells per 1.3 mg gelatin microcarrier
per site; group 2: 89000 cells per 2.1 mg
gelatin microcarrier per site; group 3:
1.8 mg gelatin microcarrier per site; group
4: 2.6 mg gelatin microcarrier per site;
group 5: placebo control) were sacrificed to
study toxic effects of intracranial xeno-
grafts of hRPE cells on gelatin microcar-
riers and microcarriers alone.
Sacrifice was after 6 and 16 months. No
treatment-related mortality, clinical ab-
normality or ophthalmologic abnormality,
or changes in body weight or relative or-
gan weights were observed. No migration
or replication of hRPE cells was found at
the 6 or 16 months sacrifice. Findings
were limited to nonspecific inflammatory
lesions in the meninges, cranial bones and
brain surface, probably secondary to intra-
cranial injection. Lesions of chronic in-
flammation, glial fibrosis and granuloma
formation were similar in all treatment
groups, with continued healing at month
16. No test article-related pathology was
found. The maximal tested dose of Sphera-
mine (443350 hRPE cells + 10.5mg gela-
tin) was well tolerated.
14.3 Spheramine 337
14.3.4.2 Animal Efficacy Studies
Microcarrier-bound hRPE cells (dose esti-
mated as 1000 cells per 150 microcarriers)
produced a significant reduction in apomor-
phine-induced circling when implanted
into the homolateral striatum of 6-OHDA-
lesioned Sprague-Dawley rats. This effect
was achieved without immunosuppression
and was sustained for the duration of the
18-week study. Unattached hRPE cells pro-
duced only a transient effect.
Similarly, a dose of around 2000 hRPE
cells per roughly 0.06 mg gelatin microcar-
rier injected intrastriatally into 6-OHDA
rats produced a significant reduction in
circling behavior of about 53% at day 29
post-implant.
In a study conducted to obtain a better
understanding of the functional and im-
munological consequences of Spheramine
implantation in the presence and absence
of systemic immunosuppression, 6-OHDA
hemiparkinsonian rats were intrastriatally
implanted with around 2000 cells attached
to gelatin microcarriers and randomized
into two groups, with and without cyclo-
sporin A. The behavioral results from this
short-term study showed a therapeutic
benefit and no difference between groups,
indicating that immunosuppression is not
necessary.
Of three Macaca mulatta monkeys im-
planted unilaterally into the striatum with
five deposits of around 10000 cells on gel-
atin microcarriers each, Monkey 1 demon-
strated about 80% improvement in the
UPDRS that was sustained throughout the
8-month experiment, Monkey 2 demon-
strated about a 60% improvement by 6
months post-implant and the third animal
could not be used for efficacy assessments
as the parkinsonism evoked by its MPTP
lesions was of insufficient severity. This
animal had bilateral surgery serving as a
safety experiment. The results of this
study clearly demonstrate a robust and
persistent effect of microcarrier-bound
hRPE cells in the MPTP hemiparkinson-
ian monkey model in the absence of im-
munosuppression.
Implantation of a total Spheramine dose
consisting of around 280000 human pig-
ment retinal epithelial cells on gelatin mi-
crocarriers distributed over five sites in the
striatum of MPTP-induced hemiparkinson-
ian M. mulatta monkeys produced a statis-
tically significant response at 12 months
post-implant in comparison to surgical
controls. A low-dose Spheramine group
(around 60000 total cells) and a group re-
ceiving microcarriers only were not statisti-
cally distinguishable from the surgical con-
trol group.
In summary, efficacy studies in animal
models of PD showed substantial and sus-
tained efficacy of Spheramine implanted
into the striata of animals with substance-
induced parkinsonism, in the absence of
immunosuppression.
14.3.5
Pilot Study in Humans
Six patients, three men and three women,
average age 52 years, with idiopathic
Hoehn and Yahr Stage III/IV PD were en-
rolled in a 1-year, open-label, single-center
pilot study to evaluate the safety and effi-
cacy of Spheramine.
Each patient was implanted with around
325000 hRPE cells on gelatin microcar-
riers (Spheramine) in the post-commissur-
al putamen contralateral to his/her most
affected side. All six patients have com-
pleted the 12-month study, and are now
enrolled in a protocol extension for contin-
ued safety and efficacy monitoring [28].
The longest follow-up duration for any pa-
tient is 48 months post-surgery at the time
of writing.
14 Spheramine

14.3.5.1 Safety Data
Spheramine was well tolerated in all pa-
tients and no serious adverse events were
considered to be definitely, probably or
possibly related to Spheramine. One oc-
currence of moderate depression and sui-
cidal ideations in a subject with a history
of depressive episodes was assessed as
being unlikely to be related to Sphera-
mine. This complication, a reoccurrence
of a previous depression, occurred ap-
proximately 14 months post-implantation,
following a change in antidepressant med-
ication. After reinstituting the original
antidepressant medication and increasing
the dosage, the symptoms resolved within
1 month.
All six patients experienced at least one
adverse event, all of which were mild or
moderate in severity. Records of adverse
events occurring in two or more subjects
(33% or more) reported dyskinesia (five
subjects), headache (five subjects), arthral-
gia (three subjects), depression (three sub-
jects), and two subjects each with upper
respiratory infection, dizziness, limb pain,
dystonia, hypoanesthesia, migraine, par-
esthesia, aggravated PD symptoms (longer
OFF times), hallucination, insomnia,
menopause and postural hypotension. One
patient had a small (approximately
4 mm 7 mm) asymptomatic cerebral
hemorrhage adjacent to the implant. The
majority (80%) of adverse events were con-
sidered to be not related to Spheramine
and no adverse events were assessed as
being definitely related to Spheramine. Ad-
verse events judged possibly or probably
related to Spheramine included dyskinesia,
visual hallucinations, hyperphagia, arthral-
gia, muscle twitching, limb pain, akinesia,
bradykinesia and overdose (of L-DOPA).
These events generally occurred when the
patients had taken their antiparkinsonian
medication and resolved spontaneously or
were responsive to dose reduction of anti-
parkinsonian medication.
Overall Dyskinesia Rating Scale scores at
the 12-month assessment were the same or
lower than at baseline for all patients. No
OFF state dyskinesias were observed.
14.3.5.2 Efficacy Data
Improvement in the primary outcome
measure, the UPDRS Motor score in the
practically defined OFF state (off all anti-
parkinsonian medication for at least 12
hours), was observed in all patients at 12
months and persisted until month 24.
Fig. 14.5 depicts the changes in the pri-
mary efficacy endpoint from baseline to
month 24 in the six individuals. However,
since baseline values differ greatly in these
patients, the percent improvement per pa-
tient may be more informative (see
Fig. 14.6).
Fig. 14.7 gives an overview of the time
course of mean percent improvements over
2 years in all patients. As can be seen, mean
improvements (n=6) were subject to fluc-
tuations over 2 years of follow-up, but re-
mained of the same order of magnitude.
14.3.5.3 Secondary Efficacy Parameters
Total UPDRS scores for the group showed
a mean improvement of 44% (51 points)
in comparison to baseline over 12 months.
Other secondary outcome measures in-
clude the UPDRS Motor ON (41%),
UPDRS Activities of Daily Living (ADL)
(OFF, 39%; ON, 41%), and Schwab and
England Physician Rated ADL (OFF, 41%;
ON, 9%).
Another parameter that was intended as
a safety assessment also deserves discus-
sion in the context of efficacy. As men-
tioned above, none of the six patients had
deterioration of pre-existing dyskinesia at
14.3 Spheramine 339
12 months. After 2 years of follow-up, dys-
kinesias were decreased as shown in Table
14.1. At month 24, mean dyskinesia values
had improved by 0.5 points on the Rush
Dyskinesia scale and by 2.2 points on the
UPDRS Part IV scale. This may be inter-
preted as support for the hypothesis that
continuous L-DOPA release as provided by
implanted hRPE cells may partially reverse
dyskinesias (ee Fig. 14.8).
14 Spheramine

Fig. 14.5 Patient UPDRS Motor OFF scores, primary outcome measure.
Fig. 14.6 Individual patient
percent improvement in the
primary outcome measure,
UPDRS Motor OFF scores,
from baseline.
Table 14.1 Dyskinesias as rated in the UPDRS and Rush Dyskinesia Scale.
Baseline 12 months 24 months
UPDRS IV, items 3235, dyskinesias 5.03.1 3.21.6 2.82.0
Rush Dyskinesia Rating Scale 1.31.0 0.70.8 0.70.8
After 2 years of follow-up, it is not suffi-
cient to look at UPDRS improvement alone.
PD is a progressive disease and a mean de-
terioration by 10.65 points on the UPDRS
scale is the expected disease progress per
year [64]. Clinical experts recommend that
in pilot studies for futility any deterioration
in UPDRS by less than 7.45 points per year
should be regarded as a potential clinically
significant therapeutic effect.
Estimating a more conservative average
deterioration of 10% per year, the improve-
ment of the six pilot study patients would
read as seen in Fig. 14.9. The percent im-
provement of the six individuals over 2
years would range between 51 and 45% if
the value was determined starting from
the adjusted baseline UPDRS, in contrast
to the nominal 1930% calculated from
the actual baseline values that were valid 2
years earlier (see Fig. 14.9).
Applying the same discussion to L-DOPA
medication usage is not as straightforward
as it requires hypothetical assumptions. In
de-novo patients, the daily doses of L-DOPA
or L-DOPA equivalents need adjustment by
14.3 Spheramine 341
Fig. 14.7 UPDRS Motor OFF scores,
from baseline.
Fig. 14.8 Mean dyskinesia values
at baseline, and at 12- and 24-
month follow-up.
a dose increase of approximately 100 mg
each year during the first 4 years after onset
of L-DOPA therapy [65]. However, this in-
crease in daily L-DOPA doses cannot be ex-
trapolated into later stages of disease, as pa-
tients will not tolerate indefinite L-DOPA
doses. On the contrary, as described above,
L-DOPA tolerability decreases with the pro-
gression of disease. The six patients in the
pilot study had optimized oral anti-PD med-
ication prior to undergoing surgery, i.e.,
they took their maximum tolerable dose
(see Fig. 14.11).
As is evident from Fig. 14.11, patient
medication was reduced by some
125 mg day
1
at month 12, but started to
go up again at month 24. It should be
clear that this increase does not automati-
cally imply loss of efficacy of the surgical
treatment. If we make a hypothetical extra-
polation of dose adaptation in these pa-
tients were they not Spheramine treated
(and would they support unlimited doses
of L-DOPA, which is not a realistic as-
sumption since these patients received
their maximum tolerable dose prior to sur-
14 Spheramine

Fig. 14.9 Patient UPDRS Motor
OFF scores, primary outcome
measure at baseline, baseline ad-
justed for natural disease course,
and 12 and 24 months post-treat-
ment.
Fig. 14.10 Group (n=6) per-
cent improvement in the pri-
mary outcome measure,
UPDRS Motor OFF scores,
from baseline and from base-
line adjusted for natural
course.
gery), L-DOPA dose adjustment would be
as shown in Fig. 14.12.
The time course of L-DOPA usage in
the six pilot study patients suggests two
conclusions: (1) patients may tolerate high-
er doses of L-DOPA than before Sphera-
mine treatment and (2) considering the
natural course of disease, L-DOPA de-
mand in Spheramine-treated patients may
be lower than would be expected from nat-
ural history data.
14.4
Randomized, Double-blind,
Placebo-controlled Multicenter Study
of the Safety, Tolerability and Efficacy
of Spheramine Implanted Bilaterally
into the Postcommissural Putamen
of Patients with Advanced PD
The summary of animal experiment safety
and efficacy data and the favorable result
from the pilot study of six patients with
unilateral Spheramine implants over a
minimum of 3 years, both with regard to
safety and efficacy, gave us confidence and
optimism to extend the clinical develop-
ment program for Spheramine.
14.4 Randomized, Double-blind, Placebo-controlled Multicenter Study 343
Fig. 14.11 Average dopaminergic
medication usage, L-DOPA equiva-
lents, mg day
1
(n=6).
Fig. 14.12 Hypothelical L-dopa dose
increase if indefinite doses were
tolerated.
From the experience with fetal mesen-
cephalic brain transplantation programs,
where hundreds of patients were treated
in open trials before investigators became
aware of serious side-effects prohibiting
the program continuation, it is clear that
obtaining controlled data is mandatory
from the beginning of a clinical develop-
ment program.
An extensive and thorough discussion of
ethical questions regarding sham surgeries
in trials for surgical treatment of PD pre-
ceded the realization of Freed and Fahns
placebo-controlled fetal mesencephalic
transplantation trial [66, 67].
The Spheramine placebo-controlled trial
is believed to fulfill the prerequisites for
ethical conduct of such studies that were
specified in the course of the discussion
and were confirmed by the New England
Journal of Medicine Sounding Board:
Preclinical and clinical signals that the
expected effect is large enough to justify
interventions also in placebo patients.
Minimum invasiveness and risk for pla-
cebo or sham procedures.
Reliable blinding to ensure data quality.
Statistical and clinical study design up
to the state of the art.
This argumentation could successfully be
mediated to ethic committees and institu-
tional review boards. Eleven of 12 boards
in the USA and 11 of 12 in Europe ap-
proved the protocol. All review boards
were very committed to the topic and took
considerable efforts to come to an agree-
ment in their committee discussions. One
refusal was received from a center in Den-
ver, Colorado, USA, and one from the Uni-
versity of Pamplona, Spain.
The patient population to be investigated
in this study is men and women aged 30
70 years with PD who have had symptoms
for at least 5 years, who are in Hoehn and
Yahr stages III and IV of the disease, and
who are responsive to dopaminergic treat-
ment, but have fluctuations and insuffi-
cient symptom control under optimized
pharmacotherapy.
The total sample size will be 68 patients,
34 of these will be assigned to Spheramine
therapy and 34 will receive placebo treat-
ment.
14.4.1
Spheramine Treatment
Spheramine treatment in this study will
consist of bilateral administration of
325000 cells on microcarriers per side,
administered by a neurosurgeon via stereo-
tactic neurosurgery. Five needle tracts going
through one burr hole on each side will be
needed to inject Spheramine into the post-
commissural putamen (see Fig. 14.13).
14.4.2
Dosing
At this point, it may be useful to spend a
few thoughts on dosing of Spheramine
and why bilateral surgery is performed in
this study, in contrast to what was done in
the pilot study.
The dose of 325000 cells applied in the
pilot study in six patients was derived
from animal experiments demonstrating
that in M. mulatta monkeys, a deposit of
around 60000 cells per tract, in five unilat-
eral tracts, reduced MPTP-induced parkin-
sonian symptoms in the treated animals,
whereas a dose of around 12000 cells per
tract did not.
The dose of 325000 cells administered
unilaterally prompted substantial improve-
ment of up to around 60% in UPDRS Mo-
tor Score in defined OFF in all six patients
treated in a pilot study and no serious
safety concerns. Transient signs of over-
14 Spheramine

dosing (hallucinations, peak-dose dyskine-
sias) could be reversed by reduction of oral
medication. Since the dose of 325000 cells
per side was found to be safe, it would not
be ethical to have patients undergo the
risk of surgery and anesthesia for a lower
dose that would potentially be ineffective.
Determination of a minimal effective dose
as in a conventional drug development
program is therefore not warranted in this
cell therapy program.
Identification of a maximum tolerable
dose is also not ethical in a treatment that
is irreversible. Defining a maximum toler-
able dose per se includes the risk of over-
dosing. In treatments where such overdos-
ing would be irreversible, sufficient safe
distance from a maximum dose must be
respected.
In experimental interventional PD thera-
pies other than Spheramine, bilateral sur-
gery is thought to provide better symptom
control than unilateral treatment. It is re-
garded as standard procedure for surgical
PD therapies. Although PD patients have
asymmetric symptoms at the onset of their
disease, symptoms will be symmetric in
stages Hoehn and Yahr III and IV. Bilat-
eral surgery is thought to provide sym-
metric symptom control and to avoid po-
tential symptom asymmetry that could re-
sult should only one side be treated. The
latter might create an intractable dilemma
where either one body side is optimally
treated and the other is insufficiently con-
trolled or one side is overdosed while the
other is sufficiently treated.
Although Spheramines effect in the
brain is locally restricted due to its small dif-
fusion diameter, the dose of 325000 cells to
be administered bilaterally (650000 cells to-
tal per patient) represents a small functional
dose increase per body side as a small effect
to the homolateral side must be expected.
This small functional dose increase is con-
sidered safe because all patients will con-
tinue to take oral anti-PD medication and
a sufficient buffer of oral dopaminergics
that can be reduced to counteract potential
overdosing is warranted.
Fig. 14.13 Stereotactic placement of Spheramine into the putamen.
14.4.3
Placebo Treatment
Patients assigned to placebo therapy will re-
ceive a sham surgical procedure including
skin incisions and drilling of incomplete
burr holes on both sides of the skull. The
burr hole will penetrate the exterior lamina
compacta of the skull and the spongiosa,
but will leave the lamina compacta interna
and the meninges intact. No penetration
of the brain will occur and nothing will be
injected. The procedure ensures that both
the patient and his or her treating neurolo-
gist cannot judge what therapy was admi-
nistered by palpation of the skull.
All other circumstances and conditions
of neurosurgical treatment will be the
same as for Spheramine patients. A stereo-
tactic frame will be fixed to the patients
head, parts of the head will be shaven for
surgery, and imaging and preparation in
the operating room as well as anesthesia,
both method and duration, will be the
same for placebo and Spheramine pa-
tients.
14.4.4
Safety Steps
In order to allow for efficient safety moni-
toring of the study, an independent data
monitoring committee composed of inter-
national experts in movement disorder
and stereotactic neurosurgery will analyze
safety data at predefined intervals during
the study. The first safety analysis will be
performed when 1-month data from 12 pa-
tients is complete, the second will be with 1-
month data from 36 patients, and the third
and last when all patients have reached the
1-month time point after surgery.
The first and second safety analyses
have been performed to date, with no
safety concerns resulting in discontinua-
tion, hold or major protocol amendments
of the trial.
14.4.5
Study Procedures
14.4.5.1 Blinding
Blinding of the patient and the treating
neurologist investigator is the biggest chal-
lenge in this clinical trial. Sophisticated
blinding instructions were established, and
all investigators and their staff had exten-
sive training to become familiar with the
blinding rules.
With the support of an experienced con-
tract research organization, an electronic
access system secured by passwords was
created to guarantee that only the neuro-
surgeon in charge, the radiologist and the
central magnetic resonance imaging (MRI)
evaluator have access to postsurgical MRI
images.
Stringent rules for the information flow
between the neurosurgeon and the treat-
ing neurologist were established, and their
compliance controlled and audited with
the most minute care.
The key feature to ensure maintenance
of the blind is the geographic separation
of the treating neurologist and the neuro-
surgeon in charge of a patient. This mini-
mizes interactions of the respective staff.
Patients and their caregivers must travel
from their home town where they are fol-
lowed by the neurologist investigator to a
city where the neurosurgeon in charge re-
sides and they return after a stay of 710
days.
In addition, all case report forms con-
taining potentially unblinding information
and MRI images must be kept sequestered
from the remaining study documentation,
except if required for patient safety. This
information enters the database only im-
mediately before unblinding of the trial.
14 Spheramine

14.4.5.2 Assessment Tools
Recommendations made by an interna-
tional expert group for study protocols in-
vestigating surgical therapies for PD (CAP-
SIT protocol) [68] were followed wherever
appropriate.
14.4.5.3 Efficacy Parameters
The primary efficacy endpoint, as in most
PD studies, is the UPDRS Part III (Motor
Score) measured in OFF, which means
after the patient has not taken any anti-PD
medication for at least 12 h. Secondary ef-
ficacy endpoints are the total UDPRS both
in ON (while the patient is on his normal
anti-PD medication regimen) and OFF,
the UPDRS Part III in ON, relative on
and off times (see below), the amount of
L-DOPA reduction, the ADL Subscale of
the UPDRS, and Quality of Life as as-
sessed by the PDQ-39, SF-36 and EQ-5D
questionnaires. [In the literature on PD,
the terms on and off are used both for
(1) the description of periods with good
and poor symptom control, respectively,
that occur despite regular medication in-
take, and (2) for characterization of the pa-
tients status with regard to therapeutic
drug levels in the blood. In the context of
this article, on and off are used in the
lower case for the characterization of fluc-
tuations, and ON and OFF (expressed
as defined ON and defined OFF) are
used in the upper case for describing the
patients pharmacologic status.]
14.4.5.4 Safety Parameters
The following parameters will be listed and
analyzed both for the final analysis and for
the interim safety evaluations by the inde-
pendent data monitoring committee:
Adverse events
Laboratory variables
Vital signs
Electrocardiogram
Physical and neurological examination
Concomitant medication
Brain MRI
Videotaping (with particular attention to
late-running or OFF dyskinesias)
UPDRS, in particular dyskinesia rating
14.4.5.5 Duration and Unblinding
In accordance with the CAPSIT protocol,
the study includes a pre-treatment period
of 3 months during which all inclusion
and exclusion criteria for patients are care-
fully checked, risk factors for surgery are
carefully excluded or controlled, optimized
and stable anti-PD medication is main-
tained, and the diagnosis of sporadic or
idiopathic PD is reconfirmed at least 3
times prior to surgery.
Prior to signing an informed consent
form, the patient must make use of at
least 6 days reflection time. He or she will
complete a total of five visits prior to sur-
gery, three of them 2-day visits with
UPDRS assessment in ON and in OFF.
He or she will then travel to the site of
the neurosurgeon in charge (e.g., patients
from Emory University, Atlanta travel to
Chicago or European patients from Dres-
den, Germany travel to Innsbruck, Aus-
tria), where they receive Spheramine or
sham surgery, whichever their treatment
assignment is. They return under the care
of the treating neurologist after 610 days
and complete a total of seven visits in the
first year of follow-up (five of them 2-day
visits).
Twelve months after the last patient has
had surgery, the study will be unblinded.
The total duration of follow-up, however, is
24 months. Four more 2-day visits must
be completed in the second year to ensure
appropriate duration of safety assessments.
The rationale behind these split endpoints
(12 months for efficacy, and 24 months for
safety and sustained efficacy) lies in the
potential deterioration of patients during a
long follow-up. At least placebo patients
should have the option of obtaining other
surgical therapies if their condition has
worsened so much that alternative surgery
is warranted, in the opinion of the treating
neurologist. It should be noted that pa-
tients are informed of the availability of
deep brain stimulation before they are en-
rolled into the trial and will normally not
consider deep brain stimulation an option
for themselves; however, they could
change their mind if they experience se-
vere deterioration of their condition.
If safety and efficacy of Spheramine is
confirmed in this study, placebo patients
will be offered Spheramine therapy after
the trial is complete, i.e., 24 months after
the last patient has been treated.
Patients will be asked to participate in a
long-term follow-up study under a separate
protocol covering the time of 5 years after
surgery.
So far, 36 patients have been treated in
this trial and their safety assessment by
the independent data monitoring commit-
tee was favorable. In all, 68 individuals are
enrolled into the study, 32 of them are
waiting to undergo surgery which restarted
in September 2004. Surgeries were fina-
lized in February 2005. Efficacy data from
this trial can thus be expected in the first
or second quarter 2006.
14.5
Summary and Outlook
Spheramine cell therapy for PD represents
an approach for achieving physiological,
even and continuous substitution of dopa-
mine in the brain of PD patients who can
no longer expect satisfactory symptom
control by oral medications and who are
not suitable for, or are not willing to un-
dergo, other surgical therapies. Sphera-
mine is administered in one session of
stereotactic neurosurgery; no follow-up op-
erations or tuning sessions, such as for
deep brain stimulation, will be required.
The immediate risk of surgery is thought
to be the same as for deep brain stimula-
tion, but the cumulative risk, including
that caused by hardware remaining in the
brain and repeated surgery in the case of
electric stimulator implantation, may be
lower.
At this point in time, promising results
from a pilot study in six patients followed
over more than 36 months require confir-
mation. Preliminary safety results from an
ongoing double-blind, placebo-controlled
study in 68 patients are encouraging.
The Spheramine approach does not
claim to provide neuroprotection or neu-
roregeneration, as has been asserted for
many other programs which failed to
prove efficacy. However, it is conceivable
that timely guarantee of continuous dopa-
minergic stimulation, before treatment
complications develop, may extend the
timespan of adequate and satisfactory
symptom control in PD patients, making
Spheramine an efficient and promising
modern biopharmaceutical.
References
1 Coombs BD, Best A, Brown MS, Miller DE,
Corboy J, Baier M, Simon JH. Multiple sclero-
sis pathology in the normal and abnormal ap-
pearing white matter of the corpus callosum
by diffusion tensor imaging. Multiple Sclerosis
2004, 10, 392397.
2 Calne DB, Mizuno Y. The neuromythology of
Parkinsons Disease. Parkinsonism Relat Disord
2004, 10, 319322.
14 Spheramine

3 Tanner CM. Is the cause of Parkinsons dis-
ease environmental or hereditary? Evidence
from twin studies. Adv Neurol 2003, 91, 133
142.
4 Strickland D, Bertoni JM. Parkinsons preva-
lence estimated by a state registry. Mov Disord
2004, 19, 318323.
5 Wooten GF, Currie LJ, Bovbjerg VE, Lee JK,
Patrie J. Are men at greater risk for Parkin-
sons disease than women? J Neurol Neurosurg
Psychiatry 2004, 75, 637639.
6 Zhang ZX, Roman GC. Worldwide occurrence
of Parkinsons disease: an updated review.
Neuroepidemiology 1993, 12, 195208.
7 de la Fuente-Fernandez R, Calne DB. Evidence
for environmental causation of Parkinsons
disease. Parkinsonism Relat Disord 2002, 8,
235241.
8 Powers KM, Smith-Weller T, Franklin GM,
Longstreth WT Jr, Swanson PD, Checkoway
H. Parkinsons disease risks associated with
dietary iron, manganese, and other nutrient
intakes. Neurology 2003, 60, 17611766.
9 Greenamyre JT, Betarbet R, Sherer TB. The
rotenone model of Parkinsons disease: genes,
environment and mitochondria. Parkinsonism
Relat Disord 2003, 9 (Suppl 2), S5964.
10 Collier TJ, Steece-Collier K, Kordower JH. Pri-
mate models of Parkinsons disease. Exp Neu-
rol 2003, 183, 258262.
11 Muller T, Buttner T, Gholipour AF, Kuhn W.
Coenzyme Q10 supplementation provides
mild symptomatic benefit in patients with Par-
kinsons disease. Neurosci Lett 2003, 341, 201
204.
12 Chen H, Zhang SM, Hernan MA, Schwarz-
schild MA, Willett WC, Colditz GA, Speizer
FE, Ascherio A. Nonsteroidal anti-inflamma-
tory drugs and the risk of Parkinson disease.
Arch Neurol 2003, 60, 10591064.
13 Braak H, Del Tredici K, Rub U, de Vos RA,
Jansen Steur EN, Braak E. Staging of brain
pathology related to sporadic Parkinsons dis-
ease. Neurobiol Aging 2003, 24, 197211.
14 Przuntek H, Muller T, Riederer P. Diagnostic
staging of Parkinsons disease: conceptual as-
pects. J Neural Transm 2004, 111, 201216.
15 Jenner P. Oxidative stress in Parkinsons dis-
ease. Ann Neurol 2003, 53 (Suppl 3), S2636.
16 Friedlander RM. Apoptosis and caspases in
neurodegenerative diseases. N Engl J Med
2003, 348, 13651375.
17 Tieu K, Perier C, Vila M, Caspersen C, Zhang
HP, Teismann P, Jackson-Lewis V, Stern DM,
Yan SD, Przedborski S. L-3-hydroxyacyl-CoA
dehydrogenase II protects in a model of Par-
kinsons disease. Ann Neurol 2004, 56, 5160.
18 Shen J, Cookson MR. Mitochondria and dopa-
mine; new insights into recessive parkinson-
ism. Neuron 2004, 43, 301304.
19 Persad AS, Stedeford T, Tanaka S, Chen L,
Banasik M. Parkinsons disease and CYP2D6
polymorphism in Asian populations: a meta-
analysis. Neuroepidemiology 2003, 22, 357361.
20 Nutt JG, Carter JH, Sexton GJ. The dopamine
transporter: importance in Parkinsons disease.
Ann Neurol 2004, 55, 766773.
21 Singleton AB, Farrer M, Johnson J, Singleton
A, Hague S, Kachergus J, Hulihan M, Peura-
linna T, Dutra A, Nussbaum R, Lincoln S,
Crawley A, Hanson M, Maraganore D, Adler
C, Cookson MR, Muenter M, Baptista M,
Miller D, Blancato J, Hardy J, Gwinn-Hardy
K. a-Synuclein locus triplication causes Par-
kinsons disease. Science 2003, 302, 841.
22 Meredith GE, Halliday GM, Totterdell S. A
critical review of the development and impor-
tance of proteinaceous aggregates in animal
models of Parkinsons disease: new insights
into Lewy body formation. Parkinsonism Relat
Disord 2004, 10, 191202.
23 Lohmann E, Periquet M, Bonifati V, Wood
NW, De Michele G, Bonnet AM, Fraix V,
Broussolle E, Horstink MW, Vidailhet M, Ver-
pillat P, Gasser T, Nicholl D, Teive H, Raskin
S, Rascol O, Destee A, Ruberg M, Gasparini
F, Meco G, Agid Y, Durr A, Brice A, French
Parkinsons Disease Genetics Study Group;
European Consortium on Genetic Susceptibili-
ty in Parkinsons Disease. How much pheno-
typic variation can be attributed to parkin
genotype? Ann Neurol 2003, 54, 176185.
24 Farrer M, Chan P, Chen R, Tan L, Lincoln S,
Hernandez D, Forno L, Gwinn-Hardy K, Pet-
rucelli L, Hussey J, Singleton A, Tanner C,
Hardy J, Langston JW. Lewy bodies and par-
kinsonism in families with parkin mutations.
Ann Neurol 2001, 50, 293300.
25 McNaught KS, Olanow CW. Proteolytic stress:
a unifying concept for the etiopathogenesis of
Parkinsons disease. Ann Neurol 2003, 53
(Suppl 3), S7384.
26 McNaught KS, Perl DP, Brownell AL, Olanow
CW. Systemic exposure to proteasome inhibi-
References 349
tors causes a progressive model of Parkinsons
disease. Ann Neurol 2004, 56, 149162.
27 Hunot S, Hirsch EC. Neuroinflammatory pro-
cesses in Parkinsons disease. Ann Neurol
2003, 53 (Suppl 3), S4958.
28 Hartmann A, Hunot S, Hirsch EC. Inflamma-
tion and dopaminergic neuronal loss in Par-
kinsons disease: a complex matter. Exp Neurol
2003, 184, 561564.
29 Ringheim GE, Conant K. Neurodegenerative
disease and the neuroimmune axis (Alzhei-
mers and Parkinsons disease, and viral infec-
tions). J Neuroimmunol 2004, 147, 4349.
30 Hornykiewicz O. Dopamine miracle: from
brain homogenate to dopamine replacement.
Mov Disord 2002, 17, 501508.
31 Factor SA, Molho ES. Transient benefit of
amantadine in Parkinsons disease: the facts
about the myth. Mov Disord 1999, 14, 515
517.
32 Bonuccelli U. Comparing dopamine agonists
in Parkinsons disease. Curr Opin Neurol 2003,
16 (Suppl 1), S1319.
33 Schapira AH, Olanow CW. Neuroprotection in
Parkinson disease: mysteries, myths, and mis-
conceptions. J Am Med Ass 2004, 291, 358364.
34 Nutt JG. Long-term L-DOPA therapy: chal-
lenges to our understanding and for the care
of people with Parkinsons disease. Exp Neurol
2003, 184, 913.
35 Shulman LM. Levodopa toxicity in Parkinson
disease: reality or myth? Reality practice pat-
terns should change. Arch Neurol 2000, 57,
406407.
36 Metman LV, Gillespie M, Farmer C, Bibbiani
F, Konitsiotis S, Morris M, Shill H, Bara-Jime-
nez W, Mouradian MM, Chase TN. Continu-
ous transdermal dopaminergic stimulation in
advanced Parkinsons disease. Clin Neurophar-
macol 2001, 24, 163169.
37 Linazasoro G. Recent failures of new potential
symptomatic treatments for Parkinsons dis-
ease: causes and solutions. Mov Disord 2004,
19, 743754.
38 de Bie RM, de Haan RJ, Schuurman PR, Esse-
link RA, Bosch DA, Speelman JD. Morbidity
and mortality following pallidotomy in Parkin-
sons disease: a systematic review. Neurology
2002, 58, 10081012.
39 Okun MS, Vitek JL. Lesion therapy for Parkin-
sons disease and other movement disorders:
update and controversies. Mov Disord 2004,
19, 375389.
40 Kumar R, Lang AE, Rodriguez-Oroz MC, Lo-
zano AM, Limousin P, Pollak P, Benabid AL,
Guridi J, Ramos E, van der Linden C, Vande-
walle A, Caemaert J, Lannoo E, van den Ab-
beele D, Vingerhoets G, Wolters M, Obeso JA.
Deep brain stimulation of the globus pallidus
pars interna in advanced Parkinsons disease.
Neurology 2000, 55 (12 Suppl 6), S34S39.
41 Kumar R, Lozano AM, Kim YJ, Hutchison
WD, Sime E, Halket E, Lang AE. Double-blind
evaluation of subthalamic nucleus deep brain
stimulation in advanced Parkinsons disease.
Neurology 1998, 51, 850855.
42 Saint-Cyr JA, Trepanier LL, Kumar R, Lozano
AM, Lang AE. Neuropsychological conse-
quences of chronic bilateral stimulation of the
subthalamic nucleus in Parkinsons disease.
Brain 2000, 123, 20912108.
43 Freed CR, Greene PE, Breeze RE, Tsai WY,
DuMouchel W, Kao R, Dillon S, Winfield H,
Culver S, Trojanowski JQ, Eidelberg D, Fahn
S. Transplantation of embryonic dopamine
neurons for severe Parkinsons disease. N Engl
J Med 2001, 344, 710719.
44 Olanow CW, Goetz CG, Kordower JH, Stoessl
AJ, Sossi V, Brin MF, Shannon KM, Nauert
GM, Perl DP, Godbold J, Freeman TB. A dou-
ble-blind controlled trial of bilateral fetal ni-
gral transplantation in Parkinsons disease.
Ann Neurol 2003, 54, 403414.
45 Fink JS, Schumacher JM, Ellias SL, Palmer
EP, Saint-Hilaire M, Shannon K, Penn R,
Starr P, VanHorne C, Kott HS, Dempsey PK,
Fischman AJ, Raineri R, Manhart C, Dins-
more J, Isacson O. Porcine xenografts in Par-
kinsons disease and Huntingtons disease pa-
tients: preliminary results. Cell Transplant
2000, 9, 273278.
46 Drucker-Colin R, Verdugo-Diaz L. Cell trans-
plantation for Parkinsons disease: present sta-
tus. Cell Mol Neurobiol 2004, 24, 301316.
47 Dismore J, Ratliff J, Deacon T, Pakzaban P,
Jacoby G, Galpern W, Isacson O. Embryonic
stem cells differentiated in vitro as a novel
source for transplantation. Cell Transplant
1996, 5, 131143.
48 Thomson JA, Odorico JS. Human embryonic
stem cells and embryonic germ cell lines.
Trends Biotechnol 2000, 18, 5357.
49 Buzanska L, Machaj EK, Zablocka B, Podja Z,
Domanska-Janik K. Human cord blood-de-
rived cells attain neuronal and glial features.
J Cell Sci 2002, 115, 21312138.
14 Spheramine

50 Kim JH, Auerbach JM, Rodriguez-Gomez JA,
Velasco I, Gvin D, Lumelsky N, Lee SH,
Nguyen J, Sanchez-Pernaute R, Bankiewicz K,
McKay RDG. Dopamine neurons from em-
bryonic stem cells function in an animal mod-
el of Parkinsons disease. Nature 2002, 418,
5056.
51 Studer L, Tabar V, McKay RDG. Transplanta-
tion of mesencephalic precursors leads to re-
covery of Parkinsonian rats. Nat Neurosci
1998, 1, 290295.
52 Song HJ, Stevens CF, Gage FH. Neural stem
cells from the hippocampus develop essential
properties of functional CNS neurons. Nat
Neurosci 2002, 5, 438445.
53 Kondziolka D, Wechsler L, Goldstein S, Melt-
zer C, Thulborn KR, Gebel J, Janetta P, DeCe-
sare S, Elder EM, McGrogan M, Reitman MA,
Bynum L. Transplantation of cultured human
neuronal cells for patients with stroke. Neu-
rology 2000, 55, 565569.
54 Barker RA. Repairing the brain in Parkinsons
disease: where next? Mov Disord 2002, 17,
233241.
55 McKay RD. Stem cell biology and neurode-
generative disease. Philos Trans R Soc Lond B
Biol Sci 2004, 359, 851856.
56 Porena M, Parziani S, Costantini E, Vespasia-
ni G, Micali F. Autologous adrenal medullary
transplant in Parkinsons disease: critical re-
view of our results in 13 patients. Neurourol
Urodyn 1996, 15, 195201.
57 Rafael H. Autotransplantation of human caro-
tid body cell aggregates for treatment of Par-
kinsons disease. Neurosurgery 2004, 54, 1035
1036.
58 Arjona V, Minguez-Castellanos A, Montoro
RJ, Ortega A, Escamilla F, Toledo-Aral JJ, Par-
dal R, Mendez-Ferrer S, Martin JM, Perez M,
Katati MJ, Valencia E, Garcia T, Lopez-Barneo
J. Autotransplantation of human carotid body
cell aggregates for treatment of Parkinsons
disease. Neurosurgery 2003, 53, 321328.
59 Hurelbrink CB, Barker RA. The potential of
GDNF as a treatment for Parkinsons disease.
Exp Neurol 2004, 185, 16.
60 Palfi S, Leventhal L, Chu Y, Ma SY, Emborg
M, Bakay R, Deglon N, Hantraye P, Aebischer
P, Kordower JH. Lentivirally delivered glial
cell line-derived neurotrophic factor increases
the number of striatal dopaminergic neurons
in primate models of nigrostriatal degenera-
tion. J Neurosci 2002, 22, 49424954.
61 Nutt JG, Burchiel KJ, Comella CL, Jankovic J,
Lang AE, Laws ER Jr, Lozano AM, Penn RD,
Simpson RK Jr, Stacy M, Wooten GF; ICV
GDNF Study Group. Randomized, double-
blind trial of glial cell line-derived neuro-
trophic factor (GDNF) in PD. Neurology 2003,
60, 6973.
62 Chase T. The significance of continuous dopa-
minergic stimulation in the treatment of Par-
kinsons disease. Drugs 1998, 55 (Suppl 1), 1
9.
63 Marmor MF, Wolfensberger TJ. The Retinal
Pigment Epithelium: Function and Disease. Ox-
ford University Press, New York, 1998.
64 Goetz CG, Stebbins GT, Blasucci LM. Differ-
ential progression of motor impairment in le-
vodopa-treated Parkinsons disease. Mov Disord
2000, 15, 479484.
65 Nutt JG, Carter JH, Lea ES, Sexton GJ. Evolu-
tion of the response to levodopa during the
first 4 years of therapy. Ann Neurol 2002, 51,
686693.
66 Macklin R. The ethical problems with sham
surgery in clinical research. N Engl J Med
1999, 341, 992996.
67 Freeman TB, Vawter DE, Leaverton PE, God-
bold JH, Hauser RA, Goetz CG, Olanow CW.
Use of placebo surgery in controlled trials of a
cellular-based therapy for Parkinsons disease.
N Engl J Med 1999, 341, 988992.
68 Defer GL, Widner H, Marie RM, Remy P, Le-
vivier M. Core assessment program for surgi-
cal interventional therapies in Parkinsons dis-
ease (CAPSIT-PD). Mov Disord 1999, 14, 572
584.
References 351
Abstract
Organ transplantation has been one of the
major medical advances of the past 30
years; however, it is becoming increasingly
apparent that the supply of organs is lim-
ited and will not improve with current
medical practice. Organogenesis repre-
sents an alternative to combat organ short-
age. Organogenesis of complex tissues,
such as the kidney, requires a coordinated
sequential transformation process, with in-
dividual stages involving time-dependent
expression of cellcell, cellmatrix, and
cellsignal interactions in three dimen-
sions. Embryonic precursor tissues are
composed of functionally diverse stem/
progenitor cell types that are organized in
spatially complex arrangements. The
theme of temporalspatial patterning of
progenitor cell interactions is programmed
in precursor tissues leading to their
growth and development. Indeed, recent
data pinpoints a window of time in hu-
man and pig kidney organogenesis that
may be optimal for transplantation into
mature recipients. Window transplants
are defined by their remarkable ability to
grow, differentiate and undergo vasculari-
zation, achieving successful organogenesis
of urine-producing miniature kidneys with
no evidence of trans-differentiation into
non-renal cell types, lack of tumorigenicity
and reduced immunogenicity, compared to
adult counterparts. In contrast, non-win-
dow transplants (earlier or later in gesta-
tion) can respectively form teratomas or
are more prone to immune rejection, and
are both less suitable for organogenesis.
Thus, when organogenesis can be success-
fully achieved in situ, it may provide the
optimal approach for replacement of or-
gans.
Abbreviations
a1,3GT a1,3-galactosyltransferase
APC antigen-presenting cells
ESRD end-stage renal disease
HBV hepatitis B virus
HCV hepatitis C virus
hDAF human decay-accelerating factor
MHC major histocompatibility complex
MSC mesenchymal stem cells
PBMC peripheral blood mononuclear
cells
PERV porcine endogenous retrovirus
SCID severe combined immunodefici-
ency
UNOS United Network for Organ Shar-
ing
353
15
Applying Human Cells to Organogenesis and Transplantation
ISBN: 3-527-31184-X
15.1
Growing Demands for Kidney Allograft
Transplantation
Organ transplantation has been one of the
major medical advances of the past 30
years. Growing number of patients can be
transplanted successfully due to the con-
tinuing progress in surgery and medical
treatments, and the development of new
immunosuppressive drugs [1]. About 1
million people have received an organ, 73
kidneys from cadaver donors have survived
more than 25 years, and 17 heart and 27
liver transplant recipients for more than
15 years [2, 3]. While the transplant com-
munity attempts to keep up with the in-
creasing demand for transplantable or-
gans, the supply continues to fall far short
of the need. This also applies to kidney
transplant programs. It has been estimated
that the number of patients with end-stage
renal disease (ESRD) for most of whom
renal transplantation is the treatment of
choice is increasing at the rate of 78%
per year in the United States [4]. The Unit-
ed Network for Organ Sharing (UNOS) da-
tabase shows that between 1988 and April
2002, the number of patients on a waiting
list for renal transplantation increased
from 13943 to 51753 patients [5]. From
1988 to 2000, the patients waiting time
has almost tripled, from a median of 400
days to more than 1100 days [5], and the
number of patients who have died every
year while awaiting a cadaveric renal trans-
plantation has increased from 736 to 2875
an increase of 290% [5]. Moreover, in
2000 only 26% of the patients on the wait-
ing list for a kidney transplant actually un-
derwent renal transplantation [5]. In most
European countries, the waiting lists for
kidney transplantation are also in a similar
or worse condition than a decade ago [2],
making the severe shortage of cadaveric or-
gan donors the major obstacle in prevent-
ing the full development of a transplant
program, and imposing a severe limit to
the number of patients who benefit from
this form of therapy worldwide. The chal-
lenge of expanding the donor pool has
forced the scientific community to explore
new avenues for organ replacement.
15.2
Alternative Sources for Human Renal
Allografts
Recently, much excitement has been fo-
cused on the development of organs, tissues
and cells for the purpose of restoring func-
tion through transplantation [6]. The idea
that replacement, repair and restoration of
function is best accomplished by cells, tis-
sues or organs that can perform the appro-
priate physiologic/metabolic duties better
than any mechanical device, recombinant
protein therapeutic or chemical compound
could be also applicable to the kidney. The
various approaches differ in source and type
of biological material applied.
15.2.1
Adult Kidney Organ Xenotransplants
The transplantation of organs from ani-
mals into humans (i.e., xenotransplanta-
tion) has been a long-sought objective in
clinical practice [7]. While it may be intui-
tive that the best xenogeneic donor for
clinical use would be a species closely re-
lated to humans, most investigators now
focus on the pig as a potential donor. The
reasons for using pigs include the follow-
ing:
Porcine organs are of an appropriate
size for use in humans.
The supply of pigs, unlike the supply of
primates, is unlimited.
Pigs can be genetically engineered,
whereas primates presently cannot.
The risk of zoonotic infection from pigs
is limited and more easily controlled.
Nevertheless, the obstructions to the use
of pig organs are significant, with the
most fearsome problems being severe re-
jection reactions and the transmission of
porcine endogenous retroviruses [8, 9]. A
pig organ, when transplanted into an un-
modified non-human primate or human,
is subject to hyperacute rejection [8, 10].
In pig-to-human transplantation, hypera-
cute rejection is initiated by the binding of
human xenoreactive IgM to the endotheli-
um of the porcine organ. More than 90%
of xenoreactive antibodies that bind to por-
cine organs are specific for Gala1-3Gal.
This antigen is synthesized by the enzyme
a1,3-galactosyltransferase (a1,3GT), which
exists in lower mammals and New World
monkeys, but is absent in Old World mon-
keys, apes and humans [1014]. Upon
binding to glycoproteins and/or glycolipids
bearing Gala1-3Gal on the endothelial sur-
face, these antibodies activate complement,
setting into motion a series of events lead-
ing to the destruction of the newly trans-
planted organ [10, 15]. In non-human pri-
mates, the hyperacute rejection of a trans-
planted pig organ can be prevented in sev-
eral ways, such as by depletion of
antibodies (e.g., by plasmapheresis or spe-
cific extracorporeal immunoadsorption of
anti-Gal antibodies), depletion of comple-
ment (e.g., with cobra-venom factor), or
the use of an organ that expresses a hu-
man complement regulatory protein, such
as decay-accelerating factor (hDAF) [16].
Nevertheless, the return or continuing
presence of anti-Gal antibodies ultimately
leads to a delayed form of antibody-
mediated rejection, acute humoral xeno-
graft rejection. An alternative approach to
overcoming the hyperacute rejection of xe-
nografts makes use of transgenic donors
expressing 1,2-fucosyltransferase (H-trans-
ferase), an enzyme competing with the
a1,3-galactosyltransferase, in endothelial
cells where this gene is normally not ex-
pressed [17]. The expression of the trans-
gene leads to altered expression of Gal-
a1,3Gal. More recently, the insertion of
such modifying genes (transgenesis) has
been replaced by disruption of the gene
for a1,3-galactosyltransferase in a donor
cell by homologous recombination and
subsequent nuclear transfer to generate an
animal from that modified cell [18]. In
July, 2002, the first homozygous galactosyl-
transferase-knockout pigs were success-
fully bred [19]. With the availability of ga-
lactosyltransferase-knockout pigs, the natu-
ral and elicited anti-Gal antibody problem
may be overcome. To what extent these
modifications will be sufficient for allow-
ing xenografting of organs in primates
and in man with classic immunosuppres-
sion remains to be determined.
15.2.2
Applying Embryonic Renal Progenitors
for Organogenesis and Transplantation
15.2.2.1 Mammalian Kidney Development
In early embryonic life, there is a distinct
stage which represents the direct develop-
mental origin of the mature kidney [20,
21] (Fig. 15.1a, b). This stage begins at 5
weeks of human gestation when a branch
of the Wolffian duct the ureteric bud
invades the metanephrogenic mesen-
chyme. Mutually inductive events cause
the ureteric bud to branch serially (a pro-
cess termed branching morphogenesis)
to form the collecting ducts, renal pelvis,
ureter, and bladder trigone, while the renal
mesenchyme undergoes mesenchymal-to-
epithelial conversion to form glomeruli,
proximal and distal tubules, and loops of
Henl. Only at 9 weeks of human gesta-
tion will primitive glomeruli appear. In the
developing human kidney, mesenchymal
cells are induced into the nephrogenic
pathway to form nephrons until 34 weeks
of gestation.
The existence of renal embryonic stem
cells in the metanephric mesenchyme is
supported by several types of evidence.
Previously, the presence of a single meta-
nephric mesenchymal cell that can gener-
ate all the epithelial elements of the ne-
phron, excluding the collecting tubule, was
established using a lineage marker [22].
This indicated the presence of renal
epithelial stem cells. Moreover, recent in
vitro [23] and in vivo [2426] data suggest
that cells residing in the metanephric mes-
enchyme have, under various experimental
conditions, the potential to differentiate, in
addition to renal epithelia, into other pro-
fessional cell types that participate in kid-
ney organogenesis (myofibroblasts, smooth
muscle, endothelium), as well as non-renal
derivatives including cartilage, bone and
blood.
Thus, the metanephric mesenchyme
contains multipotent progenitors or em-
bryonic renal stem cells with the ability to
generate, in concert with the ureteric bud,
many cell types in the mature kidney.
While the ultimate goal is the identifica-
tion of a single nephrogenic stem cell that
can build the entire organ, these precursor
tissues offer excellent starting material to
regenerate renal structures.
Over the past few years, we have studied
the transplantability of human embryonic
renal precursors in murine hosts [25, 27
29]. Because of the difficulties in obtaining
sufficient numbers of human embryos
(see Part I, Chapter 11), as well as the
ethical problems involved with the use of
human embryonic tissue we (as well as
others) have also used alternative embryo-
nic donor tissue obtained from rodents
[3034] and pigs [25, 35].
15.2.2.2 Defining a Gestational Window
for Optimal Growth and
Differentiation of Kidney
Progenitors Free of Teratoma Risk
If embryonic kidney precursors were to be
used for transplantation, extensive differ-
entiation into nephrons, together with the
organization of a collecting system re-
quired for drainage of urine, must take
place in vivo. That is, most of the meta-
nephric mesenchymal cells should convert
into nephron epithelia and also form glo-
meruli after grafting, while ureteric buds
should undergo branching morphogenesis.
In theory, this could be more difficult to
achieve for undifferentiated precursors de-
rived from early embryos, which contain
primarily metanephric blastema and a few
Fig. 15.1 (a) A simplified scheme of an embryonic
renal progenitor unit consisting of metanephric
mesenchymal stem cells and ureteric buds, which
cross-talk via growth factors (GFs) and their re-
ceptors, molecules of the extracellular matrix
(ECM) and specific integrins, proto-oncogenes
and specific ligands and give rise to the differen-
tiated cell types of the adult kidney (see text).
(b) Histology of early human kidney development.
Panel A: early human kidney precursor structures
(embryonic day 42): metanephros (mt), direct and
permanent precursor of the adult kidney; mesone-
phros (ms), transient precursor; gonad (go). Pa-
nels B and C: magnifications of the early metane-
phros; shown in panel B is a derivative of the
Wolffian duct (arrow), and in panel C a ureteric
bud (u) and the condensing metanephric me-
senchymal stem cells (cm). Panel D: a differen-
tiated fetal kidney harboring glomeruli (arrows)
and tubuli.
3
ureteric buds, compared to later-gestation
kidneys which have already differentiated
to a certain extent. Additional considera-
tions should take into account properties
that are inherent to undifferentiated and
pre-differentiated progenitor cells, such as
the ability to differentiate into several
lineages [36] and to undergo malignant
transformation in the form of embryonic
tumors after transplantation [37].
Evidence for the tremendous capacity of
kidney precursors to grow and differenti-
ate after transplantation comes from stud-
ies with murine [3034] and human tis-
sues [25, 2729]. We initially transplanted
human fetal kidney fragments derived
from mid-gestation pregnancies (1422
weeks) into immunodeficient mice and
showed that they exhibit rapid growth and
development [27, 28]. Because branching
and nephrogenesis continue to occur in
the outer rim of the human kidney (the
nephrogenic cortex) until 34 weeks, these
grafts continued to mature in vivo. We
then compared the differentiational capaci-
ty of kidney precursors obtained at differ-
ent time points of gestation [25]. While
whole-organ grafting can be applied for
the early kidney precursors, the develop-
ment of areas of graft necrosis prevents
the use of this methodology for later-gesta-
tion kidneys, and they must be implanted
in fragments, as previously described [27].
For organogenesis, the early undifferen-
tiated kidney precursors (78 weeks hu-
man gestation) were advantageous over
later-gestation kidneys, significantly grow-
ing more in size and differentiating into a
larger number of mature glomeruli and
tubules [25] (Fig. 15.2). These findings are
in accordance with the results of several
investigators [3034], who transplanted
embryonic day 15 (E14-E15) mouse and
rat kidney precursors and showed that
the latter form mature glomeruli and
tubules, as well as organized cortex and
medulla.
To increase the donor pool, we have also
analyzed the differentiational capacity of
pig kidney precursors, obtained at differ-
Fig. 15.2 Growth and differentiation of an early
human kidney precursor after transplantation into
mice (8 weeks post-transplant). (a) Macroscopic
view; note massive growth and the formed shape
of a kidney (arrow). (b) Histology (hematoxylin &
eosin staining; original magnification 10); note
the preserved architecture and differentiation into
layers of glomeruli and tubules. (c) Higher magni-
fication (original 40), showing developed glomer-
uli and tubules.
ent time points of gestation, in immuno-
deficient mice [25]. Similar to the human
transplants, we could show that grafting of
early pig kidney precursors, obtained from
E27-E28 embryos, achieves organogenesis
and leads to better growth and differentia-
tion compared to later-gestation pig kid-
neys. Because a pig gestation affords avail-
ability of very young embryos, we could
perform grafting of very early pig kidney
precursors (E21-E25) in immunodeficient
mice. Here, grafts did not differentiate ex-
clusively along the nephric lineage and
other differentiated derivatives such as car-
tilage, bone, blood vessels and myofibro-
blasts could be found. As stated before,
this finding complements recent in vitro
[23] and in vivo [24] data to suggest that
cells residing in the metanephric mesen-
chyme are pluripotent and can trans-differ-
entiate into non-renal derivatives especially
of mesodermal origin. Thus, we could de-
fine a window of opportunity for trans-
plantation where successful organogenesis
is achieved only when applying human or
porcine stage-specific kidney precursors.
To gain insight into the molecular sig-
nals which stimulate the human kidney
precursors to grow and develop after trans-
plantation, we determined to what extent
their transcriptional program resembles
that involved in induction of the normal
human kidney or the transformation into
an embryonic kidney malignancy (Wilms
tumor) [29]. Grafts originating from a 10-
week human gestation kidney were har-
vested at specific time points and cDNA
was hybridized onto nylon arrays repre-
senting 1200 genes. Gene expression pro-
files were compared to those obtained for
developing human gestation kidneys and
Wilms tumor. Strikingly, many of the de-
tails of the molecular program required to
generate and build a human nephron after
implantation into the mouse recipient
were similar, in a global sense, to normal
kidney induction. First, most of the ne-
phrogenesis genes, classified under cell
cycle regulators, transcription and growth
factors, signaling, transport, adhesion and
extracellular matrix molecules, which were
induced in the normal process were also
observed in the developing grafts. Aberrant
gene expression in the developing trans-
plants included a small group of mole-
cules which function in oxidative stress
and indicate possible early ischemia fol-
lowing the grafting procedure. Second,
comparison of temporal expression pro-
files demonstrated that the time-course for
development of normal human kidneys is
applicable to that for development of trans-
planted kidney precursors. Our approach
could also identify specific genes, includ-
ing growth factors, the expression levels of
which were lower in the transplants com-
pared to the normal kidneys at specific
time points. Development of strategies
aimed at increasing the levels of such
genes, might further enhance the differen-
tiational capacity of the developing grafts.
15.2.2.3 Host versus Donor Vascularization
of Kidney Precursors in Host
Animals
Transplanted kidney precursors rely on the
development of a vascular network in situ
for their engraftment and growth. To im-
prove the ability of the grafts to sustain an-
giogenesis in a foreign microenvironment,
it is of great importance to determine the
extent of vascularization and whether it is
derived from host or donor endothelial
cells. Moreover, hyperacute rejection is
thought to be mediated via the interaction
of preformed antibodies circulating in sus-
ceptible hosts with antigens present on the
endothelium of a xenotransplanted donor
organ [8]. In addition, donor endothelial
cells act as antigen-presenting cells (APC)
to mediate cellular rejection [38] and,
therefore, represent an immunological bar-
rier for xeno- and allo-transplantation.
Classical studies with murine metaneph-
ric tissue grown in organ culture or on the
avian chorio-allantoic membrane have sug-
gested that kidney endothelia arise from an-
giogenic ingrowth of extrinsic vessels [39].
However, when Hynik et al. [40] grafted
E11-E12 kidneys from normal mice into
anterior eye chambers of host transgenic
mice expressing beta-galactosidase in every
cell, they found beta-galactosidase activity
in the peripheral vessels, but not in glomer-
ular endothelial cells. Furthermore, Robert
et al. [41] grafted E12 kidneys immunola-
beled for the vascular endothelial growth
factor receptor, flk-1, into kidney cortices
of the adult and newborn beta-galactosidase
transgenic mice. In this model, cells ex-
pressing flk-1 but lacking galactosidase ac-
tivity are indicative of donor vasculogenic
angioblasts. Thus, when implanted into
adult recipients, the grafts exhibited only
few glomeruli containing host-derived en-
dothelium, whereas a majority of glomeruli
grafted into newborns contained host-de-
rived cells. These results suggested that,
rather than ingrowth of vessels into the de-
veloping kidney, endothelial precursors re-
siding in the metanephrogenic mesen-
chyme give rise, at least in part, to glomeru-
lar capillaries and microcirculation of the
developed kidney [21, 42]. In contrast, Ro-
gers et al. [34], using a non-cross-reactive
anti-mouse PECAM-1 antibody to stain for
host-type endothelial cells, showed that
upon xenoimplantation of E15 rat kidney
precursors into recipient mice, host-derived
cells could be detected in external vessels as
well as in some glomeruli.
We undertook the same approach to
analyze for mouse-derived endothelial cells
expressing PECAM-1 in developing trans-
plants of human and pig kidney precur-
sors [25]. Similar to Rogers et al. [34], we
found positive PECAM-1 immunoreactivity
in external vessels, as well as developing
glomeruli and small capillaries of grafts of
early human (78 weeks) and porcine
(E27-E28) kidney precursors, while grafts
of later-gestation kidneys had significantly
reduced host-derived vessel counts com-
prised mainly of external vessels. Our re-
sults suggested differential staining which
was dependent on the gestational age of
the donor at the time of grafting.
Fig. 15.3 Scheme demonstrating the origin of
blood vessels in the developing kidney trans-
plants. Upon transplantation of early embryonic
kidney precursors, a predominance of host-derived
vessels are observed (left). In contrast, when later-
gestation kidneys are transplanted, a predomi-
nance of donor-derived vessels is found (right).
Because later-gestation kidneys have al-
ready developed a vascular network includ-
ing vascularized glomeruli (donor-derived),
while the early kidney precursors are less
vascularized, prior to transplantation, re-
cipient mice contribute more to vasculo-
genesis of the latter, including the forma-
tion of the microcirculation, after grafting
(Fig. 15.3).
It seems prudent to conclude that trans-
planted early human and pig kidney pre-
cursors develop as chimeric organs in
which blood vessels are of both donor and
host origin, while external vessels required
for graft maintenance are mostly host-de-
rived. The exact relationship between do-
nor- and host-derived endothelial cells in
the formation of the microcirculation is
currently unknown.
15.2.2.4 Decreased Immunogenicity
of Kidney Precursors
It has been known for over four decades
that embryonic tissues are less immuno-
genic compared with their adult counter-
parts [43]. Nevertheless, additional studies
have indicated that different organs of fetal
origin have a variable propensity to be re-
jected [44]. Accordingly, mid-gestational rat
fetal kidney survived moderately in
outbred allograft experiments [44, 45],
while other fetal tissues such as skin,
small intestine, pancreas and liver were all
rapidly rejected [4648]. Moreover, observa-
tions in outbred models that fetal rat kid-
neys are less immunogenic, and that rejec-
tion was dependent on the age of the fetal
kidney, were confirmed in a fully allo-
geneic inbred model [49]. Similarly, pro-
longed survival was also exhibited by fetal
mouse renal grafts transplanted into adult
congenic mice [50]. In a series of intrigu-
ing experiments, Rogers et al. showed that
E15 rat kidney precursors can survive in
outbred adult rat hosts [32], and can also
be transplanted across defined major histo-
compatibility complex (MHC) barriers in
rats [33]. Furthermore, following allotrans-
plantation, E28 miniature pig kidney pre-
cursors underwent growth and differentia-
tion growth of nephrons over a 2-week per-
iod, without the need for immunosuppres-
sion of pig hosts [35].
To study the immunogenicity of the grow-
ing kidneys, we used the trimera model
which we previously developed in our labo-
ratory to investigate human immune re-
sponses [51]. The trimera technology makes
use of supra-lethal irradiation and radiopro-
tection with bone marrow from the severe
combined immunodeficiency (SCID)
mouse, so as to generate an immediate im-
mune space which enables the expansion of
human peripheral blood mononuclear cells
(PBMC), as well as the engraftment of dif-
ferent human tissues. Unlike SCID mice,
the recipient mice have, prior to irradiation,
preformed lymph nodes in which the in-
fused human T and B cells can form folli-
cles and interact so as to mount immune
responses, including very rapid antibody
switching and the generation of primary
antigen-specific human cytotoxic T lympho-
cytes [5254]. Thus, we initially used this
model to make fully human antibodies
against hepatitis B and hepatitis C [51, 55].
The trimera was also used as a model for
hepatitis B virus (HBV) and hepatitis C
virus (HCV) infection [51, 56]. Following
implantation of human liver fragments un-
der the kidney capsule, it became possible
to infect the mice with HBV or HCV and
to assess the infection by quantitative poly-
merase chain reaction (PCR). Thereby, the
model became useful for the assessment
of anti-hepatitis agents, including newly
made human anti-hepatitis antibodies.
More recently, we developed the trimera
model for the study of human kidney allo-
graft rejection (Fig. 15.4). Human kidney
fragments were transplanted, in conjunc-
tion with human lymphocytes, and the al-
loreactivity against the kidney implant by
the infused human T cells could be moni-
tored at both cellular and molecular levels
[57, 58]. Thus, infusion of allogeneic human
PBMC into the peritoneal cavity of animals
bearing a fragment of human adult kidney
implanted under the mouse kidney capsule,
leads to rapid destruction of the latter by a
humanized cellular rejection process.
The implantation of a 14-week human fetal
kidney fragment under the kidney capsule,
instead of the adult one, was associated with
reduced human T-cell infiltration and tissue
destruction, as well as with continued graft
growth in the first month after PBMC infu-
sion, but delayed graft rejection eventually
was observed during the second month
[27, 28]. More recently, by systematic titra-
tion of gestational age of the donor kidney
tissue in this model, we were able to define
a window in human embryonic develop-
ment when allogeneic human PBMC do
not recognize the kidney grafts and they
grow and develop to the same extent as
their counterparts not challenged by human
PBMC. This occurs when human early kid-
ney precursors (78 weeks of gestation) are
transplanted [25]. The same kind of analysis
performed for adult, mid-gestation and em-
bryonic porcine kidney tissues, similarly
demonstrated that early pig kidney precur-
sors (E27-E28) can escape rejection by xeno-
genic human PBMC [25]. While the very
early pig kidney precursors (E21-E25) are
also immune-privileged, they do not readily
differentiate into nephrons. Thus, our defi-
nition of the earliest time point in human
or pig renal gestation at which normal dif-
ferentiation and subsequent kidney func-
tion are possible, pinpoints the ideal time
for harvesting the tissue least prone to im-
mune rejection.
Interestingly, in contrast to allo-trans-
plantation experiments [32, 33] and those
carried out in the humanized animal
model [25, 27, 28], studies in which E15
rat [34] or E27-E28 pig [25, 35] kidney pre-
cursors were xenotransplanted into immu-
nocompetent mice resulted in instant graft
rejection. Nevertheless, these grafts were
salvaged when several short protocols of
co-stimulatory blockade were applied to
prevent rejection [25, 34, 35]. Moreover, we
could demonstrate that under co-stimula-
tory blockade the grafts of early pig kidney
precursors survive better than adult coun-
terparts, proving once more their reduced
immunogenicity in a stringent animal
model [25].
Fig. 15.4 Reduced immuno-
genicity of the early kidney
precursors (E28) compared to
later-gestation kidneys (E42)
after transplantation might be
related to decreased antigen-
presenting cells (APC) in the
early progenitors.
What are the mechanisms that underlie
the reduced immunogenicity of the hu-
man early kidney precursors? There are
several options: First, as stated before, the
early kidney precursors attract more host
vessels than later-gestation kidneys and
may therefore be less prone to cellular re-
jection and hyperacute rejection mediated
by donor endothelial cells [8, 38]. Second,
the direct pathway of rejection where pro-
fessional APC (dendritic cells) residing in
the graft that initiates the rejection pro-
cess, is impaired. It is possible that at the
time point of grafting the early kidney pre-
cursors, professional APC which are of he-
matopoietic origin, have not yet settled in
the developing tissue (see Fig. 15.4). This
is supported by the observations in the
humanized mouse model, in which only
the direct pathway of rejection is operative,
showing that early kidney precursors es-
cape immune destruction [25]. In addition,
the 78 weeks human kidney precursors
do not express crucial co-stimulatory mole-
cules, such as CD40 and B7-1, even after
implantation in the trimera model in con-
junction with human PBMC. In contrast,
developing human kidney tissue obtained
at later time points and implanted in the
trimera model under the same conditions,
exhibited up-regulation of these co-stimula-
tory molecules. These results suggest that
APC are either non-functional or absent
from the early human kidney precursors
[25]. Clearly, an incomplete CD40L: CD40
interaction can lead to a defective T helper
1 immune response [59], which has been
observed in developing human fetal kidney
grafts after human PBMC infusion [28].
Moreover, in rodents, Rogers et al. [33] have
shown that allogeneic hosts fail to accept
E15 rat kidney precursors when the latter
are transplanted in conjunction with ma-
ture rat skin, which serves as a source for
autologous professional APC.
While the importance of APC and host
vasculature levels in early embryonic kid-
ney precursor tissue cannot be underesti-
mated, the observed reduced immunogeni-
city may reflect progressive development
of a complex array of cell surface mole-
cules and soluble factors that determine
immune recognition in the fetal organ.
We profiled global gene expression in de-
veloping and adult human kidneys by
DNA microarrays, screened for immune-
related genes, and established that the de-
velopment of immunological maturity in
the human kidney is a rather late event in
gestation [25]. Altogether, the developing
kidneys (representing gestational time
points through which the developing
transplants progress) are restricted in mul-
tiple factors that determine immune recog-
nition. Thus, 13 of 57 immune genes
that are significantly up-regulated in adult
versus fetal kidney tissues belong to the
HLA class I and class II systems. In addi-
tion, molecules which mediate trafficking
of leukocytes into the graft, such as the
chemokines RANTES and MCP-1 [60], the
adhesion molecule E-selectin [61], pro-in-
flammatory cytokines such as osteopontin
[62] and complement genes known to be
associated with innate immunity [63], may
all be responsible for the reduced immu-
nogenicity of the developing kidneys.
15.2.2.5 Functionality of Kidney Precursors
after Transplantation
In-vivo differentiation of human kidney
precursors into functional mature ne-
phrons after transplantation is critical if
they were to be applicable as donor tissue
in clinical practice.
Studies in mice demonstrated glomeru-
lar filtration in donor nephrons of murine
metanephric tissue implanted into the re-
nal cortex of neonatal mice using fluores-
cein isothiocyanate (FITC)-labeled dextran
as a marker of filtration into the proximal
tubules [30]. Similarly, intravenous injec-
tions of antilaminin IgG into rats trans-
planted with fetal rat kidney tissue, re-
sulted in labeling of glomerular basement
membranes in the subcapsular grafted kid-
neys, confirming perfusion of the grafts
[31]. Because previous transplantation
studies of renal precursors obtained from
murine embryos [30] were unable to dem-
onstrate that donor nephrons become in-
corporated into the collecting system of
hosts, Rogers et al. [32] transplanted E15
rat kidney precursors in the omentum of
rat hosts (intra-abdominal grafting), so as
to render possible surgical anastomosis be-
tween the ureter (a ureteric bud derivative)
which develops as part of the transplant
and the hosts ureter (ureterostomy).
Following anastomosis, developing
grafts were shown to produce urine and to
clear inulin infused into hosts circulation.
Furthermore, inulin clearance was shown
to be increased by constant infusion of
IGF-1 into hosts, begun 4 weeks after the
implantation of E15 rat kidney precursors
[64]. If the timetable for development of
normal rodent kidneys is applicable to that
for development of such grafts, the en-
hancement of inulin clearance resulting
from IGF-1 administration occurred after
nephrogenesis was complete (3 weeks fol-
lowing birth).
We analyzed tubular function in devel-
oping human kidney grafts established in
immunodeficient mice, with
99
Te-DMSA
renography, the uptake of which occurs
through the peritubular side [29]. Positive
uptake of the radioisotope was demon-
strated only in tubules that matured in de-
veloped grafts. To further determine kid-
ney functionality, we measured levels of
urea nitrogen and creatinine in cyst fluid
collected from cysts arising from trans-
plants of the early human and pig kidney
precursors [25]. Large cysts were mostly
found in transplants established in the ab-
domen, and therefore were not limited by
the renal subcapsular space. Fluid was de-
rived by insertion of a microcatheter into
the developing renal grafts. The average
levels of urea nitrogen and creatinine were
higher in cyst fluid compared with those
found in the sera of transplanted mice, in-
dicating that the human and pig trans-
plants had produced urine. Levels of urea
nitrogen and creatinine in the cyst fluid
were significantly lower compared with na-
tive bladder urine. The dilute urine in the
cyst fluid is compatible with a reduced ca-
pacity of the fetal kidney to concentrate
urine. Thus, similar to previous reports
[3033], we were unable to demonstrate a
connection between donor human and pig
nephrons to the collecting system of hosts.
However, our transplants did integrate into
the hosts microenvironment and use its
blood vessels, and urine was produced sep-
arately from the native kidneys. Further ex-
perimentation should be developed to pro-
duce adequate urinary anastomosis and di-
version of blood supply to the kidney
grafts sufficient to correct biochemical
aberrations in a uremic individual. In-
creasing the number of transplants and/or
administering specific human growth fac-
tors might support functional replace-
ment.
15.2.2.6 In-vitro Propagation of Kidney
Precursors
Recent advances in the understanding of
the molecular biology of rodent renal de-
velopment have enabled the separate cul-
ture of the components of the developing
rat kidney, namely the ureteric bud and
the metanephric mesenchyme. Function-
ally recombining subcultures of each of
these embryonic precursor tissues might
lead to the formation of neokidneys. In
this context, Steer et al. [65] took advan-
tage of recently identified factors that di-
rect ureteric bud branching morphogen-
esis [66, 67] and metanephric mesenchyme
induction [68, 69], and showed that the
isolated rat ureteric bud and mesenchyme
can be recombined in vitro; moreover, the
resultant structure is morphologically and
architecturally indistinguishable from a
normal embryonic rat kidney precursor.
In addition, the whole rat kidney rudiment
in organ culture or the cultured isolated
rat ureteric bud can be partitioned into
smaller fragments, and these subfractions
can be propagated through several genera-
tions. These generations do not appear to
be different from their progenitors. The
subsequent generations of isolated rat ure-
teric bud can be recombined with fresh rat
mesenchyme. Within the recombined rat
neokidney, the ureteric buds branch into
the mesenchyme and seem to induce the
mesenchyme to epithelialize and form ne-
phrons in a normal manner. The nascent
tubular nephrons in the recombination ex-
periments form contiguous connections
with limbs of the branched ureteric bud,
thereby leading to an intact aqueduct be-
tween tubule and collecting system.
Considering the limited availability of
human fetal tissue, this method for sub-
culturing and propagating whole rat meta-
nephric rudiments in vitro might be appli-
cable to humans, and provide a large num-
ber of human kidney precursors derived
from a single donor. Clearly, porcine early
kidney precursors could afford an unlim-
ited source for renal transplantation.
Furthermore, the development of a large
population of pig renal primordia derived
from a single progenitor could potentially
be manipulated in vitro prior to recombi-
nation, for example, in ways that would
enhance acceptance of the xenotransplants
in immunocompetent hosts.
15.2.2.7 Generation of Histocompatible
Kidney Precursors using Nuclear
Transfer
Nuclear transfer denotes the introduction
of a nucleus from an adult donor cell into
an enucleated oocyte to generate a cloned
embryo [70] (see Part I, Chapter 11). When
transferred to the uterus of a female
recipient, this embryo has the potential to
grow into an infant that is a clone of the
adult donor cell a process termed repro-
ductive cloning. However, when explanted
in culture, this embryo can give rise to
embryonic stem cells that have the poten-
tial to become any or almost any type of
cell present in the adult body (therapeutic
cloning). Embryonic stem cells derived
from nuclear transfer are genetically iden-
tical to the donors cells, thus eliminating
the risk of immune rejection and the re-
quirement for immunosuppression [70].
Consequently, the ability to turn cloned
primordial stem cells into more complex
functional structures such as kidneys
would potentially overcome both immune
rejection and organ shortage.
Lanza et al. [71] investigated the use of
nuclear transfer to generate functional re-
nal structures. Renal cells obtained from
an early-stage cloned bovine fetus (em-
bryonic day 56, cloned from adult bovine
fibroblasts) were used to generate func-
tional immune-compatible renal tissues.
The cloned renal cells were expanded in vi-
tro, seeded onto collagen-coated cylindrical
polycarbonate membranes to form renal
devices, and implanted back into the nu-
clear donor animal without immune de-
struction. The embryonic precursor cells
organized themselves into glomeruli- and
tubule-like structures with the ability to
excrete toxic metabolic waste products
through a urine-like fluid, establishing
once more their high capacity to regener-
ate functional renal structures. Because
the cloned cells were derived from early-
stage fetuses, this approach is not an ex-
ample of therapeutic cloning and would
not be undertaken in humans. Further-
more, studies showing that cloned animals
have common abnormalities, regardless of
the type of donor cell or the species used,
and that these abnormalities correlate with
both subtle and gross errors in the nuclear
DNA leading to aberrant gene expression,
might limit this strategy for therapeutic
applications [70, 72].
15.2.3
Organogenesis by Embryonic Renal
Progenitors and Additional Stem Cell
Sources
Stem cells are defined by two major crite-
ria pluripotentality and self-renewal ca-
pacity. Recent developments in the field of
stem cell research indicate their enormous
potential as a source of tissue for regenera-
tive therapies. The success of such applica-
tions will depend on the precise properties
and potentials of stem cells isolated either
from embryonic, fetal or adult tissues.
Perhaps the most characterized stem cell
is the one residing in the adult bone mar-
row that is, the hematopoietic stem cell
which gives rise to all blood cell types [73].
In addition, mesenchymal stem cells
(MSC) are multipotent cells that can be
isolated from adult bone marrow and be
induced in vitro and in vivo to differentiate
into a variety of mesenchymal tissues, in-
cluding bone, cartilage, tendon, fat, bone
marrow stroma, and muscle [74].
Recently, it has been suggested that
adult bone marrow-derived stem cells can
cross boundaries and give rise to a broader
array of differentiated cell types that is,
turning blood into liver, brain, pancreas,
skin, intestine and, eventually, kidney [75
82]. Because bone marrow-derived stem
cells can be withdrawn from the patients
own marrow (or even from the blood),
they can serve as a source for autologous
transplantation. However, this approach re-
mains controversial for several reasons.
First, following the infusion of adult bone
marrow-derived stem cells, several reports
show a lack of donor cell engraftment in
parenchymal organs and doubt the actual
existence of adult stem cell plasticity [83,
84]. Second, other investigators that do
show a donor cell phenotype in parenchy-
mal cells, suggest that it occurs through
stem cell fusion and not by trans-differen-
tiation and generation of cells de novo [85
87]. Thus, the term regeneration is used
erroneously in this context, and should be
replaced by reparative if indeed that
will be shown to be the result of stem cell
fusion. Third, whether its trans-differentia-
tion or cell fusion, the efficiency of these
processes under basal conditions and
even when inflicting tissue injury is
rather low, underscoring the functional ca-
pacity of the adult stem cells. For example,
several studies conducted in the kidney
have shown that bone marrow-derived
cells adopt the phenotype of proximal tu-
bular cells during acute tubular injury [81,
82], as well as that of glomerular endothe-
lial and mesangial cells during glomerular
injury [8890], but hardly report on a func-
tional benefit [81]. Even if proven function-
ally to contribute to the healing of kidneys
under these and possibly other settings,
bone marrow-derived cells are not likely to
be beneficial when parenchymal cell loss
due to chronic renal damage does not
allow cell repair or regeneration. Thus,
when ESRD develops, other strategies of
tissue replacement which permit reconsti-
tution of renal structures and organogen-
esis are necessitated to overcome the
shortage for donor organs.
Can a kidney be grown from stem cells?
In theory, human embryonic stem (ES)
cells which are derived from blastocysts
can form derivatives of all three germ
layers and give rise to all cell types of the
body [91]. Nevertheless, their in vivo use is
limited; transplantation of human embryo-
nic stem cells directly into adoptive hosts
results in teratoma growth [92], and they
have to be initially programmed and differ-
entiated in vitro into a specific cell lineage
prior to transplantation [9395]. This pro-
cess does not confer purity of a single dif-
ferentiated cell type, and possibly pre-
serves the tumorigenic potential [96].
Nevertheless, while we emphasized the
role of whole embryonic renal progenitor
tissue, as opposed to purified renal stem
cells, it is likely that progress in this area
of investigation might eventually lead to a
synthesis, taking advantage of the benefits
which are offered by each of these promis-
ing approaches. For example, because it is
extremely doubtful whether human ES
cells can be used as a starting material for
creating complex functional three-dimen-
sional organs (e.g., the kidney) in addition
to their potential in generating individual
cells, it might be much easier to derive re-
nal progenitors from human embryonic
stem cells, which in turn could be grafted
and, as shown, give rise to the mature kid-
ney cells [25]. Moreover, if performed with
the assistance of nuclear transfer and ther-
apeutic cloning, issues of immunogenicity
could be completely resolved [71]. Alterna-
tively, the embryonic renal progenitor tis-
sue could be initially implanted so as to di-
rect growth of in vitro propagated kidney
committed stem cells, by providing a
three-dimensional embryonic architecture
and appropriate stroma, possibly leading
to an improved tissue-engineered kidney.
15.3
Conclusions
Regenerative medicine is focused on the
development of cells, tissues and organs
for the purpose of restoring function
through transplantation [6]. The general
thought that replacement, repair and re-
storation of function is best accomplished
by cells, tissues or organs that can perform
the appropriate physiologic/metabolic du-
ties better than any mechanical device (see
Part I, Chapter 12), recombinant protein
therapeutic or chemical compound could
be also applicable to the kidney (Table
15.1). In this regard, the use of stem cells
as a starting material offers new and
powerful strategies for future tissue devel-
opment and engineering [97]. Currently,
adult and embryonic stem cells are being
investigated to replace individual cells in a
diseased organ [98], namely insulin-pro-
ducing pancreatic cells for type 1 diabetes
mellitus or dopamine-producing brain
cells for Parkinsons disease (see Part I,
Chapter 14), but are not close to support-
ing the creation of whole organs, or even
significant parts of them. The kidney is no
exception; stem cell therapy to replace re-
nal proximal tubular cells or glomerular
mesangial cells might delay the events
leading to the deterioration of the organ
and, consequently, decrease patients de-
mands for organ transplantation. However,
once chronic renal failure ensues, neokid-
neys will be needed to combat organ
shortage. Vascularized organ xenotrans-
plants, which are now genetically engi-
neered not to elicit hyperacute rejection,
represent one potential option [10].
T
a
b
l
e
1
5
.
1
P
u
b
l
i
s
h
e
d
s
t
r
a
t
e
g
i
e
s
f
o
r
r
e
n
a
l
c
e
l
l
r
e
p
l
a
c
e
m
e
n
t
u
s
i
n
g
b
i
o
l
o
g
i
c
m
a
t
e
r
i
a
l
.
S
t
a
r
t
i
n
g
m
a
t
e
r
i
a
l
M
e
t
h
o
d
B
i
o
l
o
g
i
c
m
e
c
h
a
n
i
s
m
T
a
r
g
e
t
e
d
c
e
l
l
P
o
s
s
i
b
l
e
i
n
d
i
c
a
t
i
o
n
R
e
f
e
r
e
n
c
e
M
o
u
s
e
a
d
u
l
t
H
S
C
I
n
t
r
a
v
e
n
o
u
s
t
r
a
n
s
p
l
a
n
t
a
t
i
o
n
T
r
a
n
s
d
i
f
f
e
r
e
n
t
i
a
t
i
o
n
?
C
e
l
l
f
u
s
i
o
n
?
E
p
i
t
h
e
l
i
a
l
s
t
e
m
c
e
l
l
?
P
r
o
x
i
m
a
l
t
u
b
u
l
a
r
c
e
l
l
s
A
c
u
t
e
t
u
b
u
l
a
r
n
e
c
r
o
s
i
s
8
1
,
8
2
M
o
u
s
e
a
d
u
l
t
B
M
D
C
I
n
t
r
a
v
e
n
o
u
s
t
r
a
n
s
p
l
a
n
t
a
t
i
o
n
B
o
n
e
m
a
r
r
o
w
h
e
m
a
n
g
i
o
b
l
a
s
t
?
G
l
o
m
e
r
u
l
a
r
e
n
d
o
t
h
e
l
i
a
l
c
e
l
l
s
A
c
u
t
e
/
s
u
b
a
c
u
t
e
g
l
o
m
e
r
u
l
a
r
i
n
j
u
r
y
8
8
M
o
u
s
e
a
d
u
l
t
M
S
C
I
n
t
r
a
v
e
n
o
u
s
t
r
a
n
s
p
l
a
n
t
a
t
i
o
n
P
l
a
s
t
i
c
i
t
y
?
G
l
o
m
e
r
u
l
a
r
m
e
s
a
n
g
i
a
l
c
e
l
l
s
A
c
u
t
e
/
s
u
b
a
c
u
t
e
g
l
o
m
e
r
u
l
a
r
i
n
j
u
r
y
8
9
,
9
0
E
m
b
r
y
o
n
i
c
k
i
d
n
e
y
p
r
e
c
u
r
-
s
o
r
s
(
m
u
r
i
n
e
,
p
i
g
,
h
u
m
a
n
)
S
u
b
c
a
p
s
u
l
a
r
/
a
b
d
o
m
i
n
a
l
g
r
a
f
t
i
n
g
M
e
s
e
n
c
h
y
m
a
l
-
e
p
i
t
h
e
l
i
a
l
c
o
n
v
e
r
-
s
i
o
n
B
r
a
n
c
h
i
n
g
m
o
r
p
h
o
g
e
n
e
s
i
s
A
n
g
i
o
g
e
n
e
s
i
s
N
e
o
-
k
i
d
n
e
y
s
E
S
R
D
2
5
,
2
7
2
9
3
0
3
5
R
a
t
e
m
b
r
y
o
n
i
c
k
i
d
n
e
y
p
r
e
c
u
r
s
o
r
s
(
U
B
-
M
M
)
I
n
v
i
t
r
o
p
r
o
p
a
g
a
t
i
o
n
M
e
s
e
n
c
h
y
m
a
l
-
e
p
i
t
h
e
l
i
a
l
c
o
n
v
e
r
-
s
i
o
n
B
r
a
n
c
h
i
n
g
m
o
r
p
h
o
g
e
n
e
s
i
s
N
e
o
-
k
i
d
n
e
y
s
E
S
R
D
6
5
B
o
v
i
n
e
e
m
b
r
y
o
n
i
c
r
e
n
a
l
c
e
l
l
s
N
u
c
l
e
a
r
t
r
a
n
s
f
e
r
a
r
t
i
f
i
c
i
a
l
m
e
m
b
r
a
n
e
s
e
e
d
i
n
g
S
u
b
c
u
t
a
n
e
o
u
s
t
r
a
n
s
p
l
a
n
t
a
-
t
i
o
n
C
e
l
l
o
r
g
a
n
i
z
a
t
i
o
n
N
e
o
-
k
i
d
n
e
y
s
E
S
R
D
7
1
H
S
C
:
h
e
m
a
t
o
p
o
i
e
t
i
c
s
t
e
m
c
e
l
l
s
;
M
S
C
:
m
e
s
e
n
c
h
y
m
a
l
s
t
e
m
c
e
l
l
s
;
B
M
D
C
:
b
o
n
e
m
a
r
r
o
w
-
d
e
r
i
v
e
d
c
e
l
l
s
.
Organogenesis is an additional alternative.
Organogenesis of complex tissues, such as
the kidney, requires a coordinated sequen-
tial transformation process, with individual
stages involving time-dependent expres-
sion of cellcell, cellmatrix, and cellsig-
nal interactions in three dimensions. Pre-
cursor tissues are composed of function-
ally diverse stem/progenitor cell types that
are organized in spatially complex arrange-
ments. When obtained at gestational-spe-
cific time points, the theme of temporal-
spatial patterning of progenitor cell inter-
actions is programmed in precursor tis-
sues, leading to their optimal growth and
development. The findings summarized
here raise hope that translation of organo-
genesis for the purposes of organ replace-
ment is within reach. Nevertheless, for
renal replacement therapy (neokidney)
further experimentation needs to be devel-
oped to enhance the function of the early
human and porcine kidney grafts once
matured. Producing a prolonged adequate
urinary anastomosis of the growing kidney
and deriving a blood supply sufficient to
correct biochemical aberrations in a ure-
mic individual, is an important goal. In-
creasing the number of transplants and/or
administering specific human growth fac-
tors might support functional replacement.
Considering the limited availability of hu-
man fetal tissue, a method for subcultur-
ing and propagating whole metanephric
rudiments in vitro that was recently devel-
oped in rats [65] might provide a large
number of human kidney progenitors de-
rived from a single donor. Alternatively,
early pig kidney precursors could afford an
unlimited source for renal transplantation,
provided that risk for porcine endogenous
retrovirus (PERV) could be eliminated [9].
Large-animal models are needed to test
the relevance of these strategies for trans-
plantation in humans and for successful
continuation of this avenue of a promising
new class of modern biopharmaceuticals.
References
1 Perico N, Ruggenenti P, Scalamogna M, Re-
muzzi G. Tackling the shortage of donor kid-
neys: how to use the best that we have. Am. J.
Nephrol. 2003;23(4):245259.
2 Matesanz R: International figures on organ
donation and transplantation 2000. Newslett.
Transplant. 2001;6:141.
3 Cecka JM: The UNOS Renal Transplant Regis-
try; in: Cecka JM, Terasaki PI (Eds): Clinical
Transplants 2001. Los Angeles, UCLA Immu-
nogenetics Center, 2002, pp. 118.
4 Renal Data System: USRDS 1998 annual data
report. Bethesda, MD: National Institute of
Diabetes and Digestive Kidney Diseases. NIH
Publ 98-3176 1998, April, p. 39.
5 UNOS Critical Data: Waiting list. http://www
unos org 2002. Last update Dec 2001.
6 Fodor WL. Tissue engineering and cell based
therapies, from the bench to the clinic: the po-
tential to replace, repair and regenerate. Re-
prod. Biol. Endocrinol. 2003; 1(1):102.
7 Cascalho M, Ogle BM, Platt JL. Xenotrans-
plantation and the future of renal replace-
ment. J. Am. Soc. Nephrol. 2004;15(5):1106
1112.
8 Cascalho M, Platt JL. The immunological bar-
rier to xenotransplantation. Immunity 2001;
14(4):437446.
9 Blusch JH, Patience C, Martin U. Pig endo-
genous retroviruses and xenotransplantation.
Xenotransplantation 2002;9(4):242251.
10 Cooper DK, Gollackner B, Sachs DH. Will the
pig solve the transplantation backlog? Annu.
Rev. Med. 2002;53:133147.
11 Galili U, Mandrell RE, Hamadeh RM, et al.
The interaction between the human natural
anti-galactosyl IgG (anti-Gal) and bacteria of
the human flora. Infect. Immun. 1998;
57:17301737
12 Oriol R, Ye Y, Koren E, Cooper DKC. Carbo-
hydrate antigens of pig tissues reacting with
human natural antibodies as potential targets
for hyperacute vascular rejection in pig-to-
man organ xenotransplantation. Transplanta-
tion 1993;56:14331442
References 369
13 Good AH, Cooper DKC, Malcolm AJ, et al.
Identification of carbohydrate structures
which bind human anti-porcine antibodies:
implications for discordant xenografting in
man. Transplant. Proc. 1992;24:559562.
14 Cooper DKC, Koren E, Oriol R. Oligosaccha-
rides and discordant xenotransplantation.
Immunol. Rev. 1994;141:315318.
15 Rose AG, Cooper DKC. Venular thrombosis is
the key event in the pathogenesis of antibody-
mediated cardiac rejection. Xenotransplantation
2000;7:3141.
16 Cooper DK. Clinical xenotransplantation
how close are we? Lancet 2003;362(9383):557
559.
17 Cozzi E, White DJG. The generation of trans-
genic pigs as potential organ donors for hu-
mans. Nat. Med. 1995;1:964966.
18 Lai L, Kolber-Simonds D, Park KW, et al. Pro-
duction of a-1,3-galactosyltransferase knockout
pigs by nuclear transfer cloning. Science
2001;295:10891092.
19 Phelps CJ, Koike C, Vaught TD, et al. Produc-
tion of a-1,3-galactosyltransferase-deficient
pigs. Science 2003;299:411414.
20 Woolf AS. Embryology of the kidney. In:
Pediatric Nephrology, 4th edn. Barratt TM,
Avner A, Harmon W (Eds). Baltimore,
Williams & Wilkins, 1999, pp. 119.
21 Al-Awqati Q, Oliver JA. Stem cells in the kid-
ney. Kidney Int. 2002;61:387395.
22 Herzlinger D, Koseki C, Mikawa T, et al. Me-
tanephric mesenchyme contains multipotent
stem cells whose fate is restricted after induc-
tion. Development 1992;114:565572.
23 Oliver JA, Barasch J, Yang J, Herzlinger D,
et al. Metanephric mesenchyme contains em-
bryonic renal stem cells. Am. J. Physiol. Renal
Physiol. 2002;283:F799F809.
24 Almeida-Porada G, El Shabrawy D, Porada C,
et al. Differentiative potential of human meta-
nephric mesenchymal cells. Exp. Hematol.
2002;30:14541462.
25 Dekel B, Burakova T, Arditti FD, et al. Human
and porcine early kidney precursors as a new
source for transplantation. Nat. Med.
2003;9:5360.
26 Dekel B, Hochman E, Sanchez MJ, et al. Kid-
ney, blood and endothelium: Developmental
expression of stem cell leukemia during ne-
phrogenesis. Kidney Int. 2004; 65:11621169.
27 Dekel B, Burakova T, Ben-Hur H, et al. En-
graftment of human kidney tissue in rat ra-
diation chimera: II. Human fetal kidneys dis-
play reduced immunogenicity to adoptively
transferred human peripheral blood mono-
nuclear cells and exhibit rapid growth and de-
velopment. Transplantation 1997;64:15501558.
28 Dekel B, Marcus H, Herzel BH, Bocher WO,
Passwell JH, Reisner Y. In-vivo modulation of
the allogeneic immune response by human
fetal kidneys: the role of cytokines, chemo-
kines, and cytolytic effector molecules. Trans-
plantation 2000;69:14701478.
29 Dekel B, Amariglio N, Kaminski N, et al. En-
graftment and differentiation of human meta-
nephroi into functional mature nephrons after
transplantation into mice is accompanied by a
profile of gene expression similar to normal
human kidney development. J. Am. Soc.
Nephrol. 2002;13(4):977990.
30 Woolf AS, Palmer SJ, Snow ML, Fine LG.
Creation of a functioning chimeric mamma-
lian kidney. Kidney Int. 1990;38:991997.
31 Abrahamson DR, St. John PL, Pillion DL,
Tucker DC. Glomerular development in intra-
ocular and intra renal grafts of fetal kidneys.
Lab. Invest. 1991;64:629639.
32 Rogers SA, Lowell JA, Hammerman NA,
Hammerman MR. Transplantation of develop-
ing metanephroi into adult rats. Kidney Int.
1998;54:2737.
33 Rogers SA, Liapis H, Hammerman MR.
Transplantation of metanephroi across the
major histocompatibility complex in rats.
Am. J. Physiol. Regul. Integr. Comp. Physiol.
2001;280:R132R136.
34 Rogers SA, Hammerman MR. Transplantation
of rat metanephroi into mice. Am. J. Physiol.
Regul. Integr. Comp. Physiol. 2001;280:R1865
1869.
35 Rogers SA, Talcott M, Hammerman MR.
Transplantation of pig metanephroi. ASAIO J.
2003;49(1):4852.
36 Galli R, Gritti A, Bonfanti L, Vescovi AL. Neu-
ral stem cells: an overview. Circ. Res.
2003;92(6):598608.
37 Erdo F, Buhrle C, Blunk J, et al. Host-depen-
dent tumorigenesis of embryonic stem cell
transplantation in experimental stroke. J. Cer-
eb. Blood Flow Metab. 2003;23(7):780785.
38 Kreisel, D, Krupnick AS, Gelman AE, Engels
FH, Popma SH, Krasinskas AM. Non-hemato-
poietic allograft cells directly activate CD8+ T
cells and trigger acute rejection: An alternative
mechanism of allorecognition. Nat. Med.
2002;8:233239.
39 Sariola H, Ekblom P, Lehtonen E, Saxen L.
Differentiation and vascularization of the me-
tanephric kidney grafted on the chorioallan-
toic membrane. Dev. Biol. 1983;96:427435.
40 Hyink DP, Tucker DC, St. John PL, et al. En-
dogenous origin of glomerular endothelial
and mesangial cells in grafts of embryonic
kidneys. Am. J. Physiol. 1996;270:F886F889.
41 Robert B, St. John PL, Hyink DP, Abraham-
son DR. Evidence that embryonic kidney cells
expressing flk-1 are intrinsic, vasculogenic an-
gioblasts. Am. J. Physiol. 1996;271:F744F753.
42 Woolf AS, Loughna S. Origin of glomerular
capillaries: is the verdict in? Exp. Nephrol.
1998;6:1721.
43 Medawar PB. Some immunological and endo-
crinological problems raised by the evolution
of viviparity in vertebrates. Symp. Soc. Exp.
Biol. 1953;7:320323.
44 Foglia RP, Dipreta J, Statter MB, Donahoe
PK. Fetal allograft survival in immunocompe-
tent recipients is age-dependent and organ-
specific. Ann. Surg. 1986;204:402410.
45 Foglia RP, LaQuaglia M, DiPreta J, Donahoe
PK. Can fetal and newborn allografts survive
in an immunocompetent host? J. Pediatr. Surg.
1986;21:608703.
46 Barker CF, Billingham RE. Extension of skin
homograft survival by prevention of graft-host
skin contact. Transplant. Proc. 1973;5:153158.
47 Deutsch AA, Arensman R, Levey R, Folkman J.
The effect of antilymphocyte serum on fetal rat
intestine transplanted as free subcutaneous
homografts. J. Pediatr. Surg. 1974;9:2934.
48 Brown J, Clarck WR, Makoff RK. Pancreas
transplantation for diabetes mellitus. Ann. In-
tern. Med. 1978;89:951965.
49 Velasco AL, Hegre OD. Decreased immuno-
genicity of fetal kidneys: the role of passenger
leukocytes. J. Pediatr. Surg. 1989;24:5963.
50 Statter MB, Fahrner KJ, Barksdale EM, Parks
DE, Flavell RA, Donahoe PK. Correlation of
fetal kidney and testis congenic graft survival
with reduced major histocompatibility com-
plex burden. Transplantation 1989;47:651660.
51 Reisner Y, Dagan S. The Trimera mouse: gen-
erating human monoclonal antibodies and an
animal model for human diseases. Trends Bio-
technol. 1998;16:242246.
52 Burakova T, Marcus H, Canaan A, et al. En-
grafted human T and B lymphocytes form
mixed follicles in lymphoid organs of human/
mouse and human/rat radiation chimera.
Transplantation 1997;63:11661171.
53 Marcus H, David M, Canaan A, et al. Hu-
man/mouse radiation chimera are capable of
mounting a human primary humoral re-
sponse. Blood 1995;86:398406.
54 Segall H, Lubin I, Marcus H, Canaan A, Reis-
ner Y. Generation of primary antigen-specific
human cytotoxic T lymphocytes in human/
mouse radiation chimera. Blood 1996;88:721
730.
55 Eren R, Lubin I, Terkieltaub D, Ben-Moshe O,
et al. Human monoclonal antibodies specific
to hepatitis B virus generated in a human/
mouse radiation chimera: the Trimera system.
Immunology 1998;93(2):154161.
56 Galun E, Burakova T, Ketzinel M, et al. Hepa-
titis C virus viremia in SCID?SBNX mouse
chimera. J. Infect. Dis. 1995;172(1):2530.
57 Dekel B, Burakova T, Marcus H, et al. Engraft-
ment of human kidney tissue in rat radiation
chimera: I. A new model of human kidney al-
lograft rejection. Transplantation 1997;64:1541
1548.
58 Dekel B, Bocher WO, Marcus H, Yussim A,
Reisner Y. Acute cellular rejection of human
renal tissue by adoptive transfer of allogeneic
human peripheral blood mononuclear cells
into chimeric rats: sequential gene expression
of cytokines, chemokines and cytolytic effector
molecules, and their regulation by CTLA-4-Ig.
Int. Immunol. 1999;11:16731683.
59 Howland KC, Ausubel LJ, London CA, Abbas
AK. The roles of CD28 and CD40 ligand in T
cell activation and tolerance. J. Immunol.
2000;164(9):44654470.
60 Nelson PJ, Krensky AM. Chemokines, chemo-
kine receptors, and allograft rejection. Immu-
nity 2001;14:377386.
61 Tedder TF, Steeber DA, Chen A, Engel P. The
selectins: vascular adhesion molecules. FASEB
J. 1995;9:866873.
62 ORegan AW, Nau GJ, Chupp GL, Berman JS.
Osteopontin (Eta-1) in cell mediated immu-
nity: Teaching an old dog new tricks. Immu-
nology Today 2000;21:475478.
63 Pratt JR, Basheer SA, Sacks SH. Local synthe-
sis of complement component C3 regulates
acute renal transplant rejection. Nat. Med.
2002;8:582587.
64 Rogers SA, Powell-Braxton L, Hammerman
MR. Insulin-like growth factor I regulates re-
References 371
nal development in rodents. Dev. Genet.
1999;24:293298.
65 Steer DL, Bush KT, Meyer TN, Schwesinger
C, Nigam SK. A strategy for in vitro propaga-
tion of rat nephrons. Kidney Int. 2002;62(6):
19581965.
66 Sakurai H, Bush KT, Nigam SK. Identification
of pleiotrophin as a mesenchymal factor in-
volved in ureteric bud branching morphogen-
esis. Development 2001;128:32833293.
67 Qiao J, Bush KT, Steer DL, et al. Multiple fi-
broblast growth factors support growth of the
ureteric bud but have different effects on
branching morphogenesis. Mech. Dev.
2001;109:123135.
68 Plisov SY, Yoshino K, Dove LF, et al. TGF beta
2, LIF and FGF2 cooperate to induce nephro-
genesis. Development 2001;128:10451057.
69 Barasch J, Yang J, Ware CB, et al. Mesenchy-
mal to epithelial conversion in rat metane-
phros is induced by LIF. Cell 1999;99:377386.
70 Hochedlinger K, Jaenisch R. Nuclear trans-
plantation, embryonic stem cells, and the po-
tential for cell therapy. N. Engl. J. Med.
2003;349(3):275286.
71 Lanza RP, Chung HY, Yoo JJ, et al. Genera-
tion of histocompatible tissues using nuclear
transplantation. Nat. Biotechnol. 2002;20:689
696.
72 Booth PJ, Viuff D, Tan S, et al. Numerical
chromosome errors in day 7 somatic nuclear
transfer bovine blastocysts. Biol. Reprod.
2003;68(3):922928.
73 Kondo M, Wagers AJ, Manz MG, et al. Biol-
ogy of hematopoietic stem cells and progeni-
tors: implications for clinical application.
Annu. Rev. Immunol. 2003;21:759806.
74 Fibbe W. Mesenchymal stem cells. A potential
source for skeletal repair. Ann. Rheum. Dis.
2002;61(Suppl. 2):ii29ii31.
75 Lagasse E, Connors H, Al-Dhalimy M, et al.
Purified hematopoietic stem cells can differ-
entiate into hepatocytes in vivo. Nat. Med.
2000;6:12291234.
76 Mezey E, Chandross KJ, Harta G, Maki RA,
McKercher SR. Turning blood into brain: cells
bearing neuronal antigens generated in vivo
from bone marrow. Science 2000;290:1779
1782.
77 Ianus A, Holz GG, Theise ND, Hussain MA.
In vivo derivation of glucose-competent pan-
creatic endocrine cells from bone marrow
without evidence of cell fusion. J. Clin. Invest.
2003;111:843850.
78 Badiavas EV, Abedi M, Butmarc J, Falanga V,
Quesenberry P. Participation of bone marrow
derived cells in cutaneous wound healing.
J. Cell. Physiol. 2003;196:245250.
79 Okamoto R, Yajima T, Yamazaki M, et al.
Damaged epithelia regenerated by bone mar-
row-derived cells in the human gastrointesti-
nal tract. Nat. Med. 2002;8:10111017.
80 Poulsom R, Forbes SJ, Hodivala-Dilke K, et al.
Bone marrow contributes to renal parenchy-
mal turnover and regeneration. J. Pathol.
2001;195:229235.
81 Kale S, Karihaloo A, Clark PR, Kashgarian M,
Krause DS, Cantley LG. Bone marrow stem
cells contribute to repair of the ischemically
injured renal tubule. J. Clin. Invest. 2003;
112:4249.
82 Lin F, Cordes K, Li L, et al. Hematopoietic
stem cells contribute to the regeneration of re-
nal tubules after renal ischemiareperfusion
injury in mice. J. Am. Soc. Nephrol. 2003;
14:11881199.
83 Wagers AJ, Sherwood RI, Christensen JL,
Weissman IL. Little evidence for developmen-
tal plasticity of adult hematopoietic stem cells.
Science 2002;297(5590):22562259.
84 Choi JB, Uchino H, Azuma K, et al. Little evi-
dence of transdifferentiation of bone marrow-
derived cells into pancreatic beta cells. Diabe-
tologia 2003;46(10):13661374.
85 Terada N, Hamazaki T, Oka M, et al. Bone
marrow cells adopt the phenotype of other
cells by spontaneous cell fusion. Nature
2002;416:542545.
86 Wang X, Willenbring H, Akkari Y, et al. Cell
fusion is the principal source of bone-marrow-
derived hepatocytes. Nature 2003;422:897901.
87 Alvarez-Dolado M, Pardal R, Garcia-Verdugo
JM, et al. Fusion of bone-marrow-derived cells
with Purkinje neurons, cardiomyocytes and
hepatocytes. Nature 2003;425(6961):968973.
88 Rookmaaker MB, Smits AM, Tolboom H, et
al. Bone-marrow-derived cells contribute to
glomerular endothelial repair in experimental
glomerulonephritis. Am. J. Pathol.
2003;163(2):553562.
89 Ito T, Suzuki A, Imai E, Okabe M, Hori M.
Bone marrow is a reservoir of repopulating
mesangial cells during glomerular remodel-
ing. J. Am. Soc. Nephrol. 2001;12:26252635.
90 Cornacchia F, Fornoni A, Plati AR, et al. Glo-
merulosclerosis is transmitted by bone mar-
row-derived mesangial cell progenitors.
J. Clin. Invest. 2001;108:16491656.
91 Jones JM, Thomson JA. Human embryonic
stem cell technology. Semin. Reprod. Med.
2000;18(2):219223.
92 Amit M, Carpenter MK, Inokuma MS, et al.
Clonally derived human embryonic stem cell
lines maintain pluripotency and proliferative
potential for prolonged periods of culture.
Dev. Biol. 2000;227:271278.
93 Zhang SC, Wernig M, Duncan ID, Brustle O,
Thomson JA. In vitro differentiation of trans-
plantable neural precursors from human em-
bryonic stem cells. Nat. Biotechnol.
2001;19(12):11291133.
94 Sottile V, Thomson A, McWhir J. In vitro
osteogenic differentiation of human ES cells.
Cloning Stem Cells 2003;5(2):149155.
95 Rambhatla L, Chiu CP, Kundu P, Peng Y, Car-
penter MK. Generation of hepatocyte-like cells
from human embryonic stem cells. Cell Trans-
plant. 2003;12(1):111.
96 Vogel G. Stem cells: new excitement, persis-
tent questions. Science 2000;290:1672.
97 Hwang WS, et al. Evidence of a pluripotent
human embryonic stem cell line derived from a
cloned blastocyst. Science 2004;303:16691674.
98 Daley GQ, Goodell MA, Snyder EY. Realistic
prospects for stem cell therapeutics. Hematol-
ogy (Am. Soc. Hematol. Educ. Program)
2003;398418.
References 373
ISBN: 3-527-31184-X
Part II
Biopharmaceuticals and Their Mode of Action
Abstract
Zymogen activation is a critical step in the
regulation of many important biological
processes like coagulation, fibrinolysis or
the complement system. The endogenous
proteolytic activation of serine proteinase
zymogens like plasminogen or prothrom-
bin follows the classical mechanism, in
which a specific cleavage leads to the in-
sertion of the newly formed N-terminus
into a binding cleft in the proteinase. This
interaction triggers the conformational
change that completes the folding of the
proteinase and activates the enzyme by the
formation of functional substrate binding
subsites and the oxyanion hole. Due to
their high concentration in the blood of
the host, these zymogens of the coagula-
tion and fibrinolysis systems are attractive
targets for pathogenic bacteria. Pathogens
modulate the activity of these key pro-en-
zymes either directly through cleavage by
proteinases produced by the pathogen or
indirectly by release of effector molecules
which form complexes with host zymo-
gens. The latter mechanism involves the
insertion of the N-terminus of the bacterial
cofactor into the preformed activation
pocket of the zymogen, leading to the con-
formational activation of the zymogen
without cleavage of the activation peptide
bond. The resulting uncontrolled proteolyt-
ic activity can result in tissue damage or
in the generation of emboli, leading to
stroke and myocardial infarction. Only re-
cently has the role of cofactor-induced zy-
mogen activation been adequately recog-
nized and new studies may provide a basis
for development of biopharmaceutical
therapies adjunctive to antibiotics
based on the inhibition of pathological
processes invoked by these bacterial pro-
teins.
377
1
Mechanisms of Serine Proteinase Activation:
Insights for the Development of Biopharmaceuticals
for Coagulation and Fibrinolysis
Rainer Friedrich
ISBN: 3-527-31184-X
Quid pro Quo Lysis vs. Coagulation in the Fine-tuned Balance
of the Clotting Cascade
Abbreviations
PSGN post-streptococcal glomerulone-
phritis
SAK staphylokinase
SC staphylocoagulase
SK streptokinase
t-PA tissue-type plasminogen activator
u-PA urokinase-type plasminogen acti-
vator
1.1
Introduction
1.1.1
Activation of Trypsin-like Serine Proteinases
The activation of a precursor form of an
active enzyme is a central regulation
mechanism in many important biological
processes including blood coagulation, fi-
brinolysis and the complement system.
Most trypsin-like serine proteinases are
synthesized as inactive precursors (pro-en-
zymes or zymogens) that must either bind
to a specific cofactor to develop substantial
catalytic activity and/or be activated by lim-
ited proteolytic processing, which induces
a conformational change to create func-
tional catalytic machinery and removes an
N-terminal peptide or entire N-terminal
domains [1]. As a consequence of the acti-
vation, the catalytic activity of the enzyme
is usually enhanced by several orders of
magnitude. Some proteinase zymogens,
such as single-chain tissue-type plasmino-
gen activator (t-PA), already show weak,
but significant, activity before activation
cleavage [2]. The so-called zymogenicity
of the pro-enzyme is a measure for the in-
crease in catalytic efficiency after activa-
tion. While the precursors of the digestive
enzymes trypsin or chymotrypsin are al-
most completely inactive (with a 10
4
- to
10
6
-fold activity increase [3, 4]), the zymo-
genicity of t-PA is only 510 [5]. Because
of its importance and topicality, one such
compound, i.e., DSPA, will be presented
by Oliver Kops from Paion in the upcom-
ing edition of Modern Biopharmaceuticals.
In cascades like coagulation, fibrinolysis
and complement activation, a given protei-
nase activates another pro-proteinase in an
amplification cascade that provides a num-
ber of regulatory steps; cofactors here of-
ten play crucial roles by enhancing reac-
tion rates and modifying enzyme specifici-
ty [6, 7] (see also Part II, Chapter 3 and
Part III, Chapter 6). The endogenous pro-
teolytic activation of serine proteinase zy-
mogens follows the classical mechanism
(based on seminal studies of trypsinogen
[8, 9] and chymotrypsinogen [10] in the
1970s) in which the cleavage of an Arg15
Ile/Val16 peptide bond (numbers represent
topologically equivalent residues referring
to the chymotrypsinogen numbering)
leads to the insertion of the newly formed
N-terminal small hydrophobic residue into
a specific binding cleft (the activation
pocket) in the proteinase (domain) and the
formation of a strong salt bridge between
the charged N-terminal ammonium group
and the carboxylate of Asp194 (Scheme
1.1). This interaction triggers the confor-
mational change that completes the fold-
ing of the proteinase and activates the en-
zyme by formation of the substrate bind-
ing sites (subsites) and the oxyanion
hole. The inactive conformation of trypsi-
nogen and homologous zymogens is in
highly unfavorable, but reversible, equilib-
rium with an active conformation, in
most cases lying extremely on the zymo-
gen side. Bovine trypsinogen can assume
a trypsin-like (i.e., active) conformation
in the presence of dipeptides mimicking
the Ile16Val17 N-terminus of trypsin [11]
if assisted by binding of inhibitors in the
1 Mechanisms of Serine Proteinase Activation 378
active site. The studies on trypsinogen
showed that the major structural changes
resulting from the activation cleavages
were limited to a small portion (around
one-sixth) of the molecule, the activation
domain [9, 12]. These structural changes
are visualized in a video animation on the
supplementary CD-ROM. This region
comprises basically the four segments 16
19, 142152, 184193, and 216223, and is
relatively flexible in the zymogen, correlat-
ing with crystallographic disorder or signif-
icantly higher B values [13] (Scheme 1.2).
The catalytic domains of trypsin-like ser-
ine proteinases consist of two six-stranded
antiparallel b-sheets yielding b-barrels in
which the strands are arranged in a so-
called Greek key motif (b14) followed
by an antiparallel hairpin loop (b56) [10].
At the intersection of the two barrels re-
side the residues of the catalytic triad,
His57, Asp102 and Ser195 [14]. The back-
bone amides of Gly193 and Ser195 form
the oxyanion hole [15] which receives the
carbonyl group of the scissile peptide
bond.
1.1.2
Coagulation and Fibrinolysis
are Serine Proteinase Cascades
1.1.2.1 Prothrombin Activation
In response to vascular injury, the body
must tightly seal the leakage while pre-
venting unrestrained intravascular clot de-
velopment and vessel occlusion. The coag-
ulation process is a complex interplay of
the blood vessel wall, platelets and other
blood cells, as well as many soluble plas-
ma proteins (coagulation factors [16]). In
the ultimate step of the coagulation cas-
cade, the trypsin-like serine proteinase
thrombin (factor IIa) is released into the
blood stream, where it performs several es-
Scheme 1.1
Scheme 1.2
sential procoagulant functions [17]. Free a-
thrombin (the active form) converts solu-
ble fibrinogen to fibrin, which sponta-
neously polymerizes to form the fibrillar
matrix of the blood clot [18]. Thrombin
also activates factor XIII, a transglutami-
nase which thereafter covalently crosslinks
fibrin monomers, forming an insoluble
clot [19]. Binding of thrombin to its recep-
tor thrombomodulin leads to a dramatic
change in the substrate specificity of
thrombin, converting it from a procoagu-
lant to an anticoagulant and antifibrinolyt-
ic agent [20].
The thrombin zymogen, prothrombin, is
primarily synthesized in the liver and se-
creted into the blood as a 579-residue gly-
coprotein. The N-terminal so-called Gla
domain of prothrombin containing 10 c-
carboxyglutamic residues anchors the zy-
mogen to phospholipid membranes upon
calcium binding [21]. The Gla domain is
followed by two kringle domains and the
serine proteinase catalytic domain; these
major structural elements are connected
by relatively long peptides. The activation
to the active a-thrombin is performed in
vivo by the membrane-bound prothrombi-
nase consisting of factors Va and Xa as-
sembled on a phospholipid surface [22,
23], and requires cleavages at two posi-
tions, leading to a reduction of the molec-
ular weight from 71.6 to 39 kDa through
the release of the Gla and kringle domains
[24].
Following the initial structure determi-
nation of d-PheProArg chloromethyl ke-
tone-inhibited human a-thrombin [25], the
structures of more than 100 thrombin
complexes have been solved and deposited
in the Protein Data Bank. In addition, a
number of prothrombin fragments and in-
termediates, the three-dimensional struc-
ture of the immediate precursor of a-
thrombin, prethrombin-2 (pre-cleavage cat-
alytic domain), have also been reported
[26].
1.1.2.2 Plasminogen Activation
The activation of plasminogen is the key
event in the fibrinolytic system, leading to
the degradation of fibrin by the active en-
zyme plasmin and as a consequence to the
dissolution of blood clots (intravascular
proteolysis [27]). Plasmin also promotes
cell migration and tissue remodeling,
plays a key role in a variety of other activa-
tion cascades such as the activation of me-
talloproteinases, and has been implicated
in wound healing, tissue remodeling, an-
giogenesis, embryogenesis, pathogen and
tumor cell invasion, and metastasis (peri-
cellular proteolysis [28]). Both eukaryotic
cancer cells and prokaryotic pathogenic
microorganisms recruit the proteolytic ac-
tivity of plasmin to their cell surface to fa-
cilitate cell invasion and migration
through tissue layers.
Plasminogen is a modular protein that
comprises a pre-activation peptide, fol-
lowed by five kringle domains and a cata-
lytic C-terminal serine proteinase domain.
In the blood, plasminogen circulates in a
globular, closed conformation; when
bound to a surface, it adopts an extended,
open conformation that is more rapidly
activated to plasmin [29]. The physiological
plasminogen activators t-PA and uroki-
nase-type plasminogen activator (u-PA or
urokinase) form a fibrin-bound complex
with plasminogen, activating it and yield-
ing the 85-kDa two-chain (A and B) serine
proteinase plasmin. The activation of trun-
cated plasminogen derivatives, mini-plas-
minogen (kringle 5 and catalytic domain)
and micro-plasminogen (the catalytic do-
main), is slower than for full-length plas-
minogen, suggesting a role for the kringle
domains in the activation process. Elastase
cleavage yields two fragments, angiostatin
(kringle 14) and mini-plasminogen (krin-
gle 5 and the catalytic domain) [30]. The
kringles contain lysine-binding sites that
mediate the localization to fibrin and cellu-
lar surfaces, and serve as binding loci for
other plasma proteins. The structures of l-
plasminogen [31], l-plasmin [32, 33] and
some kringle domains have been pub-
lished, but as yet there is no full-length
structure of plasminogen available.
Dissolution of fibrin clots is the key strat-
egy in the short-term clinical treatment of
blood clotting disorders, especially in acute
myocardial infarction (see also Part II,
Chapter 3 and Part III, Chapter 6). Blood
clot lysis is initiated by plasminogen activa-
tion using recombinant t-PA or streptoki-
nase (SK), a plasminogen activator from
streptococci. SK is relatively inexpensive
[34], but because of its non-human origin
its use is associated with undesired im-
mune responses. Furthermore, SK and also
u-PA activate not only fibrin-bound, but also
circulating plasminogen, leading to serious
risks of hemorrhage. In contrast, t-PA is
more specific in its action, binding relative-
ly strongly to fibrin clots and preferentially
activating the plasminogen entrapped in
the clots, while it has little effect on freely
circulating plasminogen or other blood clot-
ting factors [35]. Staphylokinase (SAK), a
staphylococcal plasmin cofactor enabling
plasmin to activate plasminogen, is also
able to enhance fibrinolysis in a fibrin-de-
pendent manner [36], but causes high titers
of neutralizing antibodies from the second
week after infusion into patients; generat-
ing variants with reduced antigenicity
seems to be feasible, though [37] (see also
the Introduction to this volume, and Part
V, Chapters 1 and 2).
1.2
Bacterial Activators of Host Zymogens
A common strategy of bacterial parasites
is the exploitation and subversion of host
signaling pathways and other processes
[38]. Proteolysis is an important compo-
nent in pathogenesis and serves several
functions. Proteinases with a broad sub-
strate specificity range release amino acids
or peptides from mammalian tissues or in-
crease the vascular permeability, creating a
path for nutrients to the site of infection
[39, 40]. More specific targets for bacterial
proteinases are host proteinase cascades,
including coagulation, fibrinolysis, comple-
ment activation, phagocytosis and the kal-
likreinkinin cascade. Bacteria can activate
or inactivate these systems through their
proteinases or the release of proteinase co-
factors. The activation of host proteinase
zymogens can lead to uncontrolled proteo-
lytic activity at the infection site and sub-
stantial tissue damage [41]; tissue lesions
around the infection site may facilitate the
dissemination of bacteria through tissue
barriers.
Several invasive pathogens express plas-
minogen-binding proteins or receptors
which immobilize plasminogen on the
bacterial surface and enhance its activation
by mammalian plasminogen activators
[42]. These receptors turn the bacteria into
proteolytic organisms capable of degrading
and invading the extracellular matrix and
basement membranes. Some pathogenic
Gram-positive bacteria (such as streptococ-
ci or staphylococci) express proteins that
specifically activate the human blood coag-
ulation and fibrinolytic systems or stimu-
late host cells to secrete plasminogen acti-
vators and their inhibitors. Bacterial protei-
nase cofactors enhance the presentation of
the substrate to the enzyme; at the same
time, they modulate the specificity of their
cognate host enzyme towards other sub-
strates and inhibitors (specificity switch)
[43]. In some cases, however, bacterial acti-
vators cleave the host zymogen similar to
endogenous proteinases at its Arg15Ile/
Val16 activation site. The activator is, how-
ever, not necessarily a chymotrypsin-like
serine proteinase itself.
1.2.1
Proteolytic Activators
The surface proteinase Pla from Yersinia
pestis is the causative agent of plague [44].
Inactivation of the gene encoding Pla in Y.
pestis increases the median lethal dose of
the bacterium for mice by a 10
6
-fold. The
outer membrane protein is responsible for
two in vitro phenotypes on Y. pestis: a very
weak, probably unphysiological procoagu-
lant activity and lysis of fibrin clots. Pla
cleaves plasminogen at the same site and
with similar efficiency as t-PA or u-PA. Es-
cherichia coli and Salmonella typhimurium
carry the chromosomal Pla homologs
OmpT and PgtE. These proteins show no
or only weak plasminogen activator activity
and are likely to serve other functions in
the E. coli membrane protein metabolism
[4547], but together with Pla and SopA
from Shigella flexneri they form the so-
called omptin family of outer membrane
proteinases [48]. The crystal structure of
Pla is not known, although a structure of
OmpT, which shares 50% identical resi-
dues with Pla, was reported [49]. OmpT
and probably also Pla consist of a huge 10-
stranded antiparallel b-barrel of 70 in its
longest dimension and a diameter of about
32 at the top. The strands of about 23
residues each run at an angle of about 408
with respect to the barrel axis. The OmpT
barrel is hollow and negatively charged on
the inner wall. The extracellular part of
the molecule contains a large negatively
charged groove that harbors the active site
residues. The 18 residues within this
groove are fully conserved among all omp-
tins. Mutagenesis studies have shown that
substitution of residues Asp83, Asp85,
Asp218 and His212 leads to a 10000-fold
reduced activity of OmpT. Most likely, a
water molecule positioned between Asp83
and His212 is activated by the His212
Asp210 dyad and then performs a nucleo-
philic attack on the scissile peptide bond.
A similar mechanism might apply for the
plasminogen activator activity of Pla.
An 80-kDa proteinase from P. gingivalis
activates plasminogen and several inhibi-
tors, leading to uncontrolled degradation
of periodontal tissue [50]. The activation
mechanism of this protein remains to be
elucidated.
1.2.2
Nonproteolytic Bacterial Activators
Some bacterial zymogen activators are
non-enzymatic proteins that bind tightly to
the cognate serine proteinase zymogen
and promote the formation of a functional
active site without proteolysis of the pep-
tide bond following Arg or Lys15 (Scheme
1.3). In addition, these bacterial cofactors
offer novel docking sites for enhanced ac-
tive-site presentation of the substrate,
which can either correspond to the physio-
logical target or represent a novel specifici-
ty. The pathogen activators SK, staphylo-
coagulase (SC) and SAK are not enzymes
themselves, but form 1: 1 complexes with
plasminogen, prothrombin and plasmin,
respectively. SK and SC activate their cog-
nate zymogens nonproteolytically and con-
formationally, while SAK changes the sub-
strate specificity of plasmin, enabling it to
activate plasminogen.
1.2.2.1 SAK
SAK is a 136-amino acid protein produced
by strains of S. aureus that carry a pro-
phage with the sak gene. It is synthesized
during the late exponential growth phase
and responsible for the lysogenic conver-
sion of the bacteria [36]. A few coagulase-
negative staphylococci alternatively express
either l-hemolysin or SAK [51]. In these
strains the sak gene is carried by a convert-
ing phage which inactivates the l-hemoly-
sin gene during lysogeny. SAK production
mediates the a-defensin resistance of S.
aureus [52].
SAK is folded into a mixed five-stranded,
slightly twisted b-sheet wrapped around a
central a-helix and two short two-stranded
b-sheets opposing the central sheet [32],
the so-called b-grasp motif. The convex sur-
face of SAK nestles against the multiple-
turn structure of plasmin around Arg175.
An overall negative potential on the SAK
surface (the b3b4 loop) has a counterpart
in a positively charged patch on the plasmin
surface, which leads to SAK pre orientation
upon formation of the complex. SAK pos-
sesses a flexible N-terminal tail (residues
115), which is just long enough to reach
the active site of the cognate plasmin mole-
cule to be processed between K10 and K11.
This newly created SAK11136 could then
insert its N-terminal lysine into the lysine
binding site of kringle 5 of a substrate plas-
minogen. SAK does not alter the active site
conformation of the enzyme, but modifies
its specificity by restricting the S2 and S3/
S4 pockets (making them more similar to
t-PA and u-PA), and additionally offers an
exosite surface onto which plasminogen
can dock for dramatically enhanced presen-
tation of the plasminogen activation loop to-
wards the enzyme. In this way, SAK confers
a preference of plasmin for plasminogen
over fibrin.
The SAKplasminogen complex is enzy-
matically inactive and requires conversion
of plasminogen to plasmin. The initial
step in activation involves association of
SAK with trace amounts of plasmin
formed as a result of weak spontaneous
plasminogen activation. The SAKplasmin
complex formation is favored by the 160-
fold higher affinity of SAK for plasmin
than for plasminogen. Binding of a
2
-anti-
plasmin to the SAKplasmin complex re-
leases SAK from the complex and allows
binding to other plasmin(ogen) molecules
[53]. SAK primarily activates plasminogen
bound to fibrin without causing systemic
plasminogen activation [54] (see Fig. 1.1).
The plasminogen activation process via
SAK involves several characteristics includ-
ing proteinprotein interaction and com-
plex formation, apart from proteolytic
cleavage of SAK itself, which removes the
first 10 N-terminal residues of the full-
Scheme 1.3
length SAK and exposes K11 in the en-
zyme complex. The removal of 10 amino
acids may be essential for unmasking the
functional core of SAK to expose its full
activation potential. The N-terminal region
of SAK modulates the interaction of the
enzyme with the substrate, but may not
have any significant role in the formation
of binary complex to generate an initial
enzyme complex consisting of SAK and
plasminogen [55]. Four clustered charged
segments are important for the functional
properties of SAK; apart from the positive-
ly charged N-terminus, two discrete seg-
ments of SAK (Glu44Lys50 and Glu65
Asp69) form the core region of SAK, and
may be involved in plasminogen binding
and activation [56]. Met26 is part of a hy-
drophobic network having surface comple-
mentarity to the C-terminal region of plas-
min; this residue has been shown to be
critical for the efficient activation of plas-
minogen by SAK [57].
1.2.2.2 SK
SK is a single-chain 414-amino acid pro-
tein secreted by l-hemolytic group A, C
and G streptococci [58]. The base for its
use as a thrombolytic drug is that SK
forms a 1: 1 complex with plasminogen
and activates it without proteolytic cleav-
age. Plasminogen becomes an efficient
plasminogen activator and the newly
formed plasmin in turn can catalyze the
hydrolysis of fibrin. The overall amino acid
sequence identity between the SKs is 80
98%; the variable amino acid residues are
clustered in two regions designated V1
and V2 (residues 147218 and 244264)
[59]. SK has been implicated with the
pathogenesis of the kidney disease post-
streptococcal glomerulonephritis (PSGN).
V1 is considered the domain with which
SKs from nephritogenic strains bind to
glomerular structures and activate plasmi-
nogen in situ, thus triggering a cascade of
proteolytic processes leading to PSGN [60].
SK appears as three domains, termed a,
b and c, of similar folding, separated by
two coiled coils [33]. Domains a and b
each contain a major b-sheet of five mixed
b-strands and an a-helix, a typical structure
of the b-grasp folding class. The c domain
has only four b-strands and contains a
long coiled-coil segment instead of an a-
helix. The a domain binds to plasmin
mainly through interactions between the
b1 and b2 strands of SK a and a loop re-
gion of plasmin; SK a also interacts with
Fig. 1.1 Stereo ribbon plot of the SAKplasminogen complex.
SAK is shown in green and plasminogen in orange.
plasminogen near the catalytic triad resi-
dues His57 and Asp102, and Ser195. The
interaction of the SK b domain with plas-
min is relatively meager, but it may direct-
ly interact with the kringles of plasmino-
gen in the activator complex and with the
substrate plasminogen. The SK c domain
binds to plasmin near the activation cleav-
age site of plasminogen. A multitude of
charged and hydrophobic interactions sta-
bilizes the complex. On the plasmin side,
the calcium-binding loop (7080 loop) and
the autolysis loop (148 loop) are involved;
on the SK side, the major coiled coil re-
gion and the strands b1 and b2. The par-
ticipation of the calcium-binding loop in
this interaction suggests that the substan-
tial sequence difference observed in this
region between human and bovine plasmi-
nogen may contribute to the inability of
SK to activate the latter. There is no direct
interaction of SK with the plasmin activa-
tion loop (residues around Val16).
After formation of a binary plasmino-
genSK complex, the active site of plasmi-
nogen is exposed and functional without
cleavage of the Arg15Val16 peptide bond.
A so-called substrate plasminogen mole-
cule can then bind to the SK a domain in
the binary complex to form a ternary plas-
minogenSKplasminogen complex. The
substrate plasminogen is converted to
plasmin and subsequently released from
the ternary complex. SKplasminogen can
be converted to SKplasmin and still cata-
lyze the conversion of additional plasmino-
gen to plasmin; thus, SK changes the
specificity of the active proteinase in addi-
tion to inducing an active conformation in
the zymogen (see Fig. 1.2).
Recent studies have implicated an inter-
action of the SK N-terminal Ile1 residue
with the N-terminal binding pocket of
plasminogen as a critical step in the mech-
anism [61]. The near total loss of activity
of SK2414 could be partially restored by
specific binding of peptides based on the
N-terminal 1015 residue sequence of SK
[62]. The N-terminal sequence of SK (Ile
AlaGly) mimics that of the catalytic do-
main of plasmin (ValValGly). Together,
the results support a mechanism in which
Fig. 1.2 Stereo ribbon plot of the SKplasminogen complex.
SK is shown in shades of blue (indicating the a, b and c do-
mains) and plasminogen in orange.
the N-terminal sequence of SK interacts in
a sequence-specific manner with the N-ter-
minal binding cleft of plasminogen to trig-
ger the transition toward the active form
which is stabilized by high affinity binding
of SK to the active conformation of the
proteinase domain. This intrusion mecha-
nism is known as the molecular sexuality
hypothesis [11]. In the SKl-plasmin com-
plex crystal structure [33], this N-terminal
insertion mechanism could not be vali-
dated because in active l-plasmin the en-
dogenous insertion blocks the SK inser-
tion. A few studies support an additional
or alternative mechanism the insertion
of plasmins own Lys156 into the activation
pocket [63].
1.2.2.3 SC
SCs are proteins secreted by certain strains
of S. aureus, with molecular weights of
5477 kDa. Being able to adhere to fibrino-
gen through 58 tandem 27-amino-acid C-
terminal repeat sequences [64], they
furthermore have the unique ability to
form a 1: 1 complex with prothrombin
(staphylothrombin) and to activate it
without the usually required peptide bond
cleavages [6567]. Since the SCprothrom-
bin complex efficiently clots the major
physiological thrombin substrate, fibrino-
gen, SC action bypasses the blood coagula-
tion pathways. In vivo, SC is not required
for the initial infectivity of S. aureus [68,
69], but contributes to the pathogenesis of
acute bacterial endocarditis, characterized
by formation of vegetations on heart valves
consisting of bacteria, platelets and fibrin
[70]. Large vegetations are friable and em-
bolize, causing remote abscess formation,
and ultimately leading to heart failure,
myocardial infarction or stroke [70, 71].
Growth and fortification of the vegetation
by SC-induced fibrin deposition protects
the bacteria in the vegetation from clear-
ance by leukocytes and macrophages [72].
Coagulase-positive S. aureus causes 40
50% of neonatal endocarditis and 3040%
of endocarditis in adults between the ages
of 16 and 60 years, with a mortality rate of
2547%, even with antibiotic therapy [71,
73]. Because there are no known physio-
logical inhibitors of SC(pro)thrombin
complexes, they are resistant to conven-
tional anticoagulant therapy, except for
small molecule active-site-directed inhibi-
tors such as argatroban [74].
The sequences determined for SC from
a variety of S. aureus strains are highly ho-
mologous, but differ in the length of the
C-terminal region [75]. The primary struc-
ture includes eight repeating tandem re-
gions of 27 residues each at the C-termi-
nus. The chymotryptic fragment SC1324
lacking the repeats has essentially equiva-
lent properties to full-length SC in binding
and activating prethrombin-1 and generat-
ing fibrinogen clotting activity [76].
SC126324 lost fibrinogen clotting activity
while retaining significant affinity for pro-
thrombin and conformational activation.
SC126278 retains some affinity for pro-
thrombin, but does not activate the catalyt-
ic site or support clotting activity [77].
SC1324 consists of two strongly inter-
acting, rod-like triple-helical domains of
previously uncharacterized fold and there-
fore constitutes an entirely new class of
bacterial proteinase activators. The two do-
mains are arranged to each other at an an-
gle of around 1108 and are superimposa-
ble, pointing to a distant gene duplication
event [77]. SC contacts the cognate enzyme
primarily at two surface sites, the 148 or
autolysis loop which nestles into a groove
on the surface of SC domain 1, and the
anion binding exosite I or fibrinogen-rec-
ognition exosite, a positively charged sur-
face patch on thrombin which is required
for the recognition of the major thrombin,
fibrinogen. Paradoxically, the specificity of
thrombin for fibrinogen is even increased
in the SC(pro)thrombin complex. The so-
lution for this apparent contradiction
might be the fact that in the crystals [77]
two SC1324(pre)thrombin monomers
meet across an interface of 625
2
, and
form a crystallographic dimer via a cluster
of aromatic residues and the interlocking
effect of two protruding finger helices.
This dimerization is unique amongst the
nonproteolytic activators and generates an
environment well suited for the binding of
the intrinsically dimeric substrate fibrino-
gen via a new, enhanced exosite on the SC
surface (see Fig. 1.3).
The strict conservation of the N-terminal
peptide in all SCs sequenced so far and
the important role of the equivalent pep-
tide in SK-mediated plasminogen activa-
tion [61, 62] strongly suggested that the in-
tact N-terminus of SC is important for
prothrombin activation. In the crystal
structure of SC1324prethrombin-2 [77],
the hexapeptide Ile1Tyr6 is indeed fully
defined by electron density, with the Ile1
Val2Thr3 residues occupying the Ile16-
binding pocket of the cognate prethrom-
bin-2. The negatively charged environment
created by Asp194 and neighboring carbo-
nyl oxygens accommodates the free N-ter-
minus of Ile1, which forms a strong bur-
ied salt bridge with the latter. The activat-
ing peptide binds in a conformation neces-
sarily slightly different from the
endogenous N-terminus, but recapitu-
lates critical elements of the enzyme struc-
ture. An SC variant with an additional
methionine residue and SC2324 both still
activate prothrombin, with reduced po-
tency, however, showing a surprising pro-
miscuity in the prothrombin activation
pocket, in contrast to the SKplasminogen
complex where Ile1 is strictly required for
conformation activation [61]. This fact is
consistent with the recent observation that
SC containing an additional N-terminal
alanine clots human plasma [78]. The SC
prethrombin-2 crystal structure provided
direct proof for the molecular sexuality hy-
pothesis. As a matter of course, this mech-
anism will also be valid for the SKplasmi-
nogen activator complex as speculated be-
forehand [61] (see Fig. 1.4).
In spite of the extensive contacts be-
tween both moieties, the structure of SC1
324-bound a-thrombin is essentially unal-
tered compared to the over 100 crystal
structures of the enzyme deposited with
the Protein Data Bank. This observation
supports the contention that bacterial co-
Fig. 1.3 Stereo ribbon plot of the SCprethrombin-2 complex.
SC is shown in yellow and prethrombin-2 in orange.
factors of serine proteinases including also
SK and SAK modulate the bound enzyme
by offering a new specific surface site for
optimal presentation of the substrate(s),
not by reshaping the active-site conforma-
tion of the enzyme.
1.2.2.4 More Nonproteolytic Zymogen
Activators
SUPA or PauA, a protein from Streptococ-
cus uberis, whose 251 amino acids show
limited primary structure homology to SK,
is responsible for the pathogenesis of bo-
vine mastitis, an infection of the udder
[79]. S. uberis uses peptides released from
plasminogen from milk casein to meet its
amino acid demands [80]. The two-domain
protein lacks an equivalent to the SK c do-
main, leading to faster complex genera-
tion, but the complex is less firm and sen-
sitive to inhibition by a
2
-antiplasmin.
SUPA does not share significant nucleo-
tide or genomic homology with SK nor
SAK, but non-proteolytically activates bo-
vine plasminogen through a SK- and not a
SAK-type mechanism. The IleThrGly N-
terminus of SUPA points towards a molec-
ular sexuality activation mechanism.
Recently, a number of bacterial proteins
sharing sequence and apparently second-
ary structure homology with SC (ZAAPs)
could be identified by database searches
[77]. Some of them could be already
shown to possess at least weak prothrom-
bin activator activity [78]. Whether pro-
thrombin or other host serine proteinases
represent the actual targets for these pro-
teins remains to be elucidated.
1.3
Some Remarks on Nonproteolytic Activators
A multitude of bacterial pathogen proteins
have been characterized and crystallized
recently. Amongst them, cofactors of hu-
man serine proteinases or their zymogens
Fig. 1.4 The prethrombin-2 activation pocket in
the SC-prethrombin-2 complex. Prethrombin-2 is
shown in solid surface representation. The side-
chain of Asp194 is exempt from the surface and
shown in pink, with the oxygen atoms shown in
red. The N-terminal hexapeptide of SC is shown
in green, while the N-terminal nitrogen is shown
in blue.
play a central role in the subversion of
host pathways. The cofactors form 1: 1
complexes with their target proteinases.
Upon complex formation, the specificity
switches whereas plasmin shows a pref-
erence for extended substrates, stretching
across the entire active-site cleft like fibri-
nogen, the SKplasminogen complex, just
like the SAKplasmin complex, has activity
against narrower substrates, such as the
activation loop of plasminogen, which
stretches across the active site only from
P3 to P2'. The same accounts for throm-
bin its specificity is narrowed down from
a variety of substrates to only fibrinogen
and, with some uncertainty, factor XIII
[81] in the SC complex.
SK uses a fibrin-independent mecha-
nism of plasminogen activation, limiting
medical applications compared to SAK.
The SAKplasmin complex can, like plas-
min alone, be inhibited by a
2
-antiplasmin
in solution, restricting plasminogen activa-
tion to fibrin or cellular surfaces, while the
SKplasminogen complex in solution es-
capes inhibition, leading to systemic plas-
minogen activation. An a-domain-less SK
(SKD159) behaves similar to SAK in that
it becomes able to convert fibrin-bound
plasmin into a plasminogen activator
rather than to activate plasminogen [82],
i.e., it has lost the ability to perform the
molecular sexuality process.
The bacterial cofactors provide binding
surfaces (exosites) onto which the sub-
strate can dock in an optimal orientation
for efficient cleavage. SAK and SK a assist
in proper substrate pre-orientation and
presentation, SK b provides a further sub-
strate-anchoring site that also modulates
the interaction of plasmin with macromo-
lecular inhibitors, SK c seems to partici-
pate in the so-called binding activation
upon complex formation with plasmino-
gen. SC occupies the fibrinogen binding
exosite on (pro)thrombin and must thus
express a new, even more specific fibrino-
gen binding exosite given the enhanced
specificity of SC-bound (pro)thrombin for
fibrinogen. Similar cofactor-mediated sub-
strate-presentation mechanisms also occur
in other fibrinolytic and thrombotic reac-
tions. In fibrinolysis, fibrin and a recently
described t-PA receptor apparently play
such a cofactor role during plasminogen
activation by t-PA, assembling both reac-
tants in an optimal manner. Coagulation
factors FVIIIa and FVa may likewise ex-
pose additional surfaces for enhanced pre-
sentation of the zymogen substrates to-
wards the activating proteinases, leading to
a tremendous cleavage rate enhancement
(see also Part II, Chapter 3). In extrinsic
Xase (the complex catalyzing FX activa-
tion), the cofactor tissue factor presumably
offers extra surfaces for FX presentation to
the enzyme FVIIa and renders the en-
zyme more active. Cofactor modulation of
salt-bridge interactions, either directly or
via allosteric interactions, appears to be a
common mechanism for regulating protei-
nase activity (see Table 1.1).
Strikingly, SAK and the separate domains
of SK share the same fold (b-grasp), but
neither similarity on the sequence level
nor any functional relationship in the acti-
vation mechanism. While SAK changes
the specificity of the active proteinase plas-
min from a fibrin-degrading enzyme to a
plasminogen-activating enzyme, SK utilizes
the molecular sexuality mechanism to acti-
vate prothrombin. On the other hand, SC
exhibits a novel triple-helical fold and yet
employs the same activation (N-terminal in-
sertion) mechanism as SK. Thus, the b-
grasp motif in bacterial cofactors as well
as the molecular sexuality mechanism must
undergone convergent evolution.
It might be possible to inhibit bacterial
proliferation in strains that utilize the host
activation systems for invasion or coloniza-
tion by selectively inhibiting the formation
or action of the bacterial zymogen activa-
tion complexes. Learning from nature, this
again might lead to new and potent bio-
pharmaceuticals in the foreseeable future.
Acknowledgments
We would like to cordially acknowledge
helpful discussions with Drs. Paul E. Bock,
Wolfram Bode, Pablo Fuentes-Prior and
Peter Panizzi.
References
1 Stroud, R. M., A. A. Kossiakoff, J. L. Chambers,
Mechanisms of zymogen activation. Annu Rev
Biophys Bioeng 1977, 6, 177193.
2 Madison, E. L., et al., Converting tissue plas-
minogen activator to a zymogen: a regulatory
triad of AspHisSer. Science 1993, 262, 419
421.
3 Bode, W., The transition of bovine trypsinogen
to a trypsin-like state upon strong ligand bind-
ing. II. The binding of the pancreatic trypsin
inhibitor and of isoleucine-valine and of se-
quentially related peptides to trypsinogen and
to p-guanidinobenzoate-trypsinogen. J Mol
Biol 1979, 127, 357374.
4 Robinson, N. C., H. Neurath, K. A. Walsh, The
relation of the a-amino group of trypsin to en-
zyme function and zymogen activation. Bio-
chemistry 1973, 12, 420426.
5 Ranby, M., N. Bergsdorf, T. Nilsson, Enzy-
matic properties of the one- and two-chain
form of tissue plasminogen activator. Thromb
Res 1982, 27, 175183.
6 Davie, E. W., Ratnoff, O. D., Watterfall Se-
quence for Intrinsic Blook Clotting. Science
1964, 145, 13101312.
7 McFarlane, R. G., An Enzyme Cascade in the
Blood Clotting Mechanism, and its Function
as a Biochemical Amplifier. Nature 1964, 202,
498499.
8 Bode, W., P. Schwager, R. Huber, The transi-
tion of bovine trypsinogen to a trypsin-like
state upon strong ligand binding. The refined
crystal structures of the bovine trypsinogen
pancreatic trypsin inhibitor complex and of its
ternary complex with IleVal at 1.9 resolu-
tion. J Mol Biol 1978, 118, 99112.
9 Huber, R., W. Bode, Structural basis of the ac-
tivation and action of trypsin. Acc Chem Res
1978, 11, 114122.
10 Kraut, J., Serine proteases: structure and
mechanism of catalysis. Annu Rev Biochem
1977, 46, 331358.
11 Bode, W., R. Huber, Induction of the bovine
trypsinogentrypsin transition by peptides se-
quentially similar to the N-terminus of tryp-
sin. FEBS Lett 1976, 68, 231236.
12 Sigler, P. B., et al., Structure of crystalline chy-
motrypsin. II. A preliminary report including
a hypothesis for the activation mechanism. J
Mol Biol 1968, 35, 143164.
Table 1.1 Serine proteinase activation mechanisms
Mechanism Example
Substrate Activator (complex)
Cleavage by endogenous proteinase prothrombin prothrombinase complex
plasminogen u-PA or t-PA
Cleavage by bacterial proteinase plasminogen Pla (Y. pestis)
Cleavage by bacterial cofactorendogenous
proteinase complex (specificity switch)
plasminogen SAKplasmin complex
Conformational activation by bacterial cofactor
endogenous zymogen complex
plasminogen SKplasminogen complex
fibrinogen SCprothrombin complex
13 Wang, D., W. Bode, R. Huber, Bovine chymo-
trypsinogen A X-ray crystal structure analysis
and refinement of a new crystal form at 1.8
resolution. J Mol Biol 1985, 185, 595624.
14 Blow, D. M., Chymotrypsin: tertiary structure
and enzymic activity. Biochem J 1968, 110, 2P.
15 Robertus, J. D., et al., Subtilisin; a stereochem-
ical mechanism involving transition-state sta-
bilization. Biochemistry 1972, 11, 4293303.
16 Stubbs, M. T., W. Bode, Coagulation factors
and their inhibitors. Curr Opin Struct Biol
1994, 4, 823832.
17 Mann, K. G., The biochemistry of coagulation.
Clin Lab Med 1984, 4, 207220.
18 Blomback, B., M. Blomback, The molecular
structure of fibrinogen. Ann NY Acad Sci
1972, 202, 7797.
19 Folk, J. E., J. S. Finlayson, The epsilon-(gam-
ma-glutamyl)lysine crosslink and the catalytic
role of transglutaminases. Adv Protein Chem
1977, 31, 1133.
20 Fuentes-Prior, P., et al., Structural basis for
the anticoagulant activity of the thrombin-
thrombomodulin complex. Nature 2000, 404,
518525.
21 Nelsestuen, G. L., Role of gamma-carboxyglu-
tamic acid. An unusual protein transition re-
quired for the calcium-dependent binding of
prothrombin to phospholipid. J Biol Chem
1976, 251, 56485656.
22 Crook, M., Platelet prothrombinase in health
and disease. Blood Coag Fibrinol 1990, 1, 167
174.
23 Tracy, P. B., M. E. Nesheim, K. G. Mann, Coor-
dinate binding of factor Va and factor Xa to
the unstimulated platelet. J Biol Chem 1981,
256, 743751.
24 Mann, K. G., et al., Surface-dependent reac-
tions of the vitamin K-dependent enzyme
complexes. Blood 1990, 76, 116.
25 Bode, W., et al., The refined 1.9 crystal
structure of human alpha-thrombin: interac-
tion with d-PheProArg chloromethylketone
and significance of the TyrProProTrp inser-
tion segment. EMBO J 1989, 8, 34673475.
26 Vijayalakshmi, J., et al., The isomorphous
structures of prethrombin2, hirugen-, and
PPACK-thrombin: changes accompanying ac-
tivation and exosite binding to thrombin. Pro-
tein Sci 1994, 3, 22542271.
27 Chapman, H. A., Plasminogen activators, in-
tegrins, and the coordinated regulation of cell
adhesion and migration. Curr Opin Cell Biol
1997, 9, 714724.
28 Booth, N. A., Fibrinolysis and thrombosis.
Baillieres Clin Haematol 1999, 12, 423433.
29 Cockell, C. S., et al., Evidence that the confor-
mation of unliganded human plasminogen is
maintained via an intramolecular interaction
between the lysine-binding site of kringle 5
and the N-terminal peptide. Biochem J 1998,
333, 99105.
30 OReilly, M. S., et al., Angiostatin induces and
sustains dormancy of human primary tumors
in mice. Nat Med 1996, 2, 689692.
31 Peisach, E., et al., Crystal structure of the
proenzyme domain of plasminogen. Biochem-
istry 1999, 38, 1118011188.
32 Parry, M. A., et al., The ternary microplasmin
staphylokinasemicroplasmin complex is a
proteinasecofactorsubstrate complex in ac-
tion. Nat Struct Biol 1998, 5, 917923.
33 Wang, X., et al., Crystal structure of the cataly-
tic domain of human plasmin complexed with
streptokinase. Science 1998, 281, 16621665.
34 Parmley, W. W., Cost-effectiveness of reperfu-
sion strategies. Am Heart J 1999, 138, S142
152.
35 Medved, L., W. Nieuwenhuizen, Molecular
mechanisms of initiation of fibrinolysis by fi-
brin. Thromb Haemost 2003, 89, 409419.
36 Collen, D., Staphylokinase: a potent, uniquely
fibrin-selective thrombolytic agent. Nat Med
1998, 4, 279284.
37 Laroche, Y., et al., Recombinant staphyloki-
nase variants with reduced antigenicity due to
elimination of B-lymphocyte epitopes. Blood
2000, 96, 14251432.
38 Finlay, B. B., S. Falkow, Common themes in
microbial pathogenicity revisited. Microbiol
Mol Biol Rev 1997, 61, 136169.
39 Travis, J., J. Potempa, H. Maeda, Are bacterial
proteinases pathogenic factors? Trends Micro-
biol 1995, 3, 405407.
40 Goguen, J. D., N. P. Hoe, Y. V. Subrahmanyam,
Proteases and bacterial virulence: a view from
the trenches. Infect Ag Dis 1995, 4, 4754.
41 Lahteenmaki, K., P. Kuusela, T. K. Korhonen,
Bacterial plasminogen activators and recep-
tors. FEMS Microbiol Rev 2001, 25, 531552.
42 Herwald, H., et al., Activation of the contact-
phase system on bacterial surfaces a clue to
serious complications in infectious diseases.
Nat Med 1998, 4, 298302.
References 391
43 Parry, M. A., X. C. Zhang, I. Bode, Molecular
mechanisms of plasminogen activation: bacte-
rial cofactors provide clues. Trends Biochem Sci
2000, 25, 5359.
44 Sodeinde, O. A., et al., A surface protease and
the invasive character of plague. Science 1992,
258, 10041007.
45 Sodeinde, O. A., J. D. Goguen, Nucleotide se-
quence of the plasminogen activator gene of
Yersinia pestis: relationship to ompT of Escheri-
chia coli and gene E of Salmonella typhimuri-
um. Infect Immun 1989, 57, 15171523.
46 Lundrigan, M. D., R. M. Webb, Prevalence of
ompT among Escherichia coli isolates of hu-
man origin. FEMS Microbiol Lett 1992, 76, 51
56.
47 White, C. B., et al., A novel activity of OmpT.
Proteolysis under extreme denaturing condi-
tions. J Biol Chem 1995, 270, 1299012994.
48 Kukkonen, M., T. K. Korhonen, The omptin
family of enterobacterial surface proteases/ad-
hesins: from housekeeping in Escherichia coli
to systemic spread of Yersinia pestis. Int J Med
Microbiol 2004, 294, 714.
49 Vandeputte-Rutten, L., et al., Crystal structure
of the outer membrane protease OmpT from
Escherichia coli suggests a novel catalytic site.
EMBO J 2001, 20, 50335039.
50 Grenier, D., Degradation of host protease inhi-
bitors and activation of plasminogen by pro-
teolytic enzymes from Porphyromonas gingiva-
lis and Treponema denticola. Microbiology 1996,
142, 955961.
51 Sawicka-Grzelak, A., et al. [Production of sta-
phylokinase and hemolysin by coagulase-nega-
tive staphylococcus]. Medycyna Doswiadczalna i
Mikrobiologia 1993, 45, 710.
52 Jin, T., et al., Staphylococcus aureus resists hu-
man defensins by production of staphyloki-
nase, a novel bacterial evasion mechanism. J
Immunol 2004, 172, 11691176.
53 Lijnen, H. R., et al., Characterization of the in-
teraction between plasminogen and staphylo-
kinase. Eur J Biochem 1994, 224, 143149.
54 Sakharov, D. V., H. R. Lijnen, D. C. Rijken, In-
teractions between staphylokinase, plasmin-
(ogen), and fibrin. Staphylokinase discrimi-
nates between free plasminogen and plasmi-
nogen bound to partially degraded fibrin. J
Biol Chem 1996, 271, 2791227918.
55 Jespers, L., et al., Structural and functional ba-
sis of plasminogen activation by staphyloki-
nase. Thromb Haemost 1999, 81, 479485.
56 Silence, K., et al., Structurefunction relation-
ships in staphylokinase as revealed by clus-
tered charge to alanine mutagenesis. J Biol
Chem 1995, 270, 2719227198.
57 Kim, S. H., et al., A novel variant of staphylo-
kinase gene from Staphylococcus aureus ATCC
29213. Thromb Res 1997, 87, 387395.
58 Castellino, F. J., et al., Streptokinase. Methods
Enzymol 1976, 45, 244257.
59 Malke, H., et al., Expression and regulation of
the streptokinase gene. Methods 2000, 21,
111124.
60 Malke, H., Polymorphism of the streptokinase
gene: implications for the pathogenesis of
post-streptococcal glomerulonephritis. Int J
Med Microbiol 1993, 278, 246257.
61 Wang, S., G. L. Reed, L. Hedstrom, Zymogen
activation in the streptokinaseplasminogen
complex. Ile1 is required for the formation of
a functional active site. Eur J Biochem 2000,
267, 39944001.
62 Boxrud, P. D., et al., Streptokinase triggers
conformational activation of plasminogen
through specific interactions of the amino-ter-
minal sequence and stabilizes the active zy-
mogen conformation. J Biol Chem 2001, 276,
2608426089.
63 Terzyan, S., et al., Characterization of Lys-698
to met substitution in human plasminogen
catalytic domain. Proteins 2004, 56, 277284.
64 Heilmann, C., et al., Platelet-binding domains
in 2 fibrinogen-binding proteins of Staphylo-
coccus aureus identified by phage display. J In-
fect Dis 2002, 186, 3239.
65 Kawabata, S., et al., Enzymatic properties of
staphylothrombin, an active molecular com-
plex formed between staphylocoagulase and
human prothrombin. J Biochem 1985, 98,
16031614.
66 Kawabata, S., S. Iwanaga, Structure and func-
tion of staphylothrombin. Semin Thromb Hae-
most 1994, 20, 345350.
67 Morita, T., et al. [Activation mechanism of hu-
man prothrombin by staphylocoagulase]. Rin-
sho Byori (Jpn J Clin Pathol) 1982, Suppl, 127
136.
68 Moreillon, P., et al., Role of Staphylococcus
aureus coagulase and clumping factor in
pathogenesis of experimental endocarditis. In-
fect Immun 1995, 63, 47384743.
69 Phonimdaeng, P., et al., The coagulase of Sta-
phylococcus aureus 8325-4. Sequence analysis
and virulence of site-specific coagulase-defi-
cient mutants. Mol Microbiol 1990, 4, 393404.
70 Korzeniowski, O. M., Kaye, D., Infective endo-
carditis, in Heart Disease. A Textbook of Cardio-
vascular Medicine, Braunwald, E. (ed.). Saun-
ders, Philadelphia, 1992, pp. 10781105.
71 Durack, D. T., Infective and noninfective endo-
carditis, in The Heart: Arteries and Veins,
Hurst, J. W., C. E. Rackley, E. H. Sonnenblick,
N. K. Wenger (eds.). McGraw-Hill, New York,
1990, pp. 12301255.
72 Sawai, T., et al., Role of coagulase in a murine
model of hematogenous pulmonary infection
induced by intravenous injection of Staphylo-
coccus aureus enmeshed in agar beads. Infect
Immun 1997, 65, 466471.
73 Mylonakis, E., S. B. Calderwood, Infective en-
docarditis in adults. N Engl J Med 2001, 345,
13181330.
74 Hijikata-Okunomiya, A., N. Kataoka, Argatro-
ban inhibits staphylothrombin. J Thromb Hae-
most 2003, 1(9), 20602061.
75 Kanemitsu, K., et al., Relatedness between the
coagulase gene 3'-end region and coagulase
serotypes among Staphylococcus aureus strains.
Microbiol Immunol 2001, 45, 2327.
76 Kawabata, S., et al., Isolation and characteriza-
tion of staphylocoagulase chymotryptic frag-
ment. Localization of the procoagulant- and
prothrombin-binding domain of this protein.
J Biol Chem 1986, 261, 14271433.
77 Friedrich, R., et al., Staphylocoagulase is a
prototype for the mechanism of cofactor-in-
duced zymogen activation. Nature 2003, 425,
535539.
78 Bjerketorp, J., K. Jacobsson, L. Frykberg, The
von Willebrand factor-binding protein (vWbp)
of Staphylococcus aureus is a coagulase. FEMS
Microbiol Lett 2004, 234, 309314.
79 Rosey, E. L., et al., PauA: a novel plasminogen
activator from Streptococcus uberis. FEMS Mi-
crobiol Lett 1999, 178, 2733.
80 Kitt, A. J., J. A. Leigh, The auxotrophic nature
of Streptococcus uberis. The acquisition of es-
sential acids from plasmin derived casein pep-
tides. Adv Exp Med Biol 1997, 418, 647650.
81 Kawabata, S., et al., Difference in enzymatic
properties between alpha-thrombin-staphylo-
coagulase complex and free alpha-thrombin.
J Biochem 1985, 97, 10731078.
82 Sazonova, I. Y., et al., alpha Domain deletion
converts streptokinase into a fibrin-dependent
plasminogen activator through mechanisms
akin to staphylokinase and tissue plasminogen
activator. J Biol Chem 2004, 279, 2499425001.
References 393
Abstract
Molecular recognition systems that rely on
polyvalent interactions between the com-
ponents usually display considerably high-
er binding affinities and selectivities than
expected from the simple sum of associa-
tion energies of all the respective parts.
These favorable effects result thermodyna-
mically from the additivity of enthalpic
binding terms with concomitantly lower
entropic penalty as a result of the loss of
translational and rotational degrees of free-
dom associated with the binding of multi-
ple individual molecules. Nature uses the
principle of polyvalency ubiquitously to
modulate biochemical pathways, but only
recently has this principle found increas-
ing application for the control of carbohy-
drateprotein and proteinprotein interac-
tions in adhesion processes, cellular signal
transduction and enzyme inhibition by
synthetic small molecules. Two main
approaches were applied for inhibition of
proteases with the use of: 1) symmetrical
ligands that bind to symmetrical, multi-
subunit proteins in which each binding
head occupies an equivalent binding site;
and 2) asymmetrical ligands designed to
bind at the proteolytically active site and at
defined exo-sites. These two strategies are
exemplarily discussed with 20S protea-
some, b-tryptase and thrombin. The
knowledge derived from these studies may
well be used for the rational design of ar-
rays that mimic the natural multivalent
displays of effectors and inhibitors, and
thus for progresses in the development of
new generations of biopharmaceuticals
capable of interfering with a wide range of
(patho)physiological events.
Abbreviations
ATP adenosine triphosphate
FRE fibrinogen recognition site
NAPAP N-(2-naphtalenesulfonylglycyl)-
4-amidino-D,L-phenylalanine
piperidide
NEM N-ethylmaleinimide
PEG polyoxyethylene
PGPH post-glutamyl peptide hydrolysis
PPACK D-phenylalanyl-prolyl-arginine-
chloromethylketone
2.1
Introduction
In biology, both the specificity and the effi-
ciency of chemical reactions crucially de-
pend upon the selective recognition of
cross-talking molecules. This generally re-
395
2
Application of the Principle of Polyvalency to Protease Inhibition
Luis Moroder
ISBN: 3-527-31184-X
sults from optimal complementarity of
molecular surfaces where hydrophobic,
electrostatic, hydrogen bond and van der
Waals interactions constitute the main
forces responsible for the binding affinities
between interacting molecules. In addi-
tion, nature frequently uses multivalency
to achieve tight binding in cases where
univalent proteinligand binding is weak
[14]. In fact, molecular recognition sys-
tems that are based on multipoint interac-
tions between the components, display
considerably higher binding affinities.
These result mainly from the entropic
benefit, since the penalty for the loss of
overall rotational and translational entropy
is paid only once in the case of a multiva-
lent ligand rather then paying this penalty
in each of the monovalent binding events.
Such multivalent binding is fundamen-
tal to the regulation of many critical bio-
logical systems, and involves proteinpro-
tein and proteincarbohydrate interactions
in cellular adhesion processes as well as in
cellvirus and celltoxin binding, but also
in signal transduction pathways when ini-
tiated by multiple receptorligand contacts
at the cell surface. By mimicking nature,
this principle of multivalent binding has
become an emerging theme in drug de-
sign as it allows for significant increase of
affinity, but particularly of selectivity of the
receptorligand interactions [1, 5]. It has
found increasing application not only in
the design of molecules that modulate car-
bohydrateprotein interactions [614] and
which control cellular signal transduction
[1520], but also in the development of en-
zyme inhibitors [2131]. Among these en-
zymes, only a few proteases are suitable
for a multivalent inhibition. Indeed, until
recently only the X-ray structure of the
thrombinhirudin complex clearly revealed
two distinct domain interactions at the ac-
tive site and the fibrinogen binding site,
respectively [32]. By exploiting this infor-
mation, highly potent hirudin-mimicking
inhibitors were derived which simulta-
neously address the two different binding
sites, thus leading to a strong potentiation
of affinity and selectivity [33]. With the dis-
covery of the multicatalytic protease com-
plexes of the 20S proteasome and b-tryp-
tase, followed by their crystallographic
structure analyses, interesting new targets
became available with multiple identical or
different active sites displayed spatially in
geometrical order [3436]. This type of
architecture was compelling to attempt
the thermodynamically most attractive
approach with the design of homo- and
heterobivalent inhibitors [37, 38].
2.2
Thermodynamic Model of Bivalent Ligand
Binding
The physical mechanisms that govern
polyvalent binding of ligands to receptor
molecules have been analyzed extensively
[1, 3942]. Although quantifying the ther-
modynamic basis for increased affinity of
multivalent ligands is difficult, it is gener-
ally assumed that the contribution of the
single binding subsites to the overall free
energy of binding consists of their intrin-
sic binding energies and of a connection
Gibbs energy that represents the change
in the probability of binding that results
from the assembly of the binding subsites
into one molecule [41]. With homo-polyva-
lent ligands which interact with identical
binding subsites of a receptor molecule,
such as the homotrimeric vancomycin con-
struct that binds to a homotrimeric ligand,
the approximate additivity of the free ener-
gies of binding was confirmed [43, 44].
Using a simple bivalent system as mod-
el, the role of enthalpy and entropy in
polyvalent interactions can be rationalized
according to Whitesides and coworkers [1],
as shown in Fig. 2.1. In the case where the
two receptor binding sites are independent
and non-interfering, the enthalpy of bind-
ing is additive. However, there are cases of
synthetic dimeric ligands where the spacer
does not allow for optimal fitting of the
two binding heads to the binding subsites,
thus leading to distortions; the resulting
enthalpic strains are difficult to be quanti-
tatively evaluated.
In addition to the enthalpic gain, the
free energy of bivalent binding is strongly
affected by the entropic term. The total en-
tropic cost for complexation of two mono-
valent ligands with two subsites of the
receptor molecule is 2DS
trans
+2DS
rot
(Fig. 2.1). By connecting the two ligands
with a rigid linking group that allows for
optimal matching of ligands and binding
subsites, the entropic penalty for assem-
bling the two bivalent species is approxi-
mately half of that of two monovalent in-
teractions, since the second binding occurs
without additional cost of translational and
rotational entropy. However, such a scenar-
io is rather unrealistic, since all linking
groups are somewhat flexible and there-
fore the number of conformations as-
Fig. 2.1 Thermodynamics of binding of bivalent
ligands to two identical subsites of the receptor
molecule. DG8
m
, DH8
m
and TDS8
m
are the free
energy, enthalpy, and entropy of monovalent ligand
binding; DS8
trans
and DS8
rot, m
are the translational
and rotational entropies of the monovalent ligand
and DS8
H
2
O,m
the solvation entropy. DG8
bi
is the
free energy of bivalent binding, and DS8
conf
is the
loss in conformational entropy for the linker in
the intramolecular binding events.
sumed by the bivalent ligand before com-
plexation is greater than after complexa-
tion. Depending upon whether the re-
sulting conformational cost does not com-
pensate for the gains in the translational
and rotational entropy terms, or equals or
even exceeds these, the bivalent binding is
entropically enhanced, neutral or even di-
minished. This is reflected by the ratio of
the free energies of interaction of polyva-
lent-monovalent binding (b=K
poly
/K
mono
)
[1]. In the case of bivalent binding, a b>2
would indicate cooperative binding, b=2
entropically neutral, and b>2 entropically
disfavored binding.
2.3
Homo- and Heterobivalent Inhibitors of the
Yeast 20S Proteasome
In eukaryotes, the ATP-dependent ubiqui-
tin-proteasome pathway is the major intra-
cellular proteolytic machinery, and is
therefore key to activation or repression of
many cellular processes such as cell-cycle
progression and apoptosis. It is responsi-
ble for degradation of misfolded, damaged
and aged proteins, for the elimination of
regulatory proteins, and it also plays a cen-
tral role in the cellular immune response
by antigenic peptide processing [4547].
Because of this crucial function in vital
processes, a selective inhibition of the pro-
teasome is of promising therapeutic poten-
tial for treatment of cancer, inflammatory
disorders, and immune diseases [45, 46,
48, 49].
The multicatalytic protease complex con-
sists of a proteolytically active central core
(core particle, 20S proteasome) and of a
19S regulatory complex which is attached
to the core particle and is responsible for
substrate recognition and unfolding. The
spatial structure and the enzymatic mecha-
nisms have been elucidated only for the
core particle, whereas the organization of
the regulatory particle is still less under-
stood [34, 50]. The core particle or 20S pro-
teasome is a large, barrel-shaped protein
complex consisting of four rings, each
composed of seven subunits which are
tightly packed in an a
17
b
17
b
17
a
17
ar-
rangement. While the a subunits are re-
sponsible for substrate gating [51], the b
subunits act as the proteolytic centers, and
only the fully assembled 20S proteasome
is able to degrade, in a processive manner,
the unfolded proteins into small-sized pep-
tides [52, 53]. In eukaryotic proteasomes,
three b-type subunits (i.e. b1, b2 and b5)
are autolytically processed to generate the
protease active sites with the N-terminal
nucleophile; that is, the Thr1 residue
which is essential for activity. The other
four b subunits remain inactive. Within
the core particle, each pair of proteolyti-
cally active subunits shows a certain de-
gree of substrate specificity, where the b1
subunits are particularly responsible for
post-glutamyl peptide hydrolysis (PGPH)
that is a caspase-like activity, the b2 sub-
units exhibit trypsin-like and the b5 sub-
units chymotrypsin-like activities [5457].
The S1 pockets of these subunits are the
major specificity determinants and are ap-
propriately polar and sized to accommo-
date acidic, basic and apolar P1 side
chains, respectively, but also bind non-
complementary residues in a manner con-
sistent with the low specificity of the pro-
teasome [34, 55]. The latter property raises
the main difficulties in the design of
highly selective inhibitors, although the
surface characteristics of the substrate
binding subsites as well as the binding
modes of various synthetic and small-sized
natural inhibitors were characterized in de-
tails by X-ray structural analysis [34, 50,
5861].
Since peptide aldehydes were recognized
very early as efficient inhibitors of the pro-
teasome [62], the X-ray structure of the
20S proteasome from Thermoplasma acido-
philum [50] and subsequently from Sac-
charomyces cerevisiae [34] were resolved
with the proteases complexed with the cal-
pain I inhibitor that is, the tripeptide al-
dehyde Ac-Leu-Leu-Nle-H (1).
From crystallographic analysis of the
yeast 20S proteasome, a clear picture was
obtained of the spatial display of the active
sites on the two b rings, with the tripep-
tide aldehyde covalently linked via hemi-
acetal bonds to the Thr1 hydroxyl groups
of all six active sites (Fig. 2.2). This well-
defined geometry of the active sites was
compelling for attempts to bypass the
problem of selectivity and binding affinity
of proteasome inhibitors by exploiting the
principle of multivalent ligands. In fact,
the X-ray data allow extraction of the dis-
tances between the N-terminal Thr1 resi-
dues of the various active sites located on
one ring or on the two staggered b rings
(Fig 2.3) for the design of potential biva-
lent inhibitors that address adjacent active
sites on a b ring (transannular) or on the
two associated b rings (interannular).
2.3.1
Transannular Heterobivalent Inhibitors
The distance of 28 between the b1b2
and b1'b2' active sites corresponds almost
exactly to a nonapeptide in extended con-
formation. The X-ray structure of 20S pro-
teasome containing the Thr1Ala mutant of
the b1 subunit revealed the binding mode
of the non-processed propeptide to the
substrate binding cleft up to the adjacent
b2 active site (Fig. 2.4) [63]. Correspond-
ingly, this propeptide sequence was used
in a first approach to design bivalent inhi-
bitors bearing a glutamic acid aldehyde on
the C-terminal position to address in a
more selective manner the b1 active site
Fig. 2.2 Surface representation of half of the inner
proteolytic chamber of yeast 20S proteasome with
Ac-Leu-Leu-Nle-H bound to the active sites of
three b subunits.
Fig. 2.3 Schematic representation of the two cen-
tral b rings of yeast 20S proteasome with the
trans- and interannular distances between the
Thr1 residues of the active sites.
with its caspase-like specificity. As a poten-
tial anchor for the adjacent b2 active site
an N-terminal levulinic or 4-oxo-butyric
acid residue was selected (compounds 2
and 3 of Table 2.1), with the assumption
that access to the b2 Thr1 residue can oc-
cur even via the S' subsites as predicted by
modeling experiments and by the observed
scission of the propeptide at the Arg
10
-
Leu
9
peptide bond [37, 63]. The double-
headed inhibitors, however, were found to
inhibit only the PGPH (b1 or b1'), but not
the trypsin-like (b2 or b2') activity (Table
2.1). These results exclude a heterobivalent
binding, and clearly confirmed the diffi-
culty in concomitantly addressing two ad-
jacent active sites with a double-headed
peptide inhibitor that presents the anchor
group for the active site nucleophiles from
the primed as well as the non-primed sub-
strate binding cleft. In fact, X-ray analysis
of the 20S proteasomeinhibitor com-
plexes clearly revealed binding of the C-ter-
minal aldehyde to all six active sites in a
manner similar to Ac-Leu-Leu-Nle-H [37].
2.3.2
Interannular Homobivalent Inhibitors
To allow for an access of two anchor
groups to two identical or different active
sites from the non-primed S subsites, the
crystal structure of Ac-Leu-Leu-Nle-H
bound to b5 and b5' of the yeast 20S pro-
teasome was used as a template [34]. The
entry of substrates into the proteolytic
chamber is restricted by the bottle-neck of
the a ring, which recruits from outside
only fully unfolded linear polypeptides for
digestion. This fact significantly restricts
the choice of spacers for bivalent inhibitor
constructs. Such a spacer should mimic as
much as possible the unstructured poly-
peptide chain of an unfolded protein, and
reach a length of about 50 . Peptides of
appropriate size are known to be rapidly
degraded by the yeast proteasome, and
thus linear polyoxyethylene (PEG) chains
were selected as mimic of random-coiled
polypeptide chains [37, 64], since this poly-
mer is known to be highly solvated and
Fig. 2.4 In the b1 Thr1Ala mutant
of yeast 20S proteasome, the pro-
peptide is cleaved only at the Arg
10
-
Leu
9
peptide bond. This allowed
(using X-ray crystallography) the
binding mode of the nonapeptide
between the adjacent b1 and b2 sub-
units to be determined [63].
Table 2.1 Inhibition of 20S proteasome by transannular heterobivalent inhibitors (IC
50
, lM)
Inhibitor PGPH Trypsin-like Chymotrypsin-like
Ac-Leu-Leu-Nle-H (1) >100 >100 2.1
Lev-Lys-Lys-Gly-Glu-Val-Ser-Leu-Glu-H (2)
a)
102 >100 >100
Saa-Lys-Lys-Gly-Glu-Val-Ser-Leu-Glu-H (3)
a)
82 >100 >100
a) Lev=levulinic acid; Saa=4-oxobutyric acid residue.
unstructured. In fact, the pegylated tripep-
tide aldehydes (PEG)
1925
-Leu-Leu-Nle-H
(4) and (PEG)
1925
-Arg-Val-Arg-H (7) were
found to inhibit the proteasome with al-
most identical potency as the acetylated
tripeptide aldehydes 1 and 6 (Table 2.2).
Based on this observation, two interannu-
lar homobivalent inhibitors containing the
tripeptide aldehydes -Leu-Leu-Nle-H and
-Arg-Val-Arg-H as head groups for the b5/
b5' and b2/b2' active-site pairs, respective-
ly, were synthesized using a PEG spacer
with a statistical distribution of 1925
monomers and thus averaging the length
of about 50 . As shown in Table 2.2, with
the homobivalent inhibitors 5 and 8,
highly selective inhibition of the chymo-
trypsin- (b5) and trypsin-like (b2) activities
was achieved. The increase in potency by
two orders of magnitude when compared
to the monovalent inhibitors 4 and 7
clearly confirmed that the conformational
entropy costs derived from the flexible lin-
ker compensate the gains in translational
and rotational entropy of bivalent binding
extensively. While for a maximal entropi-
cally enhanced bivalent binding K
i, bi
val-
ues of approximately (K
i, mono
)
2
are ex-
pected, in the present case b values of 1.35
for inhibitor 5 and 1.34 for 8 were deter-
mined. The relatively small gain in free
energy of binding by the bivalent inhibi-
tors must be attributed to the high degree
of flexibility of the spacer and thus to the
loss of conformational entropy associated
with the bidentated interaction [65, 66].
Well-defined electron density maps were
obtained for the tripeptide -Leu-Leu-Nle-H
head groups by X-ray analysis of the 20S
proteasome complexed with the bivalent
inhibitor 5, whereas the PEG spacer could
not be identified, thus confirming degrees
of (Fig. 2.5). This flexibility allows the head
groups to reach the Thr1 residues from
the S subsites and thus concomitant for-
mation of the hemiacetal bonds at two ac-
tive sites is achieved.
Table 2.2 Inhibition of 20S proteasome by mono- and bivalent inhibitors (IC
50
, lM)
Ac-Leu-Leu-Nle-H (1) >100 >100 2.1
CO-Leu-Leu-Nle-H
|
PEG-COOH
(4) >100 >100 1.8
CO-Leu-Leu-Nle-H
|
PEG-CO-Leu-Leu-Nle-H
(5) >100 >100 0.017
Ac-Arg-Val-Arg-H (6) >100 6.4 >100
CO-Arg-Val-Arg-H
|
PEG-COOH
(7) >100 8.2 >100
CO-Arg-Val-Arg-H
|
PEG-CO-Arg-Val-Arg-H
(8) >100 0.071 >100
CO-Leu-Leu-Nle-H
|
(9) >100 0.097 0.031
Fig. 2.5 Upper panel: Stereoview of a section of the X-ray
structure of the yeast 20S proteasome/compound (5) adduct.
Lower panel: Schematic representation of the inhibitor linked
via hemiacetal bond to two active-site Thr1 residues.
Since the tripeptide moieties of the in-
hibitor 5 were identified in all six active
sites, as in the case of the acetylated tri-
peptide aldehyde, in the absence of sub-
strate and at the high concentration of in-
hibitor used for the soaking experiments,
the b1, b2 and b5 active sites are indeed
insufficiently selective to discriminate the
C-terminal norleucinal as the P1 residue.
Conversely, the bivalent inhibitor 8 con-
taining the tripeptide aldehyde -Arg-Val-
Arg-H was detected only in the two tryp-
sin-like b2 and b2' active sites, despite the
high concentration used. This observation
confirms a significant degree of selectivity
of this bivalent ligand for the trypsin-like
active sites.
2.3.3
Interannular Heterobivalent Inhibitors
With the construction of heterobivalent in-
hibitors that address simultaneously two
different b subunits, one molecule was ex-
pected to neutralize only one active site of
the existing pair. Correspondingly, two
molecules are required for complete inhi-
bition of one type of proteolytic activity,
although with the advantage of inhibiting
two activities concomitantly. To examine
this working hypothesis, a heterobivalent
inhibitor was synthesized containing the
tripeptide aldehydes -Leu-Leu-Nle-H and
-Arg-Val-Arg-H (9) as head groups [64]. As
expected, both the trypsin- and chymotryp-
sin-like activities were inhibited with very
similar potencies as those of the homobi-
valent inhibitors if the stoichiometry of
this type of inhibitor is taken into account
(Table 2.2).
2.3.4
Heterobifunctional Inhibitors
It has been known that the treatment of
mammalian [67, 68] or yeast proteasome
[69] with larger excesses of the thiol-rea-
gent N-ethylmaleinimide (NEM) leads to
selective inhibition of the trypsin-like activ-
ity. In the crystal structure of the yeast 20S
proteasome the conserved Cys118 residue
of the b3 subunit protrudes into the S3
subsite of the b2 active site [34], a fact that
could explain the inactivation of the tryp-
sin-like activity of proteasomes by its
chemical modification with NEM. The par-
Fig. 2.6 Schematic representation of the S subsites of the b2 active site of the yeast
20S proteasome as structural model for the design of maleoyl-b-alanyl-dipeptide al-
dehydes as a new type selective heterobifunctional inhibitors.
ticular position of the thiol group was
exploited for the design of inhibitors that
selectively address the trypsin-like activity
of the proteasome [70]. For this purpose
(as shown in Fig. 2.6), Cys118 was used to
anchor inhibitors of the peptide aldehyde
type via a thiol-reactive handle in close
proximity to the Thr1 residue of the b2 ac-
tive site for inactivation of the N-terminal
nucleophile by a hemiacetal bond. Using
the binding mode of Ac-Leu-Leu-Nle-H (1)
to the active-site of the b2 subunit [34],
modeling experiments were performed by
deleting the Ac-Leu moiety of the bound
inhibitor and replacing it by the maleini-
mide group as the thiol-reactive handle.
This group was positioned as P3 residue
into the S3 subsite in interacting distance
to the Cys118 thiol function. From a ki-
netic point of view, upon recognition and
binding of the P1P2 moiety by the active
site, the reaction of the maleinimide group
with the Cys118 thiol was expected to oc-
cur immediately, if a spacer of the correct
size and properties is applied. For this pur-
pose, the ethylene moiety was selected,
since this spacer restricts the flexibility of
the maleinimide group via its relatively
small size, although allowing for rotational
motion as required for its optimal interac-
tion with the reactive thiol group.
Based on this working assumption, the
bifunctional compounds 1012 were syn-
thesized and analyzed for their inhibitory
potencies (Table 2.3) [70]. Deletion of the
Leu residue in the calpain inhibitor I was
found significantly to decrease inhibition
of the chymotrypsin-like activity, although
inhibition of the trypsin-like activity was
retained. This could be attributed solely to
reaction of the maleinimide group with
Cys118. Consequently, an improvement of
the complementary properties of the P1 re-
sidue for the S1 subsite of the b2 subunit
with the basic residues Lys and Arg was
expected to exert a notable impact on the
inhibitory potencies. Indeed, with the bi-
functional inhibitor 12 containing the argi-
nal residue as P1, selective inhibition of
the trypsin-like activity with submicromo-
lar affinity was obtained (Table 2.3).
Since a 100-fold dilution of the inhibited
enzyme was not restoring trypsin-like ac-
tivity, the working assumption of a cova-
lent linkage of the inhibitor to Cys118 was
confirmed. Moreover, X-ray structural anal-
ysis of the yeast 20S proteasome/12 adduct
(Fig. 2.7) confirmed its exclusive binding
to the b2 active sites via hemiacetal forma-
tion with the Oc of Thr1 as well as the
deep insertion of the guanido group into
the S1 pocket and the covalent thiosuccini-
midyl linkage of the inhibitor to Cys118 of
the b3 subunit [70]. The thiol addition to
the maleinimide double bond occurs at
only one of the two possible carbon atoms
in a defined (R) configuration, and the re-
sulting thiosuccinimidyl ring is involved
Table 2.3 Inhibition of the PGPH, trypsin- and chymotrypsin-like
activities of yeast 20S proteasome by maleoyl-b-alanyl-dipeptide
aldehydes (IC
50
, lM)
Ac-Leu-Leu-Nle-H (1) >100 >100 2.1
Mal >bAla-Leu-Nle-H (10) >100 13 >100
Mal >bAla-Val-Lys-H (11) >100 3.4 >100
Mal >bAla-Val-Arg-H (12) >100 0.5 >100
in an additional hydrogen bonding net-
work which restricts the conformational
space of the ethylene spacer.
With this type of inhibitor, the basic
principle of multivalency was applied in a
new version where specific recognition of
peptide aldehydes led to a covalent graft-
ing near the active site and thus to an in-
crease of their in-loco concentration to val-
ues that make the inhibition practically ir-
reversible. However, such bifunctional in-
hibitors are of limited application in cell
biology because of the high intracellular
glutathione concentration which would im-
mediately neutralize the thiol-reactive mal-
einimide group.
2.4
Bivalent Inhibition of Mast Cell b-Tryptase
Human b-tryptase is a serine protease
which is stored in large amounts in mast
cell secretory granules, and represents the
major protein component released upon
degranulation [71]. Mast cells play a key
Fig. 2.7 Upper panel: Stereoview of a section of the X-ray
structure of the yeast 20S proteasome/Mal>bAla-Val-Arg-H
(12) adduct. Lower panel: The bound inhibitor is linked via
thiosuccinimidyl to Cys118 of the b3 subunit and as hemiace-
tal to Thr1 of the b2 subunit.
role in inflammatory responses, and suffi-
cient evidence has been accumulated that
b-tryptase is the mediator of many allergic
and inflammatory diseases [7276]. This
fact has fostered major efforts to develop
potent and selective inhibitors of this pro-
tease [38, 7782].
The crystal structure of human b-tryp-
tase confirmed its tetrameric assembly
from four quasi-equivalent monomers,
and disclosed their arrangement in a
square flat ring with the four active sites
pointing towards an oval central pore [35,
36]. This array of active sites, which is out-
lined schematically in Fig. 2.8, restricts the
access of macromolecular substrates to the
digestion chamber and prevents inhibition
by all known endogenous proteinase inhi-
bitors. The nature of this unique tetra-
meric architecture with four identical ac-
tive sites was soon recognized as being
ideally suited for a structure-based design
of homobivalent inhibitors to improve po-
tency and selectivity.
Unlike the proteasome, the b-tryptase
contains four identical active sites. Their
trypsin-like specificity results from the
Asp169 residue positioned at the bottom
of the S1 pocket, which accommodates
and binds lysine/arginine residues or re-
lated structural mimetica, as confirmed by
the X-ray structure of the human b-tryp-
tase/4-amidinophenyl pyruvate complex
[35, 36]. Correspondingly, efficient inhibi-
tion is obtained with peptidyl-arginals as
confirmed by the K
i
values of compounds
6 and 7 (Table 2.4). For compound 7,
which carries the large PEG tail, an in-
crease of the K
i
value by a factor of 2 was
observed. However, the homobivalent pro-
teasome inhibitor 8, where the PEG spacer
matches the intersubunit distances of 45
of the b-tryptase quite well, shows only the
relatively low increase in potency by a fac-
tor of 10 (Table 2.4). This may well be at-
tributed to the oversized PEG spacer, but
most reasonably to its high flexibility. In
this case the unfavorable conformational
entropy cost due to complexing the second
ligand of the bivalent inhibitor has to
largely exceed the gain in translational and
rotational entropy, thus leading to a mini-
mal effect on the free energy of binding.
With the earlier discovery of the tetra-
meric composition of human b-tryptase
[83], various classes of bibasic inhibitors
have first been synthesized rather in a
trial- and error-manner. Later, a structure-
based approach could be applied based on
the exact geometry of the tetrameric pro-
tease [74, 75]. Among the various genera-
Fig. 2.8 Schematic representation of the geometri-
cal array of the S1 subsites of human b-tryptase
with Asp169 at the bottom of each S1 binding
pocket (indicated by a star).
Table 2.4 Inhibition of human b-tryptase by
PEG-linked mono- and bivalent inhibitors
Inhibitor K
i
[nM]
Ac-Arg-Val-Arg-H 6 15
CO-Arg-Val-Arg-H
|
PEG-COOH
7 36
CO-Arg-Val-Arg-H
|
8 1.6
tions of symmetrical bibasic inhibitors re-
ported largely in the patent literature [75],
a detailed analysis of the mode of binding
has not been reported except for com-
pound 13 (CRA-205) [84]. This symmetri-
cal bibasic compound contains a (p-guani-
dino)phenyl group at either end of the
molecule (Fig. 2.9) [80]. It spans a length
of 33 in its extended conformation, and
thus very efficiently matches the shortest
distance between two vicinal S1 pockets of
the b-tryptase, as derived later by the X-ray
structure (see Fig. 2.8). Compared to phe-
nylguanidine (K
i
=63 lM), the symmetrical
compound 13 exhibits a K
i
of 620 pM,
which strongly supports a bivalent bind-
ing. Although the X-ray structure of b-tryp-
tase complexed with inhibitor 13 as ulti-
mate proof of the bivalent binding was not
resolved, the ratio of the free energies of
binding of the bivalent and the monova-
lent inhibitor (b=2.2) would suggest in
this case even a cooperative binding mode,
a fact which is rarely observed [1]. It would
indicate that for this molecule unfavorable
enthalpy and conformational entropy
changes resulting from the linker are rela-
tively small, and that the inhibitor behaves
like a rigid molecule capable of positioning
the two head groups without strains for
optimal interaction with the S1 pockets.
Per se, the molecule appears flexible, but
it exhibits low-energy conformers which
may induce a conformational preorganiza-
tion for optimal binding of both head
groups [80].
As an alternative to the more or less sol-
vated spacers generally used for the bibasic
compounds, carbohydrate templates were
examined in a structure-based design [38].
From modeling experiments, b-cyclodex-
trin appeared to be the most ideally sized
linker molecule. The distance between the
primary hydroxy groups of the sugar units
A and D is 13 (Fig. 2.10). Correspond-
ingly, this rigid carbohydrate template was
expected to greatly reduce the conforma-
tional entropy penalty, if decorated with
proper binding head groups for the S1
pockets of the protease subunits.
Since modeling experiments had sug-
gested a preferred binding of m- over p-
substituted benzene groups, and an opti-
mal display of the binding heads when (3-
aminomethyl)benzenesulfonyl-glycine (14)
is grafted to the distal positions of b-cyclo-
dextrin, the 6A,6D-dideoxy-6A,6D-diamino-
b-cyclodextrin was used to attach one or
two binding heads via amide bonds for
production of the monovalent cyclodextrin
conjugate 15 and the bivalent construct 16
(Fig. 2.10).
Upon linking one (3-aminomethyl)ben-
zenesulfonylglycine moiety to b-cyclodex-
Fig. 2.9 Chemical structure of the bivalent human b-tryptase inhibitor CRA-2059 (13) [80].
trin (15), binding of the monobasic head
group to the S1 pockets of the b-tryptase
and trypsin was practically not affected.
This suggested the absence of steric hin-
drance, while the carbohydrate core was
found to sensibly decrease the affinity for
thrombin (Table 2.5). Similar minor
changes in inhibitory potency were ob-
served for the bivalent construct 16 with
the monomeric enzyme species trypsin
and thrombin. However, in the case of the
tetrameric b-tryptase the bivalent ligand
proved to bridge very efficiently the space
between the enzyme subunits A/D and B/
C, leading to strong binding affinities with
a b factor of 1.9 (Table 2.5) [38]. This po-
tentiation of inhibitory activity clearly sup-
ports a bivalent binding of the cyclodextrin
construct 16, which was further supported
by titration of b-tryptase with this inhibitor.
For full inhibition, a stoichiometry of
1.930.1 was extracted, and this was in
full agreement with the theoretically ex-
pected value of 2. However, the decisive
proof for bivalent binding was derived
from X-ray analysis which confirmed the
presence of two cyclodextrin constructs lo-
cated between the A/D and B/C active
sites, respectively, in the correct position
Fig. 2.10 Structure of the binding head (3-aminomethyl)ben-
zenesulfonyl glycine as methyl ester (14) and of the related
b-cyclodextrin-based mono- (15) and bivalent (16) inhibitors.
Table 2.5 Inhibition of human b-tryptase with
b-cyclodextrin-based monovalent and bivalent
inhibitors
Inhibitor b-Tryptase
K
i
[lM]
Trypsin Thrombin
14 17 43 >300
15 41 27 32
16 0.0006 4.8 >160
for optimal display of the binding heads
and their insertion into the S1 pockets
without enthalpic strains (Fig. 2.11).
The bivalent inhibitor 16 contains the
freely rotating sulfonamide bonds as well
as those of the glycine spacer. These
groups confer degrees of torsional freedom
to the unbound inhibitor which are lost in
the bound state, thus causing conforma-
tional entropic loss. However, this tor-
sional freedom is important for optimal
occupancy of the S1 subsites. In fact,
merely replacing the sulfonamide bonds
in compound 16 with the planar carboxa-
mide group provokes a drastic decrease of
binding affinity, most probably as a result
of enthalpic strains [38].
Most of the optimized bibasic inhibitors
reported for b-tryptase show high affinities
that suggest homobivalent binding to the
two S1 subsites, as demonstrated by X-ray
analysis for the cyclodextrin construct 16.
That this may not always be the case, is
well evidenced by comparing the binding
mode of the bibasic compounds 17 and 18
(Fig. 2.12) as derived from X-ray analysis
[85]. Both compounds are tight-binding in-
hibitors with subnanomolar affinities and
b factors of 1.9 for 17 and 1.6 for 18. As
expected, in the X-ray structure of the b-
tryptase/17 complex, the inhibitor bridges
the protease subunits A/D and as well as
B/C, whereby the molecule assumes a sig-
moidal conformation that allows both head
groups to interact in identical mode with
the S1 pockets (Fig. 2.13). The (4-amino-
methyl)benzyl group inserts into the S1
pockets to a 2.6 distance from the
Asp189 carboxylate, and additional hydro-
gen bonds with residues of the protease
surface are coordinating the carboxamide
and ester groups, while the hydrophobic
spacer acts as an optimal tether (Figs. 2.13
and 2.14).
Very surprisingly, the X-ray structure of
the complex b-tryptase/compound 18 re-
vealed a completely different binding
mode, as shown in Fig. 2.15 [85].
Four inhibitors are bound to the four
protease subunits with the (4-amino-
methyl)benzyl group inserted into each S1
pocket, thereby forming salt bridges with
Asp189 at 2.5 distance. Again, the car-
boxamide group is involved in hydrogen
bonding interactions with the residues
Gln192, Ser214 and Ser195 of the protein
subunits. The rest of the molecule adapts
to the protein surface of the adjacent sub-
unit forming additional hydrogen bond in-
teractions, but without insertion of the sec-
ond binding head into the adjacent S1
subsite (Fig. 2.16). This exo-site binding
leads to a surprisingly strong potentiation
of the inhibitor affinity, thus simulating a
Fig. 2.11 X-ray structure of
the two subunits A (green)
and D (yellow) of the tetra-
meric b-tryptase complexed
with the bivalent inhibitor 16.
bivalent binding mode. It acts like the exo-
site binding of hirudin-like inhibitors in
the case of thrombin (see Section 2.5).
The strong effect of the mode of presen-
tation of the binding heads and the critical
length of the spacer for b-tryptase inhibi-
tors was well evidenced by a distance scan
of the A/D and B/C subunits of b-tryptase
Fig. 2.12 Chemical structure of bibasic inhibitors of human b-tryptase.
Fig. 2.13 X-ray structure of the human b-tryptase
complexed with two bibasic inhibitors 17.
Fig. 2.14 View of the binding mode of inhibitor 17
from one S1 pocket into the second S1 pocket.
BYK150640 (17)
(K
i
=0.00025 lM)
BYK76935 (18)
(K
i
=0.00076 lM)
using c[D-Asp-L-Asp] and c[D-Glu-L-Glu]
diketopiperazines as scaffolds and an in-
creasing number of CC bonds in the X
and Y spacing moieties, as shown in
Fig. 2.17. With a total number of bonds be-
tween the two basic amino groups of 31,
the maximum affinity was reached, which
decreases immediately by either decreasing
the number to 29 or increasing to 33 [82].
Such constructs of minimal changes in
length may well serve as efficient tools in
affinity chromatography of b-tryptase iso-
forms that are expected to vary only
slightly in their active-site geometries [86].
2.5
Heterobivalent Inhibition of Thrombin
Thrombin is a trypsin-like serine protease
which plays central functions in the
process of hemostasis and thrombosis.
Thrombin converts soluble circulating fi-
brinogen into clottable fibrin, and ampli-
fies its own generation through activation
of other coagulation enzymes such as
factors V and VIII [87, 88]. In addition,
thrombin activates factor XIII, which stabi-
lizes the clot by cross-linking fibrin, and
stimulates platelet secretion and aggrega-
tion. It also mediates a negative-feedback
regulation of the coagulation cascade by
activating protein C upon binding to the
endothelial cell surface protein thrombo-
modulin. Because of this key role in cata-
lyzing the procoagulant processes that lead
to clot formation, thrombin is implicated
in various diseases such as myocardial in-
farction, stroke, and pulmonary embolism
Fig. 2.15 X-ray structure of b-tryptase complexed
with four bibasic inhibitors 18.
Fig. 2.16 Left panel: View from the inside of the tetramer in
direction A/D interface. The S1 pockets are on the opposite
sides. Right panel: View of the inhibitor down to the S1 pocket.
[89]. Correspondingly, its inhibition by nat-
ural and synthetic inhibitors was recog-
nized as a primary target for the develop-
ment of successful anticoagulants [89, 90].
Tremendous efforts have been made dur-
ing the past few decades in the design and
synthesis of orally available, small-mole-
cule inhibitors for acute and chronic anti-
coagulation [8991]. However, at present
there is only limited clinical use of paren-
teral preparations [90]. An alternative anti-
coagulant for acute treatment is the 65-
amino acid residue recombinant desul-
fated hirudin, derived from the naturally
occurring thrombin inhibitor hirudin. This
was isolated from extracts of the leech Hir-
udo medicinalis [92].
X-ray analysis of thrombin complexed by
recombinant hirudin showed that this in-
hibitor binds to two distinct sites of the
protease that is, the amino-terminal tet-
rapeptide to the active site, and the C-ter-
minal tail (hirudin residues 5365) to the
fibrinogen recognition site (FRE) [32, 93].
This heterobivalent binding mode explains
the extremely high affinity and selectivity
of hirudin, with a K
i
of 21 fM [94]. De-
tailed analysis of the contributions of both
binding sites to the overall free energy of
binding clearly confirmed the additivity of
the free energies of binding of the hirudin
fragments 151 and 5265, although with-
out cooperative effects [33, 95, 96]. The
exo-site is rich in basic residues, and is
connected to the active site by a deep
groove. The cluster of positive charges
serves mainly for initial electrostatic recog-
nition of macromolecular substrates and
inhibitors rich in acidic residues [97]. It
contributes less to the binding affinity, as
well assessed by mutational studies involv-
ing the glutamic acid residues of hirudin,
although desulfation leads to a 10-fold de-
crease in inhibitory potency [94] (see Table
2.6). Conversely, in terms of binding en-
ergy, hydrophobic interactions between the
hirudin tail and the binding cleft seem to
play a dominant role [93, 98100].
The structural information derived from
these early studies was compelling for a
structure-based design of synthetic hetero-
bivalent inhibitors capable of simulta-
neously addressing both binding sites [33].
For this purpose, optimization of both
Fig. 2.17 Diketopiperazine-based bibasic inhibitors of b-tryp-
tase (19). With m=n=1 and X=Y=b-alanine (total number of
bonds: 31) obtained [82], the maximum inhibitory potency
(K
i
=10 nM) was obtained [82].
component parts of the bivalent inhibitors
was attempted. Starting with the simple
D-Phe-Pro-Arg motif of the irreversible
thrombin inhibitor D-Phe-Pro-Arg chloro-
methylketone (PPACK), which was used to
resolve the structure of thrombin by X-ray
analysis [101], the first bivalent inhibitor
was obtained by linking this tripeptide to
the hirudin tail 4865 (P53) [102] or via
the Pro-(Gly)
4
spacer to the hirudin frag-
ment 5364 (hirulog-1) [103]. Both inhibi-
tors were of high affinity and specificity
for thrombin (Table 2.6). These hirudin-
like inhibitors of the first generation that
contain the scissile Arg-Pro sequence in
the active site-binding domain as P1P1'
were then replaced by proteolytically more
resistant active-site binding motifs [98,
100, 104107]. Moreover, other active site-
binding domains were derived from well-
established thrombin inhibitors such as
Argatroban [108] or NAPAP [100]. By en-
hancing the affinity of the N-terminal do-
main and optimizing the spacer length
[109] and the FRE-binding motifs [110],
even inhibitors of pico- to femtomolar in-
hibition potencies were obtained (Table
2.6) [100, 111113], which strongly sup-
ported a heterobivalent hirudin-like bind-
ing mode. This was fully confirmed by X-
ray analysis of their complexes with
thrombin [100, 114118]. However, the
Table 2.6 Selected heterobivalent inhibitors of thrombin designed
in analogy to hirudin to address both the active site and the fibri-
nogen recognition exo-site (FRE). The N-terminal active site- and
C-terminal FRE-binding domains are in bold characters; the re-
maining part of the molecules serve as spacer.
Inhibitor K
i
Reference
VVYT-[5-52]-DGDFEEIPEEY(SO
3
H)LQ (native hirudin) 21 fM [94]
VVYT-[5-52]-DGDFEEIPEEYLQ (recombinant non-sulfated hirudin) 231 fM [94]
Ac-fPRP-QSHN-DGDFEEIPEEYLQ (P53)
a)
2.8 nM [102]
fPRP-GGGG-NGDFEEIPEEYL (hirulog-1) 2.3 nM [103]
(D)Cha-PRP-GGGG-NGDFEEIPEEYL (hirulog-B1) 77 pM [104]
fP-(h)Arg-Gly-GGGG-NGDFEEIPEEYL (hirulog-3) 7.4 nM [105]
dansyl-R-(D)Pip-NH-(CH
2
)
11
-CO-cAbu-DFEEIPEEYL (P535) 17 pM [108]
Bbs-R-(D)Pip-Amb-NH-(CH
2
)
6
-CO-GDYEPIEEA-Cha-e (P611) 0.23 pM [111]
Bbs-R-(D)Pip-Thi-NH-(CH
2
)
11
-CO-DYEEPIPEEA-Cha-e (P798) 17 fM [112]
(D)Cha-P-Apt-(Gly)
4
-DYEPIPEEA-Cha-e (P596) 46 fM [98]
Chg-R-2Nal-T-Asp-(D)Ala-Gly-bAla-PESHFGGDYEEIP-(Aib)
2
-Y-Cha-e 90 pM [113]
a) The canonical L-configured amino acid residues are presented
in upper-case letters, and the D-configured in lower-case one-
letter code. For non-canonical or synthetic amino acids, the
following abbreviations were used: Cha=(3-cyclohexyl)propa-
noic acid (cyclohexylalanine); (h)Arg=(3-amino-5-guanido)hex-
anoic acid (homo-arginine); Pip=pipecolic acid, cAbu=(4-ami-
no)butyric acid; Amb=(4-aminomethyl)benzoic acid; Thi =
(3-thienyl)propanoic acid (thienylalanine); Apt =Argw[COCH
2
]-
pyridylacetic acid; Aib=(2-amino)isobutyric acid; 2Nal =3-(2-
naphthyl)propanoic acid (2-naphthylalanine); Chg=2-amino-2'-
cyclohexylacetic acid (cyclohexylglycine); Bbs=(4-tert-butyl)ben-
zenesulfonyl.
contribution of the two binding sites to
the overall free energy of binding was not
evaluated quantitatively. Although crystal-
lographic analysis of the thrombininhibi-
tor complexes clearly revealed significant
flexibility of the linker portion, the confor-
mational entropic penalty has to account
for the generally observed lack of additivity
of the free energies of binding, as esti-
mated from the average micromolar affi-
nities of the FRE-binding domains and the
nanomolar affinities of the active site-bind-
ing domains used in these bivalent con-
structs. Nevertheless, femtomolar K
i
values
were determined for a few selected biva-
lent inhibitors, as exhibited by the natural
and recombinant hirudin (Table 2.6).
2.6
Perspectives
In the design of bivalent ligands the spacer
represents the main limitation, as is well
evidenced by the homo- and heterobivalent
inhibitors of the 20S proteasome, b-tryptase
and thrombin. It is, however, required to
bridge the binding subsites, and a certain
degree of flexibility is generally indispens-
able for a display of the binding domains
to optimal recognition and complexation
by the binding subsites of the receptor pro-
teins. The examples discussed in this chap-
ter confirm that, despite this severe handi-
cap, homo- and heterobivalent inhibitors
usually excel in their binding affinities and
foremost in the selectivity which can be
achieved, particularly when spatial struc-
tures are available for a rational design of li-
gands. An additional severe drawback, how-
ever, consists of the relatively large sizes of
such bivalent constructs which generally
do not satisfy the Pfizers rule of five [119]
or Vebers rules [120] for direct conversion
into bioavailable drugs. However, with bet-
ter insights into multivalent interactions
as a way of modulating selectively biological
effects, the rational design of arrays that
mimic the natural multivalent displays of
effectors and inhibitors may well advance
toward new generations of biopharmaceuti-
cals that are capable of interfering with a
wide range of interactions, including cell
cell, cellextracellular matrix, cellvirus,
and cell toxin binding, as well as with signal
transduction pathways.
Acknowledgments
The author gratefully acknowledges
Drs. M. Groll, U. Marquaerdt, and W. Bode
of the Max-Planck-Institute of Biochemis-
try, Martinsried, for the unpublished fig-
ures of X-ray structures.
References
1 M. Mammen, S.-K. Choi, G. M. Whitesides,
Angew. Chem. Int. Ed. Engl. 1998, 37, 27552794.
2 L. L. Kiessling, N. L. Pohl, Chem. Biol. 1996, 3,
7177.
3 L. L. Kiessling, J. E. Gestwicki, L. E. Strong,
Curr. Opin. Chem. Biol. 2000, 4, 696703.
4 D. Wright, L. Usher, Curr. Org. Chem. 2001, 5,
11071131.
5 L. L. Kiessling, L. E. Strong, J. E. Gestwicki,
Annu. Rep. Med. Chem. 2000, 35, 321330.
6 M. N. Matrosovich, FEBS Lett. 1989, 252, 14.
7 A. Spaltenstein, G. M. Whitesides, J. Am.
Chem. Soc. 1991, 113, 686687.
8 S.-K. Choi, M. Mammen, G. M. Whitesides,
J. Am. Chem. Soc. 1997, 119, 41034111.
9 S. D. Debenham, P. W. Snyder, E. J. Toone,
J. Org. Chem. 2003, 68, 58055811.
10 P. I. Kitov, H. Shimizu, S. W. Homans, D. R.
Bundle, J. Am. Chem. Soc. 2003, 125, 3284
3294.
11 G. D. Glick, P. L. Toogood, D. C. Wiley, J. J.
Skehel, J. R. Knowles, J. Biol. Chem. 1991,
266, 2366023669.
12 E. Fan, Z. Zhang, W. E. Minke, Z. Hou,
C. L. M. J. Verlinde, W. G. J. Hol, J. Am. Chem.
Soc. 2000, 122, 26632664.
13 Z. S. Zhang, E. A. Merritt, M. Ahn, C. Roach,
Z. Hou, C. L. M. J. Verlinde, W. G. J. Hol, E.
Fan, J. Am. Chem. Soc. 2002, 124, 12991
12998.
14 J. J. Lundquist, S. D. Debenham, E. J. Toone,
J. Org. Chem. 2000, 65, 82458250.
15 M. G. Carrithers, M. R. Lerner, Chem. Biol.
1996, 3, 537542.
16 J. C. Cheronis, E. T. Whalley, K. T. Nguyen,
S. R. Eubanks, L. G. Allen, M. J. Duggan, S. D.
Loy, K. A. Bonham, J. K. Blodgett, J. Med.
Chem. 1992, 35, 15631572.
17 A. P. Tamiz, B. C. Bandyopadhyay, J. R. Zhang,
J. L. Flippen-Anderson, M. Zhang, C. Z. Wang,
K. M. Johnson, S. Tella, A. P. Kozihowski,
J. Med. Chem. 2001, 44, 16151622.
18 A. H. Abadi, S. Lankow, B. Hoefgen, M. De-
cker, M. U. Kassack, J. Lehmann, Arch. Pharm.
2002, 335, 367373.
19 P. S. Portoghese, J. Med. Chem. 2001, 44,
22592269.
20 N. Schaschke, S. Fiori, E. Weyher, C. Escieut,
D. Fourmy, G. Mller, L. Moroder, J. Am.
Chem. Soc. 1998, 120, 70307038.
21 Y.-P. Pang, P. Quiram, T. Jelacic, F. Hong,
S. Brimijoin, J. Biol. Chem. 1996, 271, 23646
23649.
22 P. R. Carlier, Y.-F. Han, E. S.-H. Chow, C. P.-L.
Li, H. Wang, T. X. Lieu, H. S. Wong, Y.-P.
Pang, Bioorg. Med. Chem. 1999, 7, 351357.
23 P. R. Carlier, E. S.-H. Chow, Y.-F. Han, J. Liu,
J. El Yazal, Y.-P. Pang, J. Med. Chem. 1999, 42,
42254231.
24 Y.-F. Han, C. P.-L. Li, E. S.-H. Chow, H. Wang,
Y.-P. Pang, P. R. Carlier, Bioorg. Med. Chem.
Lett. 1999, 7, 25692575.
25 A. Mary, D. Z. Renko, C. Guillou, C. Thal,
Bioorg. Med. Chem. Lett. 1998, 6, 18351850.
26 W. G. Lewis, L. G. Green, F. Grynszpan,
Z. Radic, P. R. Carlier, P. Taylor, M. G. Finn,
K. B. Sharpless, Angew. Chem. Int. Ed. 2002,
41, 10531057.
27 P. R. Carlier, D.-M. Du, Y.-F. Han, J. Liu,
E. Perola, I. D. Williams, Y.-P. Pang, Angew.
Chem. Int. Ed. 2000, 39, 17751777.
28 D. M. Wong, H. M. Greenblatt, H. Dvir, P. R.
Carlier, Y.-F. Han, Y.-P. Pang, I. Silman, J. L.
Sussman, J. Am. Chem. Soc. 2003, 125, 363
373.
29 R. P. Lyon, J. J. Hill, W. M. Atkins, Biochemistry
2003, 42, 1041810428.
30 A. A. Profit, T. R. Lee, D. S. Lawrence, J. Am.
Chem. Soc. 1999, 121, 280283.
31 K. Parang, P. A. Cole, Pharmacol. Ther. 2002,
93, 145157.
32 T. J. Rydel, K. G. Ravichandran, A. Tulinsky,
W. Bode, R. Huber, C. Roitsch, J. W. II Fenton,
Science 1990, 249, 277280.
33 J. Dodt, Angew. Chem. Int. Ed. 1995, 34, 867
880.
34 M. Groll, L. Ditzel, J. Lwe, D. Stock, M.
Bochtler, H. D. Bartunik, R. Huber, Nature
1997, 386, 463471.
35 P. J. B. Pereira, A. Bergner, S. Macedo-Ribeiro,
R. Huber, G. Matschiner, H. Fritz, C. P. Som-
merhoff, W. Bode, Nature 1998, 392, 306311.
36 C. P. Sommerhoff, W. Bode, P. J. B. Pereira,
M. T. Stubbs, J. Strzebecher, G. P. Piechottka,
G. Matschiner, A. Bergner, Proc. Natl. Acad.
Sci. USA 1999, 96, 1098410991.
37 G. Loidl, M. Groll, H.-J. Musiol, R. Huber, L.
Moroder, Proc. Natl. Acad. Sci. USA 1999, 96,
54185422.
38 N. Schaschke, G. Matschiner, F. Zettl, U. Mar-
quaerdt, A. Bergner, W. Bode, C. P. Sommerh-
off, L. Moroder, Chem. Biol. 2001, 8, 313327.
39 D. M. Crothers, H. Metzger, Immunochemistry
1972, 9, 341-357.
40 H. P. Rappaport, J. Theor. Biol. 1979, 79, 157
165.
41 W. P. Jencks, Proc. Natl. Acad. Sci. USA 1981,
78, 4046-4050.
42 P. I. Kitov, D. R. Bundle, J. Am. Chem. Soc.
2003, 125, 1627116284.
43 J. Rao, J. Lahiri, L. Isaacs, R. M. Weis, G. M.
Whitesides, Science 1998, 280, 708711.
44 J. Rao, J. Lahiri, R. M. Weis, G. M. Whitesides,
J. Am. Chem. Soc. 2000, 122, 26982710.
45 J. Adams, Trends Mol. Med. 2002, 8, S49S54.
46 A. F. Kisselev, A. L. Goldberg, Chem. Biol.
2001, 8, 739758.
47 K. L. Rock, A. L. Goldberg, Annu. Rev. Immu-
nol. 1999, 17, 739779.
48 J. Myung, K. B. Kim, C. M. Crews, Med. Res.
Rev. 2001, 21, 245273.
49 A. L. Goldberg, K. Rock, Nat. Med. 2002, 8,
338340.
50 J. Lwe, D. Stock, B. Jap, P. Zwickl, W. Bau-
meister, R. Huber, Science 1995, 268, 533539.
51 M. Groll, M. Bajorek, A. Kohler, L. Moroder,
D. M. Rubin, R. Huber, M. H. Glickman, D.
Finley, Nature Struct. Biol. 2000, 7, 10621067.
52 M. Groll, H. Brandstetter, H. Bartunik, G.
Bourenkow, R. Huber, J. Mol. Biol. 2003, 327,
7583.
References 415
53 M. Groll, T. Clausen, Curr. Opin. Struct. Biol.
2003, 13, 665673.
54 W. Heinemeyer, M. Fischer, T. Krimmer, U.
Stachon, D. H. Wolf, J. Biol. Chem. 1997, 272,
2520025209.
55 A. K. Nussbaum, T. P. Dick, W. Keilholz, M.
Schirle, S. Stefanovic, K. Dietz, W. Heine-
meyer, M. Groll, D. H. Wolf, R. Huber, H. G.
Rammensee, H. Schild, Proc. Natl. Acad. Sci.
USA 1998, 95, 1250412509.
56 M. Groll, W. Heinemeyer, S. Jger, T. Ullrich,
M. Bochtler, D. H. Wolf, R. Huber, Proc. Natl.
Acad. Sci. USA 1999, 96, 1097610983.
57 M. Orlowski, S. Wilk, Arch. Biochem. Biophys.
2000, 383, 116.
58 M. Groll, Y. Koguchi, R. Huber, J. Kohno,
J. Mol. Biol. 2001, 311, 543548.
59 M. Groll, K. B. Kim, N. Kairies, R. Huber,
C. M. Crews, J. Am. Chem. Soc. 2000, 122,
12371238.
60 M. Groll, T. Nazif, R. Huber, M. Bogyo, Chem.
Biol. 2002, 9, 655662.
61 M. Kaiser, M. Groll, C. Siciliano, I. Assfalg-
Machleidt, E. Weyher, J. Kohno, A. G. Mil-
bradt, C. Renner, R. Huber, L. Moroder,
ChemBioChem 2004, 5, 12561266.
62 K. L. Rock, C. Gramm, L. Rothstein, K. Clark,
R. Stein, L. Dick, D. Hwang, A. L. Goldberg,
Cell 1994, 78, 761771.
63 L. Ditzel, R. Huber, K. Mann, W. Heinemeyer,
D. H. Wolf, M. Groll, J. Biol. Chem. 1998, 279,
11871191.
64 G. Loidl, H.-J. Musiol, M. Groll, R. Huber,
L. Moroder, J. Pept. Sci. 2000, 6, 3646.
65 M. Mammen, E. I. Shakhnovich, J. M. Deutch,
G. M. Whitesides, J. Org. Chem. 1998, 63,
38213830.
66 M. Mammen, E. I. Shakhnovich, G. M. White-
sides, J. Org. Chem. 1998, 63, 31683175.
67 A. J. Rivett, J. Biol. Chem. 1989, 264, 12215
12219.
68 L. R. Dick, C. R. Moomaw, G. N. Pramanik,
G. N. DeMartino, C. A. Slaughter, Biochemistry
1992, 31, 73477355.
69 C. S. Arendt, M. Hochstrasser, Proc. Natl.
Acad. Sci. USA 1997, 94, 71567161.
70 G. Loidl, M. Groll, H.-J. Musiol, L. Ditzel, R.
Huber, L. Moroder, Chem. Biol. 1999, 6, 197
204.
71 L. B. Schwartz, A.-M. A. Irani, K. Roller, M. C.
Castells, N. M. Schechter, J. Immunol. 1987,
138, 26112615.
72 G. H. Caughey, Mast Cell Proteases in Immunol-
ogy and Biology, Marcel Dekker, New York,
1995.
73 J. F. Molinari, M. Scuri, W. R. Moore, J. Clark,
R. Tanaka, W. M. Abraham, Am. J. Resp. Crit.
Care Med. 1996, 154, 649653.
74 K. D. Rice, P. A. Spengler, Curr. Opin. Drug
Disc. Dev. 1999, 2, 463474.
75 B. J. Newhouse, Drugs 2002, 5, 682688.
76 W. J. Tremaine, J. A. Brzezinski, J. A. Katz,
D. C. Wolf, T. J. Flemming, J. Mordentti,
L. C. Strenkoskji-Nix, M. C. Kurth, T. A. U. C. S.
Group, Aliment. Pharmacol. Ther. 2002, 16,
407413.
77 L. E. Burgess, B. J. Newhouse, P. Ibrahim, J.
Rizzi, M. A. Kashem, A. Hartman, B. J. Brand-
huber, C. D. Wright, D. S. Thomson, G. P. A.
Vigers, K. Koch, Proc. Natl. Acad. Sci. USA
1999, 96, 83488352.
78 S. Ono, S. Kuwahara, M. H. S. Takeuchi,
Y. Naito, T. Kondo, Bioorg. Med. Chem. Lett.
1999, 9, 32853290.
79 K. D. Rice, A. R. Gangloff, E. Y.-L. Kuo, J. M.
Dener, V. R. Wang, R. Lum, W. S. Newcomb,
C. Havel, D. Putnam, L. Cregar, M. Wong,
R. L. Warne, Bioorg. Med. Chem. Lett. 2000, 10,
23572360.
80 K. D. Rice, V. R. Wang, A. R. Gangloff, E. Y.-L.
Kuo, J. M. Dener, W. S. Newcomb, W. B.
Young, D. Putnam, L. Cregar, M. Wong, P. J.
Simpson, Bioorg. Med. Chem. Lett. 2000, 10,
23612366.
81 J. M. Dener, V. R. Wang, K. D. Rice, A. R.
Gangloff, E. Y.-L. Kuo, W. S. Newcomb, D. Put-
nam, M. Wong, Bioorg. Med. Chem. Lett. 2001,
11, 23252330.
82 N. Schaschke, A. Dominik, G. Matschiner,
C. P. Sommerhoff, Bioorg. Med. Chem. Lett.
2002, 12, 985988.
83 N. M. Schechter, G. Y. Eng, T. Selwood, D. R.
McCaslin, Biochemistry 1995, 34, 1062810638.
84 T. Selwood, K. C. Elrod, N. M. Schechter, Biol.
Chem. 2003, 384, 16051611.
85 U. Marquaerdt, Ph.D. Thesis, Ludwig-Maximi-
lians Universitt, Munich, 2002.
86 N. Schaschke, D. Gabrijelcic-Geiger,
A. Dominik, C. P. Sommerhoff, Chem. Bio.
Chem. 2005, 6, 95103.
87 E. W. Davie, K. Fujikawa, W. Kisiel, Biochemis-
try 1991, 30, 1036310370.
88 J. W. II Fenton, F. A. Ofosu, D. G. Moon,
J. M. Maraganore, Blood Coagulation Fibrino-
lysis 1991, 2, 6975.
89 J. Hauptmann, J. Strzebecher, Thromb. Res.
1999, 93, 203241.
90 K. Menear, Exp. Opin. Invest. Drugs 1999, 8,
13731384.
91 T. Steinmetzer, J. Hauptmann, J. Strze-
becher, Exp. Opin. Invest. Drugs 2001, 10,
845864.
92 F. Markwardt, Meth. Enzymol. 1970, 19, 924
932.
93 M. Grtter, J. P. Priestle, J. Rahuel, H. Gos-
senbacher, W. Bode, J. Hofsteenge, S. R.
Stone, EMBO J. 1990, 9, 23612365.
94 P. J. Braun, S. Dennis, J. Hofsteenge,
S. R. Stone, Biochemistry 1988, 27, 6517
6522.
95 S. Dennis, A. Wallace, J. Hofsteenge, S. R.
Stone, Eur. J. Biochem. 1990, 188, 6166.
96 T. Schmitz, M. Rothe, J. Dodt, Eur. J. Bio-
chem. 1991, 195, 251256.
97 A. Karshikov, W. Bode, A. Tulinsky, S. R.
Stone, Protein Sci. 1992, 1, 727735.
98 P. H. Rehse, T. Steinmetzer, Y. Li, Y. Konishi,
M. Cygler, Biochemistry 1995, 34, 1153711544.
99 S.-Y. Yue, J. DiMaio, S. Szewczuk, E. O. Puri-
sima, F. Ni, Y. Konishi, Protein Engineer.
1992, 5, 7785.
100 T. Steinmetzer, M. Renatus, S. Knzel, A.
Eichinger, W. Bode, P. Wikstrm, J. Haupt-
mann, J. Strzebecher, Eur. J. Biochem.
1999, 265, 598605.
101 W. Bode, I. Mayr, U. Baumann, R. Huber,
S. R. Stone, J. Hofsteenge, EMBO J. 1989, 8,
34673475.
102 J. DiMaio, B. Gibbs, D. Munn, J. Lefebvre,
F. Ni, Y. Konishi, J. Biol. Chem. 1990, 265,
2169821703.
103 J. M. Maraganore, P. Bourdon, J. Jablonski,
K. L. Ramachandran, J. W. II Fenton, Bio-
chemistry 1990, 29, 70957101.
104 J. I. Wittig, P. Bourdon, J. M. Maraganore,
J. W. II Fenton, Biochem. J. 1992, 287, 663
664.
105 T. Kline, C. Hammond, P. Bourdon, J. M.
Maraganore, Biochem. Biophys. Res. Com-
mun. 1991, 177, 10491055.
106 T. Steinmetzer, B. Y. Zhu, Y. Konishi, J. Med.
Chem. 1999, 42, 31093115.
107 R. Friedrich. T. Steinmetzer, R. Huber, J.
Strzebecher, W. Bode, J. Mol. Biol. 2002,
316, 859874.
108 Y. Tsuda, M. Cygler, B. F. Gibbs, A. Pedy-
czak, J. Fthire, S. Y. Yue, Y. Konishi, Bio-
chemistry 1994, 33, 1444314451.
109 Z. Szewczuk, B. F. Gibbs, S. Y. Yue, E. Purisi-
ma, A. Zdanov, M. Cygler, Y. Konishi, Bio-
chemistry 1993, 32, 33963404.
110 J. L. Krstenansky, R. J. Broersma, T. J. Owen,
M. H. Payne, M. T. Yates, S. J. T. Mao, Throm-
bosis and Haemostasis 1990, 63, 208214.
111 J. J. Slon-Usakiewicz, E. Purisima, Y. Tsuda,
T. Sulea, A. Pedyczak, J. Fthire, M. Cygler,
Y. Konishi, Biochemistry 1997, 36, 13494
13502.
112 J. J. Slon-Usakiewicz, J. Sivaraman, Y. Li,
M. Cygler, Y. Konishi, Biochemistry 2000, 39,
23842391.
113 A. Lombardi, F. Nastri, R. Della Morte, A.
Rossi, A. De Rosa, N. Staino, C. Pedone, V.
Pavone, J. Med. Chem. 1996, 39, 20082017.
114 X. Qiu, K. P. Padmanabhan, V. E. Carperos,
A. Tulinsky, T. Kline, J. M. Maraganore,
J. W. II Fenton, Biochemistry 1992, 31,
1168911697.
115 X. Qiu, M. Yin, K. P. Padmanabhan, J. L.
Krstenansky, A. Tulinsky, J. Biol. Chem.
1993, 268, 2031820326.
116 A. Zdanov, S. Wu, J. DiMaio, Y. Konishi,
Y. Li, X. Wu, B. F. P. Edwards, P. D. Martin,
M. Cygler, Proteins: Structure, Function and
Genetics 1993, 17, 252265.
117 J. Fthire, Y. Tsuda, R. Coulombe, Y. Ko-
nishi, M. Cygler, Protein Sci. 1996, 5, 1174
1183.
118 A. Lombardi, G. De Simone, F. Nastri, S.
Caldiero, R. Della Morte, N. Staino, C. Pe-
done, M. Bolognesi, V. Pavone, Protein Sci.
1999, 8, 9195.
119 C. A. Lepinski, F. Lombardo, B. W. Dominy,
P. J. Feeney, Adv. Drug Del. Rev. 1997, 23,
325.
120 D. F. Veber, S. R. Johnson, H.-Y. Cheng,
B. R. Smith, K. W. Ward, K. D. Kopple,
J. Med. Chem. 2002, 45, 26152623.
References 417
Abstract
The development of genetic engineering
during the late 1970s has opened new
pathways in the basic research of diseases,
and has allowed for the evolution of a
growing number of drugs and diagnostics
based on recombinant technologies. Today,
a number of these products have become
standards in the prevention and/or the
treatment of the disease for which they
were developed. Recombinant FVIII
(rFVIII) concentrates offer the advantages
of lower risk for blood-borne pathogen
transmission, reduced impact on the im-
mune system, and supply that is indepen-
dent of plasma availability. However, all
previously developed rFVIII concentrates
incorporate human- or animal-derived pro-
teins at some point in processing; thus,
concerns remain within the hemophilia
community regarding possible pathogen
transmission through these additives. This
chapter will describe the development, pro-
duction and clinical study programme of a
novel full-length rFVIII preparation for the
treatment of hemophilia A. This rFVIII is
the first to be processed using a plasma/al-
bumin-free method (rAHF-PFM), provid-
ing a new standard of pathogen safety for
hemophilia A patients.
Abbreviations
A1 subunit 1 of A domain
A2 subunit 2 of A domain
ADA adenosine deaminase
ADP adenosine diphosphate
APC activated protein C
AUC
048
activity under curve (interna-
(IUhr/dl) tional units h dL
1
)
BiP immunoglobulin-binding
protein
BU Bethesda Unit
BVDV bovine viral diarrhoea virus
CHO Chinese hamster ovary
CNX calnexin
CPMP Committee for Proprietary
Medicinal Products
CRT calreticulin
dCF deoxycoformycin
DHFR dihydrofolate reductase
EBL estimated blood loss
EMEA European Agency for the
Evaluation of Medicinal
Products
ER endoplasmic reticulum
ERGIC-53 endoplasmatic reticulum
Golgi intermediate compart-
ment lectinlike
F Xa activated factor X
419
3
A New Technology Standard for Safety and Efficacy in Factor VIII
Replacement Therapy: Designing an Advanced Category rFVIII
Concentrate
ISBN: 3-527-31184-X
FDA Food and Drug Administra-
tion
FVIII antihemophilic Factor VIII
GRP 78 glucose-regulated protein of
78 kDa
HAP Hamster antibody production
test
HC heavy chain
IA immunoaffinity
ICH International Conference on
Harmonisation
IE ion-exchange chromatography
IU International Units
kDa kilo Dalton
LC light chain
LMAN1 lectin mannose-binding pro-
tein type 1 (53 kDa)
MAP mouse antibody production
test
MASAC Medical and Scientific Advi-
sory Council
MCB master cell bank
MCFD2 multiple coagulation factor
deficiency 2 protein
MMV mice minute virus
NHF National Haemophilia Foun-
dation
PAGE polyacrylamide gel electro-
phoresis
pd FVIII plasma-derived factor VIII
P
i
inorganic phosphate
PL phospholipids
PRV porcine pseudorabies virus
PTP previously treated patient
PUP previously untreated patient
rAHF-PFM recombinant antihaemophilic
factorprotein free method
RAP rat antibody production test
REOV-3 reovirus type 3
rFVIII recombinant Factor VIII
RP-HPLC reverse-phase high-pressure
S/D solvent detergent
S
+
L
focus assay murine sarcoma

virus positive (S
+
), murine
leukemia
SDS-PAGE sodium dodecyl sulfate-poly-
acrylamide gel electrophoresis
UKHCDO United Kingdom Haemophi-
lia Centre Doctors Organiza-
tion
US FDA United States Food and Drug
Administration
vCJD variant Creutzfeld-Jakob dis-
ease
virus indicator cell focus assay
negative (L
)
vWF von Willebrand factor
WCB working cell bank
XC plaque assay XC cell line
plaque assay
X-MuLV xenotropic murine leukemia
virus
3.1
Introduction
The development and production of recom-
binant Factor VIII (rFVIII) using a plasma/
albumin-free method is a complex process,
owing in large part to the unusually large
and labile FVIII molecule, which requires
extensive and complex post-translational
modifications. rAHF-PFM (for definition,
see Section 3.3) is a full-length recombinant
FVIII molecule, retaining the known func-
tion of the B domain, and thus interaction
with key chaperone proteins (ERGIC-53,
calnexin and calreticulin) that have been
shown to play important roles in intracellu-
lar transport and monitoring of proper
FVIII protein folding during biosynthesis.
The Chinese hamster ovary (CHO) cell line
from which rAHF-PFM is processed is
highly characterized and capable of consis-
tently carrying out all necessary FVIII
post-translational modifications. Both the
3 A New Technology Standard for Safety and Efficacy in Factor VIII Replacement Therapy 420
full-length FVIII and the co-expressed von
Willebrand factor (vWF) are genetically
identical to those expressed in the parent
cell line used in the processing of Baxters
other rFVIII, Recombinate
TM
. However,
the CHO cells that produce rAHF-PFM
have been adapted to a protein-free culture
medium. Upon development, characteriza-
tion of the rAHF-PFM protein, validation
of the manufacturing process and extensive
testing in the clinical setting are imperative.
In the preclinical programme, rAHF-PFM
demonstrated physico-chemical and func-
tional characteristics (glycosylation, tyrosine
sulfation, thrombin activation and other
biomolecular interactions) similar to those
of Recombinate
TM
. Hemostatic efficacy
and toxicology studies with rAHF-PFM (Ad-
vate
TM
) and Recombinate
TM
in animal mod-
els also revealed highly similar results. Col-
lectively, the preclinical data predicted safety
and efficacy of rAHF-PFM comparable to
that of Recombinate in clinical scenarios.
The strategy for the rAHF-PFM clinical
programme was to show pharmacokinetic
comparability with Recombinate (rAHF),
assess efficacy in bleed prevention, episod-
ic treatment and surgical settings, and
evaluate immunogenicity and safety in var-
ious patient populations. Overall, several
separate clinical studies have been com-
pleted, are underway, or are planned in
previously treated and untreated patients
(PTPs and PUPs) with moderately severe
to severe hemophilia A. The Phase II/III
Pivotal study was completed at the end of
2002, and evaluated the pharmacokinetics,
immunogenicity, safety and efficacy of
rAHF-PFM in more than 100 PTPs in the
US, Canada, and Europe. Patients who
completed the Pivotal study in North
America and Europe were then eligible to
enroll in the ongoing Continuation study
examining longer-term therapy with
rAHF-PFM. An additional study is evaluat-
ing the efficacy and safety of rAHF-PFM
in perioperative settings in PTPs under-
going surgical or other invasive proce-
dures; administration by both bolus and
continuous infusion is permitted in this
study. Furthermore, rAHF-PFM is being
evaluated in young children less than 6
years of age in two different studies. The
pharmacokinetics, efficacy, safety, and im-
munogenicity of rAHF-PFM is being as-
sessed in the ongoing Phase II/III PTP pe-
diatric study in young children with at
least 50 exposure days to Factor VIII thera-
py. The other study is a Phase IV trial in
PUPs that is planned to enroll the first pa-
tient during 2004. This PUP study will ex-
amine the in vivo recovery, immunogenic-
ity and safety in this distinct population.
Overall, the clinical programme has
been designed to assess safety and efficacy
in a wide range of patient populations,
from newborn to pediatric and adult pa-
tients, and clinical settings.
3.1.1
Hemophilia A Therapy
Transfusion therapy for hemophilia was
first proposed in the mid-nineteenth cen-
tury, and whole blood transfusion began
early in the twentieth century. The early
evolution of plasma replacement therapy
featured rapid advances in technique and
technology that had a tremendous impact
on clinical practice. The large volumes of
blood or citrated plasma replacement re-
quired to achieve hemostasis following
major bleeding episodes evolved over time
to more manageable amounts of cryopre-
cipitate, to highly purified plasma-derived
FVIII (pdFVIII) concentrates [5], and fi-
nally to recombinant human FVIII con-
centrates (rFVIII).
All of these pharmaceutical preparations
have in common that they contain Antihe-
mophilic Factor VIII (FVIII) as the active
ingredient this is the blood clotting fac-
tor which is either deficient or absent in
individuals with classic hemophilia A.
Approximately 80% of hemophilia pa-
tients have hemophilia A; this is a congen-
ital bleeding disorder resulting from insuf-
ficient levels of FVIII coagulation activity,
and is characterized by a prolonged clot-
ting time. Because the FVIII gene that
codes for the FVIII protein is located on
the X chromosome, virtually all clinically
affected individuals are male [1].
Levels of FVIII deficiency relative to nor-
mal plasma are used to categorize the se-
verity of the disorder (see Table 3.1) [2]. Pa-
tients with mild hemophilia A may pres-
ent with bleeding only subsequent to ma-
jor trauma or surgery, while those with the
severe form of the disorder may suffer
spontaneous episodes of bleeding into
joints, muscles, and internal organs, even
in the absence of trauma. If untreated,
such bleeding may result in serious com-
plications including permanent joint, mus-
cle, and nerve damage and loss of muscu-
loskeletal function, or even death [1, 3, 4].
Another major complication in the treat-
ment of hemophilia A is the occurrence of
inhibitors against FVIII (neutralizing anti-
bodies) in about 30% of patients, usually
within the first 100 exposure days. Patients
with severe hemophilia A (FVIII levels
<1% of normal activity) are at higher risk
to develop an inhibitor.
As advances in hemostatic efficacy were
achieved, much of the focus in hemophilia
therapy and research shifted to safety and,
in particular, to the issues of blood-borne
pathogen transmission and FVIII inhibitor
development [5, 6].
The introduction of FVIII concentrates
derived from large plasma pools mandated
implementation of diagnostic and proce-
dural antiviral measures. Along with re-
finements in donor selection procedures,
advanced assays for screening donated
plasma have been introduced, and viral in-
activation and elimination techniques have
been further developed (see Table 3.2) [7].
Different viral inactivation technologies
were developed and introduced into the
manufacturing process of plasma-derived
factor VIII concentrates; these ranged
from relatively simple heat treatment to
more sophisticated and highly effective va-
por heating and solvent/detergent (S/D)
inactivation.
Newly developed purification methodolo-
gies such as purification steps with immo-
bilized monoclonal antibodies and other
affinity chromatographic techniques not
only increased the purity and specific activ-
ity of factor VIII concentrates, but also
showed at the same time the substantial
removal of potentially present pathogens
Table 3.1 Clinical classification of hemophilia A [1, 2]
Classification Severe Moderate Mild
FVIII activity <1% 15% >5 to <40%
Frequency of bleeding episodes 24 per month
(approx)
46 per year
(approx)
Uncommon
Pattern of bleeding episodes Spontaneous
Minor trauma
Surgery
Minor trauma
Surgery
Major trauma
Surgery
Adapted from White et al. [2]
which had been demonstrated in preclini-
cal validation studies. On the front of
screening of donated plasma, advanced as-
says for testing of larger panels of viral
antibodies were introduced, and ultimately
PCR testing of viral genome sequences in
plasma donations improved the quality of
plasma and plasma pools as starting mate-
rial for fractionation.
All of these measures and their mean-
ingful combination introduced into the
manufacturing process significantly in-
creased the safety margin of plasma-de-
rived FVIII concentrates [8].
A major breakthrough in FVIII replace-
ment therapy safety was achieved with the
development of recombinant FVIII con-
centrates [7]. The development of rFVIII
was also a major accomplishment in bio-
technology that required cloning, identifi-
cation and transfection of the rFVIII gene
into suitable host cell lines, and subse-
quent characterization of the expressed
proteins. An equally remarkable feat of
manufacturing expertise was required to
achieve purification, formulation and vali-
dation of a finished rFVIII therapeutic on
a commercial scale. The first rFVIII be-
came commercially available in 1992, and
other rFVIII concentrates followed. The
long record of efficacy and safety has
made recombinant FVIII concentrates the
standard for care in hemophilia A therapy
[911].
3.1.2
Rationale for Designing an Advanced
Category rFVIII Concentrate
Recombinant FVIII concentrates offer the
advantage of lower risk for blood-borne
pathogen transmission, reduced impact on
the immune system and supply that is in-
dependent of plasma availability. However,
all first- and next-generation rFVIII con-
centrates used either human-derived or an-
imal-derived additives at some point of
processing [12]. Despite the excellent
safety record of rFVIII therapeutics in gen-
eral, concerns remain within the hemophi-
Table 3.2 Effectiveness of methods for the attenuation of infectious risk in FVIII concentrates [7, 8]
Pathogen Viruses Prions Emerging/
Uncharacterized
Pathogens Lipid-
enveloped
Non-lipid-enveloped
HAV-like PV B19-like
Procedure
HIV, HBV,
HCV
Donor questionnaire Moderate Moderate Moderate Somewhat Unpredictable
Plasma testing
a)
High High Moderate No assay Unpredictable
IA/IE High High High Moderate Unpredictable
Heating in solution
b)
High Moderate Somewhat Unknown Unknown
S/D High None None Unknown Unknown
Nanofiltration High Moderate Moderate Somewhat Unpredictable
a) Includes minipools.
b) Product subjected to heat in an aqueous solution at 60 8C for 1011 h.
IA = immunoaffinity chromatography; IE = ion exchange chromatography; S/D = solvent detergent.
lia community with respect to the poten-
tial transmission of blood-borne infectious
agents through the use of human- or ani-
mal-derived additives in rFVIII concen-
trates. The possible transmission of the
still poorly understood infectious agent(s)
causing variant CreutzfeldJakob disease
(vCJD) is an example of one such source
of concern [13, 14].
Recommendations from the United King-
dom Hemophilia Centre Doctors Organiza-
tion (UKHCDO) [15] and the Medical and
Scientific Advisory Council (MASAC) [16]
of the US National Hemophilia Foundation
(NHF) have urged the development of a
rFVIII which is processed without any hu-
man- or animal-derived proteins.
The most recent UKHCDO guideline
(September 2003) states [15]: To reduce
the chance of infection by exogenous
viruses in a recombinant concentrate, con-
sideration should be given to choosing
one, where available and licensed, that is
manufactured with the least addition of
human or animal protein.
The most recent North American MA-
SAC recommendation # 41 says: Im-
proved viral inactivation and elimination
are required in coagulation products. All
efforts should be made to remove human
albumin from recombinant factor VIII
products. Increased efforts should be
made to eliminate human and bovine pro-
teins from the manufacturing process of
recombinant factor VIII products.
In an attempt to follow these recom-
mendations and to increase product safety,
a new rFVIII product was developed by
the introduction of modifications to the
fermentation, the purification process and
the drug product formulation. These elimi-
nated the requirements for human- and
animal-derived raw materials and excipi-
ents at all stages of the production pro-
cess. In order to distinguish between dif-
ferent product generations in the follow-
ing, the first product generation of rFVIII
will be named rAHF (for recombinant
Antihemophilic Factor VIII), while the
newly designed and developed product will
be referred as rAHF-PFM (for recombi-
nant Antihemophilic Factor VIII Protein
Free Manufactured).
3.1.3
FVIII Protein
The FVIII gene, located at the tip of the
long arm of the X chromosome, is relative-
ly large, spanning 186 kb to encode a
2351-amino acid, single-chain precursor
polypeptide. A signal peptide is cleaved
during biosynthesis as the protein translo-
cates into the endoplasmic reticulum (ER),
resulting in a mature protein of 2332 ami-
no acids. The FVIII protein comprises
three homology domains (A, B, C) ar-
ranged in the sequence: A1-A2-B-A3-C1-C2
(see Fig. 3.1) [17].
Before secretion, the single-chain poly-
peptide is cleaved to form a heterodimer
consisting of a heavy and light chain (see
Fig. 3.2) that circulates in plasma in the
inactive form. The light chain is consis-
tently 80 kDa in size, but limited proteoly-
sis occurs within the B domain, resulting
in variably sized heavy chains ranging in
molecular mass from 90 to 200 kDa. The
FVIII molecule is maximally activated as
thrombin cleaves at various sites within
the heavy and light chains, releasing the B
domain and resulting in a heterodimer
consisting of the A1 and A2 subunits non-
covalently bound to the light chain [18].
The importance of the B domain has
only recently been appreciated. Although
the B domain is not essential for hemo-
static function, it is the portion of the mol-
ecule that binds key intracellular chaper-
one proteins and thus helps to ensure
F
i
g
.
3
.
1
F
V
I
I
I
s
t
r
u
c
t
u
r
e
a
n
d
f
u
n
c
t
i
o
n
.
H
o
m
o
l
o
g
y
d
o
m
a
i
n
s
a
r
e
i
n
-
d
i
c
a
t
e
d
b
y
A
,
B
,
C
;
a
c
i
d
i
c
r
e
g
i
o
n
s
a
r
e
i
n
d
i
c
a
t
e
d
b
y
a
1
,
a
2
,
a
3
.
S
i
t
e
s
a
s
s
o
c
i
a
t
e
d
w
i
t
h
F
V
I
I
I
a
c
t
i
v
a
t
i
o
n
a
n
d
c
o
f
a
c
t
o
r
a
c
t
i
v
i
t
y
a
r
e
m
a
r
k
e
d
b
y
a
r
r
o
w
s
.
T
h
r
o
m
b
i
n
,
t
h
e
m
a
i
n
p
h
y
s
i
o
l
o
g
i
c
a
l
a
c
t
i
v
a
t
o
r
o
f
F
V
I
I
I
,
c
l
e
a
v
e
s
a
t
a
m
i
n
o
a
c
i
d
s
3
7
2
,
7
4
0
,
a
n
d
1
6
8
9
;
F
I
X
a
c
o
m
p
l
e
x
e
s
w
i
t
h
F
V
I
I
I
a
a
t
a
m
i
n
o
a
c
i
d
s
5
5
8
5
6
5
,
6
9
8
7
1
0
,
a
n
d
1
8
1
1
1
8
1
8
d
u
r
i
n
g
a
s
s
e
m
b
l
y
o
f
t
h
e
F
a
c
t
o
r
X
-
a
c
t
i
v
a
t
i
n
g
c
o
m
p
l
e
x
;
a
n
d
p
h
o
s
-
p
h
o
l
i
p
i
d
s
(
P
L
)
o
f
t
h
e
m
e
m
b
r
a
n
e
s
u
r
f
a
c
e
d
u
r
i
n
g
c
o
m
p
l
e
x
a
s
s
e
m
-
b
l
y
b
i
n
d
F
V
I
I
I
a
t
t
h
e
c
a
r
b
o
x
y
-
t
e
r
m
i
n
a
l
o
f
t
h
e
l
i
g
h
t
c
h
a
i
n
.
v
o
n
W
i
l
l
-
e
b
r
a
n
d
f
a
c
t
o
r
(
v
W
F
)
b
i
n
d
s
t
h
e
a
m
i
n
o
-
a
n
d
c
a
r
b
o
x
y
-
t
e
r
m
i
n
a
l
e
n
d
s
o
f
t
h
e
F
V
I
I
I
l
i
g
h
t
c
h
a
i
n
(
1
6
4
9
1
6
7
1
a
n
d
2
3
0
3
2
3
3
2
,
r
e
s
p
e
c
t
i
v
e
l
y
)
t
o
s
t
a
b
i
l
i
z
e
t
h
e
m
o
l
e
c
u
l
e
d
u
r
i
n
g
c
i
r
c
u
l
a
t
i
o
n
a
n
d
t
o
p
r
e
v
e
n
t
p
r
e
-
m
a
t
u
r
e
p
r
o
t
e
o
l
y
s
i
s
b
y
a
c
t
i
v
a
t
e
d
p
r
o
t
e
i
n
C
a
n
d
F
X
a
.
P
L
(
t
h
e
b
i
n
d
-
i
n
g
o
f
p
h
o
s
p
h
o
l
i
p
i
d
s
o
f
t
h
e
m
e
m
b
r
a
n
e
s
u
r
f
a
c
e
t
o
r
e
g
i
o
n
s
o
f
t
h
e
F
V
I
I
I
a
n
d
F
I
X
a
c
o
m
p
l
e
x
)
i
n
c
r
e
a
s
e
s
c
a
t
a
l
y
t
i
c
a
c
t
i
v
i
t
y
s
e
v
e
r
a
l
-
f
o
l
d
(
s
e
e
F
i
g
.
3
.
2
f
o
r
d
e
t
a
i
l
s
o
f
F
V
I
I
I
p
r
o
c
e
s
s
i
n
g
)
.
A
d
a
p
t
e
d
f
r
o
m
L
e
n
t
-
i
n
g
e
t
a
l
.
[
1
7
]
.
F
i
g
.
3
.
2
F
V
I
I
I
s
e
c
r
e
t
i
o
n
a
n
d
a
c
t
i
v
a
t
i
o
n
.
F
V
I
I
I
i
s
p
r
o
c
e
s
s
e
d
a
s
a
s
i
n
g
l
e
-
c
h
a
i
n
p
e
p
t
i
d
e
,
a
n
d
t
h
e
n
c
l
e
a
v
e
d
p
r
i
o
r
t
o
s
e
c
r
e
t
i
o
n
t
o
f
o
r
m
a
h
e
t
e
r
o
d
i
m
e
r
c
o
n
s
i
s
t
i
n
g
o
f
a
9
0
2
0
0
k
D
a
h
e
a
v
y
c
h
a
i
n
,
d
u
e
t
o
l
i
m
i
t
e
d
p
r
o
t
e
o
l
y
s
i
s
,
a
n
d
a
n
8
0
k
D
a
l
i
g
h
t
c
h
a
i
n
.
T
h
i
s
h
e
t
e
r
o
d
i
m
e
r
i
s
t
h
e
i
n
a
c
t
i
v
e
f
o
r
m
t
h
a
t
c
i
r
c
u
l
a
t
e
s
i
n
p
l
a
s
m
a
.
F
V
I
I
I
i
s
a
c
t
i
v
a
t
e
d
m
a
i
n
l
y
b
y
F
I
I
a
(
t
h
r
o
m
b
i
n
)
,
w
h
i
c
h
c
l
e
a
v
e
s
t
h
e
h
e
a
v
y
a
n
d
l
i
g
h
t
c
h
a
i
n
s
i
n
t
o
s
m
a
l
l
e
r
,
m
o
r
e
a
c
t
i
v
e
p
e
p
t
i
d
e
s
(
a
s
d
e
p
i
c
t
e
d
i
n
t
h
e
f
i
g
-
u
r
e
)
.
F
u
r
t
h
e
r
p
r
o
t
e
o
l
y
t
i
c
d
e
g
r
a
d
a
t
i
o
n
r
e
s
u
l
t
s
i
n
i
n
a
c
t
i
v
a
t
i
o
n
o
f
t
h
e
s
e
p
e
p
t
i
d
e
s
b
y
F
X
a
(
s
h
o
w
n
)
a
n
d
F
I
X
a
/
A
P
C
(
a
c
t
i
v
a
t
e
d
p
r
o
t
e
i
n
C
)
(
n
o
t
s
h
o
w
n
)
.
A
d
a
p
t
e
d
f
r
o
m
K
a
u
f
m
a
n
e
t
a
l
.
[
1
8
]
.
proper folding and secretion of FVIII dur-
ing biosynthesis (see below). Furthermore,
deletion of the B domain may have an im-
pact on clinical efficacy and, possibly,
neoantigenicity of the resulting truncated
rFVIII molecule [19, 20]. With this in
mind, rAHF-PFM (like rAHF) was devel-
oped to retain all regions of the natural,
full-length FVIII molecule [8].
3.1.4
FVIII Biosynthesis
The synthesis of human FVIII is thought
to occur primarily in reticuloendothelial
cells and hepatocytes, although this has
not been definitively established. FVIII
biosynthesis is a complex process that is
still being investigated. Signal peptide
cleavage of the FVIII primary translation
product yields the mature 2332-amino acid
polypeptide upon translocation into the lu-
men of the ER. Within the ER, the FVIII
is folded, and asparagine (N)-linked glyco-
sylation begins. The FVIII molecule is
heavily glycosylated, with 25 potential
sites, 19 of which are located within the B
domain. Secretion of FVIII requires inter-
action with several chaperone proteins, in-
cluding immunoglobulin-binding protein
(glucose-regulated protein of 78 kDa,
GRP78, BiP), calnexin (CNX), calreticulin
(CRT), LMAN1 (also called ERGIC-53),
and MCFD2. BiP binds the FVIII mole-
cule at a hydrophobic site within the A1
domain, while CNX and CRT both bind to
carbohydrate structures of the B domain.
Incorrectly folded FVIII molecules have a
prolonged association with these chaper-
one proteins and are eventually degraded.
Correctly folded FVIII proteins are trans-
ported to the Golgi by a protein complex
consisting of LMAN1 and MCFD2, which
also binds to the FVIII B domain. To-
gether, these chaperone proteins provide a
quality control mechanism in FVIII se-
cretion [18, 21].
In the Golgi apparatus, the FVIII pro-
tein undergoes further post-translational
modifications including complex glycosyla-
tion, sulfation, and cleavage to two chains
the FVIII heavy chain (A1-a1-A2-a2-B;
90200 kDa) and FVIII light chain (a3-A3-
C1-C2; 80 kDa) (see Fig. 3.3). The heavy
and light chains remain non-covalently
bound to each other in the presence of
copper ions. Now, the FVIII molecule is
ready to be secreted from the cell [18].
3.1.5
FVIII Interaction with von Willebrand Factor
Immediately upon secretion from the cell,
the FVIII protein associates with von
Willebrand factor (vWF), the natural FVIII
stabilizer. During circulation in the plas-
ma, vWF regulates FVIII activity in several
ways. It protects FVIII from activation by
FXa and from inactivation by activated
protein C (APC). vWF prevents binding of
FVIII to phospholipids and to activated
platelets in plasma. vWF also regulates
FVIII biosynthesis by promoting the asso-
ciation of FVIII heavy and light chains
and altering intracellular transport and se-
cretion of FVIII from the cell [22]. The
known properties of vWF were exploited
in recombinant FVIII technology: early
studies revealed that vWF, when added to
the culture medium, stabilized and pro-
tected the FVIII protein expressed from
CHO cells, resulting in dramatically en-
hanced accumulation of FVIII in the me-
dium. The need for vWF in the cell cul-
ture medium could be overcome if the cell
expressing FVIII also expressed vWF [23],
which was achieved by co-expressing full-
length FVIII along with vWF in the CHO
cell clones.
3.2
Development of rFVIII
rAHF-PFM is an improved modification,
based on the experience with rAHF, the
first-generation product. For both products
the drug substance is produced by the same
genetically engineered CHO cell line.
Clinical trials and post-marketing sur-
veillance studies of rAHF have conclu-
sively demonstrated the concentrate to be
an effective and safe treatment for hemo-
philia A [911]. Nonetheless, animal addi-
tives are used during cell culture and puri-
fication, and human albumin is used in
the final formulation for stabilization of
the purified rFVIII protein [8].
To maintain the proven pharmacokinetic
and efficacy benefits of rAHF, and to opti-
mize the purity and safety of the newly de-
Fig. 3.3 FVIII biosynthesis. Signal peptide cleav-
age and asparagine (N)-linked glycosylation of the
FVIII occurs after translation and translocation to
the lumen of the endoplasmic reticulum (ER).
Within the ER, the FVIII is folded, which requires
binding to chaperone proteins. FVIII binds immu-
noglobulin-binding protein (BiP) at the A1 do-
main, and is released in an ATP-dependent step.
Next, FVIII binds calnexin (CNX) and calreticulin
(CRT) (not shown) at the B domain. Properly
folded FVIII proteins are transported to the Golgi
by chaperone proteins, LMAN1 and MCFD2 (not
shown) for further processing; LMAN1 also binds
to the FVIII B domain. Incorrectly folded FVIII
molecules have a prolonged association with CNX
and CRT and are eventually degraded by the 26S
proteosome. In the Golgi apparatus, the FVIII pro-
tein undergoes further post-translational modifica-
tions, including complex glycosylation, sulfation,
and cleavage to two chains, the FVIII heavy and
light chains (see Fig. 3.2). The heavy and light
chains remain non-covalently bound to each other
and, finally, the FVIII molecule is ready to be se-
creted from the cell. Upon secretion, FVIII binds
vWF and is protected from degradation. Cu = cop-
per; ADP = adenosine diphosphate; ATP = adeno-
sine triphosphate; Pi = inorganic phosphate.
Adapted from Kaufman et al. [18].
signed rFVIII therapeutic, an advanced
process was developed that incorporates
technological innovations in cell culture,
purification, and formulation into the
proven large-scale methods developed and
used for more than 14 years in rFVIII pro-
cessing [8, 12].
3.2.1
Cell Line Selection
As mentioned above, for both product gen-
erations CHO cells have been chosen as
the host cell expression system. CHO cells
offer the following time-tested qualities [24]:
An ability to grow in the absence of se-
rum.
Resistance to infection by many human
viruses.
Suitability for scale-up in suspension
culture.
Capability of consistent post-transla-
tional modification (glycosylation, sulfa-
tion).
Thorough study and characterization for
more than 40 years.
To process the drug substance, a CHO
cell line was designed by using transfec-
tion and cloning methods to express both
full-length human FVIII and vWF. The
CHO cell clone (co-expressing full-length
FVIII and vWF) has been adapted to grow
in a proprietary culture medium free of
plasma, albumin, or any other human- or
animal-derived additive [8, 12].
Briefly, in the first step an expression
plasmid containing the FVIII cDNA was
co-transfected into dihydrofolate reductase
(DHFR)-deficient CHO cells (DUKX-B11)
along with a plasmid expressing a DHFR-
selectable marker to create the cell line
10A1. Methotrexate was then used to am-
plify FVIII expression. In the second step,
it was decided to co-express vWF in the
10A1 cells along with FVIII in order to sta-
bilize it. An expression plasmid containing
sequences for the full-length vWF cDNA
and an additional plasmid containing a se-
lectable adenosine deaminase (ADA)
marker were introduced into the 10A1 cell
line by protoplast fusion. Adenosine and
deoxycoformycin (dCF) were used to select
for expression of vWF in a serum- and
protein-free medium. Based on high levels
of expression of both FVIII and vWF, a
single clonal cell line was chosen for pro-
duction. This cell clone has not undergone
any genetic manipulations during its adap-
tation to a medium free of protein addi-
tives. Specifically, no genetic changes in
either the coding or promoter regions of
either FVIII or vWF have been observed
between the CHO cell lines used for rAHF
and those used for rAHF-PFM.
The resulting stable, homogeneous pop-
ulation of FVIII-expressing cells was used
to establish master (MCB) and working
(WCB) cell banks for processing the drug
substance [8, 12].
The development of the MCB and WCB
of the thoroughly characterized cells used
in the production involves extensive test-
ing for identity, growth and secretion char-
acteristics, and adventitious agents with
appropriate assays. Testing of the MCB for
adventitious agents included a battery of
tests for: sterility (bacteria/fungi), Myco-
plasma, lytic viruses; in vitro tests for
viruses (including bovine viruses); in vivo
tests for viruses; and an assay for retro-
viruses. Each WCB is tested for adventi-
tious agents on a routine basis, including
tests for sterility (bacteria/fungi), Mycoplas-
ma, lytic viruses, in vitro tests for viruses
(including bovine viruses) and in vivo tests
for viruses, and appropriate release specifi-
cations are set.
The genetic stability of the rFVIII se-
quences that have been integrated into the
CHO cells was assessed using a variety of
tests. First, the integrity of the Factor VIII
and vWF DNA sequences in the MCB and
the Post-Production Cell Bank were deter-
mined in order to verify that the coding
sequences remain stable throughout the
production process. Second, the structural
integrity of the expression cassettes was
examined using Southern blot methodolo-
gy. Third, real-time quantitative PCR was
used to determine the overall copy number
of the inserted genes, and to confirm the
genetic stability data from the Southern
blots. The results from these studies indi-
cate that roughly 25100 copies of the
rFVIII gene and a few copies of the vWF
gene are integrated into the CHO genome,
and that these sequences are stable
through at least 78 generations well be-
yond the production limit of 65 genera-
tions.
Following characterization of the cells,
the culture is expanded, separated, and
stored in frozen vials under optimal cell
bank conditions. Vials making up the
WCB provide the starting culture for each
production cycle.
3.3
Production of rFVIII
Using the large-scale production and quali-
ty control program developed for rAHF,
advanced cell culture, purification, and fi-
nal formulation techniques were developed
that optimize the production of rAHF-
PFM (see Table 3.3) [8].
3.3.1
Continuous (Chemostat) Perfusion Cell
Culture System [8]
To attain consistent and stable production
of FVIII from CHO cells adapted to cul-
ture in a protein-free medium, continuous
(chemostat) perfusion is used in the bio-
reactor culture system. An important ad-
vantage of the system is that the culture
conditions can be continuously controlled,
monitored, and optimized.
The cell culture process includes inocu-
lum build-up and continuous culture. For
initiation of inoculum build-up, a vial of
CHO cells from the WCB is thawed, and
the cells are grown in a proprietary, pro-
tein-free culture medium in a small flask.
Upon the attainment of a critical cell den-
sity, the cells are passaged to foster loga-
Table 3.3 Production and processing of rAHF-PFM [8]
Fermentation technology Continuous chemostat perfusion
Cell culture medium Protein-free medium
Purification Immunoaffinity chromatography
Monoclonal antibodies expressed in plasma/albu-
min-free conditions
Cation-exchange chromatography
Anion-exchange chromatography
Viral inactivation Solvent/detergent treatment
Stabilizer Trehalose
Bulking agent Mannitol
Reconstitution volume 5 mL
Potencies 250, 500, 1000, and 1500 IU per vial
rithmic growth. The cells are expanded
into an increasing number of roller bottles
and successively larger bioreactors. Before
transfer of cells into larger bioreactors, in-
process controls for microbial sterility and
cell culture ensure stable and safe condi-
tions throughout the culture process.
When the cells in the largest bioreactor
have reached an optimum density, contin-
uous (chemostat) perfusion culture is be-
gun. Fresh medium is added, and rFVIII-
containing conditioned medium is re-
moved at the same rate. This continuous
perfusion culture is maintained for several
weeks. Monitoring of culture conditions,
cell growth, and microbial sterility con-
tinues and provides consistency and stabil-
ity of cell density and expression of rAHF-
PFM over time. The conditioned medium
is filtered and rFVIII is then purified.
3.3.2
Purification [8]
The objectives of designing the purifica-
tion process are the following:
To achieve high specific activity and in-
tegrity of the drug substance protein.
To reduce to trace levels the impurities
from the cell culture (e.g. CHO cell pro-
tein) and from the purification processes
(e.g. murine monoclonal antibody).
To inactivate and remove any possible
(albeit highly improbable) viruses arising
from either the CHO cells used to pro-
duce FVIII, the hybridoma cells used to
produce anti-FVIII monoclonal anti-
bodies, or the cell culture media.
The techniques utilized for rAHF-PFM
are again based on purification methods
used for rAHF.
The rFVIII-containing cell culture medi-
um is first passed through a filter with a
relatively large pore size to remove the
CHO cells, before being sterile-filtered.
This is accompanied by in-process controls
for bacterial, fungal, and mycoplasmal
sterility.
Upon filtration of the continuous perfu-
sion harvest, rFVIII is purified in a step-
wise fashion, using multiple chromatogra-
phy columns that enlist distinct properties
of the rFVIII to remove impurities. The
cornerstone of the purification process is
an immunoaffinity chromatography col-
umn that employs the same monoclonal
antibody used in rAHF processing. During
this step, this monoclonal antibody specifi-
cally recognizes and selectively binds the
FVIII molecule. The vWF is separated
from the FVIII and, together with the im-
purities, is washed from the column. The
hybridoma cells that produce the monoclo-
nal antibody used in rAHF-PFM purifica-
tion have also been adapted to grow in a
cell culture medium devoid of human- or
animal-derived additives.
The immunoaffinity eluate is then sub-
jected to cation-exchange chromatography
prior to a dedicated viral inactivation step.
Cation-exchange chromatography takes ad-
vantage of the electrical charge differences
between FVIII and any impurities.
A dedicated viral inactivation step is an-
other new feature of rAHF-PFM process-
ing. Solvent/detergent treatment is a prov-
en viral inactivation method that preserves
the integrity of the structure and function
of the FVIII molecule.
Following the S/D inactivation step, the
eluate is further purified using anion-ex-
change chromatography prior to pooling,
freezing, and storage. Anion-exchange
chromatography also takes advantage of
the electrical charge differences between
FVIII and any remaining impurities.
As a result of these stringent processes,
the only proteins present in rAHF-PFM,
other than purified rFVIII, are trace quan-
tities of murine IgG, CHO cell proteins
and vWF. rAHF-PFM does not contain
therapeutic levels of vWF, and should not
be used to treat von Willebrand disease.
3.3.3
Formulation [8]
In order to develop a rFVIII concentrate
that does not contain any human- or ani-
mal-derived additives, formulation compo-
nents that mimic the stabilizing function
of albumin were identified. Each vial of
rAHF-PFM contains approximately 10 mM
histidine, 10 mM Tris, 90 mM sodium chlo-
ride, 0.010% (w/v) Tween-80, 3.2% (w/v)
mannitol, 0.8% (w/v) trehalose,
0.08 mg mL
1
reduced glutathione, and
1.7 mM calcium chloride. These non-pro-
tein stabilizers serve to increase the stability
of the molecule, preserve the integrity of the
lyophilized concentrate, and permit repro-
ducible crystallization during the freeze-dry-
ing process, and include a surfactant (poly-
Fig. 3.4 Basic rAHF-PFM processing. A rAHF-pro-
ducing CHO cell clone (expressing full-length
FVIII and vWF) has been adapted to grow in a
proprietary culture medium free of plasma, albu-
min, or any other human- or animal-derived addi-
tive. The cells are cultured in an increasing num-
ber of roller bottles and successively larger bior-
eactors. When the cells in the largest bioreactor
reach an optimum density, continuous (chemo-
stat) perfusion culture is begun. rFVIII from the
cell culture medium is purified using an immu-
noaffinity chromatography column that employs
the same monoclonal antibody used in rAHF pro-
cessing. However, the hybridoma cells that pro-
duce the monoclonal antibody used in rAHF-PFM
purification have been adapted to grow in a pro-
tein-free cell culture medium, similar to the CHO
cells. Next, the immunoaffinity eluate is further
purified using ion-exchange chromatography steps
and subjected to solvent/detergent (S/D) treat-
ment. Finally, rAHF-PFM is formulated without
any human- or animal-derived additives, and is
lyophilized after sterile filtration of the formulated
therapeutic and aseptic filling of vials [8].
Adaptation of rAHF CHO cell clone
to protein-fee medium
sorbate-80) to reduce the adsorption of
FVIII to processing surfaces and to prevent
high-molecular-weight aggregates forming
during freezing and lyophilization, and tre-
halose for maintaining the activity of FVIII
during lyophilization and storage.
Mannitol is employed as a bulking agent
to provide structural integrity to the lyophi-
lized concentrate; glutathione is used as a
reducing agent to protect against oxidation
and buffering agents; and Tris and histidine
are used to maintain a neutral pH.
After sterile filtration of the formulated
therapeutic and aseptic filling of vials, lyo-
philization is performed. Containing no
added preservatives, the finished product
is a sterile, non-pyrogenic, lyophilized
preparation of concentrated rFVIII for in-
travenous use. An overview of basic rAHF-
PFM processing is depicted in Fig. 3.4.
3.4
Pathogen Safety
For more than two decades, innovations in
hemophilia therapeutics have been driven
by concern for the safety of its pdFVIII
and rFVIII therapeutics. Since licensing,
more than 6 billion units of rAHF have
been supplied worldwide, with no con-
firmed reports of infectious disease trans-
mission to date. rAHF-PFM signifies an-
other technological advance in industrys
efforts to reduce risks for pathogen trans-
mission and to respond to the requests of
the hemophilia community. The absence
of additives of human or animal origin es-
sentially eliminates risk for transmission
of blood-borne pathogens that could arise
from these additives. Furthermore, all of
the cell banks and production cells for
preparation of rFVIII are rigorously tested
for viral or other contaminants. All steps
of the purification process, including im-
munoaffinity chromatography, ion-ex-
change chromatography, and S/D treat-
ment have been extensively validated for
their ability to clear potential viral con-
taminants [8].
3.4.1
Pathogen Safety: CHO and Hybridoma Cells
The pathogen safety for cell lines used in
the production of rAHF-PFM is discussed
below.
The CHO cell was chosen as the rAHF-
PFM host cell expression system for sev-
eral reasons, as previously described.
Among these reasons are the CHO cell
lines documented resistance to infection
by many human viruses and its ability to
grow in a plasma- and albumin-free cul-
ture [24]. Following adaptation to a pro-
tein-free medium, the cell line was sub-
jected to extensive testing for both known
and unknown adventitious viruses [8].
An essential purification measure for
the attenuation of infectious risk is immu-
noaffinity chromatography using a mono-
clonal antibody directed against FVIII. The
monoclonal antibody is expressed by hybri-
doma cells also adapted to growth in a
protein-free medium, and then subjected
to extensive testing for both known and
unknown viruses.
The cell line qualification programs for
the rFVIII-expressing CHO cells and the
Mab-expressing hybridoma cells have been
designed to provide the assurance of cell
line purity through all stages of develop-
ment and manufacturing.
The qualification programs involve ex-
tensive testing of the master, working, and
post-production cell banks, including tests
for Mycoplasma, sterility, and viruses [8].
These cell lines are subjected to an exten-
sive program of in vivo- and in vitro tests
for known and unknown viruses:
In-vivo tests for adventitious agents [8]:
Mouse antibody production test
(MAP): a broad screen test for known
viruses potentially present in test arti-
cle (both cell lines).
Hamster antibody production test
(HAP): a broad-screen test for known
viruses potentially present in test arti-
cle (for rAHF-PFM cell line only).
Rat antibody production test (RAP): a
broad-screen test for known viruses po-
tentially present in test article (for Mab-
expressing hybridoma cell line only).
Inoculation in animal test systems
(suckling mice, adult mice, guinea
pigs, and embryonated hens eggs):
generic tests for a wide range of ad-
ventitious agents (both cell lines).
In-vitro tests for adventitious agents [8]:
S
+
L
focus assay (xenotropic retro-

viruses).
XC plaque assay (ecotropic retro-
viruses).
Reverse transcriptase assay (infectious
retroviruses).
Human, primate, and murine cell lines.
3.4.2
Pathogen Safety: Purification of rAHF-PFM
In addition to the tests described above for
assuring the quality of the cells used in
production, the manufacturing process has
been validated for clearance of a range of
model viruses including:
Lipid-enveloped viruses:
Xenotropic leukemia virus (X-MuLV)
Bovine viral diarrhea virus (BVDV)
Porcine pseudorabies virus (PRV)
Non-lipid-enveloped viruses:
Mice minute virus (MMV)
Reovirus type 3 (REOV-3)
The purification scheme contributes sig-
nificantly to the margin of viral safety of
the product. Both the rAHF-PFM and the
Mab processes include a S/D treatment for
the inactivation of lipid-enveloped viruses,
as well as a variety of chromatographic
steps. The extent to which these steps are
capable of clearing model viruses was vali-
dated for both lipid-enveloped and non-lip-
id-enveloped virus removal and/or inactiva-
tion in accordance with the Committee for
Proprietary Medicinal Products (CPMP) of
the European Agency for the Evaluation of
Medicinal Products (EMEA), the United
States Food and Drug Administration
(FDA), and the International Conference
on Harmonisation (ICH) guidelines. A
panel of model viruses was selected as re-
presentative of potential infectious viruses
and the rAHF-PFM and Mab manufactur-
ing processes were validated for viral clear-
ance. Model virus reduction factors for the
Table 3.4 Model virus reduction for rAHF-PFM processing [8]
Model virus Overall viral reduction (Log
10
)
Lipid-enveloped
X-MuLV (xenotropic leukemia virus) >13.7 to >16.6
BVDV (bovine viral diarrhea virus) 7.0
PRV (porcine pseudorabies virus) 6.8
Non-lipid-enveloped
MMV (mice minute virus) 3.74.9
REOV-3 (reovirus type 3) >6.8 to >8.3
overall rAHF-PFM process is shown in Ta-
ble 3.4.
Steps in the processing of rAHF-PFM that
have a capacity for clearance of virus are im-
munoaffinity chromatography, cation-ex-
change chromatography, and S/D. X-MuLV,
a lipid-enveloped virus, is used as a model
for endogenous retroviruses, while BVDV
and PRV are representative of the Flaviviri-
dae and Herpesviridae families, respectively.
MMV and REOV-3 are included to validate
the robustness of the viral reduction pro-
cesses for non-lipid-enveloped viruses.
Specific virus reduction values achieved
with each individual viral removal/inactiva-
tion step are cumulative, so that the final
viral reduction value attained reflects mul-
tiple removal and/or inactivation proce-
dures occurring throughout the process.
In summary, the processing of rAHF-PFM
is associated with a significant viral reduc-
tion capacity for both lipid-enveloped and
non-lipid-enveloped viruses and provides
additional assurances of reliability with re-
spect to pathogen safety.
3.4.3
Pathogen Safety: Formulation [8]
The formulation of rAHF-PFM without any
human- or animal-derived additives pro-
vides further significant benefit to its overall
pathogen safety profile. A sugar-based for-
mulation of trehalose and mannitol is used
to replace human serum albumin to stabi-
lize the final formulation (see Section 3.3.3).
3.5
Quality Control
The efficacy and safety are monitored by
quality control (QC) protocols established
and implemented throughout product de-
velopment and processing [8].
3.5.1
Quality Control: Validation of CHO Cell Line
[8]
Stability of the FVIII and vWF genes was
verified in pre- and post-production cells
using several methods. No changes in the
coding or promoter regions of either gene
were observed, and structural integrity of
the genes was maintained throughout the
production process.
Process validation studies are performed
during bioreactor build-up and continuous
(chemostat) cell culture. CHO cell culture
conditions (e.g., viability, density, sterility)
are monitored carefully during each pro-
duction cycle. The parameters developed
for chemostat cell culture provide consis-
tency and stability of both cell density and
expression of rAHF-PFM.
3.5.2
Quality Control:
rAHF-PFM Final Drug Product [8]
These include:
Pathogen safety parameters (see Section
3.4).
Environmental monitoring.
Final product specifications.
3.6
Purity and Potency
rAHF-PFM has a specific activity [8] of
4000 to 10 000 IU mg
1
protein, resulting
from a very pure FVIII concentrate with
an extremely low residual level of extra-
neous protein. Biological activity (potency)
[8] is measured using a chromogenic sub-
strate assay.
3.7
Preclinical Studies
Prior to initiating a clinical development
program, extensive biochemical analyses
and preclinical evaluations assessing toxici-
ty/tolerability and hemostatic activity were
performed in animal models. The overall
preclinical strategy for rAHF-PFM was to
demonstrate the comparability with the
first-generation rFVIII, a drug with proven
hemostatic efficacy and low immunogenic-
ity [8].
3.7.1
Physico-chemical Properties [8]
Physico-chemical properties of rAHF-PFM,
including structure, identity, purity, po-
tency, and functional integrity, were de-
fined and compared to those of rAHF.
Conversion of FVIII to FVIIIa is re-
quired for functional activity. Samples of
both products (rAHF and rAHF-PFM)
were incubated with thrombin under con-
ditions leading to complete activation, and
the resulting heavy and light chain poly-
peptides were separated by RP-HPLC
(Fig. 3.5). In the resulting elution profiles,
the principal peaks represent the B do-
main; 73 kDa A3-C1-C2 light chain;
50 kDa A1 heavy chain fragment; and
Fig. 3.5 Reverse-phase high-performance liquid
chromatography (RP-HPLC) analysis of (A) rAHF
and (B) rAHF-PFM following thrombin digestion.
Samples of both products activated with thrombin
show similar elution profiles and peak areas. The
principal peaks represent the B domain; 73 kDa
A3-C1-C2 light chain; 50 kDa A1 heavy chain frag-
ment; and 43 kDa A2 heavy chain fragment. A 5-
kDa peptide degradation product derived from the
light chain is also noted [8].
43 kDa A2 heavy chain fragment. No sig-
nificant differences were observed upon
comparison of the elution profiles and
quantitation of peak areas.
The kinetics of thrombin activation was
assessed by sodium dodecyl sulfate (SDS)-
polyacrylamide gel electrophoresis (PAGE)
analysis. The resulting banding pattern is
complex and consistent with a mixture of
heterodimeric proteins (start lane) under-
going cleavage at multiple sites (Fig. 3.6).
The kinetics of the appearance and disap-
pearance of bands for both product gen-
erations are highly similar, suggesting
structural and functional equivalence.
Mass spectra analysis between both
product generations revealed no significant
differences in tyrosine sulfation an im-
portant post-translational modification
which is required for normal vWF binding
and thrombin activation.
The FVIII protein is extensively glycosy-
lated, particularly within the B domain. As-
sessment of overall glycosylation and sialic
acid content showed consistency. These
post-translational modifications impact
both function and metabolic fate of FVIII.
In addition, both products were sub-
jected to size-exclusion chromatography to
examine for the presence and extent of
protein aggregation, a process that can re-
sult in loss of product potency. FVIII ag-
gregates were not detected in either of the
two samples.
For comparative functional analyses in vi-
tro, both product generations were exam-
ined for the biomolecular interactions,
namely the binding to vWF, binding to Fac-
tor IXa as well as an aPTT-based coagula-
tion assay and a chromogenic substrate as-
say. By each of these measures, both prod-
ucts were found to be highly comparable.
The hemostatic efficacy of rAHF-PFM
and rAHF were compared in hemophilic,
exon-16 knockout mice using the tail
transection model. Bleeding rate and cu-
mulative blood loss after treatment, for
both preparations, were similar (Fig. 3.7).
Fig. 3.6 Kinetics of thrombin activation of rAHFand
rAHF-PFM SDS-polyacrylamide gel electrophoresis
(PAGE). Both products were incubated with a fixed
concentration of thrombin, and FVIII proteolysis was
visualized over 60 min. Both samples were rapidly
converted to FVIIIa at similar rates: the heavy
chain disappears with B domain cleavage and the
shortened light chain and A1 and A2 domains are
visible. HC = heavy chain; LC = light chain; A1 and
A2 = subunits of A domain [8].
The results of these efficacy studies in ani-
mals predict a comparable level of hemo-
static efficacy for both product generations
in humans.
3.7.2
Toxico-pharmacological Aspects
rAHF-PFM and its formulation vehicle
were shown to be relatively non-toxic in
acute- and repeat-dose toxicity studies and
tissue irritability studies. Irritation studies
involving intravenous or perivenous injec-
tions of formulation vehicle in rabbits pro-
duced no observable significant irritation
or microscopic inflammation. Additionally,
no acute toxicity was observed following
acute, single-dose toxicity studies testing
doses of up to 4750 IU kg
1
(rAHF-PFM)
in SpragueDawley rats and New Zealand
White rabbits. This dose was approximately
24-fold the maximum daily human dose.
Thirty-day repeat-dose toxicity studies
were carried out using the formulation ve-
hicle at repeated doses up to 40 mL kg
1
per day compared to normal saline in rats
and rabbits. No treatment-related effects
were observed in either species, showing
that the safety profile of the formulation
vehicle is comparable to that of saline.
All toxicity data closely resembled those
obtained with the first-generation product.
No adverse effects were observed in acute-
and repeat-dose toxicity studies conducted
in rats and rabbits, suggesting a similar
toxicology profile and predicting compar-
able levels of safety and tolerability in hu-
mans.
A pharmacokinetic comparison of both
product generations, and their formulation
buffers was conducted in partially heparin-
ized SpragueDawley rats using a mono-
clonal antibody capture/activity assay. The
two rFVIII concentrates exhibited a high
degree of similarity in half-life, area under
the curve (AUC), clearance rate, and mean
retention time, predicting biological activ-
ity of rAHF-PFM in humans similar to
that of rAHF.
The feasibility of rAHF-PFM delivery by
continuous infusion was demonstrated in
in vitro experiments simulating clinical ap-
plication. A CADD-1 infusion pump was
loaded with the reconstituted product at
30
o
C, in the presence and absence of hep-
arin (2 U mL
1
), at mid-potency (522 IU
per vial) and high-potency (1210 IU per
vial) doses at a 1.5 mL min
1
rate of infu-
sion. Controls were vials of the same lot
maintained at the same temperature, over
Fig. 3.7 Hemostatic efficacy of rAHF and rAHF-
PFM in a hemophilic mouse model. Comparable
hemostatic efficacy in hemophilic, exon-16 knock-
out mice was demonstrated. Mice received either
150 IU kg
1
body weight of either product, or their
respective formulation vehicles before tail transec-
tion. Cumulative blood loss assessed every 4 min
over a period of 20 min was similar; suggesting
that the hemostatic efficacy of both products was
highly comparable [8].
the same duration (Fig. 3.8). Stability and
recovery of at least 80% of baseline po-
tency was observed for up to 48 h after re-
constitution at room temperature in the
presence or absence of added heparin.
3.8
Clinical Studies
The completed clinical program will fea-
ture seven separate studies involving more
than 200 hemophilia A patients. These
clinical studies involve PTPs (adults, ado-
lescents, and children) and PUPs in Eu-
rope, North America, and Japan. As of
June 2003, the PTP Pivotal study has been
completed, four studies are ongoing, and
two are planned [8].
3.8.1
Pivotal study of rAHF-PFM in PTPs
Aged 10 Years
In the PTP Pivotal study, 111 patients aged
10 years with moderately severe to severe
hemophilia A (baseline FVIII 2%) and
prior treatment with FVIII concentrates
for at least 150 exposure days were en-
rolled at 23 sites in North America and
Europe; 108 patients received at least one
infusion of study medication [25, 27, 28].
3.8.1.1 Study Design
The primary aims of the study were to
demonstrate bioequivalence for rAHF-
PFM and rAHF, and to demonstrate safety
and efficacy of rAHF-PFM [8].
All patients were randomized to either
Parts 1 and 2 or Parts 2 and 3:
Part 1 (approximately 50% of patients):
double-blind, crossover, pharmacokinetic
comparison of rAHF and rAHF-PFM
produced at pilot-scale (rAHF-PFMpilot)
[8, 25, 27].
Part 2 (all patients): open label assess-
ment of safety, efficacy, and immuno-
genicity for a period of 75 rAHF-PFM
exposure days [28, 29].
Part 3 (approximately 50% of patients):
double-blind, crossover, pharmacokinetic
comparison of rAHF-PFM pilot and
rAHF-PFM produced at full scale in a
manufacturing facility (rAHF-PFM com-
mercial) [8, 2729].
Fig. 3.8 rAHF-PFM under simulated continuous
infusion conditions. The stability of rAHF-PFM by
continuous infusion delivery was demonstrated in
in vitro experiments simulating clinical application.
For experimental details, see text. Stability of
rAHF-PFM was observed for up to 48 h after re-
constitution at room temperature, as shown by re-
tention of at least 80% of baseline potency in the
presence or absence of added heparin (2 U mL
1
).
*Control = vials of the same lot of rAHF-PFM
maintained at the same temperature and for the
same duration [8].
3.8.1.2 Patient Demographics [8]
Age: 62 patients (56%) aged between 10
and 18 years; 49 (44%) aged >18 years.
Race: 103 (93%) Caucasians; seven (6%)
Blacks; one (1%) Asian.
Ethnicity: 10 (9%) of Hispanic origin.
Mean patient height (cm): 169.313.0
(range: 135 to 191).
Mean patient weight (kg): 65.816.7
(range: 32.2 to 108).
3.8.1.3 Pharmacokinetic Results
in Parts 1 and 3
In Part 1, the pharmacokinetic data ob-
served for rAHF and rAHF-PFMpilot were
highly similar (Table 3.5). Plasma FVIII
levels following infusion of either rAHF or
rAHF-PFMpilot were superimposable
(Fig. 3.9) [2527].
In Part 3 of the Pivotal study, pharmaco-
kinetic data with rAHF-PFM pilot and
rAHF-PFM commercial were also seen to
be highly similar (Table 3.6) [28].
Fig. 3.9 Plasma FVIII levels after infusion of rAHF-PFM and
rAHF. Patients (n = 30) from Part 1 of the PTP Pivotal study
were infused with 505 IU kg
1
rAHF-PFM and rAHF in a ran-
domized, crossover protocol. Results were highly similar and
nearly superimposable [25, 27].
Table 3.5 PTP Pivotal study Part 1, Pharmacokinetics: per-protocol analysis (n = 30) [26]
Parameter rAHF rAHF-PFM
MeanSD
AUC
048
(IUh dL
1
) 1515 (9702205) 1533 (8762642)
Adjusted recovery
a)
2.55 (1.473.89) 2.40 (1.543.88)
Half-life (h) 11.39 (7.8918.12) 11.98 (6.7424.70)
a) Adjusted recovery = IU dL
1
plasma per IU kg
1
body weight
infused.
3.8.1.4 Treatment of Bleeding Episodes:
Part 2
During the Pivotal study, 510 new bleed-
ing episodes in 83 PTPs were treated with
rAHF-PFM. The median dose per infusion
of rAHF-PFM for treatment of a bleeding
episode was 32.9 IU kg
1
(mean
38.1 IU kg
1
). The median dose per epi-
sode was 34.5 IU kg
1
. Overall, 93% of
bleeding episodes were resolved with one
or two infusions, and 81% were resolved
with only one infusion (Fig. 3.10). Hemo-
static efficacy was rated as excellent or
good in 86% of bleeding episodes. Some
45% of new bleeds were related to trauma,
32% were spontaneous, and 24% had an
undetermined etiology. The majority of
bleeding sites were joints (52%) and mus-
cles (33%). Similar results were observed
regardless of the etiology or the anatomic
site of the bleeding episode [8, 29].
Part 2 of the Pivotal study required a
rAHF-PFM prophylaxis regimen of 25
40 IU kg
1
administered three to four
times per week, for at least 75 exposure
days. (Doses >40 IU kg
1
were permitted
for specific situations, such as a Friday in-
fusion when increased physical activity
was expected over a week-end.) For all
treated patients during the first 75 expo-
sure days, a mean rate of 6.2 new bleeding
episodes per patient per year was ob-
served. However, a difference in the rate
of bleeding episodes became apparent
when the data were analyzed relative to
the patients adherence to the treatment
regimen. Adherence was defined as infus-
ing 25 IU kg
1
for at least 80% of the in-
fusions, at a frequency 3 times per week,
for at least 80% of the weeks on the study.
Adherent patients had an average of 4.3
bleeding episodes per patient per year,
whereas less adherent patients had an
average of 9.8 bleeding episodes per pa-
tient per year [29].
3.8.1.5 Safety
No serious adverse events related to rAHF-
PFM were reported in the Pivotal study. In
total, 19 non-serious adverse events in sev-
en patients were judged to be product-re-
lated; these included taste perversion,
headache, fever, diarrhea, dizziness, hot
flashes, pain in upper abdomen, pain in
lower chest, shortness of breath, sweating,
nausea, rigors, and itching in the arm
used for the infusion [8]. Five non-serious
drug-related events were mild, 12 were
moderate, and two (one high fever and
one severe headache, reported concur-
rently in one patient) were severe. No pa-
tient withdrew from the study because of
any study-drug-related adverse event [8, 25,
2729].
Table 3.6 PTP Pivotal study Part 3, Pharmacokinetics: per-protocol analysis (n = 37) [28]
Parameter rAHF-PFM
pilot
rAHF-PFM
commercial
MeanSD
AUC
048
(IUh dL
1
) 1544 (8562216) 1494 (7672392)
Adjusted recovery
a)
2.55 (1.734.05) 2.46 (1.713.41)
Half-life (h) 11.60 (7.5915.03) 11.72 (8.1417.34)
a) Adjusted recovery = IU dL
1
plasma per IU kg
1
body weight infused.
3.8.1.6 Inhibitor Development
The 108 treated patients in the PTP Pivo-
tal study accrued a median of 117 expo-
sure days. During the entire study, only
one patient tested positive for a low titer
(2.0 BU) inhibitor to FVIII following 26 ex-
posure days to rAHF-PFM. This patient, a
55-year-old man with severe hemophilia A,
displayed no symptomatic evidence of an
inhibitor. The inhibitor was undetectable
in a blood sample collected and tested 8
weeks later [8, 29].
3.8.2
Continuation Study in PTPs
In order to assess the long-term safety and
efficacy of rAHF-PFM, 82 patients who
completed the PTP Pivotal study have
been enrolled in an ongoing PTP Conti-
nuation study being conducted at 15 sites
in the US and Europe [8, 27].
3.8.2.1 Study Design [8, 27]
Part 1: open-label assessment of phar-
macokinetic parameters of rAHF-PFM
in patients who completed Parts 1 and 2
of the PTP Pivotal study.
Part 2: open-label evaluation of long-
term safety, hemostatic efficacy, and im-
munogenicity of rAHF-PFM following a
period 50 exposure days (all patients).
Pharmacokinetic data from Part 1 of the
PTP Continuation study were compared
with data for the same patients from
Part 1 of the PTP Pivotal study to allow
a comparison of the pharmacokinetics of
rAHF-PFM before and after a period
75 exposure days.
Fig. 3.10 rAHF-PFM hemostatic efficacy in PTPs. During the
PTP Pivotal study, 510 new bleeding episodes were reported.
Depicted here is the percentage of new bleeding episodes re-
solved according to the number of infusions received [29].
The data described below reflect an in-
terim analysis that includes 33 patients
who have received at least one infusion of
rAHF-PFM on this study.
3.8.2.2 Patient Demographics [8, 27]
Age: 19 (58%) were aged 1018 years; 14
(42%) were aged >18 years.
Race: 32 (97%) Caucasians, one (3%)
Black.
Ethnicity: 4 (12.1%) of Hispanic origin.
(range: 150185).
(range: 40.8105).
3.8.2.3 Treatment Regimens [8, 27]
Twenty-three patients have received a pro-
phylactic regimen consisting of 25
40 IU kg
1
administered three to four
times per week; four have followed an in-
vestigator-modified prophylactic regimen.
No patient included in the interim analysis
has followed an on-demand regimen.
3.8.2.4 Pharmacokinetics [8, 27]
For the 13 patients included in Part 1, the
statistical comparison of pharmacokinetic
parameters before and after a period of
75 exposure days with rAHF-PFM indi-
cates no difference in AUC
048 h
and ad-
justed recovery. Other pharmacokinetic pa-
rameters also appear comparable.
3.8.2.5 Hemostatic Efficacy [8, 27]
In the interim analysis, 13 of 27 patients
enrolled in Part 2 had a total of 51 bleed-
ing episodes, 49 of which occurred in pa-
tients receiving the standard prophylactic
regimen. The majority of the 49 episodes
(86%) required only one infusion, and
only three episodes (6%) have required
more than three infusions. Among the 49
episodes, response in 32 (65%) was rated
as excellent or good, and in 16 (33%) was
rated as fair. In one episode (2%), there
was no response. The bleeding episode
that was rated as having no response oc-
curred in a patient experiencing an elbow
bleed. The patient infused one dose of
rAHF-PFM (53.9 IU kg
1
) and received no
other FVIII replacement therapy; an an-
algesic was taken for elbow pain. No in-
hibitor was detected. Despite this efficacy
rating, the patient remained on the study.
3.8.2.6 Safety [8, 27]
Data from the interim analysis showed no
study-drug-related serious or non-serious
adverse events on the PTP Continuation
study. In a follow-up safety assessment
across all rAHF-PFM studies, one non-ser-
ious adverse event was reported in one pa-
tient on the PTP Continuation study. This
event (strange taste in the mouth) was
deemed possibly related to rAHF-PFM,
and was completely resolved.
3.8.2.7 Inhibitors [8, 27]
No inhibitors to FVIII have been detected
in patients included in the interim analy-
sis. A follow-up safety assessment across
all rAHF-PFM studies provided additional
data on inhibitor risk for 50 patients who
had accumulated at least 50 exposure days
to rAHF-PFM. No inhibitors were detected
in this expanded cohort from the PTP
Continuation study.
The results of the continuation study
with rAHF-PFM support the pivotal study
conclusion that rAHF-PFM is safe, non-
neoantigenic, and efficacious in the treat-
ment of hemophilia A [8].
3.8.3
Surgical Prophylaxis and Perioperative
Management of Hemostasis Study in PTPs
This is an ongoing open-label study of pa-
tients aged 5 years, with moderately se-
vere to severe hemophilia A and a history
of 150 FVIII exposure days, undergoing
major or minor surgical, dental, or other
invasive procedure (Tables 3.7 and 3.8) [8,
30, 31].
The interim data presented below reflect
a total of 42 patients enrolled, and 44 pro-
cedures [30, 31].
3.8.3.1 Patient Demographics [30]
Age: 1 (2.4%) patient aged 69 years; 11
(26.2%) aged 1018 years; and 30 (71.4%)
aged >18 years.
Race: 39 (92.8%) Caucasian; 2 (4.8%)
Black; 1 (2.4%) Asian.
Ethnicity: 2 (4.8%) of Hispanic origin.
(range: 138194).
(range: 23.8170.1).
3.8.3.2 Overview of Procedures [30]
Eight procedures were dental, and 26 were
orthopedic (17 of the latter were classified
as major). Seventeen procedures were per-
formed using continuous infusion (Table
3.8). Major procedures included total joint
replacements; total prostheses; and endo-
prostheses and arthrodeses of the hip,
knee, ankle, or shoulder.
3.8.3.3 Perioperative Hemostatic Efficacy
[8, 30, 31]
The estimate of actual intra-operative
blood loss was compared to estimated
blood loss (EBL) for 40 of the 45 planned
procedures (one procedure was excluded
because the surgery was not performed;
four procedures were missing a prediction
of blood loss). Actual blood loss was less
than the average EBL for the related proce-
Table 3.7 rAHF-PFM Surgical Study: prophylaxis and perioperative dosing and monitoring [8, 30, 31]
Treatment period Dosing and monitoring
Preoperative rAHF-PFM loading dose with FVIII post-infusion plasma level targets of
60100% for dental and 80120% for other procedures
Intra-operative Bolus or continuous infusion
a, b)
Surgeons efficacy assessment
Postoperative Bolus or continuous infusion
a, b)
Daily FVIII levels and clinical laboratory tests
Hematologists efficacy assessment at discharge
Additional surgical efficacy assessment at time of drain removal(s)
Home replacement/
rehabilitation infusions
Non-orthopedic procedures: up to 2 weeks
Orthopedic procedures: up to 6 weeks (including postoperative rehabilitation)
Efficacy assessment for new bleeding episodes
a) Continuous infusion is currently not part of the European license.
b) Continuous infusion administered at an initial dose level of
4 IU kg
1
h
1
for patients aged >12 years, and 5 IU kg
1
h
1
for those aged 512 years.
dure in 20 of 40 procedures, within pre-
dicted average and maximal EBL ranges
for 18 of 40, and greater than the maximal
EBL for two. Of these two, one had an ac-
tual loss of 2900 mL (predicted average
and maximal EBLs of 1000 mL and
2000 mL, respectively, for a total hip joint
replacement) and the other had an actual
loss of 100 mL (predicted average and
maximal EBLs of 15 mL and 50 mL, re-
spectively, for removal of a port). A sum-
mary of the hemostatic efficacy ratings for
all procedures during the intra-operative
and postoperative periods and at the time
of drain removal is provided in Table 3.9.
3.8.3.4 Safety and Immunogenicity [30]
No serious adverse events reported to
date have been deemed related to
rAHF-PFM.
Of the non-serious adverse events re-
ported, seven have been deemed related
to rAHF-PFM (four moderate, one mild,
and two severe events which consisted
of decreased coagulation FVIII level and
unspecified hematoma).
The decreased coagulation FVIII level
occurred postoperatively in a patient re-
ceiving rAHF-PFM via continuous infu-
sion for placement of total knee prosthe-
sis. The clearance rate increased 10 days
Table 3.8 PTP Surgical Study: enrollment by surgery and infusion types [30]
Type of surgery
a)
Infusion method Total no. of procedures
Bolus Continuous
b)
Major 14 4 18
Minor 5 13 18
Dental 8 0 8
Total 27 17 44
c)
a) Major surgery defined as having moderate to critical risk and/or
projected blood loss 500 mL; minor surgery characterized as
minimal to mild risk and/or <500 mL blood loss.
b) Any procedure that included any treatment by continuous infusion.
c) One of the 42 patients enrolled did not undergo the planned surgery;
three patients had two procedures each.
Table 3.9 PTP Surgical Study: assessments of perioperative efficacy [8, 30]
Hemostatic efficacy Intra-operative
n (%)
Postoperative
n (%)
Drain removal
n (%)
Excellent 22 (50) 32 (73) 6 (32)
Good 21 (48)
a)
12 (37) 10 (53)
Fair 0 0 2 (10)
b)
Unknown 1 (2) 0 1 (5)
c)
a) According to study criteria, two of these procedures should
have been rated by the surgeon(s) as fair.
b) Both of these were major orthopedic procedures: total hip
joint replacement and total endoprosthesis of the right knee.
c) Source documentation lost at site.
postoperatively, and the dosage of rAHF-
PFM was increased with administration
of supplemental boluses. At the time of
discharge, the investigator assessed the
hemostatic efficacy as good, and the re-
lated adverse event as having resolved
completely. An inhibitor to FVIII was
not detected in this patient.
A hematoma was reported in one pa-
tient who underwent placement of a to-
tal endoprosthesis in the right knee.
After three bolus infusions of rAHF-
PFM, the adverse event resolved com-
pletely. The global assessment of efficacy
was good at time of discharge, and no
further adverse events were reported in
this patient.
No inhibitors have been reported.
3.8.4
Pediatric Study in PTPs
The PTP Pediatric study is an ongoing
study evaluating the pharmacokinetics,
safety, immunogenicity, and efficacy of
rAHF-PFM in patients aged less than 6
years. The study has an enrollment goal of
at least 50 patients [32].
3.8.4.1 Study Design [32]
Part 1: evaluation of pharmacokinetic
and safety parameters following a single
dose (505 IU kg
1
) of rAHF-PFM.
Part 2: evaluation of immunogenicity,
hemostatic efficacy, in vivo recovery, and
safety of rAHF-PFM in conjunction with
on demand or prophylactic therapy regi-
mens for a period 50 exposure days.
Interim data for Part 1 of the study are
available for 14 subjects.
3.8.4.2 Patient Demographics [8, 32]
Age: 5 (35.7%) aged <3 years; none
(0%) aged 3 years; 6 (42.9%) aged 4
years; 3 (21.4%) aged 5 years.
Race: 13 (93%) Caucasian; 1 (7%) Black.
Ethnicity: 1 (7%) of Hispanic origin.
(range: 76121).
(range: 10.827.2).
These 14 patients have each been ex-
posed to a single infusion of 505 IU kg
1
rAHF-PFM; the total dose per patient
ranged from 540 to 1380 IU.
3.8.4.3 Pharmacokinetics [32]
For the per-protocol analysis population of
11 patients, the mean plasma half-life was
10.481.6 h, and adjusted recovery
2.00.5 IU dL
1
plasma per IU kg
1
body
weight infused. These data are consistent
with those seen in adolescents and older
patients, when the greater weight-adjusted
plasma volumes that affect the volume of
distribution and adjusted recovery in
young children, as well as potentially alter
half-life, are taken into account [32].
3.8.4.4 Safety and Immunogenicity [32]
As of the interim analysis, there have been
no adverse events reported, nor have there
been any apparent trends in changes in vi-
tal signs. rAHF-PFM appears to be safe
and well tolerated.
Inhibitor development will be assessed
in Part 2 of this study.
3.8.4.5 Conclusions [32]
Pharmacokinetic data are acknowledged as
key surrogates for FVIII concentrate effi-
cacy; thus, the apparent comparability of
pharmacokinetic results between patients
aged <6 years and those aged 10 years
suggests that rAHF-PFM will be effective
in the younger population.
3.8.5
Other rAHF-PFM Studies
PUP Study: this planned study is de-
signed to determine the pharmacoki-
netics, efficacy, immunogenicity, and
safety of rAHF-PFM in a minimum of
50 previously untreated patients, aged
<6 years, with moderately severe to se-
vere hemophilia A (baseline FVIII activ-
ity of 2%).
3.9
Summary
rAHF-PFM is a recombinant hemophilic
Factor VIII which is prepared by a plasma-
and albumin-free method. The newly de-
veloped process does not employ any hu-
man- or animal-derived additives in the
cell culture, purification or formulation of
the final product. This virtually eliminates
any risk of transmission of human blood-
borne viruses or other adventitious agents
that could, in theory, be introduced by the
use of animal or human proteins.
In addition, large viral safety margins
have been demonstrated in preclinical vali-
dation studies of viral inactivation and par-
titioning steps integrated into the manu-
facturing process. The preclinical and clin-
ical development strategy was to demon-
strate the comparability of rAHF-PFM
with its well-established predecessor prod-
uct rAHF. The drug substance has been
extensively characterized by in vitro and in
vivo animal studies, and has shown bio-
equivalence of the two product genera-
tions.
In clinical studies, plasma levels and
pharmacokinetics were investigated and
showed compliance with plasma-derived or
other recombinant products.
The data generated document that the
achieved amount of FVIII and the mainte-
nance of plasma levels is as expected.
In terms of efficacy, data such as the
number of infusions needed and treat-
ment outcome rAHF-PFM were effective
in preventing and controlling bleedings in
patients with severe hemophilia A. In
safety and immunogenicity studies, no ser-
ious adverse events reported to date have
been deemed related to rAHF-PFM. The
incidence rates for the most common non-
serious adverse events indicate that drug-
related adverse events were similar or even
lower in number and in nature to those
observed in clinical trials with other rFVIII
concentrates. To date, no inhibitors have
been reported. The first rAHF, Recombi-
nate
TM
, was developed by Baxter and be-
came commercially available in 1992. Its
long record of efficacy and safety has
made Recombinate
TM
the standard for
care in hemophilia A therapy [911].
rAHF-PFM, Advate
TM
, received market
authorization in 2003 in the US and in
2004 in Europe.
Clinical trials are ongoing for differ-
enced evaluation of the continuous infu-
sion model in surgery, to broaden the data-
base on safety and immunogenicity, and to
generate more evidence on pharmacoki-
netics, efficacy, safety and immunogenicity
in children previously treated and un-
treated with FVIII concentrates.
By following the needs of patients, sub-
stantial research and technical innovation
has led to an advanced category of recom-
binant FVIII concentrate which will estab-
lish a new standard for safety in FVIII re-
placement therapy.
3.9 Summary 447
References
1 B. Arun, C. M. Kessler, Clinical manifestations
and therapy of the hemophilias. In: Colman
R. W., Clowes A. W., Hirsh J, George J. N.,
Marder V. J., eds. Hemostasis and Thrombosis:
Basic Principles and Clinical Practice, 4th edn.
Philadelphia, Pa: Lippincott, Williams & Wil-
kins; 2000:815824.
2 G. C. White II, F. Rosendaal, L. M. Aledort,
J. M. Lusher, C. Rothschild, J. Ingerslev, for
the Factor VIII and Factor IX Subcommittee.
Definitions in hemophilia: recommendation
of the Scientific Subcommittee on Factor VIII
and Factor IX of the Scientific and Standardi-
zation Committee of the International Society
on Thrombosis and Hemostasis. Thromb He-
most 2001;85:560.
3 M. W. Hilgartner. Current treatment of hemo-
philic arthropathy. Curr Opin Pediatr.
2002;14:4649.
4 US Department of Health and Human Ser-
vices. Report on the Universal Data Collection
Program (UDC). Atlanta, GA: Hematologic
Diseases Branch, Centers for Disease Control
and Prevention; 2002:4.
5 M. W. Hilgartner. The need for recombinant
factor VIII: historical background and ratio-
nale. Semin Hematol 1991;28(suppl 1):69.
6 P. M. Mannucci, E. G. Tuddenham. The hemo-
philias from royal genes to gene therapy. N
Engl J Med 2001;344:17731779.
7 W. K. Hoots. History of plasma-product safety.
Transfus Med Rev 2001;15(suppl 1):310.
8 Data on file, Baxter Healthcare Corporation.
9 G. C. White II, S. Courter, G. L. Bray, M. Lee,
E. Gomperts, and the Recombinate Previously
Treated Patient Study Group. A multicenter
study of recombinant factor VIII (Recombina-
te
TM
) in previously treated patients with he-
mophilia A. Thromb Hemost 1997;77:660667.
10 G. L. Bray, E. Gomperts, S. Carter, et al., for
the Recombinate Study Group. A multicenter
study of recombinant factor VIII (Recombi-
nate): safety, efficacy, and inhibitor risk in pre-
viously untreated subjects with hemophilia A.
Blood 1994;83:24282435.
11 R. Gruppo, G. L. Bray, P. Schroth, et al. Safety
and immunogenicity of recombinant factor
VIII (Recombinate
TM
) in previously untreated
subjects (PUPs): a 6.5 year update. Thromb
Hemost 1997; (suppl):162. Abstract PD-663.
12 A. Mitterer, W. Mundt, F. Dorner, E. Bjornson.
Development of an advanced category recom-
binant Factor VIII, antihemophilic factor (re-
combinant) plasma/albumin-free method
(rAHF-PRM). Poster presented at ISTH, Bir-
mingham, July 1218, 2003.
13 P. Brown. Transfusion medicine and spongi-
form encephalopathy. Transfusion 2001;41:433
436.
14 P. Brown. Bovine spongiform encephalopathy
and variant Creutzfeldt-Jakob disease. Br Med
J 2001;322:841844.
15 United Kingdom Hemophilia Centre Doctors
Organisation (UKHCDO). Guidelines on the
selection and use of therapeutic products to
treat hemophilia and other hereditary bleed-
ing disorders. Hemophilia 2003;9:123.
16 National Hemophilia Foundation. MASAC
recommendations concerning the treatment
of hemophilia and other bleeding disorders.
MASAC Recommendations. Revised November
2002; No. 135.
17 P. J. Lenting, J. A. van Mourik, K. Mertens.
The life cycle of coagulation factor VIII in
view of its structure and function. Blood
1998;92:39833996.
18 R. J. Kaufman, S. E. Antonarakis, P. J. Fay. Fac-
tor VIII and hemophilia A. In: Colman R. W.,
Clowes A. W., Hirsh J., George J. N., Marder
V. J., eds. Hemostasis and Thrombosis: Basic
Principles and Clinical Practice, 4th edn. Phila-
delphia, PA: Lippincott, Williams & Wilkins;
2000:135156.
19 R. A. Gruppo, D. Brown, M. M. Wiles, R. J. Na-
vickis. Comparative effectiveness of full-length
and B-domain deleted factor VIII for prophy-
laxis a meta-analysis. Hemophilia
2003;9:251260.
20 ReFacto
Antihemophilic factor (Recombi-

nant) Prescribing Information. Wyeth Phar-
maceuticals.
21 B. Zhang, M. A. Cunningham, W. C. Nichols,
et al. Bleeding due to disruption of a cargo-
specific ER-to-Golgi transport complex. Nat.
Genet. 2003;34(2):220225.
22 R. J. Kaufman, S. W. Pipe. Regulation of factor
VIII expression and activity by von Willebrand
factor. Thromb Hemost 1999;82(2):201208.
23 R. J. Kaufman, L. C. Wasley, M. V. Davies, R. J.
Wise, D. I. Israel, A. J. Dorner. Effect of von
Willebrand factor coexpression on the synthe-
sis and secretion of factor VIII in Chinese
hamster ovary cells. Mol Cell Biol
1989;9(3):12331242.
24 E. Gomperts, R. Lundblad, R. Adamson. The
manufacturing process of recombinant factor
VIII, Recombinate. Transfus Med Rev
1992;6:247251.
25 M. Tarantino, I. Lopez, L. Navale, G. L. Bray.
Pharmacokinetic evaluation of a new protein-
free manufactured recombinant FVIII, rAHF-
PFM. Poster presented at WFH, Seville, 2002.
26 rAHF-PFM European Public Assessment Re-
port (EPAR) scientific discussion; http://
www.emea.eu.int/index/indexh1.htm.
27 A. Shapiro, M. Tarantino, I. Warrier, et al. On-
going Evaluation of an Advanced Category Re-
combinant FVIII Processed Using a Plasma/
Albumin-Free Method, rAHF-PFM. Poster
presented at National Hemophilia Foundation,
Salt Lake City, November 68, 2003.
28 M. D. Tarantino, L. M. Navale, G. L. Bray, B. M.
Ewenstein. Clinical evaluation of a new gen-
eration recombinant FVIII, plasma/albumin
free method (rAHF-PFM). Blood
2002;100(suppl):493.
29 M. D. Tarantino, L. Navale, G. Bray, B. Ewen-
stein. Clinical evaluation of a new generation
recombinant FVIII, plasma/albuminfree
method (rAHF-PFM), in previously treated pa-
tients. Poster presented at: Annual Meeting of
the American Society of Hemophilia; Decem-
ber 610, 2002; Philadelphia, PA.
30 J. Astermark, C. Negrier, P. Schroth et al.
Clinical evaluation of an advanced category
recombinant FVIII, antihemophilic factor
(recombinant) plasma/albumin-free method
(rAHF-PFM) in surgical settings. Poster pre-
sented at ISTH, Birmingham, July 1218,
2003.
31 C. Negrier, J. Astermark, I. Pabinger, J. Di
Paola et al. Surgical Evaluation of rAHF-PFM,
an Advanced Category Recombinant Antihe-
mophilic Factor Prepared Using a Plasma/Al-
bumin-Free Method. Blood 2003;102(11):Ab-
stract 2944.
32 K. Hoots, V. Blanchette, A. Shapiro, L. Navale
et al. Clinical evaluation of an advanced cate-
gory recombinant FVIII, antihemophilic factor
(recombinant) plasma/albumin-free method
(rAHF-PFM) in pediatric previously treated
patients. Poster presented at ISTH, Birming-
ham, July 1218, 2003.
References 449
Abstract
An investigation of human-use drugs ap-
proved world-wide since 1981 shows that a
substantial proportion fall into the catego-
ry of biological agents. On further subdivi-
sion, a significant number of these are in
fact compounds that either are natural
products (small- to medium-sized polypep-
tides) or are derivatives of such natural
materials that have been produced via re-
combinant/biotechnologies and have been
expressed via fermentative processes in
both prokaryotic and eukaryotic systems.
In addition, in the realm of cancer and po-
tential anti-HIV compounds, a number of
agents in preclinical studies are based
upon natural products that are either poly-
saccharides, glycopeptides, or peptides, all
of which are produced at some stage dur-
ing their synthesis via genetic means.
These include significant numbers of po-
tential agents in diseases other than those
of man that should be considered for bio-
technological intervention in the widest
sense. This chapter will not be exhaustive,
but will cover selected areas showing the
multiplicity of materials that have been
commercialized (Section 4.1) or may po-
tentially be obtained (Section 4.2) by such
processes, thus extending the definition of
biopharmaceuticals beyond conventional
usage.
Abbreviations
BHK baby hamster kidney cells
CSF colony-stimulating factor
CV-N cyanovirin-N
EGFR endothelial growth factor receptor
EPO erythropoietin
factor
hCG human chorionic gonadotropin
HSV herpes simplex virus
IFN interferon
IL-2 interleukin-2
LH luteinizing hormone
451
ISBN: 3-527-31184-X
4
Biopharmaceutical Drugs from Natural Sources
Errare Humanum Est What Causes Cancer
and How to Selectively Fight Tumors
LHRH luteinizing hormone-releasing
hormone
MAbs monoclonal antibodies
NCE New Chemical Entity
RIP ribosome-inactivating protein
RSV respiratory syncytial virus
t-PA tissue plasminogen activator
4.1
Biotechnologically Produced Proteins
and Peptides as Approved Drugs
4.1.1
Introduction
Over the past 20 years, from about the
time that recombinant insulin was intro-
duced by Lilly (as Humulin
), the number
and types of biopharmaceuticals have in-
creased on an almost yearly basis when
counted as materials that have been ap-
proved by the regulatory authorities in the
USA, Canada and the European Union
(EU). In fact, if one looks at the website of
the US biopharmaceutical trade group
PhRMA (http://www.pharma.org), then
close to 400 biotech medicines are cur-
rently undergoing trials in the USA, with
the majority (close to 50%) being directed
against cancer in one or more of its many
manifestations. Likewise, a significant
number are directed towards infectious
disease, autoimmune disease, and HIV. It
should be pointed out that a significant
proportion of all materials in the areas of
cancer and infectious disease are vaccines
of one type or another, and that the second
largest class of biopharmaceuticals under
development are monoclonal antibodies
(MAbs), with the largest numbers being
directed towards cancer and autoimmune
diseases.
In this chapter, we will not cover vac-
cines of any type but rather will place em-
phasis on the use of proteins and peptides
as pharmaceutical agents in their own
right, in the modifications that can be
made to such agents in order to increase
their potential utility. We will, however, in-
clude data on selected antibodies particu-
larly in the area of cancer and autoim-
mune diseases, where some very recent
agents have begun to show how biotech-
nology can materially aid in the derivation
of new pharmaceutical agents. We will
also show some of the modifications that
have been made to (relatively) small mole-
cules either directly from, or related to
products from natural sources that have
permitted them to be used in spite of sig-
nificant negative physico-chemical and/or
pharmacologic properties in their native
state. Finally, we will discuss the potential
for the use of peptidic compounds that are
derived from other than human sources as
either leads to novel agents or as agents in
their own right.
4.1.2
Why Biotechnology/Natural Products
as a Route to Novel Treatments?
If one looks at the sources of drugs ap-
proved in the time period 19812002 (the
latest time for which there is a compila-
tion of all drugs), then of the 1031 New
Chemical Entities (NCEs) identified by
Newman et al. [1], which was not exhaus-
tive in dealing with biologicals, 125 were
classified as biologicals (or B) and 29
as vaccines (or V). There was also a
small number of peptidic agents that fell
into the categories of natural products
(N) or derived from a natural product
(ND), where were generally 40 amino
acid residues or less, usually made by syn-
thesis. What is significant however, is that
aside from the Lilly Pharmaceutical Com-
pany and the vaccine divisions of then ex-
isting pharmaceutical companies (a large
number of whom have now combined), al-
most all of the other agents were discov-
ered by and in some cases, developed by
what now would be classified as biotech-
nology (also known as biopharmaceutical)
companies.
There are some fairly recent compila-
tions of biotechnology-derived drugs, in
particular those by Walsh from the Univer-
sity of Limerick in Ireland, in two reports
in 2000 and 2003 [2, 3], followed by a re-
cent update in 2004 [4] (see also the Intro-
duction of this book).
In 2002, Walsh [5] defined a biopharma-
ceutical as follows: A biopharmaceutical is
a protein or nucleic acid based pharmaceu-
tical substance used for therapeutic or in
vivo diagnostic purposes, which is pro-
duced by means other than direct extrac-
tion from a native (non-engineered) biolog-
ical source.
He contrasted this definition with the
term biotechnological medicine which
could include semisynthetic compounds
originally obtained from nature (paclitaxel
being a good example) or antibiotics iso-
lated from fermentation broths.
How then would one define the antibiot-
ic daptomycin which was originally pro-
duced by fermentation of the native organ-
ism, and which has been successfully ob-
tained from a genetically modified host
organism? Likewise, the production of
epothilone C and D by modification of the
epothilone A and B produced by deletion
of the cytochrome P
450
gene responsible
for the epoxidation? In both cases, the
compounds required were then produced
by transfer of the gene complex into an-
other organism for larger-scale fermenta-
tion.
Definitions therefore are relatively fluid
phrases, depending upon the individual
and the context, as exemplified by a recent
review on biologic pharmaceuticals [6].
4.1.3
Biopharmaceutical Drugs
(Defined in a Broader Sense)
In the following tables, we have listed
those agents that we can confirm as being
used in commerce, together with suitable
comments as to their provenance. We have
included (where available) the other names
under which the base agent has been com-
mercialized, but have usually not listed
slightly different mixtures of agents where
the components are similar to the original,
the differences usually being designed to
extend patent lives. The information has
been complied from a variety of sources
but, with two exceptions, every compound
has been checked by use of the Prous In-
tegrity
database. Hence, there will be dis-

crepancies between these data and other
compilations.
We have elected to use very broad-
brush definitions as to the Base Class of
agent types, in order to avoid large num-
bers of definitions. Thus, the 159 agents
listed fall into the following categories. An-
tibodies (mono- and polyclonal) (see Part
V, Chapters 1 and 2); anticoagulants (both
modified heparins and protein C); antihe-
mophilics (mainly Factors VII and VIII)
(see Part II, Chapters 1 and 3 and Part III,
Chapter 7); antithrombotics (mainly tissue
plasminogen activators and hirudins); en-
zymes (glycosidases and urate oxidases)
(see Part II, Chapter 6); hematopoietic
agents (EPO, GM-CSF, etc.) (see Part VIII,
Chapter 3); insulins (see Part IV, Chapter
13, and Part VI, Chapter 4); interferons;
peptides and proteins (other than those
listed in other categories).
Each table follows a general format, with
country of first approval/year; the generic
name; the tradename first approved; other
tradenames; source and methods of prepa-
ration where obtainable; and the disease
entity first approved. In order to make the
tables simpler to use, the compounds are
arranged in alphabetical order of their first
tradename and, if desired, a searchable
Excel spreadsheet is available from the
authors.
Following each table, a short discussion
of certain compounds in the particular
table will be presented.
4.1.4
Antibodies (Table 4.1)
These comprise 21 agents, and range from
a simple murine MAb directed against the
CD3 antigen of T cells (Orthoclone
OKT3
) from 1986 used as an immuno-

suppressant, through humanized MAbs
such as Synagis
in 1998 directed against

a protein in the coat of human respiratory
syncytial virus (RSV), to the very recent ex-
amples in the anti-cancer field of Erbi-
tux
, the multiple treatment system

known under the name of Zevalin
, Bex-
xar
and, in particular, Avastin
, the first
anti-vascular endothelial growth factor
(VEGF) agent approved, to Mylotarg
, in
which a very potent anti-tumor antibiotic
was covalently attached to an antiCD33
MAb.
In addition, three other very important
firsts in this field were the recent ap-
provals of Amevive
and Raptiva
for the
treatment of psoriasis, and of Xolair
as
an anti-asthmatic agent two diseases not
conventionally thought of as being targets
for antibody therapy. All of these recent
(post-1998) agents show how protein engi-
neering in its widest sense can be utilized
for the production of drugs against impor-
tant diseases. There is also one other
MAb-based agent, but it is listed under an-
tithrombotics.
4.1.5
Anticoagulants (Table 4.2)
These comprise 13 agents, and range from
some of the earliest work leading to the
approval in Germany in 1983 of Throm-
bate III
, a concentrate of human anti-

thrombin III (see Part IV, Chapter 11)
from human plasma through the use of
dermatan sulfate in Italy in 1999 and even
human-derived protein C from Japan
(Anact C
) and the corresponding equiva-

lent, Ceprotin
in the EU in 2001, to the

variety of chemically modified heparins
the so-called parins which are derived
from porcine and bovine heparin in order
to obtain low molecular-weight (LMW) ma-
terials (3.5 to 8 kDa average). These latter
compounds led to the production of a syn-
thetic sulfated hexacarbohydrate (Quixi-
dar
), approved in the USA in 2002, that

directly inhibits Factor Xa, a mechanism
that the LMW heparins are reputed also to
inhibit to some extent. One of the LMW
heparins, Sandoparin
, which was ap-

proved in Austria in 1989, is one of the
two agents not listed in the Prous Integri-
ty
database, but it is shown as being in

use in articles in PubMed.
4.1.6
Antihemophilia Agents (Table 4.3)
The 11 agents show how the sourcing of
such agents, all of which are based around
Factors VII, VIII or IX, has gone from iso-
lation of purified materials from pooled
human plasma in 1982 (Haemate HS
)
which was first approved in Germany in
1982, through later versions (improved
purification systems) such as Bioclate
in
T
a
b
l
e
4
.
1
A
n
t
i
b
o
d
i
e
s
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
2
0
0
3
;
U
S
A
a
l
e
f
a
c
e
p
t
a
n
t
i
b
o
d
y
A
m
e
v
i
v
e
R
e
c
o
m
b
i
n
a
n
t
d
i
m
e
r
i
c
f
u
s
i
o
n
p
r
o
t
e
i
n
o
f
C
D
2
-
b
i
n
d
i
n
g
p
o
r
t
i
o
n
o
f
L
F
A
-
3
l
i
n
k
e
d
t
o
t
h
e
F
c
p
o
r
t
i
o
n
o
f
h
u
m
a
n
I
g
G
1
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
A
n
t
i
p
s
o
r
i
a
t
i
c
2
0
0
4
;
U
S
A
b
e
v
a
c
i
z
u
m
a
b
a
n
t
i
b
o
d
y
A
v
a
s
t
i
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
z
e
d
m
u
r
i
n
e
-
b
a
s
e
d
M
A
b
t
h
a
t
i
s
s
p
e
c
i
f
i
c
f
o
r
V
E
G
F
.
C
a
n
c
e
r
2
0
0
3
;
U
S
A
t
o
s
i
t
u
m
o
m
a
b
a
n
t
i
b
o
d
y
B
e
x
x
a
r
A
n
t
i
-
C
D
2
0
m
o
u
s
e
M
A
b
l
a
b
e
l
e
d
w
i
t
h
1
3
1
I
p
r
o
d
u
c
e
d
i
n
a
n
a
n
t
i
b
i
o
t
i
c
-
f
r
e
e
m
a
m
-
m
a
l
i
a
n
c
e
l
l
l
i
n
e
.
T
r
e
a
t
m
e
n
t
i
n
v
o
l
v
e
s
b
o
t
h
f
r
e
e
a
n
d
l
a
b
e
l
e
d
M
A
b
s
N
H
L
2
0
0
1
;
U
S
A
a
l
e
m
t
u
z
u
m
a
b
a
n
t
i
b
o
d
y
C
a
m
p
a
t
h
M
a
b
C
a
m
p
a
t
h
H
u
m
a
n
i
z
e
d
M
A
b
t
h
a
t
r
e
c
o
g
n
i
z
e
s
t
h
e
C
D
w
5
2
a
n
t
i
g
e
n
o
n
n
o
r
m
a
l
a
n
d
m
a
l
i
g
-
n
a
n
t
B
a
n
d
T
c
e
l
l
s
A
n
t
i
-
c
a
n
c
e
r
1
9
9
1
;
N
e
t
h
e
r
l
a
n
d
s
H
A
-
1
A
a
n
t
i
b
o
d
y
C
e
n
t
o
x
i
n
H
u
m
a
n
I
g
M
M
A
b
t
o
e
n
d
o
t
o
x
i
n
.
W
i
t
h
-
d
r
a
w
n
a
f
t
e
r
2
y
e
a
r
s
S
e
p
s
i
s
2
0
0
3
;
S
w
i
t
z
e
r
l
a
n
d
c
e
t
u
x
i
m
a
b
a
n
t
i
b
o
d
y
E
r
b
i
t
u
x
H
u
m
a
n
i
z
e
d
m
u
r
i
n
e
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
E
G
F
R
.
P
r
o
d
u
c
e
d
i
n
a
m
u
r
i
n
e
m
y
e
l
o
m
a
c
u
l
t
u
r
e
i
n
s
e
r
u
m
-
f
r
e
e
m
e
d
i
a
a
n
d
p
u
r
i
f
i
e
d
b
y
a
f
f
i
n
i
t
y
,
i
o
n
-
e
x
c
h
a
n
g
e
a
n
d
g
e
l
f
i
l
t
r
a
t
i
o
n
c
h
r
o
m
a
t
o
g
r
a
p
h
i
e
s
C
a
n
c
e
r
1
9
9
8
;
U
S
A
t
r
a
s
t
u
z
u
m
a
b
a
n
t
i
b
o
d
y
H
e
r
c
e
p
t
i
n
A
n
t
i
-
H
E
R
2
/
n
e
u
/
E
r
b
B
2
h
u
m
a
n
i
z
e
d
m
u
r
i
n
e
M
A
b
A
n
t
i
-
c
a
n
c
e
r
2
0
0
3
;
U
S
A
a
d
a
l
i
m
u
m
a
b
a
n
t
i
b
o
d
y
H
u
m
i
r
a
T
r
u
d
e
x
a
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
z
e
d
m
u
r
i
n
e
b
a
s
e
d
M
A
b
t
h
a
t
i
s
s
p
e
c
i
f
i
c
f
o
r
T
N
F
-
a
l
p
h
a
,
c
r
e
a
t
e
d
b
y
p
h
a
g
e
d
i
s
p
l
a
y
a
n
d
i
s
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
A
n
t
i
-
a
r
t
h
r
i
t
i
c
2
0
0
0
;
U
S
A
g
e
m
t
u
z
u
m
a
b
o
z
o
g
a
m
i
c
i
n
a
n
t
i
b
o
d
y
M
y
l
o
t
a
r
g
H
u
m
a
n
i
z
e
d
I
g
G
4
a
n
t
i
C
D
3
3
a
n
t
i
b
o
d
y
h
P
6
7
.
6
c
o
n
j
u
g
a
t
e
d
t
o
N
-
a
c
e
t
y
l
-
g
a
m
m
a
c
a
l
i
c
h
e
a
m
i
c
i
n
A
n
t
i
-
c
a
n
c
e
r
T
a
b
l
e
4
.
1
(
c
o
n
t
i
n
u
e
d
)
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
9
9
;
U
S
A
d
e
n
i
l
e
u
k
i
n
d
i
f
t
i
t
o
x
a
n
t
i
b
o
d
y
O
n
t
a
k
O
n
z
a
r
F
u
s
i
o
n
p
r
o
t
e
i
n
o
f
t
h
e
r
e
c
e
p
t
o
r
b
i
n
d
i
n
g
d
o
m
a
i
n
o
f
d
i
p
h
t
h
e
r
i
a
t
o
x
i
n
c
o
u
p
l
e
d
t
o
I
L
-
2
A
n
t
i
-
c
a
n
c
e
r
1
9
8
6
;
U
S
A
m
u
r
o
m
o
n
a
b
-
C
D
3
a
n
t
i
b
o
d
y
O
r
t
h
o
c
l
o
n
e
O
K
T
3
M
u
r
i
n
e
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
t
h
e
C
D
3
a
n
t
i
g
e
n
o
f
h
u
m
a
n
T
c
e
l
l
s
I
m
m
u
n
o
-
s
u
p
p
r
e
s
s
a
n
t
1
9
9
5
;
G
e
r
m
a
n
y
m
a
b
1
7
-
1
a
a
n
t
i
b
o
d
y
P
a
n
o
r
e
x
M
u
r
i
n
e
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
a
n
e
p
i
t
h
e
l
i
a
l
s
u
r
f
a
c
e
g
l
y
c
o
p
r
o
t
e
i
n
A
d
j
u
v
a
n
t
/
c
o
l
o
-
r
e
c
t
a
l
c
a
n
c
e
r
2
0
0
3
;
U
S
A
e
f
a
l
i
z
u
m
a
b
a
n
t
i
b
o
d
y
R
a
p
t
i
v
a
X
a
n
e
l
i
m
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
z
e
d
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
C
D
-
1
1
a
n
d
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
P
s
o
r
i
a
s
i
s
1
9
9
8
;
U
S
A
i
n
f
l
i
x
i
m
a
b
a
n
t
i
b
o
d
y
R
e
m
i
c
a
d
e
C
e
n
T
N
F
,
c
A
2
,
A
v
a
k
i
n
e
H
u
m
a
n
i
z
e
d
m
u
r
i
n
e
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
h
u
m
a
n
T
N
F
C
r
o
h
n
s
d
i
s
e
a
s
e
1
9
9
7
;
U
S
A
r
i
t
u
x
i
m
a
b
a
n
t
i
b
o
d
y
R
i
t
u
x
a
n
M
a
b
T
h
e
r
a
H
u
m
a
n
i
z
e
d
m
u
r
i
n
e
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
C
D
2
0
A
n
t
i
-
c
a
n
c
e
r
1
9
9
8
;
S
w
i
t
z
e
r
l
a
n
d
b
a
s
i
l
i
x
i
m
a
b
a
n
t
i
b
o
d
y
S
i
m
u
l
e
c
t
H
u
m
a
n
i
z
e
d
m
u
r
i
n
e
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
C
D
2
5
(
I
L
-
2
r
e
c
e
p
t
o
r
)
I
m
m
u
n
o
-
s
u
p
p
r
e
s
s
a
n
t
1
9
9
8
;
U
S
A
p
a
l
i
v
i
z
u
m
a
b
a
n
t
i
b
o
d
y
S
y
n
a
g
i
s
H
u
m
a
n
i
z
e
d
m
u
r
i
n
e
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
R
S
V
p
r
o
t
e
i
n
F
R
S
V
1
9
8
4
;
F
r
a
n
c
e
a
n
t
i
-
t
h
y
m
o
c
y
t
e
g
l
o
b
u
l
i
n
a
n
t
i
b
o
d
y
T
h
y
m
o
-
g
l
o
b
u
l
i
n
e
T
h
y
m
o
g
l
o
b
u
l
i
n
P
o
l
y
c
l
o
n
a
l
r
a
b
b
i
t
a
n
t
i
t
h
y
m
o
c
y
t
e
a
n
t
i
b
o
d
i
e
s
I
m
m
u
n
o
-
s
u
p
p
r
e
s
s
a
n
t
2
0
0
3
;
U
S
A
o
m
a
l
i
z
u
m
a
b
a
n
t
i
b
o
d
y
X
o
l
a
i
r
H
u
m
a
n
i
z
e
d
m
u
r
i
n
e
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
I
g
E
,
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
A
n
t
i
-
a
s
t
h
m
a
t
i
c
1
9
9
7
;
U
S
A
d
a
c
l
i
z
u
m
a
b
a
n
t
i
b
o
d
y
Z
e
n
a
p
a
x
H
u
m
a
n
i
z
e
d
m
u
r
i
n
e
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
C
D
2
5
(
I
L
-
2
r
e
c
e
p
t
o
r
)
I
m
m
u
n
o
-
s
u
p
p
r
e
s
s
a
n
t
2
0
0
2
;
U
S
A
i
b
r
i
t
u
m
o
m
a
b
t
i
u
e
t
a
n
a
n
t
i
b
o
d
y
Z
e
v
a
l
i
n
I
n
-
1
1
1
Z
e
v
a
l
i
n
,
Y
-
9
0
Z
e
v
a
l
i
n
H
u
m
a
n
i
z
e
d
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
C
D
2
0
,
p
r
o
d
u
c
e
d
i
n
C
H
O
c
e
l
l
s
a
n
d
i
s
p
a
r
t
o
f
a
f
o
u
r
-
p
a
r
t
t
r
e
a
t
m
e
n
t
w
i
t
h
R
i
t
u
x
a
n
a
n
d
t
h
e
r
a
d
i
o
l
a
b
e
l
e
d
Z
e
v
a
l
i
n
s
N
H
L
T
a
b
l
e
4
.
2
A
n
t
i
c
o
a
g
u
l
a
n
t
s
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
9
6
;
U
K
c
e
r
t
o
p
a
r
i
n
s
o
d
i
u
m
a
n
t
i
c
o
a
g
u
l
a
n
t
A
l
p
h
a
p
a
r
i
n
T
r
o
p
a
r
i
n
P
r
o
d
u
c
t
f
r
o
m
i
s
o
a
m
y
l
n
i
t
r
i
t
e
d
e
g
r
a
d
a
t
i
o
n
o
f
p
o
r
c
i
n
e
h
e
p
a
r
i
n
.
R
e
l
a
t
i
v
e
m
o
l
e
c
u
l
a
r
m
a
s
s
e
s
r
a
n
g
e
f
r
o
m
5
7
k
D
a
D
V
T
;
a
n
t
i
c
o
a
g
u
l
a
n
t
2
0
0
1
;
J
a
p
a
n
a
c
t
i
v
a
t
e
d
p
r
o
t
e
i
n
C
c
o
n
c
n
a
n
t
i
c
o
a
g
u
l
a
n
t
A
n
a
c
t
C
C
o
n
c
e
n
t
r
a
t
e
o
f
p
r
o
t
e
i
n
C
f
r
o
m
h
u
m
a
n
p
l
a
s
m
a
D
V
T
;
a
n
t
i
c
o
a
g
u
l
a
n
t
2
0
0
2
;
U
S
A
f
o
n
d
a
p
a
r
i
n
s
o
d
i
u
m
a
n
t
i
c
o
a
g
u
l
a
n
t
A
r
i
x
t
r
a
Q
u
i
x
i
d
a
r
S
y
n
t
h
e
t
i
c
s
u
l
f
a
t
e
d
h
e
x
a
c
a
r
b
o
h
y
d
r
a
t
e
t
h
a
t
i
n
h
i
b
i
t
s
c
o
a
g
u
l
a
t
i
o
n
F
a
c
t
o
r
X
a
D
V
T
;
a
n
t
i
c
o
a
g
u
l
a
n
t
2
0
0
1
;
E
U
p
r
o
t
e
i
n
C
c
o
n
c
e
n
-
t
r
a
t
e
a
n
t
i
c
o
a
g
u
l
a
n
t
C
e
p
r
o
t
i
n
H
u
m
a
n
p
r
o
t
e
i
n
C
c
o
n
c
e
n
t
r
a
t
e
p
u
r
i
f
i
e
d
b
y
M
A
b
P
u
r
p
u
r
a
f
u
l
m
i
n
a
n
s
1
9
9
3
;
G
e
r
m
a
n
y
r
e
v
i
p
a
r
i
n
s
o
d
i
u
m
a
n
t
i
c
o
a
g
u
l
a
n
t
C
l
i
v
a
r
i
n
C
l
i
v
a
r
i
n
1
7
5
0
,
C
l
i
v
a
r
i
n
e
,
L
o
w
m
o
r
i
n
P
r
o
d
u
c
t
f
r
o
m
n
i
t
r
o
u
s
a
c
i
d
d
e
g
r
a
d
a
t
i
o
n
o
f
p
o
r
c
i
n
e
h
e
p
a
r
i
n
.
R
e
l
a
t
i
v
e
m
o
l
e
c
u
l
a
r
m
a
s
s
e
s
r
a
n
g
e
f
r
o
m
3
.
5
5
.
1
5
k
D
a
D
V
T
;
a
n
t
i
c
o
a
g
u
l
a
n
t
1
9
9
3
;
I
t
a
l
y
p
a
r
n
a
p
a
r
i
n
s
o
d
i
u
m
a
n
t
i
c
o
a
g
u
l
a
n
t
F
l
u
x
u
m
L
o
w
h
e
p
a
,
M
i
n
i
d
a
l
t
o
n
,
O
P
/
L
M
W
H
L
M
W
h
e
p
a
r
i
n
p
r
o
d
u
c
e
d
b
y
h
y
d
r
o
g
e
n
p
e
r
o
x
i
d
e
/
c
o
p
p
e
r
s
a
l
t
t
r
e
a
t
m
e
n
t
o
f
b
o
v
i
n
e
o
r
p
o
r
c
i
n
e
h
e
p
a
r
i
n
.
M
a
s
s
a
v
e
r
a
g
e
M
W
4
6
k
D
a
D
V
T
;
a
n
t
i
c
o
a
g
u
l
a
n
t
1
9
8
5
;
G
e
r
m
a
n
y
d
a
l
t
e
p
a
r
i
n
s
o
d
i
u
m
a
n
t
i
c
o
a
g
u
l
a
n
t
F
r
a
g
m
i
n
P
r
o
d
u
c
t
f
r
o
m
n
i
t
r
o
u
s
a
c
i
d
d
e
g
r
a
d
a
t
i
o
n
o
f
p
o
r
c
i
n
e
h
e
p
a
r
i
n
.
R
e
l
a
t
i
v
e
m
o
l
e
c
u
l
a
r
m
a
s
s
e
s
r
a
n
g
e
f
r
o
m
5
.
6
6
.
4
k
D
a
D
V
T
;
a
n
t
i
c
o
a
g
u
l
a
n
t
1
9
8
5
;
G
e
r
m
a
n
y
n
a
d
r
o
p
a
r
i
n
c
a
l
c
i
u
m
a
n
t
i
c
o
a
g
u
l
a
n
t
F
r
a
x
i
p
a
r
i
n
e
F
r
a
x
i
p
a
r
i
n
a
,
F
r
a
x
o
d
i
,
S
e
l
e
p
a
r
i
n
a
P
r
o
d
u
c
t
f
r
o
m
n
i
t
r
o
u
s
a
c
i
d
d
e
g
r
a
d
a
t
i
o
n
o
f
p
o
r
c
i
n
e
h
e
p
a
r
i
n
.
R
e
l
a
t
i
v
e
m
o
l
e
c
u
l
a
r
m
a
s
s
e
s
r
a
n
g
e
f
r
o
m
3
.
6
5
.
0
k
D
a
D
V
T
;
a
n
t
i
c
o
a
g
u
l
a
n
t
1
9
8
7
;
F
r
a
n
c
e
e
n
o
x
a
p
a
r
i
n
a
n
t
i
c
o
a
g
u
l
a
n
t
L
o
v
e
n
o
x
C
l
e
x
a
n
e
,
K
l
e
x
a
n
e
L
M
W
h
e
p
a
r
i
n
p
r
o
d
u
c
e
d
b
y
a
l
k
a
l
i
n
e
d
e
p
o
l
y
m
e
r
i
z
a
t
i
o
n
o
f
t
h
e
b
e
n
z
y
l
e
s
t
e
r
d
e
r
i
v
a
t
i
v
e
p
o
r
c
i
n
e
h
e
p
a
r
i
n
.
M
a
s
s
a
v
e
r
a
g
e
M
W
3
.
5
5
.
5
k
D
a
D
V
T
;
a
n
t
i
c
o
a
g
u
l
a
n
t
T
a
b
l
e
4
.
2
(
c
o
n
t
i
n
u
e
d
)
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
9
9
;
I
t
a
l
y
d
e
r
m
a
t
a
n
s
u
l
f
a
t
e
a
n
t
i
c
o
a
g
u
l
a
n
t
M
i
s
t
r
a
l
D
e
r
m
a
t
a
n
s
u
l
f
a
t
e
,
n
o
n
-
r
e
c
o
m
b
i
n
a
n
t
D
V
T
;
a
n
t
i
c
o
a
g
u
l
a
n
t
1
9
9
7
;
U
S
A
a
r
d
e
p
a
r
i
n
s
o
d
i
u
m
a
n
t
i
c
o
a
g
u
l
a
n
t
N
o
r
m
i
f
l
o
C
e
n
t
a
x
a
r
i
n
L
M
W
h
e
p
a
r
i
n
o
b
t
a
i
n
e
d
b
y
p
e
r
o
x
i
d
e
d
e
g
r
a
d
a
t
i
o
n
o
f
p
o
r
c
i
n
e
h
e
p
a
r
i
n
w
i
t
h
a
v
e
r
a
g
e
r
e
l
a
t
i
v
e
m
a
s
s
i
n
t
h
e
r
a
n
g
e
o
f
5
.
5
6
.
5
k
D
a
D
V
T
;
a
n
t
i
c
o
a
g
u
l
a
n
t
1
9
8
9
;
A
u
s
t
r
i
a
L
M
W
h
e
p
a
r
i
n
a
n
t
i
c
o
a
g
u
l
a
n
t
S
a
n
d
o
p
a
r
i
n
L
M
W
h
e
p
a
r
i
n
,
n
o
w
n
o
t
i
n
u
s
e
D
V
T
;
a
n
t
i
c
o
a
g
u
l
a
n
t
1
9
8
3
;
G
e
r
m
a
n
y
a
n
t
i
t
h
r
o
m
b
i
n
I
I
I
a
n
t
i
c
o
a
g
u
l
a
n
t
T
h
r
o
m
b
a
t
e
I
I
I
A
t
e
n
a
t
i
v
,
K
y
b
e
r
n
i
n
H
u
m
a
n
a
n
t
i
t
h
r
o
m
b
i
n
I
I
I
c
o
n
c
e
n
t
r
a
t
e
a
n
t
i
c
o
a
g
u
l
a
n
t
T
a
b
l
e
4
.
3
A
n
t
i
h
e
m
o
p
h
i
l
i
a
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
2
0
0
3
;
U
S
A
r
A
H
F
-
P
F
M
a
n
t
i
h
e
m
o
p
h
i
l
i
c
A
d
v
a
t
e
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
c
o
a
g
u
l
a
t
i
o
n
f
a
c
t
o
r
V
I
I
I
p
r
o
d
u
c
e
d
b
y
a
p
l
a
s
m
a
/
a
l
b
u
m
i
n
-
f
r
e
e
m
e
t
h
o
d
H
e
m
o
p
h
i
l
i
a
1
9
9
1
;
U
S
A
f
a
c
t
o
r
I
X
h
u
m
a
n
a
n
t
i
h
e
m
o
p
h
i
l
i
c
A
l
p
h
a
N
i
n
e
-
S
D
N
o
n
a
f
a
c
t
H
i
g
h
-
p
u
r
i
t
y
h
u
m
a
n
f
a
c
t
o
r
I
X
f
r
o
m
n
a
t
u
r
a
l
s
o
u
r
c
e
s
.
V
i
r
u
s
e
s
i
n
a
c
t
i
v
a
t
e
d
b
y
o
r
g
a
n
i
c
s
o
l
v
e
n
t
/
d
e
t
e
r
g
e
n
t
t
r
e
a
t
m
e
n
t
.
H
e
m
o
p
h
i
l
i
a
1
9
9
7
;
U
S
A
r
h
-
F
a
c
t
o
r
I
X
a
n
t
i
h
e
m
o
p
h
i
l
i
c
B
e
n
e
f
i
x
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
F
a
c
t
o
r
I
X
p
r
o
d
u
c
e
d
i
n
C
H
O
c
e
l
l
s
H
e
m
o
p
h
i
l
i
a
1
9
8
2
;
G
e
r
m
a
n
y
h
u
m
a
n
f
a
c
t
o
r
V
I
I
I
a
n
t
i
h
e
m
o
p
h
i
l
i
c
H
a
e
m
a
t
e
H
S
H
a
e
m
a
t
e
P
,
H
u
m
a
t
e
P
H
u
m
a
n
f
a
c
t
o
r
V
I
I
I
a
n
d
v
o
n
W
i
l
l
e
b
r
a
n
d
f
a
c
t
o
r
i
s
o
l
a
t
e
d
f
r
o
m
p
o
o
l
e
d
p
l
a
s
m
a
H
e
m
o
p
h
i
l
i
a
1
9
9
9
;
N
Z
B
A
Y
-
1
4
-
2
2
2
2
a
n
t
i
h
e
m
o
p
h
i
l
i
c
K
o
g
e
n
a
t
e
F
S
H
e
l
i
x
a
t
e
F
S
,
K
o
g
e
n
a
t
e
S
F
H
u
m
a
n
r
e
c
o
m
b
i
n
a
n
t
f
a
c
t
o
r
V
I
I
I
e
x
p
r
e
s
s
e
d
i
n
b
a
b
y
h
a
m
s
t
e
r
k
i
d
n
e
y
c
e
l
l
s
a
n
d
p
u
r
i
f
i
e
d
b
y
m
o
d
i
f
i
e
d
c
h
r
o
m
a
t
o
g
r
a
p
h
y
a
n
d
t
h
e
n
s
t
a
b
i
l
i
z
e
d
w
i
t
h
s
u
c
r
o
s
e
H
e
m
o
p
h
i
l
i
a
1
9
8
7
;
U
S
A
f
a
c
t
o
r
V
I
I
I
;
C
,
m
a
b
p
u
r
i
f
i
e
d
a
n
t
i
h
e
m
o
p
h
i
l
i
c
M
o
n
o
c
l
a
t
e
-
P
B
i
o
c
l
a
t
e
,
M
o
n
o
c
l
a
t
e
H
u
m
a
n
f
a
c
t
o
r
V
I
I
I
i
s
o
l
a
t
e
d
f
r
o
m
p
o
o
l
e
d
p
l
a
s
m
a
H
e
m
o
p
h
i
l
i
a
1
9
9
5
;
S
w
i
t
z
e
r
l
a
n
d
f
a
c
t
o
r
V
I
I
a
a
n
t
i
h
e
m
o
p
h
i
l
i
c
N
o
v
o
S
e
v
e
n
N
i
a
S
t
a
s
e
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
F
a
c
t
o
r
V
I
I
f
r
o
m
h
u
m
a
n
c
l
o
n
e
l
a
m
b
d
a
H
V
I
I
2
4
6
3
,
p
r
o
d
u
c
e
d
i
n
C
H
O
c
e
l
l
s
H
e
m
o
p
h
i
l
i
a
1
9
9
9
;
G
e
r
m
a
n
y
m
o
r
o
c
t
o
c
o
g
a
l
f
a
a
n
t
i
h
e
m
o
p
h
i
l
i
c
R
e
f
a
c
t
o
R
e
f
a
c
t
o
A
F
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
F
a
c
t
o
r
V
I
I
I
w
i
t
h
t
h
e
B
-
d
o
m
a
i
n
d
e
l
e
t
e
d
,
p
r
o
d
u
c
e
d
i
n
C
H
O
c
e
l
l
s
H
e
m
o
p
h
i
l
i
a
1
9
9
3
;
U
S
A
r
e
c
o
m
b
i
n
a
n
t
f
a
c
t
o
r
V
I
I
I
a
n
t
i
h
e
m
o
p
h
i
l
i
c
K
o
g
e
n
a
t
e
H
e
l
i
x
a
t
e
,
H
e
l
i
x
a
t
e
N
e
x
G
e
n
H
u
m
a
n
r
e
c
o
m
b
i
n
a
n
t
f
a
c
t
o
r
V
I
I
I
e
x
p
r
e
s
s
e
d
i
n
b
a
b
y
h
a
m
s
t
e
r
k
i
d
n
e
y
c
e
l
l
s
a
n
d
p
u
r
i
f
i
e
d
b
y
M
A
b
a
f
f
i
n
i
t
y
c
h
r
o
m
a
t
o
g
r
a
p
h
y
H
e
m
o
p
h
i
l
i
a
1
9
9
2
;
J
a
p
a
n
f
a
c
t
o
r
I
X
a
n
t
i
h
e
m
o
p
h
i
l
i
c
N
o
v
a
c
t
M
P
u
r
i
f
i
e
d
h
u
m
a
n
f
a
c
t
o
r
I
X
H
e
m
o
p
h
i
l
i
a
1
9
9
2
;
S
w
e
d
e
n
f
a
c
t
o
r
V
I
I
I
a
n
t
i
h
e
m
o
p
h
i
l
i
c
R
e
c
o
m
b
i
n
a
t
e
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
F
a
c
t
o
r
V
I
I
I
,
p
r
o
d
u
c
e
d
i
n
a
n
a
n
i
m
a
l
c
e
l
l
l
i
n
e
H
e
m
o
p
h
i
l
i
a
the USA in 1987 and AlphaNine-SD
in
the USA and Novact M
in Japan in 1991
and 1992, respectively. These were rapidly
followed by a variety of recombinant pro-
teins expressed in mammalian cell lines
[baby hamster kidney (BHK) cells and Chi-
nese hamster ovary (CHO) cells in general]
(see Part IV, Chapters 1, 2, and 12), with a
variety of downstream purification tech-
niques being utilized, including immuno-
sorption against specific MAbs to reduce
contaminants, leading to the latest ap-
proved agent, Advate
which was intro-

duced in the USA in 2003 (and in the EU)
by Baxter (see Part II, Chapter 3). As ex-
emplified in the case study, this material
is produced and purified in the absence of
any human or animal-derived raw materi-
als, thus eliminating any risk of biological
contamination from components of the
process.
4.1.7
Antithrombotic Agents (Table 4.4)
These 16 agents, as in the case of the anti-
hemophilic compounds, range from iso-
lated biological materials from animal
sources (polydeoxyribonucleotides, Nora-
vid
, approved in Italy in 1986), through a

mixture of bacterial streptokinase/human
plasminogen treated with acylating agents
(Eminase
launched in the UK in 1987)

to native tissue plasminogen activators (t-
PAs) from normal human cells (Plasvata
,
launched in Japan in 1991). Contempora-
neously with the launch of Eminase, the
first recombinant human t-PA expressed
in CHO cells was launched in the USA in
1987, followed over the subsequent years
by a variety of similar materials. However,
what is notable within the overall group
was the advent of two recombinant ver-
sions of hirudin from leeches, perhaps the
first antithrombotic actually used in man,
Refludan
in Germany in 1997 and An-

giomax
in New Zealand in 1999, plus

the one antibody-derived therapy, Reo-
Pro
, a humanized MAb directed against

human integrins which was introduced in
the USA in 1995.
4.1.8
Protein-based Agents (Table 4.5)
As with the antibodies, these 24 agents
cover a wide range of disease entities and,
in common with other agents, range from
proteins isolated from a variety of sources
such as Aralast
, an alpha-1 antitrypsin
from human plasma launched in the USA
in 2003, and which had an earlier version,
Prolastin
, launched in the USA in 1987,

follicle-stimulating hormone (FSH) from
human urine as Metrodin HP
, in Swit-
zerland in 1982 for infertility, Biostim
, a
glycoprotein from Klebsiella pneumoniae
used as an immunomodulator launched in
France in 1985, at a comparable time to
the use of Gliptide
in Italy, a sulfated
glycopeptide isolated from porcine duode-
num as an anti-ulcer medication. Follow-
ing these early materials, the first launch
of a recombinant protein was probably
that of Gentel
, an ophthalmic formula-
tion of epidermal growth factor in Switzer-
land in 1987. This is the other agent that
could not be identified in the Prous Integ-
rity database, but it is listed in the Annual
Reports of Medicinal Chemistry under
drugs marketed in 1987. An inspection of
Table 4.5 shows that since then, a multi-
plicity of human proteins has been
approved, usually following expression in
E. coli or in CHO cells.
It should be noted that quite recently,
the first two agents that can effect bone
growth have appeared on the market:
Novos
(bone morphogenic protein 7) in

Australia in 2001; and InFuse
(bone mor-
T
a
b
l
e
4
.
4
A
n
t
i
t
h
r
o
m
b
o
t
i
c
s
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
8
7
;
U
S
A
a
l
t
e
p
l
a
s
e
a
n
t
i
t
h
r
o
m
b
o
t
i
c
A
c
t
i
v
a
s
e
A
c
t
i
l
y
s
e
,
A
c
t
i
v
a
c
i
n
,
C
a
t
h
f
l
o
A
c
t
i
v
a
s
e
,
G
R
T
P
A
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
t
i
s
s
u
e
p
l
a
s
m
i
n
o
g
e
n
a
c
t
i
v
a
t
o
r
,
g
l
y
c
o
f
o
r
m
a
l
p
h
a
p
r
o
d
u
c
e
d
i
n
C
H
O
c
e
l
l
s
T
h
r
o
m
b
o
s
i
s
1
9
9
9
;
N
Z
b
i
v
a
l
r
u
d
i
n
a
n
t
i
t
h
r
o
m
b
o
t
i
c
A
n
g
i
o
m
a
x
A
n
g
i
o
x
,
H
i
r
u
l
o
g
S
y
n
t
h
e
t
i
c
p
e
p
t
i
d
e
c
o
r
r
e
s
p
o
n
d
i
n
g
t
o
t
h
e
5
3
6
4
s
e
q
u
e
n
c
e
o
f
d
e
s
u
l
f
o
h
i
r
u
d
i
n
T
h
r
o
m
b
o
s
i
s
1
9
9
8
;
J
a
p
a
n
m
o
n
t
e
p
l
a
s
e
a
n
t
i
t
h
r
o
m
b
o
t
i
c
C
l
e
a
c
t
o
r
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
t
-
P
A
w
i
t
h
8
4
C
y
s
r
e
p
l
a
c
e
d
b
y
8
4
S
e
r
T
h
r
o
m
b
o
s
i
s
1
9
8
7
;
U
K
a
n
i
s
t
r
e
o
l
a
s
e
a
n
t
i
t
h
r
o
m
b
o
t
i
c
E
m
i
n
a
s
e
I
m
i
n
a
s
e
,
M
u
l
t
i
l
a
s
e
A
d
e
f
i
n
e
d
m
i
x
t
u
r
e
o
f
b
a
c
t
e
r
i
a
l
s
t
r
e
p
t
o
k
i
n
a
s
e
,
h
u
m
a
n
L
y
s
-
p
l
a
s
m
i
n
o
g
e
n
a
n
d
a
n
o
n
-
s
p
e
c
i
f
i
c
a
n
i
s
o
y
l
a
t
i
o
n
o
f
t
h
e
c
o
m
b
i
n
a
t
i
o
n
T
h
r
o
m
b
o
s
i
s
1
9
9
6
;
J
a
p
a
n
n
a
t
e
p
l
a
s
e
a
n
t
i
t
h
r
o
m
b
o
t
i
c
M
i
l
y
z
e
r
T
e
p
a
s
e
M
i
x
t
u
r
e
o
f
r
e
c
o
m
b
i
n
a
n
t
N
-
[
N
2
(
N
-
g
l
y
c
y
l
-
L
-
a
l
a
n
y
l
)
-
L
-
a
r
g
i
n
y
l
)
-
p
l
a
s
m
i
n
o
g
e
n
a
c
t
i
v
a
t
o
r
,
b
e
t
a
g
l
y
c
o
f
o
r
m
a
n
d
p
l
a
s
m
i
n
o
g
e
n
a
c
t
i
v
a
t
o
r
,
b
e
t
a
g
l
y
c
o
f
o
r
m
T
h
r
o
m
b
o
s
i
s
1
9
8
6
;
I
t
a
l
y
d
e
f
i
b
r
o
t
i
d
e
a
n
t
i
t
h
r
o
m
b
o
t
i
c
N
o
r
a
v
i
d
P
r
o
c
i
c
l
i
d
e
P
o
l
y
d
e
o
x
y
r
i
b
o
n
u
c
l
e
o
t
i
d
e
s
f
r
o
m
m
a
m
m
a
l
i
a
n
o
r
g
a
n
s
w
i
t
h
M
W
i
n
t
h
e
1
5
3
0
k
D
a
r
a
n
g
e
T
h
r
o
m
b
o
s
i
s
1
9
9
1
;
J
a
p
a
n
s
i
l
t
e
p
l
a
s
e
a
n
t
i
t
h
r
o
m
b
o
t
i
c
P
l
a
s
m
i
n
e
r
R
e
c
o
m
b
i
n
a
n
t
N
-
[
N
2
(
N
-
g
l
y
c
y
l
-
L
-
a
l
a
n
y
l
)
-
L
-
a
r
g
i
n
y
l
)
-
p
l
a
s
m
i
n
o
g
e
n
a
c
t
i
v
a
t
o
r
,
b
e
t
a
g
l
y
c
o
f
o
r
m
e
x
p
r
e
s
s
e
d
i
n
m
u
r
i
n
e
C
1
2
7
c
e
l
l
s
T
h
r
o
m
b
o
s
i
s
1
9
9
1
;
J
a
p
a
n
t
i
s
o
k
i
n
a
s
e
a
n
t
i
t
h
r
o
m
b
o
t
i
c
P
l
a
s
v
a
t
a
H
a
p
a
s
e
N
a
t
i
v
e
t
-
P
A
f
r
o
m
h
u
m
a
n
-
d
e
r
i
v
e
d
n
o
r
m
a
l
c
e
l
l
s
T
h
r
o
m
b
o
s
i
s
T
a
b
l
e
4
.
4
(
c
o
n
t
i
n
u
e
d
)
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
9
7
;
G
e
r
m
a
n
y
l
e
p
i
r
u
d
i
n
a
n
t
i
t
h
r
o
m
b
o
t
i
c
R
e
f
l
u
d
a
n
R
e
c
o
m
b
i
n
a
n
t
1
-
L
-
L
e
u
-
2
-
L
-
T
h
r
-
6
3
-
d
e
s
u
l
f
o
h
i
r
u
d
i
n
p
r
o
d
u
c
e
d
i
n
S
a
c
c
h
a
r
o
m
y
c
e
s
c
e
r
e
v
i
s
i
a
e
T
h
r
o
m
b
o
s
i
s
1
9
9
5
;
U
S
A
a
b
c
i
x
i
m
a
b
a
n
t
i
t
h
r
o
m
b
o
t
i
c
R
e
o
P
r
o
C
e
n
t
o
R
x
H
u
m
a
n
i
z
e
d
m
u
r
i
n
e
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
h
u
m
a
n
i
n
t
e
g
r
i
n
s
T
h
r
o
m
b
o
s
i
s
1
9
9
6
;
G
e
r
m
a
n
y
r
e
t
e
p
l
a
s
e
a
n
t
i
t
h
r
o
m
b
o
t
i
c
R
e
t
a
v
a
s
e
E
c
o
k
i
n
a
s
e
,
R
a
p
i
l
y
s
i
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
t
-
P
A
(
r
e
s
i
d
u
e
s
1
7
3
5
2
7
)
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
T
h
r
o
m
b
o
s
i
s
1
9
9
8
;
G
e
r
m
a
n
y
d
e
s
i
r
u
d
i
n
a
n
t
i
t
h
r
o
m
b
o
t
i
c
R
e
v
a
s
c
R
e
c
o
m
b
i
n
a
n
t
6
3
-
d
e
s
u
l
f
o
h
i
r
u
d
i
n
e
x
p
r
e
s
s
e
d
i
n
S
a
c
c
h
a
r
o
m
y
c
e
s
c
e
r
e
v
i
s
i
a
e
T
h
r
o
m
b
o
s
i
s
1
9
9
5
;
J
a
p
a
n
d
u
t
e
p
l
a
s
e
a
n
t
i
t
h
r
o
m
b
o
t
i
c
S
o
l
c
l
o
t
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
t
i
s
s
u
e
t
y
p
e
2
t
-
P
A
T
h
r
o
m
b
o
s
i
s
1
9
9
9
;
J
a
p
a
n
p
a
m
i
t
e
p
l
a
s
e
a
n
t
i
t
h
r
o
m
b
o
t
i
c
S
o
l
i
n
a
s
e
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
t
-
P
A
w
i
t
h
a
s
i
t
e
m
u
t
a
t
i
o
n
a
t
t
h
e
k
r
i
n
g
l
e
-
2
-
s
e
r
i
n
e
p
r
o
t
e
a
s
e
l
i
n
k
a
g
e
s
i
t
e
T
h
r
o
m
b
o
s
i
s
2
0
0
0
;
U
S
A
t
e
n
e
c
t
e
p
l
a
s
e
a
n
t
i
t
h
r
o
m
b
o
t
i
c
T
N
K
a
s
e
M
e
t
a
l
y
s
e
R
e
c
o
m
b
i
n
a
n
t
m
o
d
i
f
i
e
d
h
u
m
a
n
t
-
P
A
,
v
e
r
y
r
a
p
i
d
a
c
t
i
n
g
,
p
r
o
d
u
c
e
d
i
n
C
H
O
c
e
l
l
s
T
h
r
o
m
b
o
s
i
s
1
9
9
2
;
J
a
p
a
n
p
r
o
-
u
r
o
k
i
n
a
s
e
a
n
t
i
t
h
r
o
m
b
o
t
i
c
T
h
r
o
m
b
o
l
y
s
e
T
o
m
i
e
z
e
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
p
r
o
u
r
o
k
i
n
a
s
e
,
c
l
o
n
e
p
A
3
/
p
D
2
/
p
F
1
,
g
l
y
c
o
s
y
l
a
t
e
d
T
h
r
o
m
b
o
s
i
s
T
a
b
l
e
4
.
5
P
r
o
t
e
i
n
-
b
a
s
e
d
a
g
e
n
t
s
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
2
0
0
3
;
U
S
A
A
l
p
h
a
-
1
a
n
t
i
t
r
y
p
s
i
n
p
r
o
t
e
i
n
A
r
a
l
a
s
t
R
e
s
p
i
t
i
n
H
u
m
a
n
p
l
a
s
m
a
-
d
e
r
i
v
e
d
a
l
p
h
a
-
1
a
n
t
i
-
t
r
y
p
s
i
n
,
n
o
n
-
r
e
c
o
m
b
i
n
a
n
t
E
m
p
h
y
s
e
m
a
1
9
9
9
;
G
e
r
m
a
n
y
t
a
s
o
n
e
r
m
i
n
p
r
o
t
e
i
n
B
e
r
o
m
u
n
R
e
s
i
d
u
e
s
1
1
5
7
o
f
h
u
m
a
n
T
N
F
-
a
l
p
h
a
.
R
e
c
o
m
b
i
n
a
n
t
e
x
p
r
e
s
s
i
o
n
i
n
E
.
c
o
l
i
A
n
t
i
-
c
a
n
c
e
r
1
9
8
5
;
F
r
a
n
c
e
R
U
-
4
1
7
4
0
p
r
o
t
e
i
n
B
i
o
s
t
i
m
G
l
y
c
o
p
r
o
t
e
i
n
e
x
t
r
a
c
t
e
d
f
r
o
m
K
l
e
b
s
i
e
l
l
a
p
n
e
u
m
o
n
i
a
O
1
K
2
s
e
r
o
t
y
p
e
I
m
m
u
n
o
-
m
o
d
u
l
a
t
o
r
1
9
9
2
;
J
a
p
a
n
c
e
l
m
o
l
e
u
k
i
n
p
r
o
t
e
i
n
C
e
l
e
u
k
R
e
c
o
m
b
i
n
a
n
t
i
n
t
e
r
l
e
u
k
i
n
2
(
I
L
-
2
)
f
r
o
m
t
h
e
h
u
m
a
n
c
l
o
n
e
p
T
I
L
2
-
2
1
a
A
n
g
i
o
s
a
r
c
o
m
a
1
9
9
1
;
U
K
c
b
o
t
u
l
i
n
u
m
t
y
p
e
a
t
o
x
i
n
p
r
o
t
e
i
n
D
y
s
p
o
r
t
C
l
o
s
t
r
i
d
i
u
m
b
o
t
u
l
i
n
u
m
T
y
p
e
A
t
o
x
i
n
c
o
m
p
l
e
x
e
d
w
i
t
h
h
e
m
a
g
l
u
t
i
n
i
n
.
P
r
o
d
u
c
e
d
o
r
i
g
i
n
a
l
l
y
b
y
P
o
r
t
o
n
i
n
t
h
e
U
K
b
y
f
e
r
m
e
n
t
a
t
i
o
n
B
l
e
p
h
a
r
o
s
p
a
s
m
1
9
9
8
;
U
S
A
e
t
a
n
e
r
c
e
p
t
p
r
o
t
e
i
n
E
n
b
r
e
l
R
e
c
o
m
b
i
n
a
n
t
f
u
s
i
o
n
p
r
o
t
e
i
n
o
f
h
u
m
a
n
s
o
l
u
b
l
e
p
7
5
T
N
F
r
e
c
e
p
t
o
r
(
r
e
s
i
d
u
e
s
1
2
3
5
)
l
i
n
k
e
d
t
o
t
h
e
F
c
p
o
r
t
i
o
n
o
f
h
u
m
a
n
I
g
G
1
(
r
e
s
i
d
u
e
s
2
3
6
4
6
7
)
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
A
n
t
i
-
a
r
t
h
r
i
t
i
c
1
9
8
7
;
S
w
i
t
z
e
r
l
a
n
d
e
p
i
d
e
r
m
a
l
g
r
o
w
t
h
f
a
c
t
o
r
p
r
o
t
e
i
n
G
e
n
t
e
l
R
e
c
o
m
b
i
n
a
n
t
e
p
i
d
e
r
m
a
l
g
r
o
w
t
h
f
a
c
t
o
r
(
f
o
r
o
p
h
t
h
a
l
m
i
c
u
s
e
)
V
u
l
n
e
r
a
r
y
1
9
8
5
;
I
t
a
l
y
s
u
l
g
l
y
c
o
t
i
d
e
p
r
o
t
e
i
n
G
l
i
p
t
i
d
e
S
u
l
f
a
t
e
d
g
l
y
c
o
p
e
p
t
i
d
e
i
s
o
l
a
t
e
d
f
r
o
m
t
h
e
m
u
c
i
n
o
f
p
o
r
c
i
n
e
d
u
o
d
e
n
u
m
A
n
t
i
-
u
l
c
e
r
1
9
9
4
;
F
i
n
l
a
n
d
f
o
l
l
i
t
r
o
p
i
n
a
l
f
a
p
r
o
t
e
i
n
G
o
n
a
l
-
F
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
F
S
H
i
n
i
t
s
a
l
p
h
a
g
l
y
c
o
f
o
r
m
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
H
o
r
m
o
n
e
1
9
9
4
;
U
S
A
s
o
m
a
t
o
m
e
d
i
n
-
1
p
r
o
t
e
i
n
I
g
e
f
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
I
G
F
-
1
H
o
r
m
o
n
e
T
a
b
l
e
4
.
5
(
c
o
n
t
i
n
u
e
d
)
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
9
2
;
J
a
p
a
n
t
e
c
e
l
e
u
k
i
n
p
r
o
t
e
i
n
I
m
u
n
a
c
e
B
i
o
l
e
u
k
i
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
m
e
t
I
L
-
2
A
n
t
i
-
c
a
n
c
e
r
2
0
0
2
;
U
S
A
d
i
b
o
t
e
r
m
i
n
a
l
f
a
p
r
o
t
e
i
n
I
n
F
u
s
e
I
n
d
u
c
t
O
S
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
b
o
n
e
m
o
r
p
h
o
g
e
n
i
c
p
r
o
t
e
i
n
-
2
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
B
o
n
e
f
u
s
i
o
n
2
0
0
1
;
U
S
A
a
n
a
k
i
n
r
a
p
r
o
t
e
i
n
K
i
n
e
r
e
t
A
n
t
r
i
l
A
r
e
c
o
m
b
i
n
a
n
t
n
o
n
-
g
l
y
c
o
s
y
l
a
t
e
d
h
u
m
a
n
I
L
-
1
r
e
c
e
p
t
o
r
a
n
t
a
g
o
n
i
s
t
i
s
o
l
a
t
e
d
f
r
o
m
m
o
n
o
c
y
t
e
s
a
n
d
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
A
n
t
i
-
a
r
t
h
r
i
t
i
c
2
0
0
1
;
G
e
r
m
a
n
y
l
u
t
r
o
p
i
n
a
l
f
a
p
r
o
t
e
i
n
L
u
v
e
r
i
s
L
h
a
d
i
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
L
H
,
a
l
p
h
a
g
l
y
c
o
f
o
r
m
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
H
o
r
m
o
n
e
1
9
8
2
;
S
w
i
t
z
e
r
l
a
n
d
u
r
o
f
o
l
l
i
t
r
o
p
h
i
n
p
r
o
t
e
i
n
M
e
t
r
o
d
i
n
H
P
B
r
a
v
e
l
l
e
,
F
e
r
t
i
n
e
x
,
F
e
r
t
i
n
o
r
m
H
u
m
a
n
-
s
o
u
r
c
e
d
F
S
H
I
n
f
e
r
t
i
l
i
t
y
;
f
e
m
a
l
e
1
9
9
8
;
U
S
A
o
p
r
e
l
v
e
k
i
n
p
r
o
t
e
i
n
N
e
u
m
e
g
a
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
n
t
e
r
l
e
u
k
i
n
-
1
1
i
n
c
l
u
d
i
n
g
r
e
s
i
d
u
e
s
2
1
7
8
,
h
u
m
a
n
c
l
o
n
e
p
X
M
/
I
L
-
1
1
,
l
a
c
k
i
n
g
o
r
i
g
i
n
a
l
N
-
t
e
r
m
i
n
a
l
P
r
o
,
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
T
h
r
o
m
b
o
-
c
y
t
o
p
e
n
i
a
2
0
0
1
;
A
u
s
t
r
a
l
i
a
O
P
-
1
p
r
o
t
e
i
n
N
o
v
o
s
O
P
-
1
i
m
p
l
a
n
t
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
o
s
t
e
o
g
e
n
i
c
p
r
o
t
e
i
n
-
1
,
o
r
b
o
n
e
m
o
r
p
h
o
g
e
n
i
c
p
r
o
t
e
i
n
7
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
B
o
n
e
m
o
r
p
h
o
g
e
n
e
s
i
s
1
9
9
9
;
J
a
p
a
n
m
o
b
e
n
a
k
i
n
p
r
o
t
e
i
n
O
c
t
i
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
I
L
-
1
b
e
t
a
w
i
t
h
S
e
r
a
t
7
1
f
r
o
m
h
u
m
a
n
c
l
o
n
e
p
I
L
-
1
-
1
4
A
n
t
i
-
c
a
n
c
e
r
1
9
8
7
;
U
S
A
a
l
p
h
a
-
1
a
n
t
i
t
r
y
p
s
i
n
p
r
o
t
e
i
n
P
r
o
l
a
s
t
i
n
P
r
o
l
a
s
t
i
n
a
,
Z
e
m
a
i
r
a
P
u
r
i
f
i
e
d
h
u
m
a
n
a
l
p
h
a
1
-
a
n
t
i
t
r
y
p
s
i
n
A
n
t
i
-
e
m
p
h
y
s
e
m
i
c
T
a
b
l
e
4
.
5
(
c
o
n
t
i
n
u
e
d
)
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
8
9
;
D
e
n
m
a
r
k
i
n
t
e
r
l
e
u
k
i
n
-
2
p
r
o
t
e
i
n
P
r
o
l
e
u
k
i
n
M
a
c
r
o
l
i
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
I
L
-
2
w
i
t
h
N
-
t
e
r
m
i
n
a
l
A
l
a
a
n
d
C
y
s
1
2
5
m
u
t
a
t
e
d
t
o
S
e
r
,
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
A
n
t
i
-
c
a
n
c
e
r
1
9
9
6
;
E
U
f
o
l
l
i
t
r
o
p
i
n
b
e
t
a
p
r
o
t
e
i
n
P
u
r
e
g
o
n
P
u
r
e
g
o
n
P
e
n
,
F
o
l
l
i
s
t
i
m
,
F
o
l
l
i
s
t
i
m
A
Q
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
F
S
H
i
n
i
t
s
b
e
t
a
g
l
y
c
o
f
o
r
m
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
H
o
r
m
o
n
e
1
9
9
8
;
U
S
A
b
e
c
a
p
l
e
r
m
i
n
p
r
o
t
e
i
n
R
e
g
r
a
n
e
x
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
P
D
G
F
-
B
e
x
p
r
e
s
s
e
d
i
n
S
a
c
c
h
a
r
o
m
y
c
e
s
c
e
r
e
v
i
s
i
a
e
D
i
a
b
e
t
i
c
f
o
o
t
u
l
c
e
r
s
1
9
9
5
;
J
a
p
a
n
m
e
c
a
s
e
r
m
i
n
p
r
o
t
e
i
n
S
o
m
a
z
o
n
R
e
c
o
m
b
i
n
a
n
t
I
G
F
-
1
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
A
n
t
i
-
d
i
a
b
e
t
i
c
2
0
0
1
;
U
S
A
d
r
o
t
r
e
c
o
g
i
n
a
l
f
a
(
a
c
t
i
v
a
t
e
d
)
p
r
o
t
e
i
n
X
i
g
r
i
s
Z
o
v
a
n
t
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
a
c
t
i
v
a
t
e
d
p
r
o
t
e
i
n
C
e
x
p
r
e
s
s
e
d
i
n
a
m
a
m
m
a
l
i
a
n
c
e
l
l
l
i
n
e
A
n
t
i
s
e
p
s
i
s
phogenic protein 2) in the USA in 2002,
both being expressed in CHO cells.
4.1.9
Interferons (Table 4.6)
Aside from the insulin preparations (see
below), the 15 interferons could be consid-
ered one of the earliest attempts to use
biotechnological means (in all of the defi-
nitions of the term) to obtain human pro-
teins for use as drugs. Both recombinant-
derived and natural-sourced materials were
approved for use within a year of each
other, with the recombinant material,
Frone
being launched in Italy in 1985,

closely followed by a natural-sourced mate-
rial, Wellferon
in the UK in 1986. An in-

spection of Table 4.6 shows that this type
of leap-frogging between natural-sourced
and recombinant-sourced materials contin-
ued over the next few years, with some
modified agents being obtained from both
sources.
Recently however, the production of pe-
gylated materials using recombinant pro-
duction of the base molecule followed by
chemical derivatization with polyethylene
glycol has generated agents that are not
only simpler to administer but will also ex-
tend the patent lives of the base agents
(see Part VI, Chapters 1 and 2). Examples
are the relatively recent introduction of Pe-
gintron
in the EU in 2000 and of Pega-

sys
in Switzerland in 2001, both for the

treatment of hepatitis.
4.1.10
Insulins (Table 4.7)
Even though recombinant insulins have
been available since 1982, when Humulin
N
was launched in the USA, inspection

of the 19 insulins shown in Table 4.7, even
as late as 1998, a neutral porcine insulin
was launched in the UK (Pork Atropid
).
However, as can be seen from Table 4.7,
many recombinant versions of human in-
sulin have been launched over the years
(see Part IV, Chapter 13 and Part VI,
Chapter 4), with two new recombinant ver-
sions in 2004. One of these Levemir
was launched in the UK, being a chemical

modification whereby specific lysines have
been acylated with long-chain fatty acids
[7] in order to prolong albumin binding
and thus generate a long-acting insulin.
The other version Apidra
was
launched in the USA, being modified in
the B-chain by substitutions at B3 and
B29.
4.1.11
Hematopoietic Agents (Table 4.8)
The 11 agents listed in Table 4.8 are effec-
tively divided into two proteins, either ery-
thropoietin (EPO) or colony-stimulating
factor (CSF). Although later than insulin
in terms of obtaining them by recombi-
nant techniques, the claim to fame of
EPO-derived materials is that they
launched the most profitable biotechnol-
ogy company, Amgen, and permitted it to
become one of the second level pharma-
ceutical houses and the first in terms of
sales of biotechnology companies. Thus, in
2003, the sales figures for the Amgen-de-
rived EPO-based drugs amounted to over
US$ 9 billion, when the various trade-
names were combined, though these fig-
ures were not all accounted for by Am-
gens sales. Similarly, the CSF-based drugs
just from Amgen amounted to over US$
2.5 billion for the same time period (see
Part VIII, Chapter 3).
As with insulin, longer-acting versions
of both EPO and CSF have been engi-
neered, with Aransep
being the EPO de-

rivative and Neulasta
the pegylated ver-

T
a
b
l
e
4
.
6
I
n
t
e
r
f
e
r
o
n
s
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
9
1
;
U
S
A
i
n
t
e
r
f
e
r
o
n
,
g
a
m
m
a
-
1
b
i
n
t
e
r
f
e
r
o
n
A
c
t
i
m
m
u
n
e
I
m
m
u
k
i
n
,
I
m
u
k
i
n
1
N
a
t
u
r
a
l
s
o
u
r
c
e
d
I
m
m
u
n
o
-
s
t
i
m
u
l
a
n
t
1
9
8
7
;
I
t
a
l
y
i
n
t
e
r
f
e
r
o
n
a
l
f
a
i
n
t
e
r
f
e
r
o
n
A
l
f
a
f
e
r
o
n
e
O
i
f
,
V
e
l
d
o
n
a
N
a
t
u
r
a
l
h
u
m
a
n
l
y
m
p
h
o
b
l
a
s
t
o
i
d
a
l
p
h
a
i
n
t
e
r
f
e
r
o
n
H
e
p
a
t
i
t
i
s
1
9
8
9
;
U
S
A
i
n
t
e
r
f
e
r
o
n
a
l
f
a
-
n
3
i
n
t
e
r
f
e
r
o
n
A
l
f
e
r
o
n
N
A
l
f
e
r
o
n
L
D
O
,
A
l
t
e
m
o
l
,
C
e
l
l
f
e
r
o
n
N
a
t
u
r
a
l
h
u
m
a
n
l
e
u
k
o
c
y
t
e
i
n
t
e
r
f
e
r
o
n
;
a
l
p
h
a
-
n
3
v
a
r
i
a
n
t
w
i
t
h
a
p
p
r
o
x
i
m
a
t
e
l
y
1
6
6
a
m
i
n
o
a
c
i
d
s
w
i
t
h
m
o
l
e
c
u
l
a
r
w
e
i
g
h
t
s
i
n
t
h
e
r
a
n
g
e
1
6
2
7
k
D
a
A
n
t
i
v
i
r
a
l
1
9
9
6
;
U
S
A
i
n
t
e
r
f
e
r
o
n
,
b
e
t
a
-
1
a
i
n
t
e
r
f
e
r
o
n
A
v
o
n
e
x
R
e
b
i
f
R
e
c
o
m
b
i
n
a
n
t
1
6
6
a
m
i
n
o
a
c
i
d
r
e
s
i
d
u
e
s
p
r
o
d
u
c
e
d
b
y
C
H
O
c
e
l
l
s
t
o
g
i
v
e
h
u
m
a
n
i
n
t
e
r
f
e
r
o
n
b
e
t
a
,
w
i
t
h
g
l
y
c
o
s
y
l
a
t
i
o
n
a
t
A
s
n
8
0
g
i
v
i
n
g
a
p
p
r
o
x
1
1
%
o
f
C
/
H
b
y
w
e
i
g
h
t
M
u
l
t
i
p
l
e
s
c
l
e
r
o
s
i
s
1
9
9
3
;
U
S
A
i
n
t
e
r
f
e
r
o
n
,
b
e
t
a
-
s
e
r
i
n
t
e
r
f
e
r
o
n
B
e
t
a
s
e
r
o
n
B
e
t
a
f
e
r
o
n
N
o
n
-
g
l
y
c
o
s
y
l
a
t
e
d
m
o
d
i
f
i
e
d
h
u
m
a
n
i
n
t
e
r
f
e
r
o
n
b
e
t
a
.
C
y
s
1
7
m
u
t
a
t
e
d
t
o
S
e
r
.
P
r
o
d
u
c
e
d
i
n
E
.
c
o
l
i
M
u
l
t
i
p
l
e
s
c
l
e
r
o
s
i
s
1
9
9
2
;
J
a
p
a
n
i
n
t
e
r
f
e
r
o
n
,
g
a
m
m
a
-
1
a
i
n
t
e
r
f
e
r
o
n
B
i
o
g
a
m
m
a
R
e
c
o
m
b
i
n
a
n
t
i
n
t
e
r
f
e
r
o
n
g
a
m
m
a
-
1
a
A
n
t
i
-
c
a
n
c
e
r
1
9
8
5
;
I
t
a
l
y
i
n
t
e
r
f
e
r
o
n
b
e
t
a
i
n
t
e
r
f
e
r
o
n
F
r
o
n
e
F
e
r
o
n
,
R
-
F
r
o
n
e
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
n
t
e
r
f
e
r
o
n
b
e
t
a
H
e
p
a
t
i
t
i
s
1
9
9
7
;
U
S
A
i
n
t
e
r
f
e
r
o
n
a
l
f
a
c
o
n
-
1
i
n
t
e
r
f
e
r
o
n
I
n
f
e
r
g
e
n
A
d
v
a
f
e
r
o
n
,
I
n
f
e
r
a
x
R
e
c
o
m
b
i
n
a
n
t
c
o
n
s
e
n
s
u
s
h
u
m
a
n
l
y
m
p
h
o
b
l
a
s
t
i
n
t
e
r
f
e
r
o
n
a
l
p
h
a
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
A
n
t
i
v
i
r
a
l
T
a
b
l
e
4
.
6
(
c
o
n
t
i
n
u
e
d
)
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
8
6
;
U
S
A
i
n
t
e
r
f
e
r
o
n
a
l
f
a
-
2
b
i
n
t
e
r
f
e
r
o
n
I
n
t
r
o
n
A
A
l
f
a
t
r
o
n
o
l
,
D
e
p
o
I
n
t
e
r
f
e
r
o
n
a
l
p
h
a
,
V
i
r
a
f
e
r
o
n
R
e
c
o
m
b
i
n
a
n
t
I
F
N
-
a
l
p
h
a
2
b
f
r
o
m
t
h
e
c
l
o
n
e
H
i
f
-
S
N
2
0
6
w
i
t
h
s
u
b
s
t
i
t
u
t
i
o
n
a
t
2
3
(
A
r
g
)
a
n
d
3
4
(
H
i
s
)
f
r
o
m
a
n
E
.
c
o
l
i
e
x
p
r
e
s
s
i
o
n
s
y
s
t
e
m
C
a
n
c
e
r
,
h
e
p
a
t
i
t
i
s
1
9
9
7
;
J
a
p
a
n
i
n
t
e
r
f
e
r
o
n
g
a
m
m
a
-
n
1
i
n
t
e
r
f
e
r
o
n
O
G
a
m
m
a
1
0
0
N
o
r
m
a
l
h
u
m
a
n
i
n
t
e
r
f
e
r
o
n
g
a
m
m
a
A
n
t
i
f
u
n
g
a
l
2
0
0
1
;
S
w
i
t
z
e
r
l
a
n
d
p
e
g
i
n
t
e
r
f
e
r
o
n
a
l
f
a
-
2
a
i
n
t
e
r
f
e
r
o
n
P
e
g
a
s
y
s
R
e
c
o
m
b
i
n
a
n
t
i
n
t
e
r
f
e
r
o
n
a
l
f
a
-
2
a
f
r
o
m
a
n
E
.
c
o
l
i
e
x
p
r
e
s
s
i
o
n
s
y
s
t
e
m
l
i
n
k
e
d
t
o
p
o
l
y
e
t
h
y
l
e
n
e
g
l
y
c
o
l
H
e
p
a
t
i
t
i
s
2
0
0
0
;
E
U
p
e
g
i
n
t
e
r
f
e
r
o
n
a
l
f
a
-
2
b
i
n
t
e
r
f
e
r
o
n
P
e
g
i
n
t
r
o
n
V
i
r
a
f
e
r
o
n
P
e
g
,
V
i
r
t
r
o
n
R
e
c
o
m
b
i
n
a
n
t
i
n
t
e
r
f
e
r
o
n
a
l
f
a
-
2
b
f
r
o
m
a
n
E
.
c
o
l
i
e
x
p
r
e
s
s
i
o
n
s
y
s
t
e
m
l
i
n
k
e
d
t
o
p
o
l
y
e
t
h
y
l
e
n
e
g
l
y
c
o
l
H
e
p
a
t
i
t
i
s
1
9
8
9
;
G
e
r
m
a
n
y
i
n
t
e
r
f
e
r
o
n
,
g
a
m
m
a
i
n
t
e
r
f
e
r
o
n
P
o
l
y
f
e
r
o
n
I
m
m
u
n
e
r
o
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
n
t
e
r
f
e
r
o
n
g
a
m
m
a
A
n
t
i
-
i
n
f
l
a
m
-
m
a
t
o
r
y
1
9
8
6
;
U
S
A
i
n
t
e
r
f
e
r
o
n
a
l
f
a
2
a
i
n
t
e
r
f
e
r
o
n
R
o
f
e
r
o
n
-
A
R
o
c
e
r
o
n
-
A
,
A
l
p
h
a
f
e
r
o
n
,
C
a
n
f
e
r
o
n
A
R
e
c
o
m
b
i
n
a
n
t
i
n
t
e
r
f
e
r
o
n
a
l
f
a
2
a
w
i
t
h
L
y
s
2
3
a
n
d
H
i
s
3
4
s
u
b
s
t
i
t
u
t
i
o
n
s
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
A
n
t
i
-
c
a
n
c
e
r
1
9
8
6
;
U
K
i
n
t
e
r
f
e
r
o
n
a
l
f
a
-
n
1
i
n
t
e
r
f
e
r
o
n
W
e
l
l
f
e
r
o
n
S
u
m
i
f
e
r
o
n
M
P
N
a
t
u
r
a
l
h
u
m
a
n
i
n
t
e
r
f
e
r
o
n
a
l
p
h
a
H
e
p
a
t
i
t
i
s
T
a
b
l
e
4
.
7
I
n
s
u
l
i
n
s
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
2
0
0
4
;
U
S
A
i
n
s
u
l
i
n
g
l
u
l
i
s
i
n
e
i
n
s
u
l
i
n
A
p
i
d
r
a
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
n
s
u
l
i
n
w
i
t
h
s
u
b
s
t
i
t
u
t
i
o
n
i
n
t
h
e
B
3
(
L
y
s
)
a
n
d
B
2
9
(
G
l
u
)
p
o
s
i
t
i
o
n
s
D
i
a
b
e
t
e
s
1
9
9
6
;
U
S
A
i
n
s
u
l
i
n
l
i
s
p
r
o
i
n
s
u
l
i
n
H
u
m
a
l
o
g
L
i
p
r
o
l
o
g
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
n
s
u
l
i
n
w
h
e
r
e
t
h
e
B
c
h
a
i
n
a
m
i
n
o
a
c
i
d
s
a
t
2
8
a
n
d
2
9
a
r
e
i
n
v
e
r
t
e
d
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
D
i
a
b
e
t
e
s
1
9
9
9
;
U
S
A
b
i
p
h
a
s
i
c
i
n
s
u
l
i
n
l
i
s
p
r
o
i
n
s
u
l
i
n
H
u
m
a
l
o
g
M
I
X
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
n
s
u
l
i
n
w
h
e
r
e
t
h
e
B
c
h
a
i
n
a
m
i
n
o
a
c
i
d
s
a
t
2
8
a
n
d
2
9
a
r
e
i
n
v
e
r
t
e
d
w
i
t
h
d
i
f
f
e
r
e
n
t
p
r
o
p
o
r
t
i
o
n
s
o
f
a
p
r
o
t
a
m
i
n
e
s
u
s
p
e
n
s
i
o
n
D
i
a
b
e
t
e
s
1
9
8
5
;
U
S
A
h
u
m
a
n
i
n
s
u
l
i
n
Z
n
s
u
s
p
e
n
s
i
o
n
i
n
s
u
l
i
n
H
u
m
u
l
i
n
L
H
u
m
u
l
i
n
L
e
n
t
e
A
m
i
x
t
u
r
e
o
f
c
r
y
s
t
a
l
l
i
n
e
a
n
d
a
m
o
r
p
h
o
u
s
r
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
n
s
u
l
i
n
o
b
t
a
i
n
e
d
f
r
o
m
a
n
E
.
c
o
l
i
e
x
p
r
e
s
s
i
o
n
s
y
s
t
e
m
D
i
a
b
e
t
e
s
1
9
9
2
;
U
S
A
b
i
p
h
a
s
i
c
h
u
m
a
n
i
n
s
u
l
i
n
i
n
s
u
l
i
n
H
u
m
u
l
i
n
M
H
u
m
u
l
i
n
5
0
/
5
0
,
7
0
/
3
0
,
M
1
,
M
2
,
M
3
,
M
4
,
M
5
A
m
i
x
t
u
r
e
o
f
i
s
o
p
h
a
n
e
a
n
d
s
o
l
u
b
l
e
r
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
n
s
u
l
i
n
o
b
t
a
i
n
e
d
f
r
o
m
a
n
E
.
c
o
l
i
e
x
p
r
e
s
s
i
o
n
s
y
s
t
e
m
D
i
a
b
e
t
e
s
1
9
8
2
;
U
S
A
i
s
o
p
h
a
n
e
i
n
s
u
l
i
n
i
n
s
u
l
i
n
H
u
m
u
l
i
n
N
H
u
m
u
l
i
n
I
,
I
n
s
u
m
a
n
b
a
s
a
l
A
m
i
x
t
u
r
e
o
f
c
r
y
s
t
a
l
l
i
n
e
z
i
n
c
r
e
c
o
m
b
i
n
a
n
t
)
h
u
m
a
n
i
n
s
u
l
i
n
o
b
t
a
i
n
e
d
f
r
o
m
a
n
E
.
c
o
l
i
e
x
p
r
e
s
s
i
o
n
s
y
s
t
e
m
a
n
d
p
r
o
t
a
m
i
n
e
s
u
l
f
a
t
e
D
i
a
b
e
t
e
s
1
9
8
5
;
U
K
h
u
m
a
n
i
n
s
u
l
i
n
z
i
n
c
s
u
s
p
e
n
s
i
o
n
i
n
s
u
l
i
n
H
u
m
u
l
i
n
Z
n
H
u
m
u
l
i
n
U
,
U
m
u
l
i
n
e
z
i
n
c
C
r
y
s
t
a
l
l
i
n
e
z
i
n
c
(
r
e
c
o
m
b
i
n
a
n
t
)
h
u
m
a
n
i
n
s
u
l
i
n
o
b
t
a
i
n
e
d
f
r
o
m
a
n
E
.
c
o
l
i
e
x
p
r
e
s
s
i
o
n
s
y
s
t
e
m
D
i
a
b
e
t
e
s
T
a
b
l
e
4
.
7
(
c
o
n
t
i
n
u
e
d
)
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
9
2
;
F
r
a
n
c
e
h
u
m
a
n
n
e
u
t
r
a
l
i
n
s
u
l
i
n
i
n
s
u
l
i
n
I
n
s
u
m
a
n
R
a
p
i
d
H
u
m
u
l
i
n
R
,
H
u
m
u
l
i
n
S
,
I
n
s
u
m
a
n
N
e
u
t
r
a
l
(
r
e
c
o
m
b
i
n
a
n
t
)
h
u
m
a
n
i
n
s
u
l
i
n
o
b
t
a
i
n
e
d
f
r
o
m
a
n
E
.
c
o
l
i
e
x
p
r
e
s
s
i
o
n
s
y
s
t
e
m
D
i
a
b
e
t
e
s
2
0
0
0
;
G
e
r
m
a
n
y
i
n
s
u
l
i
n
g
l
a
r
g
i
n
e
i
n
s
u
l
i
n
L
a
n
t
u
s
O
p
t
i
s
u
l
i
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
n
s
u
l
i
n
w
i
t
h
m
o
d
i
f
i
c
a
t
i
o
n
s
i
n
A
a
n
d
B
c
h
a
i
n
s
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
D
i
a
b
e
t
e
s
2
0
0
4
;
U
K
i
n
s
u
l
i
n
d
e
t
e
r
m
i
r
i
n
s
u
l
i
n
L
e
v
e
m
i
r
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
n
s
u
l
i
n
w
i
t
h
a
c
y
l
a
t
e
d
l
y
s
i
n
e
s
f
o
r
m
e
d
b
y
a
c
y
l
a
t
i
o
n
o
f
e
p
s
i
l
o
n
a
m
i
n
o
g
r
o
u
p
s
o
f
s
p
e
c
i
f
i
c
L
y
s
r
e
s
i
-
d
u
e
s
D
i
a
b
e
t
e
s
2
0
0
2
;
I
r
e
l
a
n
d
b
i
p
h
a
s
i
c
i
n
s
u
l
i
n
a
s
-
p
a
r
t
i
n
s
u
l
i
n
N
o
v
o
M
i
x
3
0
N
o
v
o
m
i
x
5
0
,
N
o
v
o
m
i
x
7
0
,
N
o
v
o
r
a
p
i
d
3
0
M
i
x
M
i
x
t
u
r
e
o
f
r
e
c
o
m
b
i
n
a
n
t
i
n
s
u
l
i
n
s
,
i
n
s
u
l
i
n
a
s
p
a
r
t
a
n
d
i
n
s
u
l
i
n
a
s
p
a
r
t
w
i
t
h
p
r
o
t
a
m
i
n
e
D
i
a
b
e
t
e
s
1
9
9
9
;
U
K
i
n
s
u
l
i
n
a
s
p
a
r
t
i
n
s
u
l
i
n
N
o
v
o
R
a
p
i
d
N
o
v
o
l
o
g
,
N
o
v
o
L
o
g
7
0
/
3
0
F
l
e
x
P
e
n
,
N
o
v
o
L
o
g
F
l
e
x
P
e
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
n
s
u
l
i
n
w
i
t
h
a
n
A
s
p
m
o
d
i
f
i
c
a
t
i
o
n
a
t
p
o
s
i
t
i
o
n
2
8
i
n
t
h
e
B
c
h
a
i
n
e
x
p
r
e
s
s
e
d
i
n
S
a
c
c
h
a
r
o
m
y
c
e
s
c
e
r
e
v
i
s
i
a
e
D
i
a
b
e
t
e
s
1
9
9
1
;
U
S
A
h
u
i
n
s
u
l
i
n
Z
n
s
u
s
p
e
n
s
i
o
n
i
n
s
u
l
i
n
N
o
v
o
l
i
n
R
A
c
t
r
a
p
i
d
,
H
u
m
a
n
A
c
t
r
a
p
i
d
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
n
s
u
l
i
n
e
x
p
r
e
s
s
e
d
i
n
S
a
c
c
h
a
r
o
m
y
c
e
s
c
e
r
e
v
i
s
i
a
e
D
i
a
b
e
t
e
s
1
9
9
1
;
U
S
A
h
u
i
n
s
u
l
i
n
Z
n
s
u
s
p
e
n
s
i
o
n
3
0
/
7
0
i
n
s
u
l
i
n
N
o
v
o
l
i
n
L
M
o
n
o
t
a
r
d
,
H
u
m
a
n
M
o
n
o
t
a
r
d
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
n
s
u
l
i
n
,
z
i
n
c
s
u
s
p
e
n
s
i
o
n
,
m
i
x
t
u
r
e
o
f
p
h
y
s
i
c
a
l
f
o
r
m
s
f
o
l
l
o
w
i
n
g
e
x
p
r
e
s
s
i
o
n
i
n
S
a
c
c
h
a
r
o
m
y
c
e
s
c
e
r
e
v
i
s
i
a
e
D
i
a
b
e
t
e
s
T
a
b
l
e
4
.
7
(
c
o
n
t
i
n
u
e
d
)
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
9
1
;
U
S
A
h
u
i
n
s
u
l
i
n
Z
n
s
u
s
/
p
r
o
t
a
m
i
n
e
i
n
s
u
l
i
n
N
o
v
o
l
i
n
N
H
u
m
a
n
I
n
s
u
l
a
t
a
r
d
,
I
n
s
u
l
a
t
a
r
d
N
P
H
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
n
s
u
l
i
n
e
x
p
r
e
s
s
e
d
i
n
S
a
c
c
h
a
r
o
m
y
c
e
s
c
e
r
e
v
i
s
i
a
e
a
n
d
t
h
e
n
e
u
t
r
a
l
z
i
n
c
s
u
s
p
e
n
s
i
o
n
m
i
x
e
d
w
i
t
h
p
r
o
t
a
m
i
n
e
s
u
l
f
a
t
e
D
i
a
b
e
t
e
s
1
9
9
8
;
U
K
p
o
r
c
i
n
e
n
e
u
t
r
a
l
i
n
s
u
l
i
n
i
n
s
u
l
i
n
P
o
r
k
A
t
r
a
p
i
d
P
o
r
c
i
n
e
n
e
u
t
r
a
l
i
n
s
u
l
i
n
D
i
a
b
e
t
e
s
1
9
8
2
;
U
K
p
o
r
c
i
n
e
i
s
o
p
h
a
n
e
i
n
s
u
l
i
n
i
n
s
u
l
i
n
P
o
r
k
I
n
s
u
l
a
t
a
r
d
P
o
r
c
i
n
e
i
s
o
p
h
a
n
e
i
n
s
u
l
i
n
D
i
a
b
e
t
e
s
1
9
8
2
;
U
K
b
i
p
h
a
s
i
c
p
o
r
c
i
n
e
i
n
s
u
l
i
n
i
n
s
u
l
i
n
P
o
r
k
M
i
x
t
a
r
d
3
0
P
o
r
c
i
n
e
n
e
u
t
r
a
l
a
n
d
i
s
o
p
h
a
n
e
i
n
s
u
l
i
n
D
i
a
b
e
t
e
s
1
9
8
6
;
U
S
A
s
o
l
u
b
l
e
i
n
s
u
l
i
n
i
n
s
u
l
i
n
V
e
l
o
s
u
l
i
n
B
R
H
u
m
a
n
V
e
l
o
s
u
l
i
n
,
O
r
g
a
s
u
l
i
n
e
r
a
p
i
d
e
S
e
m
i
s
y
n
t
h
e
t
i
c
m
o
d
i
f
i
c
a
t
i
o
n
o
f
p
o
r
c
i
n
e
i
n
s
u
l
i
n
D
i
a
b
e
t
e
s
T
a
b
l
e
4
.
8
H
e
m
a
t
o
p
o
i
e
t
i
c
a
g
e
n
t
s
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
2
0
0
1
;
U
S
A
d
a
r
b
e
p
o
i
e
t
i
n
a
l
f
a
h
e
m
a
t
o
p
o
i
e
s
i
s
A
r
a
n
e
s
p
N
e
s
p
r
o
R
e
c
o
m
b
i
n
a
n
t
a
n
a
l
o
g
u
e
o
f
h
u
m
a
n
E
P
O
w
i
t
h
f
i
v
e
a
m
i
n
o
a
c
i
d
s
u
b
s
t
i
t
u
t
i
o
n
s
(
3
0
L
-
A
s
n
,
3
2
L
-
T
h
r
,
8
7
L
-
V
a
l
,
8
8
L
-
A
s
n
,
9
0
L
-
T
h
r
)
g
i
v
i
n
g
f
i
v
e
N
-
l
i
n
k
e
d
C
/
H
c
h
a
i
n
s
,
a
M
W
o
f
3
8
k
D
a
a
n
d
a
5
2
%
C
/
H
c
o
n
t
e
n
t
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
H
e
m
a
t
o
p
o
i
e
s
i
s
2
0
0
2
;
E
U
e
p
o
e
t
i
n
d
e
l
t
a
h
e
m
a
t
o
p
o
i
e
s
i
s
D
y
n
e
p
o
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
E
P
O
e
x
p
r
e
s
s
e
d
u
s
i
n
g
T
r
a
n
s
K
a
y
r
o
t
i
c
T
h
e
r
a
p
i
e
s
g
e
n
e
a
c
t
i
v
a
t
i
o
n
p
r
o
c
e
s
s
.
T
h
i
s
p
r
o
t
e
i
n
,
E
P
O
d
e
l
t
a
,
i
s
c
u
r
r
e
n
t
l
y
h
e
l
d
u
p
b
y
p
a
t
e
n
t
i
n
f
r
i
n
g
e
m
e
n
t
s
u
i
t
s
H
e
m
a
t
o
p
o
i
e
s
i
s
1
9
8
8
;
U
S
A
e
r
y
t
h
r
o
p
o
i
e
t
i
n
a
l
f
a
h
e
m
a
t
o
p
o
i
e
s
i
s
E
p
o
g
e
n
E
p
o
a
d
e
,
E
p
o
p
e
n
,
E
p
r
e
x
,
E
r
y
p
o
,
E
s
p
o
,
G
l
o
b
u
r
e
n
,
P
r
o
c
r
i
t
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
1
1
6
5
E
P
O
u
s
i
n
g
h
u
m
a
n
c
l
o
n
e
l
a
m
b
d
a
H
E
P
O
F
L
1
3
,
t
h
e
g
l
y
c
o
f
o
r
m
i
s
a
l
p
h
a
a
n
d
e
x
p
r
e
s
s
e
d
i
n
a
m
a
m
m
a
l
i
a
n
c
e
l
l
l
i
n
e
H
e
m
a
t
o
p
o
i
e
s
i
s
1
9
9
0
;
J
a
p
a
n
e
r
y
t
h
r
o
p
o
i
e
t
i
n
b
e
t
a
h
e
m
a
t
o
p
o
i
e
s
i
s
E
p
o
g
i
n
M
a
r
o
g
e
n
,
N
e
o
r
e
c
o
r
m
o
n
,
R
e
c
o
r
m
o
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
1
1
6
5
E
P
O
u
s
i
n
g
h
u
m
a
n
c
l
o
n
e
l
a
m
b
d
a
H
E
P
O
F
L
1
3
,
t
h
e
g
l
y
c
o
f
o
r
m
i
s
b
e
t
a
a
n
d
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
H
e
m
a
t
o
p
o
i
e
s
i
s
1
9
9
2
;
S
w
e
d
e
n
m
o
l
g
r
a
m
o
s
t
i
m
h
e
m
a
t
o
p
o
i
e
s
i
s
L
e
u
c
o
m
a
x
R
e
c
o
m
b
i
n
a
n
t
C
S
F
-
2
f
r
o
m
h
u
m
a
n
c
l
o
n
e
p
H
G
2
5
H
e
m
a
t
o
p
o
i
e
s
i
s
1
9
9
1
;
U
S
A
s
a
r
g
r
a
m
o
s
t
i
m
h
e
m
a
t
o
p
o
i
e
s
i
s
L
e
u
k
i
n
e
P
r
o
k
i
n
e
R
e
c
o
m
b
i
n
a
n
t
1
2
7
r
e
s
i
d
u
e
C
S
F
-
2
f
r
o
m
h
u
m
a
n
c
l
o
n
e
p
H
G
2
5
p
r
o
d
u
c
e
d
i
n
S
a
c
c
h
a
r
o
m
y
c
e
s
c
e
r
e
v
i
s
i
a
e
a
n
d
s
u
b
s
t
i
t
u
t
i
n
g
p
o
s
i
t
i
o
n
2
3
H
e
m
a
t
o
p
o
i
e
s
i
s
T
a
b
l
e
4
.
8
(
c
o
n
t
i
n
u
e
d
)
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
9
1
;
J
a
p
a
n
m
i
r
i
m
o
s
t
i
m
h
e
m
a
t
o
p
o
i
e
s
i
s
L
e
u
k
o
p
r
o
l
C
o
s
t
i
l
a
t
e
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
c
l
o
n
e
p
3
A
C
S
F
-
6
9
y
i
e
l
d
i
n
g
a
1
2
1
4
r
e
s
i
d
u
e
p
r
o
t
e
i
n
H
e
m
a
t
o
p
o
i
e
s
i
s
2
0
0
2
;
U
S
A
p
e
g
f
i
l
g
a
s
t
r
i
m
h
e
m
a
t
o
p
o
i
e
s
i
s
N
e
u
l
a
s
t
a
N
e
u
p
o
p
e
g
C
o
v
a
l
e
n
t
c
o
n
j
u
g
a
t
e
o
f
r
e
c
o
m
b
i
n
a
n
t
m
e
t
h
i
o
n
y
l
h
u
m
a
n
G
-
C
S
F
a
n
d
m
o
n
o
m
e
t
h
o
x
y
p
o
l
y
e
t
h
y
l
e
n
e
g
l
y
c
o
l
H
e
m
a
t
o
p
o
i
e
s
i
s
1
9
9
1
;
U
S
A
f
i
l
g
r
a
s
t
i
m
h
e
m
a
t
o
p
o
i
e
s
i
s
N
e
u
p
o
g
e
n
G
r
a
n
,
G
r
a
n
u
l
o
k
i
n
e
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
G
-
C
S
F
,
h
u
m
a
n
c
l
o
n
e
1
0
3
4
w
i
t
h
a
n
e
x
t
r
a
N
-
t
e
r
m
i
n
a
l
M
e
t
,
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
H
e
m
a
t
o
p
o
i
e
s
i
s
1
9
9
1
;
J
a
p
a
n
l
e
n
o
g
a
s
t
r
i
m
h
e
m
a
t
o
p
o
i
e
s
i
s
N
e
u
t
r
o
g
i
n
G
r
a
n
o
c
y
t
e
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
G
-
C
S
F
,
h
u
m
a
n
c
l
o
n
e
1
0
3
4
g
l
y
c
o
s
y
l
a
t
e
d
H
e
m
a
t
o
p
o
i
e
s
i
s
1
9
9
4
;
J
a
p
a
n
n
a
r
t
o
g
r
a
s
t
i
m
h
e
m
a
t
o
p
o
i
e
s
i
s
N
e
u
-
U
p
R
e
c
o
m
b
i
n
a
n
t
G
-
C
S
F
f
r
o
m
h
u
m
a
n
c
l
o
n
e
1
0
3
4
w
h
e
r
e
t
h
e
N
-
t
e
r
m
i
n
a
l
a
m
i
n
o
a
c
i
d
s
h
a
v
e
b
e
e
n
m
u
t
a
t
e
d
H
e
m
a
t
o
p
o
i
e
s
i
s
sion of CSF. There is one interesting mod-
ification of EPO known as Dynepo
, an
epoetin delta expressed using a phage dis-
play technique that was approved in the
EU in 2002 but which, at the time of writ-
ing, has been held up by patent infringe-
ment suits in both the EU and the US
(see Part V, Chapter 2).
4.1.12
Enzymes (Table 4.9)
These nine agents range from Adagen
,
which is bovine-derived adenosine deami-
nase treated with polyethylene glycol and
introduced in 1990 in the USA, through
human placental beta-glycocerebroside
with a chemical remodeling of the carbo-
hydrate side chains (Ceredase
), launched
in the USA in 1991, to its recombinant de-
rivative, Cerezyme
launched in 1994 also

in the USA. However, of significant impor-
tance in 2001 were the launches of two re-
combinant alpha-galactosidases which dif-
fered in their glycoforms, as treatments
for Fabrys disease. These were Fabrazy-
me
, launched in Germany, and Repla-

gal
, launched in Sweden. With the prob-

lems associated with gene therapies for
treatment of disease where enzymes are
lacking, it is probable that more enzymes
designed to ameliorate such diseases will
be forthcoming in the near future.
4.1.13
Peptidic Agents (Table 4.10)
Of the 20 compounds shown in Table
4.10, all but five are produced by biotech-
nological means, predominantly expres-
sion in E. coli or in Saccharomyces cerevisiae
(see Part IV, Chapters 12 and 13). The
others are relatively simple peptides that
are made synthetically, but are based upon
the parental natural product. An inspec-
tion of Table 4.10 shows that a significant
number of the recombinant proteins are
variations on human growth hormone,
with some elimination/substitution of
amino acid residues, as can be seen in
comparison of Geref
with Nutropin
, or
alternatively, esterified with polyethylene
glycol as in Somavert
.
The remainder of the compounds cover
a wide range of activities, from parts of
human parathyroid hormone through hu-
man chorionic gonadotropin (hCG) to the
very interesting congestive heart failure
treatment with a portion of the human B-
type natriuretic peptide, Natrecor
.
4.1.14
Anti-tumor Agents, Direct and Indirect
(Table 4.11)
In Table 4.11, we have listed those agents
that have been used as anti-tumor agents
to date. Of the 18 agents, all arranged as
in the other tables alphabetically by initial
trade name, seven are antibodies in their
own right (e.g., Avastin
, Campath
,
Erbitux
, Herceptin
) (see Part I, Chap-

ter 5), and two are examples of the magic
bullet concept in one way or another, with
Mylotarg
being an example of a very po-

tent anti-tumor antibiotic being linked to a
targeting MAb. In contrast, Ontak
is a
fusion protein of the receptor domain of
diphtheria toxin linked to interleukin-2
(IL-2). What is also of immediate impor-
tance is the recent (2002 and 2004) ap-
proval of two radiolabeled MAb-based
therapies (of the seven MAbs), that rely on
treatment with both labeled and unlabeled
antibodies in sequence.
It should also be noted that, as with
small molecules, there is now the begin-
ning of a trend towards generating anti-
bodies directed towards signal transduc-
tion pathways, with the first example
T
a
b
l
e
4
.
9
E
n
z
y
m
e
s
.
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
9
0
;
U
S
A
p
e
g
a
d
e
m
a
s
e
b
o
v
i
n
e
e
n
z
y
m
e
A
d
a
g
e
n
R
e
a
c
t
i
o
n
p
r
o
d
u
c
t
o
f
r
e
c
o
m
b
i
n
a
n
t
a
d
e
n
o
s
i
n
e
d
e
a
m
i
n
a
s
e
w
i
t
h
s
u
c
c
i
n
i
c
a
n
h
y
d
r
i
d
e
a
n
d
s
u
b
s
e
q
u
e
n
t
e
s
t
e
r
i
f
i
c
a
t
i
o
n
w
i
t
h
p
o
l
y
e
t
h
y
l
e
n
e
g
l
y
c
o
l
m
o
n
o
m
e
t
h
y
l
e
t
h
e
r
I
m
m
u
n
o
-
s
t
i
m
u
l
a
n
t
2
0
0
3
;
U
S
A
l
a
r
o
n
i
d
a
s
e
e
n
z
y
m
e
A
l
d
u
r
a
z
y
m
e
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
8
-
L
-
h
i
s
t
i
d
i
n
e
a
l
p
h
a
-
L
-
i
d
u
r
o
n
i
d
a
s
e
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
P
o
l
y
s
a
c
c
h
a
r
i
d
o
-
s
i
s
1
9
9
1
;
U
S
A
a
l
g
l
u
c
e
r
a
s
e
e
n
z
y
m
e
C
e
r
e
d
a
s
e
H
u
m
a
n
p
l
a
c
e
n
t
a
l
t
i
s
s
u
e
b
e
t
a
g
l
y
c
o
-
c
e
r
e
b
r
o
s
i
d
a
s
e
t
h
a
t
i
s
c
a
r
b
o
h
y
d
r
a
t
e
r
e
m
o
d
e
l
e
d
G
a
u
c
h
e
r
s
d
i
s
e
a
s
e
1
9
9
4
;
U
S
A
i
m
i
g
l
u
c
e
r
a
s
e
e
n
z
y
m
e
C
e
r
e
z
y
m
e
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
b
e
t
a
g
l
y
c
o
-
c
e
r
e
b
r
o
s
i
d
a
s
e
f
r
o
m
C
H
O
c
e
l
l
s
,
4
9
5
c
h
a
n
g
e
d
t
o
L
-
H
i
s
a
n
d
m
o
d
i
f
i
c
a
t
i
o
n
s
m
a
d
e
t
o
C
/
H
c
h
a
i
n
s
a
t
1
9
a
n
d
1
4
6
G
a
u
c
h
e
r
s
d
i
s
e
a
s
e
2
0
0
1
;
G
e
r
m
a
n
y
a
g
a
l
s
i
d
a
s
e
b
e
t
a
e
n
z
y
m
e
F
a
b
r
a
z
y
m
e
F
a
b
r
a
s
e
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
a
l
p
h
a
-
g
a
l
a
c
-
t
o
s
i
d
a
s
e
c
l
o
n
e
l
a
m
b
d
a
A
G
1
8
,
g
l
y
c
o
f
o
r
m
b
e
t
a
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
F
a
b
r
y
d
i
s
e
a
s
e
2
0
0
1
;
G
e
r
m
a
n
y
r
a
s
b
u
r
i
c
a
s
e
e
n
z
y
m
e
F
a
s
t
u
r
t
e
c
E
l
i
t
e
k
R
e
c
o
m
b
i
n
a
n
t
u
r
a
t
e
o
x
i
d
a
s
e
p
r
o
d
u
c
e
d
i
n
a
S
a
c
c
h
a
r
o
m
y
c
e
s
c
e
r
e
v
i
s
i
a
e
f
e
r
m
e
n
t
a
t
i
o
n
A
n
t
i
-
h
y
p
e
r
u
r
i
c
e
m
i
a
1
9
9
4
;
U
S
A
p
e
g
a
s
p
a
r
g
a
s
e
e
n
z
y
m
e
O
n
c
a
s
p
a
r
A
s
p
a
r
a
g
i
n
a
s
e
c
o
u
p
l
e
d
t
o
p
o
l
y
e
t
h
y
l
e
n
e
g
l
y
c
o
l
A
n
t
i
-
c
a
n
c
e
r
1
9
9
4
;
U
S
A
d
o
r
n
a
s
e
a
l
f
a
e
n
z
y
m
e
P
u
l
m
o
z
y
m
e
A
E
R
x
-
d
o
r
n
a
s
e
a
l
f
a
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
d
e
o
x
y
r
i
b
o
-
n
u
c
l
e
a
s
e
1
,
h
u
m
a
n
c
l
o
n
e
1
8
-
1
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
C
y
s
t
i
c
f
i
b
r
o
s
i
s
2
0
0
1
;
S
w
e
d
e
n
a
g
a
l
s
i
d
a
s
e
a
l
f
a
e
n
z
y
m
e
R
e
p
l
a
g
a
l
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
a
l
p
h
a
-
g
a
l
a
c
t
o
s
i
d
a
s
e
c
l
o
n
e
l
a
m
b
d
a
A
G
1
8
,
g
l
y
c
o
f
o
r
m
a
l
p
h
a
p
r
o
d
u
c
e
d
f
r
o
m
a
m
a
m
m
a
l
i
a
n
c
e
l
l
l
i
n
e
F
a
b
r
y
d
i
s
e
a
s
e
T
a
b
l
e
4
.
1
0
P
e
p
t
i
d
i
c
a
g
e
n
t
s
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
9
5
;
F
r
a
n
c
e
h
G
H
p
e
p
t
i
d
e
B
i
o
T
r
o
p
i
n
G
r
o
w
j
e
c
t
,
S
c
i
T
r
o
p
i
n
,
T
e
v
-
T
r
o
p
i
n
,
Z
o
m
a
c
t
o
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
g
r
o
w
t
h
h
o
r
m
o
n
e
H
o
r
m
o
n
e
(
d
w
a
r
f
i
s
m
)
1
9
9
9
;
E
U
c
e
t
r
o
r
e
l
i
x
p
e
p
t
i
d
e
C
e
t
r
o
t
i
d
e
S
y
n
t
h
e
t
i
c
L
H
R
H
a
n
t
a
g
o
n
i
s
t
H
o
r
m
o
n
e
1
9
8
5
;
S
p
a
i
n
e
l
c
a
t
o
n
i
n
p
e
p
t
i
d
e
D
i
a
t
i
n
E
l
c
i
t
o
n
i
n
I
n
j
.
2
0
S
,
T
u
r
b
o
c
a
l
c
i
n
s
p
r
a
y
E
s
t
e
r
i
f
i
e
d
e
e
l
c
a
l
c
i
t
o
n
i
n
C
a
l
c
i
u
m
m
e
t
a
b
o
l
i
s
m
2
0
0
2
;
U
S
A
t
e
r
i
p
a
r
a
t
i
d
e
p
e
p
t
i
d
e
F
o
r
t
e
o
F
o
s
t
e
o
R
e
c
o
m
b
i
n
a
n
t
1
3
4
r
e
s
i
d
u
e
s
o
f
h
u
m
a
n
p
a
r
a
t
h
y
r
o
i
d
h
o
r
m
o
n
e
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
H
o
r
m
o
n
e
1
9
9
8
;
U
S
A
s
e
r
m
o
r
e
l
i
n
a
c
e
t
a
t
e
p
e
p
t
i
d
e
G
e
r
e
f
R
e
c
o
m
b
i
n
a
n
t
1
2
9
r
e
s
i
d
u
e
s
o
f
h
u
m
a
n
g
r
o
w
t
h
h
o
r
m
o
n
e
-
r
e
l
e
a
s
i
n
g
h
o
r
m
o
n
e
H
o
r
m
o
n
e
1
9
9
3
;
D
e
n
m
a
r
k
g
l
u
c
a
g
o
n
r
D
N
A
p
e
p
t
i
d
e
G
l
u
c
a
G
e
n
G
l
u
c
a
g
o
n
G
N
o
v
o
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
g
l
u
c
a
g
o
n
e
x
p
r
e
s
s
e
d
i
n
S
a
c
c
h
a
r
o
m
y
c
e
s
c
e
r
e
v
i
s
i
a
e
H
o
r
m
o
n
e
1
9
8
7
;
U
S
A
s
o
m
a
t
r
o
p
i
n
p
e
p
t
i
d
e
H
u
m
a
t
r
o
p
e
C
r
e
s
c
o
r
m
o
n
,
G
e
n
o
t
r
o
p
i
n
,
G
e
n
o
t
r
o
p
i
n
M
i
n
i
q
u
i
c
k
,
U
m
a
t
r
o
p
e
,
G
e
n
e
t
o
n
o
r
m
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
g
r
o
w
t
h
h
o
r
m
o
n
e
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
H
o
r
m
o
n
e
2
0
0
1
;
U
S
A
n
e
s
i
r
i
t
i
d
e
p
e
p
t
i
d
e
N
a
t
r
e
c
o
r
N
o
r
a
t
a
k
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
B
-
t
y
p
e
n
a
t
r
i
u
r
e
t
i
c
p
e
p
t
i
d
e
,
r
e
s
i
d
u
e
s
1
3
2
C
o
n
g
e
s
t
i
v
e
h
e
a
r
t
f
a
i
l
u
r
e
T
a
b
l
e
4
.
1
0
(
c
o
n
t
i
n
u
e
d
)
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
9
7
;
U
S
A
s
o
m
a
t
r
o
p
i
n
p
e
p
t
i
d
e
N
o
r
d
i
t
r
o
p
i
n
N
o
r
d
i
t
r
o
p
i
n
N
o
r
d
i
f
l
e
x
,
N
o
r
d
i
t
r
o
p
i
n
S
i
m
p
l
e
X
x
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
g
r
o
w
t
h
h
o
r
m
o
n
e
H
o
r
m
o
n
e
(
d
w
a
r
f
i
s
m
)
1
9
9
4
;
U
S
A
s
o
m
a
t
o
t
r
o
p
i
n
p
e
p
t
i
d
e
N
u
t
r
o
p
i
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
g
r
o
w
t
h
h
o
r
m
o
n
e
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
H
o
r
m
o
n
e
1
9
9
6
;
U
S
A
s
o
m
a
t
r
o
p
i
n
r
e
c
o
m
b
i
n
a
n
t
p
e
p
t
i
d
e
N
u
t
r
o
p
i
n
A
Q
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
g
r
o
w
t
h
h
o
r
m
o
n
e
a
q
u
e
o
u
s
f
o
r
m
u
l
a
t
i
o
n
H
o
r
m
o
n
e
2
0
0
0
;
U
S
A
S
M
P
-
1
4
0
p
e
p
t
i
d
e
N
u
t
r
o
p
i
n
D
e
p
o
t
P
r
o
L
e
a
s
e
-
r
h
G
H
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
g
r
o
w
t
h
h
o
r
m
o
n
e
u
s
i
n
g
P
r
o
r
e
l
e
a
s
e
m
i
c
r
o
e
n
c
a
p
s
u
l
a
t
i
o
n
,
b
u
t
w
i
t
h
d
r
a
w
n
i
n
2
0
0
4
H
o
r
m
o
n
e
(
d
w
a
r
f
i
s
m
)
1
9
8
5
;
I
t
a
l
y
s
a
l
m
o
n
c
a
l
c
i
t
o
n
i
n
p
e
p
t
i
d
e
O
s
t
e
o
t
o
n
i
n
a
C
a
l
o
g
e
n
,
F
o
r
c
a
l
c
i
t
o
n
i
n
,
M
i
a
c
a
l
c
i
c
,
M
i
a
c
a
l
c
i
c
N
S
,
S
i
c
a
l
,
T
o
n
o
c
a
l
c
i
n
,
F
o
r
t
i
c
a
l
R
e
c
o
m
b
i
n
a
n
t
s
a
l
m
o
n
c
a
l
c
i
t
o
n
i
n
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
H
o
r
m
o
n
e
2
0
0
1
;
I
t
a
l
y
c
h
o
r
i
o
g
o
n
a
d
o
-
t
r
o
p
i
n
a
l
f
a
p
e
p
t
i
d
e
O
v
i
d
r
e
l
O
v
i
d
r
e
l
l
e
,
O
v
i
t
r
e
l
l
e
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
c
h
o
r
i
o
g
o
n
a
d
o
-
t
r
o
p
i
n
e
x
p
r
e
s
s
e
d
i
n
C
H
O
c
e
l
l
s
H
o
r
m
o
n
e
1
9
8
5
;
U
S
A
s
o
m
a
t
r
e
m
p
e
p
t
i
d
e
P
r
o
t
r
o
p
i
n
S
o
m
a
t
o
n
o
r
m
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
g
r
o
w
t
h
h
o
r
m
o
n
e
w
i
t
h
a
n
e
x
t
r
a
N
-
t
e
r
m
i
n
a
l
M
e
t
.
E
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
H
o
r
m
o
n
e
1
9
8
9
;
U
K
r
e
c
D
N
A
s
o
m
a
t
o
t
r
o
p
i
n
p
e
p
t
i
d
e
S
a
i
z
e
n
S
e
r
o
t
i
m
,
Z
o
r
b
t
i
v
e
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
g
r
o
w
t
h
h
o
r
m
o
n
e
H
o
r
m
o
n
e
T
a
b
l
e
4
.
1
0
(
c
o
n
t
i
n
u
e
d
)
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
8
8
;
S
w
i
t
z
e
r
l
a
n
d
o
c
t
r
e
o
t
i
d
e
p
e
p
t
i
d
e
S
a
n
d
o
s
t
a
t
i
n
L
o
n
g
a
s
t
a
t
i
n
L
A
R
,
L
o
n
g
a
s
t
a
t
i
n
a
,
O
n
c
o
L
A
R
,
S
a
n
d
o
s
t
a
t
i
n
L
A
R
S
y
n
t
h
e
t
i
c
g
r
o
w
t
h
h
o
r
m
o
n
e
r
e
l
e
a
s
e
i
n
h
i
b
i
t
o
r
A
n
t
i
-
s
e
c
r
e
t
o
r
y
2
0
0
3
;
U
S
A
p
e
g
v
i
s
o
m
a
n
t
p
e
p
t
i
d
e
S
o
m
a
v
e
r
t
T
r
o
v
e
r
t
R
e
c
o
m
b
i
n
a
n
t
m
o
d
i
f
i
e
d
h
u
m
a
n
g
r
o
w
t
h
h
o
r
m
o
n
e
(
a
m
i
n
o
a
c
i
d
s
u
b
s
t
i
t
u
t
i
o
n
s
)
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
a
n
d
l
i
n
k
e
d
t
o
p
o
l
y
e
t
h
y
l
e
n
e
g
l
y
c
o
l
A
c
r
o
m
e
g
a
l
y
1
9
9
6
;
C
h
i
n
a
t
h
y
m
a
l
f
a
s
i
n
p
e
p
t
i
d
e
Z
a
d
a
x
i
n
S
y
n
t
h
e
t
i
c
v
e
r
s
i
o
n
o
f
a
n
a
t
u
r
a
l
p
r
o
d
u
c
t
H
e
p
a
t
i
t
i
s
1
9
8
7
;
U
K
g
o
s
e
r
e
l
i
n
p
e
p
t
i
d
e
Z
o
l
a
d
e
x
S
y
n
t
h
e
t
i
c
s
m
a
l
l
p
e
p
t
i
d
e
L
H
R
H
a
g
o
n
i
s
t
H
o
r
m
o
n
e
T
a
b
l
e
4
.
1
1
A
n
t
i
t
u
m
o
r
a
g
e
n
t
s
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
2
0
0
4
;
U
S
A
b
e
v
a
c
i
z
u
m
a
b
a
n
t
i
b
o
d
y
A
v
a
s
t
i
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
i
z
e
d
m
u
r
i
n
e
b
a
s
e
d
M
A
b
t
h
a
t
i
s
s
p
e
c
i
f
i
c
f
o
r
V
E
G
F
A
n
t
i
-
c
a
n
c
e
r
1
9
9
9
;
G
e
r
m
a
n
y
t
a
s
o
n
e
r
m
i
n
p
r
o
t
e
i
n
B
e
r
o
m
u
n
R
e
s
i
d
u
e
s
1
1
5
7
o
f
h
u
m
a
n
T
N
F
-
a
l
p
h
a
.
R
e
c
o
m
b
i
n
a
n
t
e
x
p
r
e
s
s
i
o
n
i
n
E
.
c
o
l
i
A
n
t
i
-
c
a
n
c
e
r
2
0
0
3
;
U
S
A
t
o
s
i
t
u
m
o
m
a
b
a
n
t
i
b
o
d
y
B
e
x
x
a
r
A
n
t
i
-
C
D
2
0
m
o
u
s
e
M
A
b
l
a
b
e
l
e
d
w
i
t
h
1
3
1
I
p
r
o
d
u
c
e
d
i
n
a
n
a
n
t
i
b
i
o
t
i
c
-
f
r
e
e
m
a
m
m
a
l
i
a
n
c
e
l
l
l
i
n
e
.
T
r
e
a
t
m
e
n
t
i
n
v
o
l
v
e
s
b
o
t
h
f
r
e
e
a
n
d
l
a
b
e
l
e
d
M
A
b
s
N
H
L
1
9
9
2
;
J
a
p
a
n
i
n
t
e
r
f
e
r
o
n
,
g
a
m
m
a
-
1
a
i
n
t
e
r
f
e
r
o
n
B
i
o
g
a
m
m
a
R
e
c
o
m
b
i
n
a
n
t
i
n
t
e
r
f
e
r
o
n
g
a
m
m
a
-
1
a
A
n
t
i
-
c
a
n
c
e
r
2
0
0
1
;
U
S
A
a
l
e
m
t
u
z
u
m
a
b
a
n
t
i
b
o
d
y
C
a
m
p
a
t
h
M
a
b
C
a
m
p
a
t
h
H
u
m
a
n
i
z
e
d
M
A
b
t
h
a
t
r
e
c
o
g
n
i
z
e
s
t
h
e
C
D
w
5
2
a
n
t
i
g
e
n
o
n
n
o
r
m
a
l
a
n
d
m
a
l
i
g
n
a
n
t
B
a
n
d
T
c
e
l
l
s
A
n
t
i
-
c
a
n
c
e
r
1
9
9
2
;
J
a
p
a
n
c
e
l
m
o
l
e
u
k
i
n
p
r
o
t
e
i
n
C
e
l
e
u
k
R
e
c
o
m
b
i
n
a
n
t
i
n
t
e
r
l
e
u
k
i
n
2
(
I
L
-
2
)
f
r
o
m
t
h
e
h
u
m
a
n
c
l
o
n
e
p
T
I
L
2
-
2
1
a
A
n
t
i
-
c
a
n
c
e
r
2
0
0
3
;
S
w
i
t
z
e
r
l
a
n
d
c
e
t
u
x
i
m
a
b
a
n
t
i
b
o
d
y
E
r
b
i
t
u
x
H
u
m
a
n
i
z
e
d
m
u
r
i
n
e
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
E
G
F
R
.
P
r
o
d
u
c
e
d
i
n
a
m
u
r
i
n
e
m
y
e
l
o
m
a
c
u
l
t
u
r
e
i
n
s
e
r
u
m
-
f
r
e
e
m
e
d
i
a
a
n
d
p
u
r
i
f
i
e
d
b
y
a
f
f
i
n
i
t
y
,
i
o
n
-
e
x
c
h
a
n
g
e
a
n
d
g
e
l
f
i
l
t
r
a
t
i
o
n
c
h
r
o
m
a
t
o
g
r
a
p
h
i
e
s
A
n
t
i
-
c
a
n
c
e
r
1
9
9
8
;
U
S
A
t
r
a
s
t
u
z
u
m
a
b
a
n
t
i
b
o
d
y
H
e
r
c
e
p
t
i
n
I
g
G
1
(
h
u
m
a
n
-
m
o
u
s
e
M
A
b
r
h
u
M
a
b
H
E
R
2
g
a
m
m
a
1
c
h
a
i
n
a
n
t
i
h
u
m
a
n
p
1
8
5
c
-
e
r
b
B
2
r
e
c
e
p
t
o
r
)
A
n
t
i
-
c
a
n
c
e
r
1
9
9
2
;
J
a
p
a
n
t
e
c
e
l
e
u
k
i
n
p
r
o
t
e
i
n
I
m
u
n
a
c
e
B
i
o
l
e
u
k
i
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
m
e
t
I
L
-
2
A
n
t
i
-
c
a
n
c
e
r
T
a
b
l
e
4
.
1
1
(
c
o
n
t
i
n
u
e
d
)
D
a
t
e
;
C
o
u
n
t
r
y
G
e
n
e
r
i
c
n
a
m
e
B
a
s
e
c
l
a
s
s
T
r
a
d
e
n
a
m
e
O
t
h
e
r
t
r
a
d
e
n
a
m
e
s
M
a
n
u
f
a
c
t
u
r
i
n
g
m
e
t
h
o
d
s
/
S
o
u
r
c
e
s
/
T
r
e
a
t
m
e
n
t
s
D
i
s
e
a
s
e
i
n
d
i
c
a
t
i
o
n
1
9
8
6
;
U
S
A
i
n
t
e
r
f
e
r
o
n
a
l
f
a
-
2
b
i
n
t
e
r
f
e
r
o
n
I
n
t
r
o
n
A
A
l
f
a
t
r
o
n
o
l
,
D
e
p
o
I
n
t
e
r
f
e
r
o
n
a
l
p
h
a
,
V
i
r
a
f
e
r
o
n
R
e
c
o
m
b
i
n
a
n
t
I
F
N
-
a
l
p
h
a
2
b
f
r
o
m
t
h
e
c
l
o
n
e
H
i
f
-
S
N
2
0
6
w
i
t
h
s
u
b
s
t
i
t
u
t
i
o
n
a
t
2
3
(
A
r
g
)
a
n
d
3
4
(
H
i
s
)
f
r
o
m
a
n
E
.
c
o
l
i
e
x
p
r
e
s
s
i
o
n
s
y
s
t
e
m
A
n
i
t
-
c
a
n
c
e
r
,
h
e
p
a
t
i
t
i
s
2
0
0
0
;
U
S
A
g
e
m
t
u
z
u
m
a
b
o
z
o
g
a
m
i
c
i
n
a
n
t
i
b
o
d
y
M
y
l
o
t
a
r
g
H
u
m
a
n
i
z
e
d
I
g
G
4
a
n
t
i
C
D
3
3
a
n
t
i
b
o
d
y
h
P
6
7
.
6
c
o
n
j
u
g
a
t
e
d
t
o
N
-
a
c
e
t
y
l
-
g
a
m
m
a
c
a
l
i
c
h
e
a
m
i
c
i
n
A
n
t
i
-
c
a
n
c
e
r
1
9
9
9
;
J
a
p
a
n
m
o
b
e
n
a
k
i
n
p
r
o
t
e
i
n
O
c
t
i
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
I
L
-
1
b
e
t
a
w
i
t
h
S
e
r
a
t
7
1
f
r
o
m
h
u
m
a
n
c
l
o
n
e
p
I
L
-
1
-
1
4
A
n
t
i
-
c
a
n
c
e
r
1
9
9
4
;
U
S
A
p
e
g
a
s
p
a
r
g
a
s
e
e
n
z
y
m
e
O
n
c
a
s
p
a
r
A
s
p
a
r
a
g
i
n
a
s
e
c
o
u
p
l
e
d
t
o
p
o
l
y
e
t
h
y
l
e
n
e
g
l
y
c
o
l
A
n
t
i
-
c
a
n
c
e
r
1
9
9
9
;
U
S
A
d
e
n
i
l
e
u
k
i
n
d
i
f
t
i
t
o
x
a
n
t
i
b
o
d
y
O
n
t
a
k
O
n
z
a
r
F
u
s
i
o
n
p
r
o
t
e
i
n
o
f
t
h
e
r
e
c
e
p
t
o
r
b
i
n
d
i
n
g
d
o
m
a
i
n
o
f
d
i
p
h
t
h
e
r
i
a
t
o
x
i
n
c
o
u
p
l
e
d
t
o
I
L
-
2
A
n
t
i
-
c
a
n
c
e
r
1
9
8
9
;
D
e
n
m
a
r
k
i
n
t
e
r
l
e
u
k
i
n
-
2
p
r
o
t
e
i
n
P
r
o
l
e
u
k
i
n
M
a
c
r
o
l
i
n
R
e
c
o
m
b
i
n
a
n
t
h
u
m
a
n
I
L
-
2
w
i
t
h
N
-
t
e
r
m
i
n
a
l
A
l
a
a
n
d
C
y
s
1
2
5
m
u
t
a
t
e
d
t
o
S
e
r
,
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
A
n
t
i
-
c
a
n
c
e
r
1
9
9
7
;
U
S
A
r
i
t
u
x
i
m
a
b
a
n
t
i
b
o
d
y
R
i
t
u
x
a
n
M
a
b
T
h
e
r
a
H
u
m
a
n
i
z
e
d
m
u
r
i
n
e
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
C
D
2
0
A
n
t
i
-
c
a
n
c
e
r
1
9
8
6
;
U
S
A
i
n
t
e
r
f
e
r
o
n
a
l
f
a
2
a
i
n
t
e
r
f
e
r
o
n
R
o
f
e
r
o
n
-
A
R
o
c
e
r
o
n
-
A
,
A
l
p
h
a
f
e
r
o
n
,
C
a
n
f
e
r
o
n
A
R
e
c
o
m
b
i
n
a
n
t
i
n
t
e
r
f
e
r
o
n
a
l
f
a
2
a
w
i
t
h
L
y
s
2
3
a
n
d
H
i
s
3
4
s
u
b
s
t
i
t
u
t
i
o
n
s
e
x
p
r
e
s
s
e
d
i
n
E
.
c
o
l
i
A
n
t
i
-
c
a
n
c
e
r
2
0
0
2
;
U
S
A
i
b
r
i
t
u
m
o
m
a
b
t
i
u
e
t
a
n
a
n
t
i
b
o
d
y
Z
e
v
a
l
i
n
I
n
-
1
1
1
Z
e
v
a
l
i
n
,
Y
-
9
0
Z
e
v
a
l
i
n
H
u
m
a
n
i
z
e
d
M
A
b
d
i
r
e
c
t
e
d
a
g
a
i
n
s
t
C
D
2
0
,
p
r
o
d
u
c
e
d
i
n
C
H
O
c
e
l
l
s
a
n
d
i
s
p
a
r
t
o
f
a
f
o
u
r
-
p
a
r
t
t
r
e
a
t
m
e
n
t
w
i
t
h
R
i
t
u
x
a
n
a
n
d
t
h
e
r
a
d
i
o
l
a
b
e
l
e
d
Z
e
v
a
l
i
n
s
A
n
t
i
-
c
a
n
c
e
r
being Herceptin
(see Part I, Chapter 5)

directed against the HER/2-neu receptor,
and the most recent one being Avastin
,
directed towards VEGF.
Among the remainder, four of the five
proteins are variants on interleukins, with
the fifth (Beromun
) being a recombinant
shortened version of tumor necrosis factor-
alpha (TNF-alpha). Only one enzyme is
listed; this is a pegylated version of an old
treatment, where asparaginase (see Part II,
Chapter 6) has been coupled to polyethy-
lene glycol for better pharmacodynamics.
The remainder of the 18 are composed of
three recombinant interferons, including
two of the earliest.
4.1.15
Concluding Comments
As can be seen from an inspection of Tables
4.14.11, these 159 agents are produced by a
wide range of methods, from total synthesis
in the case of small peptides with non-ribo-
somally expressed amino acids, through iso-
lation of proteins from a variety of organ-
isms both prokaryotic and eukaryotic, to ex-
pression systems ranging from E. coli and
S. cerevisiae to CHO and BHK cells in both
regular and serum-free media. In addition,
a number of earlier agents have been mod-
ified by treatment with polyethylene glycol
(see Part VI, Chapter 2) or with other chem-
ical agents such as long-chain fatty acids in
order to alter their pharmacodynamics in a
favorable manner (see Part VII, Chapter 1).
Such modifications will possibly become
more usual in the future, as we will work
on other human peptides and small pro-
teins, and as the knowledge of cellular me-
tabolism and methods of delivery increases
in later years (see Part VI, Chapter 1).
4.2
Potential Agents from Non-mammalian
Sources as Leads to Novel Therapies
4.2.1
Introduction
Bioactive proteins from natural product ex-
tracts have been isolated based on a broad
spectrum of potentially useful characteris-
tics, including anti-cancer, anti-fungal,
anti-viral and anti-bacterial activities [8].
These proteins possess a wide variety of
structural and functional motifs, but can
be grouped into particular classes based
on similarities in their three-dimensional
structural or primary amino acid se-
quence. Recently, many reviews have been
published which closely examine defined
groups of these proteins, including defen-
sins [9], ribosome-inactivating proteins
[10], cyclotides [11], and antimicrobial pep-
tides from food proteins [12].
Here, we will endeavor to review many
of the recent discoveries of unique bioac-
tive proteins derived from natural sources
such as plants, marine invertebrates, and
insects. Rather than conduct an in-depth
study of an individual structural group of
these proteins, we will survey the diversity
of these proteins and organize them based
on their biological activity. We will not dis-
cuss mammalian bioactive proteins, as
these have recently been reviewed exten-
sively elsewhere [1316]. Instead, we will
concentrate on those proteins from natural
sources that have demonstrated activity in
heterologous systems (i.e., plant proteins
active against human disease).
4.2.1
Anti-cancer Proteins
Although many different non-ribosomal
cyclic peptides with anti-cancer activity
4.2 Potential Agents from Non-mammalian Sources as Leads to Novel Therapies 481
have been isolated from mainly marine or-
ganisms, few proteins with anti-cancer ac-
tivity have been described in the literature.
The majority of these proteins fall into two
groups ribosome-inactivating proteins
(RIPs) and lectins though these can of-
ten overlap. RIPs are separated into single
chain (e.g., trichsanthin) and double chain
(e.g., ricin) classes, with the double-chain
class often possessing both a RIP and a
lectin functionality [17, 18]. For example,
Viscum album (mistletoe) lectin has been
used clinically in Europe for both its cyto-
toxic and biological response modifying
activities [19, 20], and actually combines
both functions in two disulfide-linked
chains, the A-chain being a RIP and the B-
chain a sialic acid-specific lectin [21]. Simi-
larly, the recently reported anti-tumor pro-
tein aralin from the Japanese Angelica tree
Aralia elata [22] is also a dual-functioning
heterodimer.
Cytotoxic homodimeric RIPs with no
lectin activity are also found in nature, one
example being the protein panaxagin,
which was isolated from Panax ginseng
[23]. Although single-chain RIPs are also
potently cytotoxic, certain members of this
group (e.g., GAP31 and MAP30) are able
to inhibit the growth of human breast tu-
mor xenografts in mice at doses below the
toxicity threshold [24]. Whether or not any
member of this class of proteins has the
specificity necessary for systemic therapeu-
tics will require further research.
The reverse activity is also found in na-
ture, where there are cytotoxic lectins, with
no ribosome-inactivating activity; an exam-
ple is the lectin from the plant Cheloido-
nium majus [25]. Fungi produce several cy-
totoxic lectins, including the 81 kDa pro-
tein from Pleurotus ostreatus (the oyster
mushroom) that displays potent activity in
mice against both sarcomas and hepato-
mas [26], and the 32 kDa lectin from the
mushroom Volvariella volvacea which is re-
ported to have anti-tumor activity due to
cell-cycle arrest [27] mediated by the ex-
pression of cyclin kinase inhibitors. Lec-
tins from the sponge Haliclona cratera [28]
and the marine mollusks Aplysia kurodai
and Dolabella auricularia [29] have also
been reported selectively to kill tumor
cells, though in the latter two cases the
small peptides identified are from ingested
cyanophytes [30].
Proteins from sponges that do not bind
to sugars have also been reported to have
anti-tumor activity. Thus, the lytic 21 kDa
protein from the sponge Tethya ingalli,
previously isolated from T. lycinurium
[31], was reported selectively to kill hu-
man ovarian cancer cells with an
EC
50
=0.16 lg mL
1
[32]. The anti-tumor
protein pachymatismin (molecular weight
46 kDa) was isolated from the sponge Pa-
chymatisma johnstonii, and inhibited the
growth of human lung carcinoma cells
(IC
50
=0.82.0 lg mL
1
) via a unique mech-
anism which involved cell-cycle inhibition
at the G
0
/G
1
phase [33, 34].
4.2.2
Anti-fungal Proteins
Anti-fungal proteins have been isolated
from a wide range of taxa including other
fungi, bacteria, plants, and insects. Several
small, cysteine-rich antifungal proteins
have been shown to be secreted from fila-
mentous members of the Ascomycetes
[35], generally displaying only inhibitory
activity against other ascomycetes, with no
activity against prokaryotic organisms, nor
do they demonstrate any significant struc-
tural homology to the anti-fungal plant de-
fensins (discussed below), though the tar-
get organisms for both of these groups of
compounds apparently are similar. Exam-
ples are AFP from Aspergillus giganteus
[36], ANAFP from A. niger [37], and NAF
from Penicillium nalgiovense [38]. From the
prokaryotes, examples include the pseudo-
mycins and syringomycins produced by
Pseudomonas syringae [39, 40] which have
both a peptide and a lipid component.
The plant defensins were originally de-
scribed as a group of cysteine-rich anti-
fungal peptides that cause pore formation
in cell membranes, and which show struc-
tural homology to both insect and mam-
malian defensins [41, 42]. Plant defensins
contain from 4554 amino acids, and are
unusual in that, although they permeabi-
lize cell membranes, they act specifically
against fungal cell membranes with little
activity against bacteria and none against
plant or human cells [43, 44]. This selectiv-
ity has led to the successful genetic engi-
neering of fungus-resistant potatoes with a
derivative of the plant defensin Hs-AFP1
[45]. In addition to the defensins, there are
recent reports of anti-fungal proteins from
many plants including lilin (14 kDa) from
the bulbs of Lilium brownii [46], ocatin
(18 kDa) from the Andean tuber Oxalis tu-
berose [47], and five related proteins from
Malva parviflora [48, 49]. Many of these
more recently discovered anti-fungal pro-
teins, including Pf1 and Pf2 isolated from
the seeds of Passiflora edulis, have been
found to be utilized both for energy stor-
age and anti-fungal activity [50].
The phylum Arthropoda, which includes
the class Insecta, is a rich source of novel
compounds. These include drosomycin, an
anti-fungal protein [51] produced by Droso-
phila melanogaster, which shares significant
sequence and structural homology with
the plant anti-fungal protein Rs-AFP2,
from the radish Raphanus sativus [52]. The
cecropins, *3537 amino acid peptides,
were originally isolated from the moth
Hyalopora cecropia based on their anti-bac-
terial activity [53], but were later reported
also to have anti-fungal activity [54]. In-
sects have also been reported to produce
anti-fungal defensins such as heliomicin,
an unusual defensin from the moth larvae
Heliothis virescens [55], which interacts with
specific glucosylceramides on the fungal
cell wall [56]. For further information, the
reader should consult the recent review by
Theis and Stahl [54].
4.2.3
Anti-viral Proteins
Aside from the report of RC-183
(*10 kDa) from the edible mushroom Ro-
zites caperata, which inhibits both herpes
simplex virus 1 (HSV-1) and HSV-2
(IC
50
5 lM) [57], examples are rare as
there are very few (if any) examples of
anti-viral proteins produced by bacteria,
yeast, or fungi aside from the Cyanophyta.
Organisms from this phylum have been
shown to produce several interesting anti-
viral proteins, with the best studied being
cyanovirin-N (CV-N), an 11 kDa protein
produced by Nostoc ellipsosporum [57, 58].
CV-N possesses a novel primary amino
acid sequence which bears no significant
homology to any known protein. Further
indicating the novelty of this protein was
the fact that the three-dimensional struc-
ture of CV-N, elucidated by both NMR [59]
and X-ray crystallography [60], represented
a new superfamily of protein folds. CV-N
displays potent anti-HIV activity, with EC
50
values generally in the 110 nM range
[58]. The activity of CV-N is based on its
ability to inhibit viral fusion and entry
mediated through specific interactions
with the HIV envelope glycoproteins
gp120 [58] and gp41 [61]. CV-N was shown
to bind to these proteins through interac-
tions with the high-mannose oligosaccha-
rides oligomannose-8 and -9 [62, 63]. Sub-
sequent studies have shown that this pro-
tein uses a similar mechanism in inacti-
vating both the Ebola and influenza
viruses [64, 65]. CV-N has been licensed to
several companies, and is currently in pre-
clinical development for therapeutic and
prophylactic applications against both HIV
and influenza.
A second anti-HIV protein, scytovirin
(*9 kDa), was isolated from the terrestrial
cyanophyte Scytonema varium [66]. This
protein bore no homology to CV-N, but
did function in a similar manner against
HIV, was shown to bind to similar carbo-
hydrate structures on gp120, but was later
determined to bind to different substruc-
tures of high-mannose oligosaccharides
[67]. Another protein, MVL (13 kDa), was
first isolated from the cyanophyte Microcys-
tis viridis by following its activity as a man-
nan-binding lectin [68]. More recently, this
dimeric protein was shown to have the
ability to block a model system of HIV fu-
sion [69].
With the macroalgae, the dearth of re-
ports on anti-viral proteins is even more
pronounced, with only one report cur-
rently in the literature, dealing with the
isolation and purification of the *13 kDa
protein griffithsin from the red algae Grif-
fithsia sp. (Rhodophyta). This protein has
an unique 121-amino acid sequence [70],
and was shown to inhibit the cytopathic ef-
fects of HIV against human T lympho-
blasts, with EC
50
values below 1 nM.
Unlike microorganisms and macroalgae,
plants have been reported to produce
many anti-viral proteins. As with the anti-
cancer proteins described earlier, many of
the anti-viral proteins so far described are
RIPs. One such example of an isolated sin-
gle-chain RIP is quinqueginsin, which was
isolated from Panax ginseng based on its
antiviral activity [71]. Other RIPs that are
active against HIV [72] are trichosanthin
and TAP 29, isolated from the tubers of
Trichosanthes kirilowii [73, 74], MAP 30,
from Momordica charantia [75], and PAP,
from Phytolacca americana [76].
From a mechanistic aspect, the RIPs
GAP 31 (from Gelonium multiflorum) [77],
luffin (from Luffa cylindrica) and saporin
(from Saponaria officinalis) [78] were re-
ported to inhibit the HIV-1 integrase pro-
tein.
A very interesting group of plant-derived
HIV-inhibitory proteins, originally isolated
from the tropical tree Chassalia parvifolia
(Rubiaceae), were named the circulins [79].
These small cyclic proteins share a com-
mon disulfide linkage pattern, termed a
cysteine knot motif, with other members
of the cyclotide family [11, 80]. The circu-
lins were found to inhibit HIV-1, with
EC
50
values ranging from 40 to 275 nM
[81, 82]. The precise mechanism of action
for the circulins and other plant cyclotides
is still being studied, but preliminary re-
sults suggest that these proteins interact
directly with cell membranes. Additional
cyclotides have been discovered in the
plant families Rubiaceae and Violiaceae,
including the kalata peptides from Olden-
landia affinis [83, 84], palicourein from Pa-
licourea condensata [85], cycloviolacins from
Viola odorata [86], varv peptides from Viola
arvensis [87], and the cycloviolins from Leo-
nia cymosa [88]. Additional studies on the
cyclotides have shown that, in addition to
their anti-viral activity, members of this
class of peptides also possess insecticidal
properties [89].
Anti-HIV activity has also been reported
for several mannose- and N-acetyl glucosa-
mine-specific plant lectins, including those
from Urtica dioica [90], Myrianthus holstii
[91], Concanavalia sp. (concanavalin A,
Con A) [92], Narcissus sp. [93], and an un-
usual gymnosperm-derived protein from
Cycas revoluta [94] amongst many others
recently reviewed [95, 96]. The mechanism
of these non-catalytic, carbohydrate-bind-
ing proteins [97] has been reported as spe-
cific binding to the HIV envelope protein
gp120, thereby inhibiting viral attachment
and entry [98, 99].
Marine-sourced antiviral proteins,
though less numerous than those of
plants, are also represented in the litera-
ture. The protein niphatevirin, a 19 kDa
anti-HIV glycoprotein isolated from the
sponge Niphates erecta is one such exam-
ple. Niphatevirin was shown specifically to
interact with the cellular receptor CD4 in
a manner that inhibited HIV by prevent-
ing cellcell fusion and syncytium forma-
tion, with an EC
50
=10 nM [100]. A second
anti-HIV protein, isolated from the Haplo-
sclerid sponge Adocia sp., bound to both
CD4 and gp120, and this protein, adocia-
virin, was found to be a 37-kDa disulfide-
linked homodimer that inhibited diverse
strains of HIV, with EC
50
s ranging from
0.4 to >400 nM [101]. More recently, a do-
main of the aggregation factor of the
sponge Microciona prolifera (MAF) was re-
ported to bind to gp120 and protect T-lym-
phoblastoid cells from infection with HIV
(EC
50
0.12 lg mL
1
) [102]. One additional
anti-viral protein was recently discovered
in Penaeid shrimp; this was a large pro-
tein (*74 kDa) which was identified as a
hemocyanin and had activity against a
variety of fish viruses, with EC
50
values of
*5 lg mL
1
[103].
4.2.4
Anti-bacterial Proteins
The most commonly reported biological
activity for proteins derived from natural
products is anti-bacterial activity. It appears
as if every living organism has developed
some mechanism by which to prevent in-
fection by or direct competition with bacte-
ria, including other bacteria. One such
group of proteinaceous compounds pro-
duced in bacteria is the bacteriocins, ribo-
somally produced antibiotic peptides and
proteins [104] which include the lantibio-
tics (produced by Gram-positive bacteria),
and the microcins (produced by Gram-neg-
ative bacteria) [105, 106]. Both groups are
generally 2040 amino acids in length
[107], and their anti-bacterial activity is
mediated by their ability to form pores in
the cytoplasmic membranes of susceptible
microorganisms [106, 108]. This mecha-
nism of action is common to many differ-
ent classes of anti-bacterial peptides from
a variety of source organisms, and has
been discussed in detail in a recent review
[109].
Plants have been reported to produce
many different classes of anti-bacterial
peptides, including RIPs [110], defensins
[111], lectins [112], and cyclotides [11].
Some members of these classes have been
mentioned previously in this review sec-
tion; thus, they will not be repeated here.
The thionins are a group of small plant
proteins which are generally 4547 amino
acids in length, with four disulfide bonds.
The most prominent member in this class
is purothionin, originally isolated from
wheat endosperm (Graminae) [113]. Thio-
nins have been isolated from other plant
families, including the Loranthaceae (i.e.,
viscotoxins) [114] and the Leguminosae
(i.e., the fabatins) [115]. In addition to
their ability to form pores in a variety of
cell membranes [116], the thionins have
been reported selectively to form disulfide
bridges with other proteins [117, 118].
Two distinct classes of anti-microbial
peptides have recently been reported to be
produced in Solanum tuberosum (potato).
The snakin proteins (StSN1 and 2) are cys-
teine-rich basic proteins which weigh
*7 kDa and show activity against a variety
of plant bacterial pathogens including Cla-
vibacter michiganensis and Rhizobium meli-
loti [119, 120]. Another novel anti-bacterial
protein from potato, AP
1
, was made up of
343 amino acids and displayed potent ac-
tivity against the bacterium Ralstonia sola-
nacearum [121]. This proteins structure
was dissimilar to that of other plant anti-
microbial proteins in that it contained an
ATP-binding domain and strong homology
to an acid phosphatase from Mesorhizo-
bium loti. Two other unique anti-bacterial
proteins were isolated from Capsella bursa-
pastoris (shepherds purse). These small,
histidine-rich, cysteine-free proteins (she-
pherdin I and II) were reported to display
potent anti-bacterial activity against Gram-
negative bacteria, but were inactive against
all Gram-positive bacteria against which
they were tested [122]. Other, more recent,
reports have described anti-microbial pep-
tides Sp-AMP and Fa-AMP from the gym-
nosperm Pinus sylvestris [123] and the food
crop Fagopyrum esculentum (buckwheat)
[124], respectively.
The term defensin was first used to de-
scribe anti-microbial peptides isolated
from human neutrophils [125] which have
since been found as a part of the innate
immune system in both vertebrates and
invertebrates [126]. Defensins, which gen-
erally are small, cysteine-rich peptides of
approximately 30 amino acids, have been
isolated from a variety of insects including
examples from the beetle Oryctes rhinoceros
[127], the termite Pseudacanthotermes spini-
ger [128], and the mosquito Anopheles gam-
biae [129]. In addition, anti-microbial pep-
tides from Drosophila melanogaster [130],
and several proline-rich peptides from the
Hymenoptera and the Hemiptera (i.e., api-
daecins and pyrrhocoricin, respectively)
[131] and the cecropins, originally isolated
from Hyalophora cecropia [53] have been
reported. A more recent publication de-
scribes a new glycine-rich antibacterial
protein (*10 kDa), isolated from the spi-
der Acanthoscurria gomesiana, with no
structural similarities to the defensins or
any other plant or animal antimicrobial
peptide, but with moderate activity against
the Gram-negative bacterium E. coli [132].
Amphibian skin has also been a source
of many anti-bacterial peptides including
bombesins and bombinins, bradykinins,
dermorphins and caruleins [133, 134]. One
particularly interesting group of these pep-
tides is the magainins, which were isolated
from the skin of the toad Xenopus laevis
[135]. Magainins are *2 kDa, pore-form-
ing peptides [136] that form a-helices in
lipid bilayers [137] and have bactericidal,
fungicidal, and virucidal activity [138]. Ma-
gainin-based creams for the topical treat-
ment of skin ulcers have been tested in
clinical trials in humans but, at present
have not been approved by the FDA
(USA).
Several species of mussel, including My-
tilus edulis and M. galloprovincialis, have
been reported to produce anti-bacterial
peptides [139, 140], as well as lectins that
are toxic to various marine Vibrio species
[141]. Another mollusk, Dolabella auricula-
ria (sea hare) was recently reported to pro-
duce a novel, 33-amino acid anti-bacterial
peptide named dolabellanin B2 [142]. The
ascidian Styela clava has been reported to
produce a-helical antimicrobial peptides
called clavanins, and 32-amino acid phe-
nylalanine-rich antibacterial peptides called
styelins [143, 144]. In addition, horseshoe
crabs from the genera Tachypleus have
been reported to produce a variety of anti-
bacterial peptides and proteins ranging
from 2.3 to 42 kDa in size, including ta-
chyplesins, tachystatins, and polyphemu-
sins [145, 146].
Crustaceans produce anti-microbial pro-
teins, with two examples being the penaei-
dins from the hemolymph of the pacific
white shrimp, Litopenaeus vannamei [147],
and an 11.5-kDa protein from the crab
Carcinus maenas [148], shown to be active
against Gram-positive bacteria. Anti-bacte-
rial peptides have also been isolated from
the skin of fish, including the oncorhyn-
cins from Oncorhynchus mykiss (rainbow
trout) [149], and the pleurocidins from
Pseudopleuronectes americanus (winter
flounder) [150]. The Atlantic Salmon, Sal-
mo salar, was reported to produce antibac-
terial proteins (20.7 kDa) in its liver that
showed moderate activity against E. coli,
appearing to disrupt membrane integrity
in the susceptible bacteria [151].
4.2.5
Other Biological Activities
Other discoveries have been made of
bioactive proteins from natural sources
that do not easily fall into any of the pre-
vious categories of biological activity. Some
of these proteins may be only isolated oc-
currences, whilst others represent signifi-
cant research and preclinical efforts. A few
particularly interesting proteins will be de-
scribed briefly here.
An unusual protein has been reported
from the secretions of the caterpillar Lono-
mia achelous [152] that causes a bleeding
syndrome which may be mediated by spe-
cific interactions with the blood coagula-
tion Factor V. Thus, this finding may lead
to new areas of research into thrombosis
(see Part II, Chapters 1 and 3 and Part III,
Chapter 6). Anti-freeze proteins have been
isolated from several organisms, including
a *2 kDa peptide recently discovered
from the arctic sponge Homaxinell balfour-
ensis [153]. Research into this family of
proteins may yield advances in our ability
to store organs for transplantation for
more extended periods of time (see Part I,
Chapter 15). Another peptide, isolated
from the skin of the Australian frog Litoria
leseuri, was shown specifically to inhibit
neuronal nitric oxide synthase at low mi-
cromolar concentrations, thereby providing
a unique bioprobe to further examine the
role of nitric oxide in cell signaling [154].
Interesting polypeptides have also been
isolated from the venom of arachnids and
arthropods. Recent reviews on the ion-
channel toxins from scorpions [155], neu-
rotoxins from spider venom [156, 157] and
their effects on the cardiovascular system
[158] have been published.
Finally, an outstanding example of re-
search into marine bioactive peptides is
that of the conotoxin peptides derived
from the venom of marine snails of the
genus Conus [159]. The initial propeptide
is processed by both post-translational
modification and proteolytic cleavage into
a variety of small peptides which are 10
50 amino acids in length [160, 161]. Re-
search into potential therapeutic applica-
tions of the conotoxins has shown that a
number of these peptides interact uniquely
with ion channels to induce a wide variety
of pharmacological effects in mammalian
systems [162164]. General reviews of the
conotoxins are available [160, 165], as well
as more-detailed reviews of their structures
[166] and potential therapeutic uses [166
168].
Lastly, Ziconotide (Prialt
), a conotoxin-
derived peptide that despite being made by
total synthesis is identical to the natural
product, was approved at the end of De-
cember, 2004 by the US FDA for the treat-
ment of chronic pain.
4.2.6
Concluding Comments
As research into the bioactive constituents
of natural product extracts continues, it is
certain that more new and unusual pro-
teins will be discovered. The structural,
biochemical, and functional diversity of
these proteins provides exciting opportu-
nities for future research. The identifica-
tion of new anti-microbial peptides con-
tinues to be an area of active research, and
leads the wave of potential biopharmaceu-
ticals from non-mammalian sources,
though whether any of these agents will
make good drug candidates is still the sub-
ject of much debate [169]. What is certain
however is that, as we delve deeper into
the uncharted proteome of these organ-
isms, we will continue to be surprised by
the unique and potentially useful proteins
that we encounter.
4.3
Overall Concluding Comments
The studies reported in both sections of
this chapter demonstrate both the current
potential (Section 4.1) and future potential
(Section 4.2) of the search for biologically
active peptides and proteins, and the abil-
ity to express these agents in homologous
and/or heterologous hosts. The ability to
manipulate gene sequences to produce
subtle modifications of existing active
agents, and the potential for semi-synthe-
sis to modify the basic properties of the
initial agents, is amply demonstrated in
the discussions of these compounds, and
it is to be hoped that such modified com-
pounds will continue to demonstrate the
advantages of working with Mother Na-
tures Pharmacopoiea.
The real message is that for a very large
number of disease entities not only in
man, but also in veterinary medicine and
perhaps, as importantly, in crop protection
the potential for large-scale production
of these agents by biotechnological means
is limited only by imagination. The basic
compounds are there!
References
1 Newman, D. J., Cragg, G. M., Snader, K. M.
Natural products as sources of new drugs over
the period 19812002. J. Nat. Prod. 2003, 66,
10221037
2 Walsh, G. Biopharmaceutical benchmarks.
Nature Biotech. 2000, 18, 831833
3 Walsh, G. Biopharmaceutical benchmarks
2003. Nature Biotech. 2003, 21, 865870
4 Walsh, G. Second-generation biopharmaceuti-
cals. Eur. J. Pharmaceut. Biopharmaceut. 2004,
58, 185196
5 Walsh, G. Biopharmaceuticals and biotechnol-
ogy medicines: an issue of nomenclature. Eur.
J. Pharm. Sci. 2002, 15, 135138
6 Projan, S. J., Gill, D., Lu, Z., Herrmann, S. H.
Small molecules for small minds? The case
for biologic pharmaceuticals. Expert Opin.
Biol. Ther. 2004, 4, 13451350
7 Kurtzhals, P., Havelund, S., Jonassen, I.,
Kiehr, B., Ribel, U., Markussen, J. Albumin
binding and time of action of acylated insu-
lins in various species. J. Pharm. Sci. 1996,
85, 304308
8 OKeefe, B. R. Biologically active proteins from
natural product extracts. J. Nat. Prod. 2001,
64, 13731381
9 Thomma, B. P., Cammue, B. P., Thevissen, K.
Mode of action of plant defensins suggests
therapeutic potential. Curr. Drug Targets Infect.
Disord. 2003, 3, 18
10 Nielsen, K., Boston, R. S. Ribosome-inactivat-
ing proteins: a plant perspective. Annu. Rev.
Plant Physiol. Plant Mol. Biol. 2001, 52, 785
816
11 Craik, D. J., Daly, N. L., Waine, C. The cystine
knot motif in toxins and implications for drug
design. Toxicon 2001, 39, 4360
12 Pellegrini, A. Antimicrobial peptides from
food proteins. Curr. Pharm. Des. 2003, 9,
12251238
13 Bardan, A., Nizet, V., Gallo, R. L. Antimicro-
bial peptides and the skin. Expert Opin. Biol.
Ther. 2004, 4, 543549
14 Zanetti, M. Cathelicidins, multifunctional pep-
tides of the innate immunity. J. Leukoc. Biol.
2004, 75, 3948
15 Otvos, L., Jr. The short proline-rich antibacter-
ial peptide family. Cell. Mol. Life Sci. 2002, 59,
11381150
16 Yang, D., Biragyn, A., Kwak, L. W., Oppen-
heim, J. J. Mammalian defensins in immunity:
more than just microbicidal. Trends Immunol.
2002, 23, 291296
17 Stirpe, F., Barbieri, L., Battelli, M. G., Soria,
M., Lappi, D. A. Ribosome-inactivating pro-
teins from plants: present status and future
prospects. Biotechnology (N Y) 1992, 10, 405
412
18 Girbes, T., Ferreras, J. M., Iglesias, R., Citores,
L., De Torre, C., Carbajales, M. L., Jimenez, P.,
De Benito, F. M., Munoz, R. Recent advances
in the uses and applications of ribosome-inac-
tivating proteins from plants. Cell. Mol. Biol.
(Noisy-le-grand) 1996, 42, 461471
19 Zarkovic, N., Vukovic, T., Loncaric, I., Miletic,
M., Zarkovic, K., Borovic, S., Cipak, A., Sabo-
lovic, S., Konitzer, M., Mang, S. An overview
on anticancer activities of the Viscum album
extract Isorel. Cancer Biother. Radiopharm.
2001, 16, 5562
20 Abdullaev, F. I., de Mejia, E. G. Antitumor effect
of plant lectins. Nat. Toxins 1997, 5, 157163
21 Muthing, J., Meisen, I., Bulau, P., Langer, M.,
Witthohn, K., Lentzen, H., Neumann, U., Pe-
ter-Katalinic, J. Mistletoe lectin I is a sialic
acid-specific lectin with strict preference to
gangliosides and glycoproteins with terminal
Neu5Ac alpha 2-6Gal beta 1-4GlcNAc resi-
dues. Biochemistry 2004, 43, 29963007
22 Tomatsu, M., Mujin, T., Shibamoto, N., Ta-
shiro, F., Ikuta, A. Production of aralin, a se-
lective cytotoxic lectin against human trans-
formed cells, in callus culture of Aralia elata.
Planta Med. 2004, 70, 469471
23 Ng, T. B., Wang, H. Panaxagin, a new protein
from Chinese ginseng possesses anti-fungal,
anti-viral, translation-inhibiting and ribonu-
clease activities. Life Sci. 2001, 68, 739749
24 Lee-Huang, S., Huang, P. L., Sun, Y., Chen,
H. C., Kung, H. F., Huang, P. L., Murphy, W. J.
Inhibition of MDA-MB-231 human breast tu-
mor xenografts and HER2 expression by anti-
tumor agents GAP31 and MAP30. Anticancer
Res. 2000, 20, 653659
25 Fik, E., Wolun-Cholewa, M., Kistowska, M.,
Warchol, J. B., Gozdzicka-Jozefiak, A. Effect of
lectin from Chelidonium majus L. on normal
and cancer cells in culture. Folia Histochem.
Cytobiol. 2001, 39, 215216
26 Wang, H., Gao, J., Ng, T. B. A new lectin with
highly potent antihepatoma and antisarcoma
activities from the oyster mushroom Pleurotus
ostreatus. Biochem. Biophys. Res. Commun.
2000, 275, 810816
27 Liua, W.-k., Ho, J. C. K., Ng, T. B. Suppression
of cell cycle progression by a fungal lectin: ac-
tivation of cyclin-dependent kinase inhibitors.
Biochem. Pharmacol. 2001, 61, 3337
28 Pajic, I., Kljajic, Z., Dogovic, N., Sladic, D.,
Juranic, Z., Gasic, M. J. A novel lectin from
the sponge Haliclona cratera: isolation, charac-
terization and biological activity. Comp. Bio-
chem. Physiol. C Toxicol. Pharmacol. 2002, 132,
213221
29 Yamazaki, M., Kisugi, J., Iijima, R. Antineo-
plastic glycoproteins in marine invertebrates.
Gan To Kagaku Ryoho 1997, 24, 14771485
30 Newman, D. J., Cragg, G. M. Marine natural
products and related compounds in clinical
and advanced clinical trials. J. Nat. Prod. 2004,
67, 12161238
31 Mangel, A., Leitao, J. M., Batel, R., Zimmer-
mann, H., Muller, W. E. G., Schroder, H. C.
Purification and characterization of a pore-
forming protein from the marine sponge
Tethya lyncurium. Eur. J. Biochem. 1992, 210,
499507
32 OKeefe, B. R., Beutler, J. A., Cardellina II,
J. H., Prather, T. R., Shoemaker, R. H., Sowder
II, R. C., Henderson, L. E., Pannell, L. K., Boyd,
M. R. Isolation of a novel Kunitz family pro-
tease inhibitor in association with Tethya he-
molysin from the sponge Tethya ingalli. J. Nat.
Prod. 1997, 60, 10941099
33 Zidane, M., Pondaven, P., Roussakis, C.,
More, M. T. Effects in vitro of pachymatismin,
a glycoprotein from the marine sponge Pachy-
matisma johnstonii, on a non-small-cell
bronchopulmonary carcinoma line (NSCLC-
N6). Anticancer Res. 1996, 16, 28052812
34 Zidane, M., More, M. T., Pondaven, P., Jaquot,
C., Riou, D., Roussakis, C. In vivo effect of pa-
chymatismin, a new marine glycoprotein, on
a human non-small-cell lung carcinoma. In
Vivo 1997, 11, 185188
35 Marx, F. Small, basic antifungal proteins se-
creted from filamentous ascomycetes: a com-
parative study regarding expression, structure,
function and potential application. Appl. Mi-
crobiol. Biotechnol. 2004, 65, 133142
36 Theis, T., Marx, F., Salvenmoser, W., Stahl, U.,
Meyer, V. New insights into the target site and
mode of action of the antifungal protein of As-
pergillus giganteus. Res. Microbiol. 2005, 156,
4756
37 Gun Lee, D., Shin, S. Y., Maeng, C. Y., Jin,
Z. Z., Kim, K. L., Hahm, K. S. Isolation and
References 489
characterization of a novel antifungal peptide
from Aspergillus niger. Biochem. Biophys. Res.
Commun. 1999, 263, 646651
38 Geisen, R. P. nalgiovense carries a gene which
is homologous to the paf gene of P. chrysogen-
um which codes for an antifungal peptide. Int.
J. Food Microbiol. 2000, 62, 95101
39 Harrison, L., Teplow, D. B., Rinaldi, M., Stro-
bel, G. Pseudomycins, a family of novel pep-
tides from Pseudomonas syringae possessing
broad-spectrum antifungal activity. J Gen. Mi-
crobiol. 1991, 137, 28572865
40 Segre, A., Bachmann, R. C., Ballio, A., Bossa,
F., Grgurina, I., Iacobellis, N. S., Marino, G.,
Pucci, P., Simmaco, M., Takemoto, J. Y. The
structure of syringomycins A1, E and G. FEBS
Lett. 1989, 255, 2731
41 Hoffman, J. A., Hetru, C. Insect defensins: in-
ducible antibacterial peptides. Immunol. Today
1992, 13, 411415
42 Ganz, T., Lehrer, R. I. Defensins. Curr. Opin.
Immunol. 1994, 6, 584589
43 Broekaert, W. F., Terras, F. R. G., Cammue,
B. P. A., Osborn, R. W. Plant defensins: novel
antimicrobial peptides as components of the
host defense system. Plant Physiol. 1995, 108,
13531358
44 Terras, F. R. G., Schoofs, H. M. E., De Bolle,
M. F. C., Van Leuven, F., Rees, S. B., Vander-
leyden, J., Cammue, B. P. A., Broekaert, W. F.
Analysis of two novel classes of plant antifun-
gal proteins from radish (Raphanus sativus L.)
seeds. J. Biol. Chem. 1992, 267, 1530115309
45 Gao, A. G., Hakimi, S. M., Mittanck, C. A.,
Wu, Y., Woerner, B. M., Stark, D. M., Shah,
D. M., Liang, J., Rommens, C. M. T. Fungal
pathogen protection in potato by expression of
a plant defensin peptide. Nat. Biotechnol.
2000, 18, 13071310
46 Wang, H., Ng, T. B. Isolation of lilin, a novel
arginine- and glutamate-rich protein with po-
tent antifungal and mitogenic activities from
lily bulbs. Life Sci. 2002, 70, 10751084
47 Flores, T., Alape-Giron, A., Flores-Diaz, M.,
Flores, H. E. Ocatin. A novel tuber storage
protein from the Andean tuber crop oca with
antibacterial and antifungal activities. Plant
Physiol. 2002, 128, 12911302
48 Wang, X., Bunkers, G. J. Potent heterologous
antifungal proteins from cheeseweed (Malva
parviflora). Biochem. Biophys. Res. Commun.
2000, 279, 669673
49 Wang, X., Bunkers, G. J., Walters, M. R., Tho-
ma, R. S. Purification and characterization of
three antifungal proteins from cheeseweed
(Malva parviflora). Biochem. Biophys. Res. Com-
mun. 2001, 282, 12241228
50 Agizzio, A. P., Carvalho, A. O., Ribeiro Sade,
F., Machado, O. L., Alves, E. W., Okorokov,
L. A., Samarao, S. S., Bloch Jr, C., Prates,
M. V., Gomes, V. M. A 2S albumin-homolo-
gous protein from passion fruit seeds inhibits
the fungal growth and acidification of the me-
dium by Fusarium oxysporum. Arch. Biochem.
Biophys. 2003, 416, 188195
51 Fehlbaum, P., Bulet, P., Michaut, L., Lagueux,
M., Broekaert, W. F., Hetru, C., Hoffmann,
J. A. Insect immunity. Septic injury of Droso-
phila induces the synthesis of a potent anti-
fungal peptide with sequence homology to
plant antifungal peptides. J. Biol. Chem. 1994,
269, 3315933163
52 Landon, C., Pajon, A., Vovelle, F., Sodano, P.
The active site of drosomycin, a small insect
antifungal protein, delineated by comparison
with the modeled structure of Rs-AFP2, a
plant antifungal protein. J. Pept. Res. 2000, 56,
231238
53 Steiner, H., Hultmark, D., Engstrom, A., Ben-
nich, H., Boman, H. G. Sequence and specific-
ity of two antibacterial proteins involved in in-
sect immunity. Nature 1981, 292, 246248
54 Theis, T., Stahl, U. Antifungal proteins: tar-
gets, mechanisms and prospective applica-
tions. Cell. Mol. Life Sci. 2004, 61, 437455
55 Lamberty, M., Ades, S., Uttenweiler-Joseph, S.,
Brookhart, G., Bushey, D., Hoffmann, J. A.,
Bulet, P. Insect immunity. Isolation from the
lepidopteran Heliothis virescens of a novel in-
sect defensin with potent antifungal activity.
J. Biol. Chem. 1999, 274, 93209326
56 Thevissen, K., Warnecke, D. C., Franois,
I. E. J. A., Leipelt, M., Heinz, E., Ott, C., Zh-
ringer, U., Thomma, B. P. H. J., Ferket,
K. K. A., Cammue, B. P. A. Defensins from in-
sects and plants interact with fungal glucosyl-
ceramides. J. Biol. Chem. 2004, 279, 3900
3905
57 Piraino, F., Brandt, C. R. Isolation and partial
characterization of an antiviral, RC-183, from
the edible mushroom Rozites caperata. Antivi-
ral Res. 1999, 43, 6778.
58 Boyd, M. R., Gustafson, K. R., McMahon, J. B.,
Shoemaker, R. H., OKeefe, B. R., Mori, T., Gu-
lakowski, R. J., Wu, L., Rivera, M. I., Lauren-
cot, C. M., Currens, M. J., Cardellina II, J. H.,
Buckheit J., R. W., Nara, P. R., Pannell, L. K.,
Sowder II, R. C., Henderson, L. E. Discovery
of cyanovirin-N, a novel human immunodefi-
ciency virus-inactivating protein that binds
viral surface envelope glycoprotein gp120:
potential applications to microbicide develop-
ment. Antimicrob. Agents Chemother. 1997, 41,
15211530
59 Bewley, C. A., Gustafson, K. R., Boyd, M. R.,
Covell, D. G., Bax, A., Clore, G. M., Gronen-
born, A. M. Solution structure of cyanovirin-N,
a potent HIV-inactivating protein. Nat. Struct.
Biol. 1998, 5, 571578
60 Yang, F., Bewley, C. A., Louis, J. M., Gustafson,
K. R., Boyd, M. R., Gronenborn, A. M., Clore,
G. M., Wlodawer, A. Crystal structure of cya-
novirin-N, a potent HIV-inactivating protein,
shows unexpected domain swapping. J. Mol.
Biol. 1999, 288, 403412
61 OKeefe, B. R., Shenoy, S. R., Xie, D., Zhang,
W., Muschik, J. M., Currens, M. J., Chaiken, I.,
Boyd, M. R. Analysis of the interaction be-
tween the HIV-inactivating protein cyanovirin-
N and soluble forms of the envelope glycopro-
teins gp120 and gp41. Mol. Pharmacol. 2000,
58, 982992
62 Bolmstedt, A. J., OKeefe, B. R., Shenoy, S. R.,
McMahon, J. B., Boyd, M. R. Cyanovirin-N de-
fines a new class of antiviral agent targeting
N-linked, high-mannose glycans in an oligo-
saccharide-specific manner. Mol. Pharmacol.
2001, 59, 949554
63 Shenoy, S. R., OKeefe, B. R., Bolmstedt, A. J.,
Cartner, L. K., Boyd, M. R. Selective interac-
tions of the human immunodeficiency virus-
inactivating protein cyanovirin-N with high-
mannose oligosaccharides on gp120 and other
glycoproteins. J. Pharmacol. Exp. Ther. 2001,
297, 704710
64 Barrientos, L. G., OKeefe, B. R., Bray, M., San-
chez, A., Gronenborn, A. M., Boyd, M. R. Cya-
novirin-N binds to the viral surface glycopro-
tein, GP1,2 and inhibits infectivity of Ebola
virus. Antiviral Res. 2003, 58, 4756
65 OKeefe, B. R., Smee, D. F., Turpin, J. A., Sau-
cedo, C. J., Gustafson, K. R., Mori, T., Blake-
slee, D., Buckheit, R., Boyd, M. R. Potent anti-
influenza activity of cyanovirin-N and interac-
tions with viral hemagglutinin. Antimicrob.
Agents Chemother. 2003, 47, 25182525
66 Bokesch, H. R., OKeefe, B. R., McKee, T. C.,
Pannell, L. K., Patterson, G. M., Gardella, R. S.,
Sowder II, R. C., Turpin, J., Watson, K., Buc-
kheit Jr, R. W., Boyd, M. R. A potent novel
anti-HIV protein from the cultured cyanobac-
terium Scytonema varium. Biochemistry 2003,
42, 25782584
67 Adams, E. W., Ratner, D. M., Bokesch, H. R.,
McMahon, J. B., OKeefe, B. R., Seeberger,
P. H. Oligosaccharide and glycoprotein micro-
arrays as tools in HIV glycobiology; glycan-de-
pendent gp120/protein interactions. Chem.
Biol. 2004, 11, 739740
68 Yamaguchi, M., Ogawa, T., Muramoto, K., Ka-
mio, Y., Jimbo, M., Kamiya, H. Isolation and
characterization of a mannan-binding lectin
from the freshwater cyanobacterium (blue-
green algae) Microcystis viridis. Biochem. Bio-
phys. Res. Commun. 1999, 265, 703708
69 Bewley, C. A., Cai, M., Ray, S., Ghirlando, R.,
Yamaguchi, M., Muramoto, K. New carbohy-
drate specificity and HIV-1 fusion blocking ac-
tivity of the cyanobacterial protein MVL:
NMR, ITC and sedimentation equilibrium
studies. J. Mol. Biol. 2004, 339, 901914
70 Mori, T., OKeefe, B. R., Sowder II, R. C.,
Bringans, S., Gardella, R., Berg, S., Cochran,
P., Turpin, J. A., Buckheit Jr, R. W., McMahon,
J. B., Boyd, M. R. Isolation and characteriza-
tion of griffithsin, a novel HIV-inactivating
protein from the red alga Griffithsia sp. J. Biol.
Chem. 2005, 280, 93459353.
71 Wang, H. X., Ng, T. B. Quinqueginsin, a novel
protein with anti-human immunodeficiency
virus, antifungal, ribonuclease and cell-free
translation-inhibitory activities from American
ginseng roots. Biochem. Biophys. Res. Commun.
2000, 269, 203208
72 Bourinbaiar, A. S., Lee-Huang, S. The activity
of plant-derived antiretroviral proteins MAP30
and GAP31 against herpes simplex virus in vi-
tro. Biochem. Biophys. Res. Commun. 1996,
219, 923929
73 McGrath, M. S., Huang, K. M., Caldwell, S. E.,
Gaston, I., Luk, K. C., Wu, P., Ng, V. L., Crowe,
S., Daniels, J., Marsh, J., Dienhart, T., Lekas,
P. V., Vennari, J. C., Yeung, H. W., Lifson, J. D.
GLQ223: an inhibitor of human immunodefi-
ciency virus replication in acutely and chroni-
cally infected cells of lymphocyte and mono-
nuclear phagocyte lineage. Proc. Natl. Acad.
Sci. USA 1989, 86, 28442848
74 Lee-Huang, S., Huang, P. L., Kung, H. F., Li,
B. Q., Huang, P. L., Huang, P., Huang, H. I.,
Chen, H. C. TAP 29: an anti-human immuno-
References 491
deficiency virus protein from Trichosanthes kir-
ilowii that is nontoxic to intact cells. Proc.
Natl. Acad. Sci. USA 1991, 88, 65706574
75 Lee-Huang, S., Huang, P. L., Nara, P. L., Chen,
H. C., Kung, H. F., Huang, P., Huang, H. I.,
Huang, P. L. MAP 30: a new inhibitor of HIV-
1 infection and replication. FEBS Lett. 1990,
272, 1218
76 Zarling, J. M., Moran, P. A., Haffar, O., Sias,
J., Richman, D. D., Spina, C. A., Myers, D. E.,
Kuebelbeck, V., Ledbetter, J. A., Uckun, F. M.
Inhibition of HIV replication by pokeweed
antiviral protein targeted to CD4+ cells by
monoclonal antibodies. Nature 1990, 347, 92
95
77 Lee-Huang, S., Huang, P. L., Huang, P. L.,
Bourinbaiar, A. S., Chen, H. C., Kung, H. F. In-
hibition of the integrase of human immuno-
deficiency virus (HIV) type 1 by anti-HIV
plant proteins MAP30 and GAP31. Proc Natl
Acad Sci USA 1995, 92, 88188822
78 Au, T. K., Collins, R. A., Lam, T. L., Ng, T. B.,
Fong, W. P., Wan, D. C. The plant ribosome
inactivating proteins luffin and saporin are po-
tent inhibitors of HIV-1 integrase. FEBS Lett.
2000, 471, 169172
79 Gustafson, K. R., Sowder II, R. C., Henderson,
L. E., Parsons, I. C., Kashman, Y., Cardellina
II, J. H., McMahon, J. B., Buckheit Jr, R. W.,
Pannell, L. K., Boyd, M. R. Circulins A and B.
Novel human immunodeficiency virus (HIV)-
inhibitory macrocyclic peptides from the tropi-
cal tree Chassalia parvifolia. J. Am. Chem. Soc.
1994, 116, 93379338
80 Gustafson, K. R., McKee, T. C., Bokesch, H. R.
Anti-HIV cyclotides. Curr. Protein Pept. Sci.
2004, 5, 331340
81 Gustafson, K. R., Walton, L. K., Sowder Jr,
R. C., Johnson, D. G., Pannell, L. K., Cardellina
II, J. H., Boyd, M. R. New circulin macrocyclic
polypeptides from Chassalia parvifolia. J. Nat.
Prod. 2000, 63, 176178
82 Derua, R., Gustafson, K. R., Pannell, L. K.
Analysis of the disulfide linkage pattern in cir-
culin A and B, HIV-inhibitory macrocyclic
peptides. Biochem. Biophys. Res. Commun.
1996, 228, 632638.
83 Saether, O., Craik, D. J., Campbell, I. D., Slet-
ten, K., Juul, J., Norman, D. G. Elucidation of
the primary and three-dimensional structure
of the uterotonic polypeptide kalata B1. Bio-
chemistry 1995, 34, 41474158
84 Craik, D. J., Daly, N. L., Bond, T., Waine, C.
Plant cyclotides: A unique family of cyclic and
knotted proteins that defines the cyclic cystine
knot structural motif. J. Mol. Biol. 1999, 294,
13271336
85 Bokesch, H. R., Pannell, L. K., Cochran, P. K.,
Sowder II, R. C., McKee, T. C., Boyd, M. R. A
novel anti-HIV macrocyclic peptide from Pali-
courea condensata. J. Nat. Prod. 2001, 64, 249
250
86 Svangard, E., Goransson, U., Smith, D., Ver-
ma, C., Backlund, A., Bohlin, L., Claeson, P.
Primary and 3-D modelled structures of two
cyclotides from Viola odorata. Phytochemistry
2003, 64, 135142
87 Goransson, U., Luijendijk, T., Johansson, S.,
Bohlin, L., Claeson, P. Seven novel macrocyc-
lic polypeptides from Viola arvensis. J. Nat.
Prod. 1999, 62, 283286
88 Hallock, Y. F., Sowder II, R. C., Pannell, L. K.,
Hughes, C. B., Johnson, D. G., Gulakowski, R.,
Cardellina II, J. H., Boyd, M. R. Cycloviolins A-
D, anti-HIV macrocyclic peptides from Leonia
cymosa. J. Org. Chem. 2000, 65, 124128
89 Jennings, C., West, J., Waine, C., Craik, D.,
Anderson, M. Biosynthesis and insecticidal
properties of plant cyclotides: the cyclic
knotted proteins from Oldenlandia affinis.
Proc. Natl. Acad. Sci. USA 2001, 98, 10614
10619
90 Balzarini, J., Neyts, J., Schols, D., Hosoya, M.,
Van Damme, E., Peumans, W., De Clercq, E.
The mannose-specific plant lectins from Cym-
bidium hybrid and Epipactis helleborine and the
(N-acetylglucosamine)-specific plant lectin
from Urtica dioica are potent and selective in-
hibitors of human immunodeficiency virus
and cytomegalovirus replication in vitro. Anti-
viral. Res. 1992, 18, 191207
91 Charan, R. D., Munro, M. H. G., OKeefe, B. R.,
Sowder II, R. C., McKee, T. C., Currens, M. J.,
Pannell, L. K., Boyd, M. R. Isolation and char-
acterization of Myrianthus holstii lectin, a po-
tent HIV-1 inhibitory protein from the plant
Myrianthus holstii. J. Nat. Prod. 2000, 63,
11701174
92 Robinson Jr, W. E., Montefiori, D. C., Mitchell,
W. M. Evidence that mannosyl residues are in-
volved in human immunodeficiency virus type
1 (HIV-1) pathogenesis. AIDS Res. Hum. Ret-
roviruses 1987, 3, 265282
93 Lopez, S., Armand-Ugon, M., Bastida, J., Vila-
domat, F., Este, J. A., Stewart, D., Codina,
C. Anti-human immunodeficiency virus type
1 (HIV-1) activity of lectins from Narcissus
species. Planta Med. 2003, 69, 109112
94 Yagi, F., Iwaya, T., Haraguchi, T., Goldstein,
I. J. The lectin from leaves of Japanese cycad,
Cycas revoluta Thunb. (gymnosperm) is a
member of the jacalin-related family. Eur. J.
Biochem. 2002, 269, 43354341
95 Astoul, C. H., Peumans, W. J., Van Damme,
E. J. M., Rouge, P. Accessibility of the high-
mannose glycans of glycoprotein gp120
from human immunodeficiency virus type 1
probed by in vitro interaction with mannose-
binding lectins. Biochem. Biophys. Res. Com-
mun. 2000, 274, 455460
96 Balzarini, J., Van Laethem, K., Hatse, S.,
Vermeire, K., De Clercq, E., Peumans, W.,
Van Damme, E., Vandamme, A. M., Bolm-
stedt, A., Schols, D. Profile of resistance of
human immunodeficiency virus to man-
nose-specific plant lectins. J. Virol. 2004, 78,
1061710627
97 Peumans, W. J., Van Damme, E. J. M. Lectins
as plant defense proteins. Plant Physiol.
1995, 109, 347352
98 De Clercq, E. Current lead natural products
for the chemotherapy of human immunode-
ficiency virus (HIV) infection. Med. Res. Rev.
2000, 20, 323349
99 Vlietinck, A. J., De Bruyne, T., Apers, S., Pi-
eters, L. A. Plant-derived leading compounds
for chemotherapy of human immunodefi-
ciency virus (HIV) infection. Planta Med.
1998, 64, 97109
100 OKeefe, B. R., Beutler, J. A., Cardellina II,
J. H., Gulakowski, R. J., Krepps, B. L., McMa-
hon, J. B., Sowder II, R. C., Henderson, L. E.,
Pannell, L. K., Pomponi, S. A., Boyd, M. R.
Isolation and characterization of niphatevirin,
a human-immunodeficiency-virus-inhibitory
glycoprotein from the marine sponge Ni-
phates erecta. Eur. J. Biochem. 1997, 245, 4753
101 OKeefe, B. R., Erim, T., Beutler, J. A., Cardel-
lina II, J. H., Gulakowski, R. J., Krepps, B. L.,
McMahon, J. B., Sowder II, R. C., Johnson,
D. G., Buckheit Jr, R. W., Halliday, S., Boyd,
M. R. Isolation and characterization of ado-
ciavirin, a novel HIV-inhibitory protein from
the sponge Adocia sp. FEBS Lett. 1998, 431,
8590
102 MacKenzie, R., Newman, D., Burger, M. M.,
Roy, R., Kuhns, W. J. Adhesion of a viral en-
velope protein to a non-self-binding domain
of the aggregation factor in the marine
sponge Microciona prolifera. Biol. Bull. 2000,
199, 209211
103 Zhang, X., Huang, C., Qin, Q. Antiviral
properties of hemocyanin isolated from
shrimp Penaeus monodon. Antiviral Res.
2004, 61, 9399
104 Kolter, R., Moreno, F. Genetics of riboso-
mally synthesized peptide antibiotics. Annu.
Rev. Microbiol. 1992, 46, 141163
105 Jack, R. W., Jung, G. Lantibiotics and micro-
cins: polypeptides with unusual chemical di-
versity. Curr. Opin. Chem. Biol. 2000, 4, 310
317
106 Ennahar, S., Sashihara, T., Sonomoto, K.,
Ishizaki, A. Class IIa bacteriocins: biosynth-
esis, structure and activity. FEMS Microbiol.
Rev. 2000, 24, 85106
107 Gaillard-Gendron, S., Vignon, D., Cotten-
ceau, G., Graber, M., Zorn, N., van Dorsse-
laer, A., Pons, A.-M. Isolation, purification
and partial amino acid sequence of a highly
hydrophobic new microcin named microcin
L produced by Escherichia coli. FEMS Micro-
biol. Lett. 2000, 193, 9598
108 Sahl, H. G., Bierbaum, G. Lantibiotics: bio-
synthesis and biological activities of un-
iquely modified peptides from gram-positive
bacteria. Annu. Rev. Microbiol. 1998, 52, 41
79
109 Shai, Y. Mode of action of membrane active
antimicrobial peptides. Biopolymers 2002, 66,
236248
110 Stirpe, F. Ribosome-inactivating proteins.
Toxicon 2004, 44, 371383
111 Thomma, B. P., Cammue, B. P., Thevissen,
K. Plant defensins. Planta 2002, 216, 193
202
112 Cowan, M. M. Plant products as antimicro-
bial agents. Clin. Microbiol. Rev. 1999, 12,
564582
113 Balls, A. K., Hale, W. S., Harris, T. H. a crys-
talline protein obtained from a lipoprotein
of wheat flour. Cereal Chem. 1942, 19, 279
288
114 Olson, T., Samuelsson, G. Purification of vis-
cotoxin Aox2 from the European Mistletoe
(Viscum album L.) by chromatography on
DEAE-cellulose. Acta Chem. Scand. 1970, 24,
720721
115 Zhang, Y., Lewis, K. Fabatins: new antimi-
crobial plant peptides. FEMS Microbiol. Lett.
1997, 149, 5964
References 493
116 Evans, J., Wang, Y. D., Shaw, K. P., Vernon,
L. P. Cellular responses to Pyrularia thionin
are mediated by Ca2+ influx and phospholi-
pase A2 activation and are inhibited by thio-
nin tyrosine iodination. Proc. Natl. Acad. Sci.
USA 1989, 86, 58495853
117 Diaz, I., Carmona, M. J., Garcia-Olmedo, F.
Effects of thionins on beta-glucuronidase in
vitro and in plant protoplasts. FEBS Lett.
1992, 296, 279282
118 Pineiro, M., Diaz, I., Rodriguez-Palenzuela,
P., Titarenko, E., Garcia-Olmedo, F. Selective
disulphide linkage of plant thionins with
other proteins. FEBS Lett. 1995, 369, 239
242
119 Segura, A., Moreno, M., Madueno, F., Moli-
na, A., Garcia-Olmedo, F. Snakin-1, a pep-
tide from potato that is active against plant
pathogens. Mol. Plant Microbe Interact. 1999,
12, 1623
120 Berrocal-Lobo, M., Segura, A., Moreno, M.,
Lopez, G., Garcia-Olmedo, F., Molina, A.
Snakin-2, an antimicrobial peptide from po-
tato whose gene is locally induced by
wounding and responds to pathogen infec-
tion. Plant Physiol. 2002, 128, 951961
121 Feng, J., Yuan, F., Gao, Y., Liang, C., Xu, J.,
Zhang, C., He, L. A novel antimicrobial pro-
tein isolated from potato (Solanum tubero-
sum) shares homology with an acid phospha-
tase. Biochem. J. 2003, 376, 481487
122 Park, C. J., Park, C. B., Hong, S. S., Lee,
H. S., Lee, S. Y., Kim, S. C. Characterization
and cDNA cloning of two glycine- and histi-
dine-rich antimicrobial peptides from the
roots of shepherds purse, Capsella bursa-pas-
toris. Plant Mol. Biol. 2000, 44, 187197
123 Asiegbu, F. O., Choi, W., Li, G., Nahalkova,
J., Dean, R. A. Isolation of a novel antimicro-
bial peptide gene (Sp-AMP) homologue
from Pinus sylvestris (Scots pine) following
infection with the root rot fungus Heterobasi-
dion annosum. FEMS Microbiol. Lett. 2003,
228, 2731
124 Fujimura, M., Minami, Y., Watanabe, K., Ta-
dera, K. Purification, characterization, and
sequencing of a novel type of antimicrobial
peptides, Fa-AMP1 and Fa-AMP2, from
seeds of buckwheat (Fagopyrum esculentum
Moench.). Biosci. Biotechnol. Biochem. 2003,
67, 16361642
125 Ganz, T., Selsted, M. E., Szklarek, D., Har-
wig, S. S., Daher, K., Bainton, D. F., Lehrer,
R. I. Defensins. Natural peptide antibiotics
of human neutrophils. J. Clin. Invest. Oct.
1985, 76, 14271435
126 Hughes, A. L. Evolutionary diversification of
the mammalian defensins. Cell. Mol. Life
Sci. 1999, 56, 94103
127 Ishibashi, J., Saido-Sakanaka, H., Yang, J.,
Sagisaka, A., Yamakawa, M. Purification,
cDNA cloning and modification of a defen-
sin from the coconut rhinoceros beetle,
Oryctes rhinoceros. Eur. J. Biochem. 1999, 266,
616623
128 Da Silva, P., Jouvensal, L., Lamberty, M., Bu-
let, P., Caille, A., Vovelle, F. Solution struc-
ture of termicin, an antimicrobial peptide
from the termite Pseudacanthotermes spiniger.
Protein Sci. 2003, 12, 438446
129 Eggleston, P., Lu, W., Zhao, Y. Genomic
organization and immune regulation of the
defensin gene from the mosquito, Anopheles
gambiae. Insect Mol. Biol. 2000, 9, 481490
130 Meister, M., Hetru, C., Hoffmann, J. A. The
antimicrobial host defense of Drosophila.
Curr. Top. Microbiol. Immunol. 2000, 248,
1736
131 Hetru, C., Hoffmann, D., Bulet, P. Antimi-
crobial peptides from insects. In Molecular
mechanisms of immune responses in insects;
Brey, P. T. and Hultmark, D. (Eds.); Chap-
man & Hall: New York, 1998; pp. 4466
132 Lorenzini, D. M., da Silva Jr, P. I., Fogaca,
A. C., Bulet, P., Daffre, S. Acanthoscurrin: a
novel glycine-rich antimicrobial peptide con-
stitutively expressed in the hemocytes of the
spider Acanthoscurria gomesiana. Dev. Comp.
Immunol. 2003, 27, 781791
133 Erspamer, V., Falconieri-Erspamer, G., Cei,
J. M. Active peptides in the skins of two
hundred and thirty American amphibian
species. Comp. Biochem. Physiol. C 1986, 85,
125137
134 Roseghini, M., Falconieri-Erspamer, G.,
Severini, C., Simmaco, M. Biogenic amines
and active peptides in extracts of the skin of
thirty-two European amphibian species.
Comp. Biochem. Physiol. C 1989, 94, 455
460.
135 Zasloff, M. Magainins, a class of antimicro-
bial peptides from Xenopus skin: isolation,
characterization of two active forms, and
partial cDNA sequence of a precursor. Proc.
Natl. Acad. Sci. USA 1987, 84, 54495453
136 Matsuzaki, K. Magainins as paradigm for
the mode of action of pore forming polypep-
tides. Biochim. Biophys. Acta 1998, 1376,
391400
137 Ramamoorthy, A., Marassi, F. M., Zasloff,
M., Opella, S. J. Three-dimensional solid-
state NMR spectroscopy of a peptide orien-
ted in membrane bilayers. J. Biomol. NMR
1995, 6, 329334
138 Bechinger, B. Structure and functions of
channel-forming peptides: magainins, cecro-
pins, melittin and alamethicin. J. Membr.
Biol. 1997, 156, 197211
139 Charlet, M., Chernysh, S., Philippe, H.,
Hetru, C., Hoffmann, J. A., Bulet, P. Innate
immunity. Isolation of several cysteine-rich
antimicrobial peptides from the blood of a
mollusc, Mytilus edulis. J. Biol. Chem. 1996,
271, 2180821813
140 Hubert, F., Noel, T., Roch, P. A member of
the arthropod defensin family from edible
Mediterranean mussels (Mytilus galloprovin-
cialis). Eur. J. Biochem. 1996, 240, 302306
141 Tunkijjanukij, S., Olafsen, J. A. Sialic acid-
binding lectin with antibacterial activity
from the horse mussel: further characteriza-
tion and immunolocalization. Dev. Comp.
Immunol. 1998, 22, 139150
142 Iijima, R., Kisugi, J., Yamazaki, M. A novel
antimicrobial peptide from the sea hare Do-
labella auricularia. Dev. Comp. Immunol.
2003, 27, 305311
143 Lee, I. H., Zhao, C., Cho, Y., Harwig, S. S.,
Cooper, E. L., Lehrer, R. I. Clavanins, alpha-
helical antimicrobial peptides from tunicate
hemocytes. FEBS Lett. 1997, 400, 158162
144 Lee, I. H., Cho, Y., Lehrer, R. I. Styelins,
broad-spectrum antimicrobial peptides from
the solitary tunicate, Styela clava. Comp. Bio-
chem. Physiol. B Biochem. Mol. Biol. 1997,
118, 515521
145 Iwanaga, S., Kawabata, S. Evolution and
phylogeny of defense molecules associated
with innate immunity in horseshoe crab.
Front. Biosci. 1998, 3, 973984.
146 Miyata, T., Tokunaga, F., Yoneya, T., Yoshika-
wa, K., Iwanaga, S., Niwa, M., Takao, T., Shi-
monishi, Y. Antimicrobial peptides, isolated
from horseshoe crab hemocytes, tachyplesin
II, and polyphemusins I and II: chemical
structures and biological activity. J. Biochem
(Tokyo) 1989, 106, 663668
147 Destoumieux, D., Bulet, P., Loew, D., Van
Dorsselaer, A., Rodriguez, J., Bachere, E. Pe-
naeidins, a new family of antimicrobial pep-
tides isolated from the shrimp Penaeus van-
namei (Decapoda). J. Biol. Chem. 1997, 272,
2839823406
148 Relf, J. M., Chisholm, J. R., Kemp, G. D.,
Smith, V. J. Purification and characterization
of a cysteine-rich 11. 5-kDa antibacterial pro-
tein from the granular haemocytes of the
shore crab, Carcinus maenas. Eur. J. Biochem.
1999, 264, 350357
149 Fernandes, J. M. O., Smith, V. J. A novel anti-
microbial function for a ribosomal peptide
from rainbow trout skin. Biochem. Biophys.
Res Commun. 2002, 296, 167171
150 Cole, A. M., Weis, P., Diamond, G. Isolation
and characterization of pleurocidin, an anti-
microbial peptide in the skin secretions of
winter flounder. J. Biol. Chem. 1997, 272,
1200812013
151 Richards, R. C., ONeil, D. B., Thibault, P.,
Ewart, K. V. Histone H1: an antimicrobial
protein of Atlantic salmon (Salmo salar).
Biochem. Biophys. Res. Commun. 2001, 284,
549555
152 Lopez, M., Gil, A., Arocha-Pinango, C. L.
The action of Lonomia achelous caterpillars
venom on human factor V. Thromb. Res.
2000, 98, 103110
153 Wilkins, S. P., Blum, A. J., Burkepile, D. E.,
Rutland, T. J., Wierzbicki, A., Kelly, M., Ha-
mann, M. T. Isolation of an antifreeze pep-
tide from the Antarctic sponge Homaxinella
balfourensis. Cell. Mol. Life Sci. 2002, 59,
22102215
154 Doyle, J., Llewellyn, L. E., Brinkworth, C. S.,
Bowie, J. H., Wegener, K. L., Rozek, T., Wab-
nitz, P. A., Wallace, J. C., Tyler, M. J. Amphi-
bian peptides that inhibit neuronal nitric ox-
ide synthase. Isolation of lesuerin from the
skin secretion of the Australian Stony Creek
frog Litoria lesueuri. Eur. J. Biochem. 2002,
269, 100109.
155 Possani, L. D., Merino, E., Corona, M., Boli-
var, F., Becerril, B. Peptides and genes cod-
ing for scorpion toxins that affect ion-chan-
nels. Biochimie 2000, 82, 861868
156 Grishin, E. Polypeptide neurotoxins from
spider venoms. Eur. J. Biochem. 1999, 264,
276280
References 495
157 Escoubas, P., Diochot, S., Corzo, G. Struc-
ture and pharmacology of spider venom
neurotoxins. Biochimie 2000, 82, 893907
158 Gueron, M., Ilia, R., Margulis, G. Arthropod
poisons and the cardiovascular system. Am.
J. Emerg. Med. 2000, 18, 708714
159 Gray, W. R., Luque, A., Olivera, B. M., Bar-
rett, J., Cruz, L. J. Peptide toxins from Conus
geographus venom. J. Biol. Chem. 1981, 256,
47344740
160 Arias, H. R., Blanton, M. P. Alpha-conotox-
ins. Int. J. Biochem. Cell. Biol. 2000, 32,
10171028
161 Craig, A. G., Bandyopadhyay, P., Olivera,
B. M. Post-translationally modified neuropep-
tides from Conus venoms. Eur. J. Biochem.
1999, 264, 271275
162 Nielsen, K. J., Schroeder, T., Lewis, R. Struc-
tureactivity relationships of omega-conotox-
ins at N-type voltage-sensitive calcium chan-
nels. J. Mol. Recognit. 2000, 13, 5570
163 McIntosh, J. M., Santos, A. D., Olivera, B. M.
Conus peptides targeted to specific nicotinic
acetylcholine receptor subtypes. Annu. Rev.
Biochem. 1999, 68, 5988
164 Favreau, P., Le Gall, F., Benoit, E., Molgo, J.
A review on conotoxins targeting ion chan-
nels and acetylcholine receptors of the verte-
brate neuromuscular junction. Acta Physiol.
Pharmacol. Ther. Latinoam 1999, 49, 257267
165 Olivera, B. M., Cruz, L. J. Conotoxins, in ret-
rospect. Toxicon 2001, 39, 714
166 Dutton, J. L., Craik, D. J. alpha-Conotoxins:
nicotinic acetylcholine receptor antagonists
as pharmacological tools and potential drug
leads. Curr. Med. Chem. 2001, 8, 327344
167 Chiang, J. S. New developments in cancer
pain therapy. Acta Anaesthesiol. Sin. 2000, 38,
3136
168 Jones, R. M., Bulaj, G. Conotoxins new vis-
tas for peptide therapeutics. Curr. Pharm.
Des. 2000, 6, 12491285
169 Diamond, G. Natures antibiotics: the poten-
tial of antimicrobial peptides as new drugs.
Biologist (London) 2001, 48, 209212.
Abstract
Targeted in situ radiotherapy is a strategy
intended to selectively eradicate dissemi-
nated cancer while sparing normal tissues
through the use of biomolecules to deliver
radionuclides that emit a-particles, b-parti-
cles or Auger electrons to tumor cells. The
field is more than 20 years old in concept,
but its potential role in the management
of malignancies has only recently been
fully appreciated. The success in treating
non-Hodgkins B-cell lymphoma (NHL)
using anti-CD20 monoclonal antibodies
conjugated to iodine-131 (
131
I-tositumo-
mab; Bexxar
) or yttrium-90 (
90
Y-tositumo-
mab tiuxetan; Zevalin
) has reinvigorated
the search for other biologically targeted
radiotherapeutic agents. In this chapter,
the basic principles of targeted in-situ
radiotherapy are introduced and important
advances are reviewed, including treat-
ment of NHL and leukemias as well as
eradication of minimal residual disease in
solid tumors using direct or pre-targeted
radioimmunotherapy (RIT). The clinical
experience in treating neuroendocrine ma-
lignancies using
90
Y-DOTATOC, an octa-
peptide analogue of somatostatin, is dis-
cussed. Finally, the application of short-
range Auger electron-emitters or a-emitters
for targeted radiotherapy of cancer is de-
scribed. Particular emphasis is placed on
111
In-DTPA-D-Phe
1
-octreotide for treat-
ment of somatostatin receptor-positive tu-
mors and
111
In-DTPA-hEGF for treatment
of epidermal growth factor receptor
(EGFR)-overexpressing breast cancer. The
exciting future of the field is exemplified
by research demonstrating surgical cleav-
age of specific gene sequences in cancer
cells using triplex-forming oligonucleotides
conjugated to Auger electron-emitters
(antigene radiotherapy).
Abbreviations
131
I-MIBG
131
I-metaiodobenzylguanidine
5-FU 5-fluorouracil
AML acute myelogenous leukemia
AMU atomic mass units
ASC autologous stem cell
BMT bone marrow transplantation
BUN blood urea nitrogen
CML chronic myeloid leukemia
CR complete remission
EC electron capture
F-FDG 18F-fluorodeoxyglucose
GBM Glioblastoma multiforme
GLUT glucose transporter
HAMA human anti-mouse antibody
HASA human anti-streptavidin anti-
bodies
497
5
Biopharmaceuticals as Targeting Vehicles for In situ Radiotherapy
of Malignancies
Raymond M. Reilly
ISBN: 3-527-31184-X
hEGF human epidermal growth factor
IFN interferon
LET linear energy transfer
LSC leukemic stem cells
MDR multi-drug resistance
MRD minimal residual disease
MTD maximum tolerated dose
NEt norepinephrine transporters
NHL non-Hodgkins B-cell lymphoma
ODG oligodendrogliomas
OS overall survival
PDRT Peptide-directed radiotherapy
PET positron-emission tomography
PFS progression-free survival
PR partial remission
PSA prostate-specific antigen
RIT radioimmunotherapy
SA streptavidin
SSR somatostatin receptors
TFO triplex-forming oligonucleotide
5.1
Introduction
Cancer is a major health issue worldwide.
The most common solid tumors are breast,
colorectal, ovarian, prostate and lung can-
cer, which account for more than 3.2 mil-
lion new cases annually, and 1.7 million
deaths each year [1]. In addition, large num-
bers of individuals are diagnosed with and
die each year from hematological malignan-
cies such as lymphomas (>166000 new
cases and 93000 deaths, respectively) or leu-
kemias (144000 new cases and 109000
deaths, respectively). Early detection com-
bined with advances in surgery and external
radiotherapy have improved the prognosis
for many patients with solid tumors in
which the disease is confined to the primary
anatomical site, but the outlook for patients
with advanced disseminated cancer remains
poor. Lymphomas and leukemias are more
responsive to radiation and systemic che-
motherapies than solid tumors, but patients
often relapse and fail salvage therapy, due to
the emergence of multi-drug resistance
(MDR) [2, 3]. MDR as well as hormonal re-
sistance are also major challenges in treat-
ing breast [4] and prostate cancer [5].
Current systemic chemotherapy poorly
differentiates between malignant and nor-
mal cells, and places patients at risk for
serious, dose-limiting and sometimes life-
threatening toxicities. Cancer biology re-
search has revealed important insights
however into how tumor cells function, es-
cape normal growth-regulatory mecha-
nisms, and develop the capability to invade
surrounding normal tissues and dissemi-
nate to vital organs, causing death. This
body of knowledge presents the opportu-
nity to design new biologically targeted
therapies for cancer that, in theory, should
be more selective for eradicating tumors
and less damaging to normal tissues.
These targeted therapies have the potential
to significantly impact the outcome of can-
cer patients, both in terms of improved
survival and quality of life. One such strat-
egy is the use of biomolecules as targeting
vehicles to deliver radionuclides selectively
to malignancies for in situ radiotherapy.
This is not an entirely new concept, since
radiolabeled monoclonal antibodies
(mAbs) directed against tumor-associated
antigens have been studied since the mid-
1980s for the treatment of cancer (i.e.,
radioimmunotherapy; RIT) [6]. However,
the potential role of targeted radiotherapy
in the management of cancer has only re-
cently been fully appreciated. In this chap-
ter, the principles of biomolecularly tar-
geted in situ radiotherapy are introduced,
and the most important recent advances
are reviewed and placed in the context of
the cancer problem.
5 Biomolecules as Targeting Vehicles for In situ Radiotherapy of Malignancies 498
5.2
Principles of Targeted In situ Radiotherapy
of Malignancies
The basic principle of targeted radiotherapy
of malignancies is that molecular transfor-
mations in cancer cells present targets for
specific interaction with biomolecules carry-
ing radionuclides, allowing selective deposi-
tion of lethal doses of DNA-damaging radia-
tion in tumors, but sparing normal tissues.
The strategy is therefore dependent on the
identification of an appropriate target and
the optimal design of a targeting vehicle-
radionuclide conjugate. Historically, the
simplest and most successful example of
targeted in situ radiotherapy of malignan-
cies is iodine-131 (
131
I), which is used to
eradicate residual primary and metastatic
thyroid cancer by taking advantage of ex-
pression of the sodium iodide symporter.
Another example of a targeted radiothera-
peutic agent is
131
I-metaiodobenzylguani-
dine (
131
I-MIBG) which is actively imported
into neuronal cells by norepinephrine trans-
porters (NE
t
) and is used for the treatment
of neuroblastoma [7]. More recently, it was
shown that fluorine-18 2-fluorodeoxyglu-
cose (
18
F-FDG), a simple radiolabeled glu-
cose analogue used for positron-emission
tomography (PET) of tumors by exploiting
up-regulation of glucose transporters
(GLUT1) and/or increased hexokinase lev-
els, can also kill breast cancer cells through
the DNA-damaging properties of the posi-
trons [8]. For the purposes of this chapter,
however, the focus will be on biomolecular
targeting vehicles, in particular mAbs that
specifically recognize cell surface tumor-as-
sociated epitopes; synthetic or endogenous
peptide growth factors which interact with
transmembrane tyrosine kinase receptors;
or triplex-forming oligonucleotides that
bind to specific oncogene sequences pres-
ent in the DNA of cancer cells.
5.2.1
Radionuclides for Targeted In situ
Radiotherapy of Malignancies
Radionuclides which may be conjugated to
targeting vehicles for in situ radiotherapy
of malignancies include a-emitters (e.g.,
211
At,
213
Bi or
225
Ac), b-emitters (e.g.,
131
I
or
90
Y) or low-energy Auger and conver-
sion electron-emitters (e.g.,
111
In,
125
I,
67
Ga and
123
I). a-Particles consist of a he-
lium nucleus (two protons and two neu-
trons), and have a mass of two atomic
mass units (AMU) and a double positive
charge. a-Particles deposit all of their en-
ergy over a very short track length (50
100 lm) in tissues, and are therefore
known as high linear energy transfer
(LET) radiation. The energy deposition of
a-particles may reach as high as
100 keV lm
1
, which is close to the theo-
retical optimal value for cell killing [9].
Radionuclides which emit b-particles are
most suitable for eradicating small clusters
of cancer cells.
b-Particles have the same mass as an
electron (1/1850th of an AMU), and a sin-
gle negative charge. The range of b-parti-
cles in tissues is much longer than for a-
particles, but is directly proportional to
their energy. For example, the range in tis-
sues of the b-particles emitted by iodine-
131 (
131
I; Eb=0.6 MeV) is about 2 mm,
while the range of the b-particles emitted
by yttrium-90 (
90
Y; Eb=2.3 MeV) is 12 mm
[9]. b-Particles deposit most of their energy
at the end of their track length, which
makes them most suitable for treating tu-
mors with a diameter of 212 mm. Smal-
ler tumors may receive a suboptimal radia-
tion absorbed dose due to deposition of
most of the energy outside the tumor vol-
ume [10]. Furthermore, the long path
length of b-particles (2001200 cell diame-
ters) is advantageous for treating large tu-
mor nodules in which there may be in-
complete targeting of malignant cells due
to heterogeneity in targeting vehicle deliv-
ery, since radioactivity targeted to tumor
cells can also kill neighboring non-targeted
cells (cross-fire effect). The cross-fire ef-
fect from b-particle emitters is responsible
however for non-specific toxicity to bone
marrow stem cells from targeted tumor
cells infiltrating the bone marrow, or sim-
ply from perfusion of the marrow by
radioactivity.
Auger electrons, which were first discov-
ered in 1929 by the French physicist,
Pierre Auger [11], are very low-energy elec-
trons emitted by radionuclides that decay
by electron capture (EC). In EC, a proton
in the nucleus captures an inner orbital
electron, decreasing the number of pro-
tons by one, and creating a vacancy in the
shell. The vacancy is filled by the decay of
an electron from a higher shell. The excess
energy released is imparted on an outer or-
bital electron, which is then ejected from
the atom, creating a doubly positive-
charged nucleus. In fact, the EC decay pro-
cess causes the release of a shower of Au-
ger electrons of discrete low energies
(<30 keV) that travel nanometer to micro-
meter distances in tissues (less than one
cell diameter). Auger electron-emitters are
high LET radiation, and their energy de-
position approaches that of a-emitters
(100 keV lm
1
). The subcellular range of
the electrons renders Auger electron-emit-
ting radionuclides most suitable for killing
single cancer cells. They are especially
damaging to DNA when the radionuclides
are internalized into the cytoplasm of the
cell, and particularly if transported to the
cell nucleus. There is no cross-fire effect
from Auger electron-emitters, theoretically
making them highly selective for killing
single targeted cancer cells while sparing
non-targeted normal cells. However, the
lack of a cross-fire effect limits their ability
to eradicate larger tumors, in which not all
cancer cells can be effectively targeted.
Conversion electrons are higher energy
and longer range electrons emitted by
radionuclides undergoing decay by inter-
nal conversion, and are sometimes emitted
by radionuclides which also emit Auger
electrons.
5.3
RIT of Non-Hodgkins B-Cell Lymphomas:
The Pre-eminent Success Story
RIT of relapsed NHL represents the pre-
eminent success story for in situ targeted
radiotherapy of malignancies (Table 5.1).
The majority of RIT trials of NHL have fo-
cused on the CD20 differentiation antigen,
a 35-kDa transmembrane glycoprotein dis-
played by >95% of B-cell lymphomas and
normal B-cells, but not present on early
progenitor B cells [12]. RIT of NHL using
131
I-labeled anti-CD20 mAbs was pio-
neered by two groups in the mid-1990s,
one at the University of Washington [13],
and a second group at the University of
Michigan [14]. The Seattle group adminis-
tered high doses (10.429.0 GBq; 280
785 mCi) of
131
I-labeled anti-CD20 mAb
B1 (murine IgG
2a
) to 29 patients with re-
lapsed NHL followed by autologous stem
cell (ASC) rescue, achieving complete re-
missions (CR) in 23 patients (79%) and
partial remission (PR) in two additional
patients, for an overall response rate of
86% [13, 15]. Furthermore, responses were
durable, with remissions lasting for 27
years in 14 of the 29 patients, suggesting
that myeloablative RIT may be curative in
some cases. Non-hematological toxicity
was mainly to the thyroid (due to in-vivo
catabolism of
131
I-mAb B1) manifested by
elevated thyroid stimulating hormone
T
a
b
l
e
5
.
1
R
e
p
r
e
s
e
n
t
a
t
i
v
e
c
l
i
n
i
c
a
l
t
r
i
a
l
s
o
f
r
a
d
i
o
i
m
m
u
n
o
t
h
e
r
a
p
y
(
R
I
T
)
o
f
n
o
n
-
H
o
d
g
k
i
n
s
B
-
c
e
l
l
l
y
m
p
h
o
m
a
R
a
d
i
o
i
m
m
u
n
o
c
o
n
j
u
g
a
t
e
T
a
r
g
e
t
D
o
s
e
N
o
.
o
f
p
a
t
i
e
n
t
s
R
e
s
p
o
n
s
e
T
o
x
i
c
i
t
y
R
e
f
e
r
e
n
c
e
C
R
P
R
O
v
e
r
a
l
l
r
e
s
p
o
n
s
e
[
%
]
9
0
Y
-
i
b
r
i
t
u
m
o
m
a
b
t
i
u
x
e
t
a
n
(
Z
e
v
a
l
i
n
)
C
D
2
0
7
.
4
1
4
.
8
M
B
q
/
k
g
5
1
1
3
2
1
6
8
H
e
m
a
t
o
l
o
g
i
c
;
H
A
M
A
2
4
9
0
Y
-
i
b
r
i
t
u
m
o
m
a
b
t
i
u
x
e
t
a
n
(
Z
e
v
a
l
i
n
)
C
D
2
0
1
4
.
8
M
B
q
k
g
1
7
3
2
2
3
6
8
0
H
e
m
a
t
o
l
o
g
i
c
;
n
a
u
s
e
a
2
7
9
0
Y
-
i
b
r
i
t
u
m
o
m
a
b
t
i
u
x
e
t
a
n
(
Z
e
v
a
l
i
n
)
C
D
2
0
1
1
.
0
M
B
q
k
g
1
3
0
1
1
1
2
7
7
H
e
m
a
t
o
l
o
g
i
c
;
n
a
u
s
e
a
2
6
9
0
Y
-
i
b
r
i
t
u
m
o
m
a
b
t
i
u
x
e
t
a
n
(
Z
e
v
a
l
i
n
)
C
D
2
0
1
4
.
8
M
B
q
k
g
1
5
7
8
3
4
7
4
H
e
m
a
t
o
l
o
g
i
c
;
n
a
u
s
e
a
,
c
h
i
l
l
s
2
9
1
3
1
I
-
a
n
t
i
-
B
1
C
D
2
0
1
0
4
0
0
2
9
0
0
0
M
B
q
2
9
2
3
2
8
6
S
e
v
e
r
e
h
e
m
a
t
o
l
o
g
i
c
(
A
S
C
r
e
s
c
u
e
)
;
e
l
e
v
a
t
e
d
T
S
H
;
c
a
r
d
i
o
p
u
l
m
o
n
a
r
y
1
5
1
3
1
I
-
t
o
s
i
t
u
m
o
m
a
b
(
B
e
x
x
a
r
)
C
D
2
0
7
5
c
G
y
(
w
h
o
l
e
b
o
d
y
)
5
9
2
0
2
2
7
1
H
e
m
a
t
o
l
o
g
i
c
;
H
A
M
A
;
e
l
e
v
a
t
e
d
T
S
H
;
m
y
e
l
o
d
y
s
p
l
a
s
i
a
1
7
1
3
1
I
-
t
o
s
i
t
u
m
o
m
a
b
(
B
e
x
x
a
r
)
C
D
2
0
6
5
7
5
c
G
y
(
w
h
o
l
e
b
o
d
y
)
4
5
1
5
1
2
5
7
H
e
m
a
t
o
l
o
g
i
c
;
f
a
t
i
g
u
e
;
n
a
u
s
e
a
;
f
e
v
e
r
;
H
A
M
A
1
9
1
3
1
I
-
t
o
s
i
t
u
m
o
m
a
b
(
B
e
x
x
a
r
)
C
D
2
0
2
0
0
0
2
7
0
0
c
G
y
(
c
r
i
t
i
c
a
l
o
r
g
a
n
s
)
3
1
2
4
3
8
7
S
e
v
e
r
e
h
e
m
a
t
o
l
o
g
i
c
(
A
S
C
r
e
s
c
u
e
)
;
m
y
e
l
o
d
y
s
p
l
a
s
i
a
2
1
1
3
1
I
-
t
o
s
i
t
u
m
o
m
a
b
(
B
e
x
x
a
r
)
C
D
2
0
6
5
7
5
c
G
y
(
w
h
o
l
e
b
o
d
y
)
2
5
8
1
1
7
6
H
e
m
a
t
o
l
o
g
i
c
;
p
r
u
r
i
t
i
s
;
c
h
i
l
l
s
;
r
a
s
h
2
3
1
3
1
I
-
t
o
s
i
t
u
m
o
m
a
b
(
B
e
x
x
a
r
)
C
D
2
0
1
7
0
0
3
1
0
0
(
c
r
i
t
i
c
a
l
o
r
g
a
n
s
)
2
7
2
3
2
9
3
S
e
v
e
r
e
h
e
m
a
t
o
l
o
g
i
c
(
A
S
C
r
e
s
c
u
e
)
;
m
y
e
l
o
d
y
s
p
l
a
s
i
a
2
2
1
3
1
I
-
t
o
s
i
t
u
m
o
m
a
b
(
B
e
x
x
a
r
)
C
D
2
0
6
5
7
5
c
G
y
(
w
h
o
l
e
b
o
d
y
)
4
1
2
0
1
1
7
6
H
e
m
a
t
o
l
o
g
i
c
;
m
y
e
l
o
d
y
s
p
l
a
s
i
a
;
e
l
e
v
a
t
e
d
T
S
H
;
H
A
M
A
1
5
0
1
3
1
I
-
h
L
L
2
C
D
2
2
5
0
1
0
0
c
G
y
(
b
o
n
e
m
a
r
r
o
w
)
1
4
1
1
1
4
H
e
m
a
t
o
l
o
g
i
c
3
0
9
0
Y
-
h
L
L
2
C
D
2
2
5
0
1
0
0
c
G
y
(
b
o
n
e
m
a
r
r
o
w
)
8
0
2
2
5
H
e
m
a
t
o
l
o
g
i
c
3
0
1
3
1
I
-
L
y
m
-
1
L
y
m
-
1
1
5
0
0
3
7
0
0
M
B
q
m
2
1
4
7
1
0
7
1
H
e
m
a
t
o
l
o
g
i
c
3
1
C
R
:
C
o
m
p
l
e
t
e
r
e
m
i
s
s
i
o
n
;
P
R
:
P
a
r
t
i
a
l
r
e
m
i
s
s
i
o
n
;
A
S
C
:
A
u
t
o
l
o
g
o
u
s
s
t
e
m
c
e
l
l
;
T
S
H
:
T
h
y
r
o
i
d
-
s
t
i
m
u
l
a
t
i
n
g
h
o
r
m
o
n
e
;
H
A
M
A
:
H
u
m
a
n
a
n
t
i
-
m
o
u
s
e
a
n
t
i
b
o
d
y
.
(TSH) in 60% of patients, despite attempts
at blocking thyroid uptake of free
131
I
using Lugols iodine and saturated solution
of KI.
The Michigan group administered much
lower patient-specific doses of
131
I-mAb B1
(34161 mCi) intended to deliver a maxi-
mum radiation absorbed dose of 25
85 cGy (rads) to the whole body (correlated
with bone marrow toxicity) in 28 patients
with relapsed NHL, achieving CR in 14 of
28 (50%) and PR in 8 of 28 (29%) patients
[14, 16]. Hematologic toxicity was dose-lim-
iting, with the maximum tolerated dose
(MTD) corresponding to a whole-body ra-
diation dose of 75 cGy (rads). Despite the
lower CR rate compared to myeloablative
RIT using
131
I-mAb B1 [13, 15], responses
were also long-lasting, with 6/28 patients
remaining in CR for 1631 months after
treatment. A follow-up report in a larger
group of 59 patients with NHL showed
that the overall response rate to
131
I-mAb
B1 was 71%, with 34% of patients achiev-
ing a CR [17]. The median duration of re-
sponse was 12 months, but seven patients
remained in CR for between 3 and 5.7
years. These early trials set the stage for
intensive clinical investigation of RIT of
NHL over the next 10 years that ultimately
led to the development and regulatory ap-
proval of two new targeted radiotherapeu-
tic agents for the disease: 1)
131
I-tositumo-
mab (
131
I-mAb B1; Bexxar
; Corixa); and
2) yttrium-90 ibritumomab tiuxetan (
90
Y-
mAb Y2B8; Zevalin
; Schering) [18].
5.3.1
131
I-tositumomab (Bexxar
)
The pivotal multicenter Phase II trial of
131
I-tositumomab (Bexxar
) enrolled 47
patients with chemotherapy-relapsed/re-
fractory NHL [19]. Patients first received
450 mg of unlabeled mAb B1, followed by
35 mg (185 MBq; 5 mCi) of
131
I-tositumo-
mab for imaging. The unlabeled mAb B1
was intended to saturate CD20 sites on
normal B cells in the blood and spleen,
thereby enhancing tumor uptake of
131
I-to-
situmomab [18]. Dosimetry was performed
by c-scintigraphy. The treatment dose was
then selected to deliver a total body radia-
tion absorbed dose of 75 cGy (rads) for pa-
tients with platelet counts
>150000 cells mm
3
, or 65 cGy for patients
with platelet counts between 100 000 and
150000 cells mm
3
. Although, the desired
radiation absorbed dose to lymphoma de-
posits should also be important in select-
ing the optimal dose for RIT, these studies
have found no correlation between the ra-
diation dose deposited in tumors and the
response to
131
I-tositumomab, suggesting
that there are mechanisms in addition to
radiation damage responsible for cytotoxi-
city [20]. The therapeutic doses of
131
I-tosi-
tumomab ranged from 1.7 to 6.5 GBq (45
to 177 mCi). The overall response rate was
57%, with 32% of patients achieving a CR.
The median duration of response was 10
months. The most frequent and dose-lim-
iting toxicity was myelosuppression mani-
fested as reversible thrombocytopenia,
neutropenia or anemia which required
platelet or red blood cell (RBC) infusions
or colony-stimulating factors in 2025% of
cases.
131
I-tositumomab is a murine anti-
body, but only one patient developed a hu-
man anti-mouse antibody (HAMA) re-
sponse, likely due to an attenuated im-
mune response in patients with NHL.
There is a long-term risk for development
of myelodysplasia, and in rare cases
acute myelogenous leukemia (AML) in pa-
tients receiving
131
I-tositumomab, but the
risk does not appear to be greater than for
patients treated with high-dose chemother-
apy, especially alkylating agents [17]. Ele-
vated TSH was found in 8.5% in patients
treated with non-myeloablative doses of
131
I-tositumomab [18].
In an effort to improve the survival of
patients with relapsed NHL, a Phase I/II
trial of myeloablative doses of
131
I-tositu-
momab in combination with etoposide, cy-
clophosphamide and ASC rescue was per-
formed in 52 patients [21]. The dose of
131
I-tositumomab was selected to achieve a
target radiation absorbed dose of 2000
2700 cGy (rads) to critical normal organs
(liver, kidneys or lungs), and ranged from
10.1 to 31.1 GBq (272840 mCi; 108
347 mg). These radiation-absorbed doses
are much higher than is tolerated by the
bone marrow (100200 cGy), and were
only feasible due to the incorporation of
ASC rescue into the protocol. The overall
survival (OS) and progression-free survival
(PFS) at 2 years were 83% and 68%, re-
spectively. This compared well with a non-
randomized control group treated with
bone marrow transplantation (BMT),
whole body irradiation, etoposide and cy-
clophosphamide (OS and PFS of 53% and
36% at 2 years, respectively). A direct com-
parison of myeloablative RIT with
131
I-tosi-
tumomab and ASC versus conventional
whole-body irradiation, high-dose chemo-
therapy and BMT for treatment of NHL
was subsequently performed in 125 pa-
tients [22]. Higher 5-year OS and PFS
were observed for RIT (67% and 48%, re-
spectively) compared to conventional BMT
(53% and 29%, respectively). Unfortu-
nately, despite the good long-term survival
achieved, most patients ultimately devel-
oped recurrent lymphoma.
5.3.2
90
Y-ibritumomab tiuxetan (Zevalin
)
90
Y-ibritumomab tiuxetan (Zevalin
) is the
murine IgG
1
j anti-CD20 mAb Y2B8 cova-
lently linked to the radiometal chelator iso-
thiocyanatobenzyl MX-DTPA (tiuxetan)
which strongly binds the b-emitting radio-
nuclide,
90
Y. An advantage of
90
Y com-
pared to
131
I as a radionuclide for RIT of
NHL is its higher b-energy (2.3 MeV ver-
sus 0.6 MeV), which provides a longer
path length in tumors (12 mm versus
2 mm), thereby achieving killing of non-
targeted tumor cells through a cross-fire
effect. This is an important issue for the
treatment of large tumor deposits in which
there may be inadequate delivery of the
agent [6]. Secondly, the absence of c-emis-
sions permits out-patient treatment, since
patients do not pose a radiation hazard to
family members or the public due to the
non-penetrating properties of b-particles
[18]. Nevertheless, it was shown recently
that
131
I-tositumomab (which emits a pe-
netrating c-photon of 364 keV) can also be
safely administered in non-myeloablative
doses (<3.77.4 GBq) on an out-patient ba-
sis, provided that appropriate radiation
safety protocols are followed [23]. The ab-
sence of c-radiation for
90
Y-ibritumomab
tiuxetan, necessitates that dosimetry stud-
ies be performed using the mAb labeled
with the c-emitter
111
In (Fig. 5.1).
In a multicenter Phase I/II trial, 51 pa-
tients with relapsed/refractory NHL were
treated with escalating doses of
90
Y-ibritu-
momab tiuxetan ranging from
7.4 MBq kg
1
(0.2 mCi kg
1
) to
14.8 MBq kg
1
(0.4 mCi kg
1
) [24]. The
antibody dose was 2 mg in each case. Sim-
ilar to the studies with
131
I-tositumomab,
patients first received an infusion of
250 mg m
2
of anti-CD20 chimeric mAb,
rituximab (Rituxan
; Roche Pharmaceuti-
cals) to pre-saturate accessible CD20 bind-
ing sites on normal B cells in the blood
and spleen. Patients then received
185 MBq (5 mCi; 1.6 mg) of
111
In-labeled
ibritumomab tiuxetan to estimate the ra-
diation absorbed dose to critical organs or
to the bone marrow from
90
Y-ibritumomab
tiuxetan (limited to <2000 cGy or 300 cGy,
respectively) by imaging. Patients with
>25% bone marrow involvement were ex-
cluded due to their high risk for bone mar-
row toxicity.
90
Y-ibritumomab tiuxetan was
administered 7 days after the rituximab in-
fusion. The overall response rate was 67%,
with 13 patients achieving a CR and 21 pa-
tients exhibiting a PR. The median dura-
tion of response was approximately 1 year.
A dose of 14.8 MBq kg
1
(4 mCi kg
1
) was
safely administered to NHL patients with
platelet counts >150000 mm
3
, but a dose
reduction to 11.1 MBq kg
1
(0.3 mCi kg
1
)
was needed for patients with platelet
counts of 100000150000 mm
3
[25]. The
maximum total dose administered was
1184 MBq (32 mCi). The toxicity was
mainly hematologic, manifested as dose-
related thrombocytopenia, neutropenia and
anemia. Only one patient developed a
HAMA response. More recent trials have
similarly demonstrated a high response
rate (overall response of 83%, and CR in
37% of patients) to
90
Y-ibritumomab tiuxe-
tan administered to patients at
11.1 MBq kg
1
(0.3 mCi kg
1
) because of
pre-existing mild thrombocytopenia
(100000150000 platelets per mm
3
) [26].
The median duration of response in these
patients was 912 months.
One of the most important studies in
targeted radiotherapy of malignancies was
a Phase III clinical trial in which 143 pa-
tients with relapsed/recurrent NHL were
randomized to receive RIT with
90
Y-ibritu-
momab tiuxetan or immunotherapy with
rituximab [27]. Usually, only about half of
NHL patients respond to rituximab [28].
The RIT protocol involved infusions of
250 mg m
2
of rituximab on days 1 and 8,
111
In-ibritumomab tiuxetan (185 MBq,
1.6 mg) on day 1 (for dosimetry studies),
and treatment with
90
Y-ibritumomab
(14.8 MBq kg
1
) on day 8. The maximum
dose of
90
Y-ibritumomab tiuxetan used was
1184 MBq (32 mCi). Patients in the rituxi-
mab arm received four-weekly infusions of
375 mg m
2
. The overall response rate was
much higher for patients receiving RIT
with
90
Y-ibritumomab tiuxetan than for pa-
tients treated with rituximab (80% versus
56%, respectively). In addition, the num-
ber of CR in the RIT group (22/73; 30%)
was double that of the group treated with
rituximab (11/70; 16%). There were no sig-
nificant differences in the duration of re-
sponse in patients treated with either
90
Y-
ibritumomab tiuxetan (14 months) or ri-
tuximab (12 months). Perhaps even more
interesting, however, was that in a study of
57 patients with refractory NHL that was
unresponsive to rituximab, 74% responded
to RIT using
90
Y-ibritumomab tiuxetan
Fig. 5.1 Whole-body images of a patient with non-
Hodgkins lymphoma at selected times after ad-
ministration of
111
In-labeled ibritumomab tiuxetan
(Zevalin
) demonstrating strong uptake of radio-

activity in regions of periaortic lymphadenopathy
(red arrows). Normal liver accumulation of Zeva-
lin
is also observed. [Reprinted from Witzig T. E.,

et al., J. Clin. Oncol. 1999; 37933803.]
[29]. The proportion of CR (15%) in this
patient population was lower than in pa-
tients not refractory to rituximab [27].
These two studies introduced the new con-
cepts that: 1) the response to immunother-
apy may be enhanced by incorporation of
a radionuclide; and 2) a lack of response
to immunotherapy does not preclude a re-
sponse to RIT, despite the fact that both
modalities exploit the same target.
5.4
Other Strategies for In situ Radiotherapy
of Non-Hodgkins Lymphoma
Other strategies for the in situ radiotherapy
of NHL which have been explored include
RIT with the anti-CD22 humanized mAb
hLL2 (Immunomedics) labeled with
131
I or
90
Y [30] or with the murine anti-HLA-DR
mAb Lym-1 labeled with
131
I [31]. The re-
sponse rate in two very small cohorts of
patients with recurrent NHL treated with
131
I- or
90
Y-hLL2 intended to deliver a max-
imum of 50100 cGy (rads) to the bone
marrow was relatively low (2/13 and 2/7
patients responding, respectively) in com-
parison to that observed for
131
I-tositumo-
mab (6070%) or
90
Y-ibritumomab tiuxe-
tan in larger patient groups (7085%) [18].
The response rate to treatment with
1.5 GBq m
2
(40 mCi m
2
) to 3.7 GBq m
2
(100 mCi m
2
) of
131
I-Lym-1 was higher,
with 10/14 patients responding. A new
approach to in situ radiotherapy of NHL
(and also multiple myeloma) may be to
target B-lymphocyte stimulator (BLyS) re-
ceptors displayed by immunoglobulin-posi-
tive B-cells [32]. Preclinical studies have
shown avid and specific localization of
125
I-BLyS in B-cell tumors in mice as well
as in normal tissues harboring large num-
bers of B cells (e.g., spleen) [33]. The 30-
fold faster blood clearance of BlyS in mice
compared to immunoglobulins (due to its
low molecular weight of 52 kDa) suggests
that treatment with
131
I-BlyS may cause
less bone marrow toxicity than RIT, since
the residence time of radiotherapeutic
agents in the blood has been directly corre-
lated with hematologic toxicity [34].
5.5
Radioimmunotherapy of AML:
Success but not Cure
Radioimmunotherapy of AML also repre-
sents a success story for in situ targeted
radiotherapy of malignancies. Despite the
fact that this strategy has been shown to
kill massive numbers (>10
12
) of leukemic
cells in patients [35], it has been unable
completely to eradicate leukemic stem
cells (LSC) in the bone marrow, which
eventually cause recurrence. RIT of AML
has focused mainly on the murine mAb
M195 or its humanized analogue,
HuM195, that specifically bind the CD33
epitope displayed by normal and leukemic
myeloid cells, but not by granulocytes or
the earliest pluripotent myeloid stem cells
[36]. CD33 is a 67-kDa member of the sia-
loadhesin family of cell surface glycopro-
teins [37]. RIT of AML was pioneered by
Dr. David Scheinbergs group at Memorial
Sloan-Kettering Hospital in New York [38].
In a Phase 1 trial, 24 patients with leuke-
mia were treated with 1850 MBq m
2
(50 mCi m
2
) to 7770 MBq m
2
(210 mCi
m
2
) of
131
I-M195. Decreased leukemic
blasts were found in the bone marrow of
89% of patients, with >99% of blasts killed
in some cases [35]. One patient developed
hepatic toxicity, but all other patients ex-
perienced insignificant non-hematologic
toxicity. Hematological toxicity was difficult
to ascertain in this patient population how-
ever due to prior disease-related and che-
motherapy-related pancytopenia, but total
doses >5 GBq (135 mCi) produced pro-
longed (>14 days) myelosuppression. In
fact, profound myelosuppression allowed
eight patients to proceed to BMT. In sub-
sequent trials, high doses of
131
I-M195 or
HuM195 (44408510 MBq m
2
) were incor-
porated into a BMT preparatory regimen
for various leukemias [39]. CR was
achieved in 28/30 patients, and 3/16 pa-
tients with AML remained in remission
for 4.5 to 8 years following BMT [38]. In
the same report, lower, non-myeloablative
doses of
131
I-M195 (1850 MBq m
2
or
2590 MBq m
2
) were administered to sev-
en patients to eradicate minimal residual
disease [39]. The median OS was 28
months, which compared favorably with
more conventional treatment regimens for
leukemia.
The high incidence of HAMA (40%) in
patients treated with
131
I-M195 necessi-
tated humanization of the antibody for
subsequent RIT trials [40]. Interestingly,
humanization of M195 enhanced its affini-
ty 3-fold for binding CD33 [41]. RIT of
AML was studied in a Phase I trial of 19
patients administered 3.7 MBq kg
1
(0.1 mCi kg
1
) to 11.1 MBq kg
1
(0.3 mCi
kg
1
) of
90
Y-HuM195 [38]. Transient hepat-
ic toxicity was observed in 11 patients. Pro-
longed, dose-related myelosuppression was
found, but all patients treated at the high-
est dose showed an absence of leukemic
cells in the bone marrow, suggesting that
RIT with
90
Y-HuM195 may be useful for
BMT preparation. A review of the use of
myeloablative RIT as a preparatory regi-
men for leukemia, lymphoma and other
malignancies has recently been published
[42].
5.5.1
Treatment of AML using HuM195 Labeled
with a-Emitters
Scheinbergs group have explored the poten-
tial for RIT of AML using HuM195 conju-
gated to short-range a-emitting radionu-
clides such as
213
Bi or
225
Ac [38]. Due to
the short range of a-particles in tissues (5
10 cell diameters), it was anticipated that
these radioimmunoconjugates would be
more selective for killing leukemia cells
and would significantly diminish non-spe-
cific toxicity to normal cells, which occurs
with mAbs conjugated to the long-range b-
particle-emitters,
131
I or
90
Y due to the
cross-fire effect.
213
Bi has a short half-life
of 46 min, and emits a single a-particle with
energy of 8 MeV as well as a c-photon of
440 keV, which is useful for dosimetric
imaging. In the first Phase I trial of a-parti-
cle RIT for leukemia, 18 patients with re-
lapsed/refractory AML or chronic myeloid
leukemia (CML) were treated with increas-
ing doses (10.4 MBq kg
1
; 0.28 mCi kg
1
)
to 37 MBq kg
1
(1 mCi kg
1
) of
213
Bi-
HuM195 [43]. Virtually all patients devel-
oped reversible myelosuppression, possibly
due to targeting and/or non-specific irradia-
tion of CD33-positive monocytes and mye-
loid progenitors in the bone marrow by
213
Bi-HuM195. Minor and transient liver
function abnormalities were noted in four
patients. Strong antileukemic effects were
observed for
213
Bi-HuM195 therapy, with
14/15 evaluable patients achieving signifi-
cant reductions in circulating leukemic
blasts, and 14/18 patients showing de-
creased bone marrow blasts. Dosimetry
studies revealed that the radiation absorbed
dose delivered to sites that harbor leukemic
cells (bone marrow, liver and spleen) was
theoretically 1000-fold greater than that for
HuM195 labeled with the b-emitters,
131
I
or
90
Y [44]. Unfortunately, no patients
achieved CR, possibly due to inadequate tar-
geting of inaccessible CD33-positive leuke-
mic cells or the sparing of CD33-negative
LSCs in the bone marrow.
One possible reason that
213
Bi-HuM195
was unable to induce CR in patients was that
the short half-life (46 min) of the radionu-
clide restricted targeting and killing to only
the most accessible leukemic cells. Schein-
bergs group attempted to address this issue
by conjugating HuM195 to
225
Ac, which has
a much longer half-life (10 days) and yields
six daughter radionuclides in its decay
scheme to stable
209
Bi; these are either a-
emitters (
221
Fr,
217
At,
213
Bi or
213
Po) or b-
emitters (
213
Bi,
209
Tl or
209
Po) [45]. Using
this targeted atomic nanogenerator
approach and the appropriate internalizing
225
Ac-conjugated mAbs, many different
types of cancer cells (leukemia, lymphoma,
breast, ovarian, neuroblastoma and pros-
tate) were killed at extremely low, becquerel
(picocurie) levels, about 1000-fold lower
than was required for
213
Bi analogues [46].
In addition, tumor regression and pro-
longed survival were achieved in mice im-
planted with LNCaP human prostate cancer
xenografts or disseminated Daudi human B-
cell lymphomas. However, a recent report by
Scheinbergs group found that
225
Ac-
HuM195 caused severe renal tubular dam-
age and anemia in cynomolgus monkeys
[47]. The MTD of
225
Ac-HuM195 in mon-
keys was <28 kBq kg
1
. Renal toxicity was
speculated to be due to the release of
213
Bi
(and other daughter radionuclides) from
the decay of
225
Ac, and their subsequent re-
distribution and accumulation by the kid-
neys. Although
225
Ac-HuM195 may yet be
studied in patients with AML, these unanti-
cipated preclinical findings represent a ma-
jor limitation to the strategy that could result
in serious and dose-limiting normal tissue
toxicity in humans. It may in fact be better
to utilize radionuclides that decay to stable
elements, so that the redistribution of decay
products does not cause toxicity to normal
tissues. Perhaps the greatest challenge in
RIT of AML, however, remains targeting
the LSC. RIT targeting CD33 kills differen-
tiated leukemic cells in the blood, but may
spare LSCs in the bone marrow, thereby al-
lowing a recurrence of the disease. Although
not yet examined, it is probable that RIT tar-
geting the CD123 epitope or other epitopes
found on LSCs [48] may be more effective
than strategies exploiting CD33.
5.6
RIT of Solid Tumors: Encouraging Results
in Minimal Residual Disease
In contrast to the success in treating hema-
tological malignancies such as lymphoma
and leukemia, RIT of the more common
solid tumors such as breast, colon, ovary
or prostate has been severely limited by tox-
icity to the bone marrow, poor tumor pene-
tration of the radioimmunoconjugates, a 3-
fold lower radiosensitivity of these malig-
nancies, and the development of an im-
mune response to the mAbs, which re-
stricted re-treatment [6]. Although a few pa-
tients have achieved CR, most experienced
PR or minor responses (MR) (Table 5.2).
One niche where RIT may be particularly
useful in solid tumors, is in the eradication
of minimal residual disease (MRD). Small
tumor deposits and micrometastases accu-
mulate disproportionately high amounts of
radioimmunoconjugates compared to larg-
er tumors, and are theoretically more re-
sponsive to treatment [49, 50]. Behr et al.
[51] reported the results of a Phase II trial
of 30 colorectal cancer patients with MRD
treated with 2.22 GBq m
2
(60 mCi m
2
) of
131
I-hMN-14 humanized anti-CEA mAb.
Twenty-one patients had chemo-refractory
small tumors (diameter 3 cm), and an-
5.6 RIT of Solid Tumors: Encouraging Results in Minimal Residual Disease 507
T
a
b
l
e
5
.
2
S
o
m
e
r
e
p
r
e
s
e
n
t
a
t
i
v
e
c
l
i
n
i
c
a
l
t
r
i
a
l
s
o
f
r
a
d
i
o
i
m
m
u
n
o
t
h
e
r
a
p
y
(
R
I
T
)
o
f
s
o
l
i
d
t
u
m
o
r
s
M
a
l
i
g
n
a
n
c
y
R
a
d
i
o
i
m
m
u
n
o
-
T
a
r
g
e
t
D
o
s
e
[
G
B
q
m
2
]
N
o
.
o
f
R
e
s
p
o
n
s
e
T
o
x
i
c
i
t
y
R
e
f
e
r
e
n
c
e
c
o
n
j
u
g
a
t
e
p
a
t
i
e
n
t
s
C
R
P
R
/
M
R
S
D
C
o
l
o
r
e
c
t
a
l
;
p
a
n
c
r
e
a
t
i
c
;
b
i
l
e
d
u
c
t
1
3
1
I
-
C
C
4
9
T
A
G
-
7
2
1
.
8
5
1
1
.
1
1
5
0
0
0
H
e
m
a
t
o
l
o
g
i
c
;
H
A
M
A
1
5
1
C
o
l
o
r
e
c
t
a
l
;
l
u
n
g
;
m
e
d
u
l
l
a
r
y
t
h
y
r
o
i
d
9
0
Y
-
c
T
8
4
.
6
6
C
E
A
0
.
1
8
0
.
8
1
2
2
0
2
3
H
e
m
a
t
o
l
o
g
i
c
;
H
A
C
A
1
5
2
C
o
l
o
r
e
c
t
a
l
c
a
n
c
e
r
(
s
m
a
l
l
v
o
l
u
m
e
d
i
s
e
a
s
e
)
1
3
1
I
-
h
M
N
-
1
4
C
E
A
2
.
2
2
3
0
0
1
1
0
H
e
m
a
t
o
l
o
g
i
c
5
1
C
o
l
o
r
e
c
t
a
l
c
a
n
c
e
r
9
0
Y
-
c
T
8
4
.
6
6
+
c
h
e
m
o
t
h
e
r
a
p
y
C
E
A
0
.
6
1
0
.
7
6
2
1
0
0
1
1
H
e
m
a
t
o
l
o
g
i
c
;
H
A
C
A
5
4
B
r
e
a
s
t
c
a
n
c
e
r
1
3
1
I
-
C
C
4
9
T
A
G
-
7
2
2
.
2
2
2
.
7
7
1
5
0
3
0
H
e
m
a
t
o
l
o
g
i
c
;
H
A
M
A
1
5
3
B
r
e
a
s
t
c
a
n
c
e
r
9
0
Y
-
B
r
E
3
M
u
c
i
n
0
.
2
3
0
.
3
4
6
0
3
0
H
A
M
A
1
5
4
O
v
a
r
i
a
n
c
a
n
c
e
r
1
7
7
L
u
-
C
C
4
9
T
A
G
-
7
2
0
.
3
7
1
.
1
1
1
2
0
1
1
H
e
m
a
t
o
l
o
g
i
c
1
5
5
O
v
a
r
i
a
n
c
a
n
c
e
r
1
3
1
I
-
M
N
-
1
4
C
E
A
1
.
1
1
1
.
8
5
1
4
1
1
0
H
e
m
a
t
o
l
o
g
i
c
;
H
A
M
A
1
5
6
O
v
a
r
i
a
n
c
a
n
c
e
r
9
0
Y
-
B
7
2
.
3
T
A
G
-
7
2
0
.
2
4
1
.
1
1
3
8
2
2
0
H
e
m
a
t
o
l
o
g
i
c
;
H
A
M
A
1
5
7
O
v
a
r
i
a
n
c
a
n
c
e
r
9
0
Y
-
C
C
4
9
+
c
h
e
m
o
t
h
e
r
a
p
y
T
A
G
-
7
2
0
.
5
2
0
.
8
9
2
0
0
2
0
H
e
m
a
t
o
l
o
g
i
c
;
H
A
M
A
1
5
8
P
r
o
s
t
a
t
e
c
a
n
c
e
r
9
0
Y
-
C
Y
T
-
3
5
6
P
S
M
A
0
.
0
7
0
.
4
4
1
2
0
0
0
H
e
m
a
t
o
l
o
g
i
c
;
H
A
M
A
1
5
9
P
r
o
s
t
a
t
e
c
a
n
c
e
r
9
0
Y
-
J
5
9
1
P
S
M
A
0
.
1
8
0
.
7
4
2
9
0
2
4
H
e
m
a
t
o
l
o
g
i
c
5
9
G
l
i
o
b
l
a
s
t
o
m
a
1
3
1
I
-
B
C
-
2
/
B
C
-
4
T
e
n
a
s
c
i
n
0
.
7
6
1
.
6
3
7
4
1
9
1
0
B
r
a
i
n
e
d
e
m
a
;
H
A
M
A
5
3
C
R
:
C
o
m
p
l
e
t
e
r
e
m
i
s
s
i
o
n
;
P
R
:
P
a
r
t
i
a
l
r
e
m
i
s
s
i
o
n
;
M
R
:
M
i
n
o
r
r
e
m
i
s
s
i
o
n
;
S
D
:
s
t
a
b
l
e
d
i
s
e
a
s
e
;
H
A
M
A
:
H
u
m
a
n
a
n
t
i
-
m
o
u
s
e
a
n
t
i
b
o
d
y
;
H
A
C
A
:
H
u
m
a
n
a
n
t
i
-
c
h
i
m
e
r
i
c
a
n
t
i
b
o
d
y
.
other nine were treated in an adjuvant set-
ting following surgical resection of liver me-
tastases with curative intent. The response
rate in patients with measurable tumors
was 16% (three PRs and eight MRs) (Fig.
5.2). In the adjuvant setting, seven of nine
patients remained disease-free for up to 3
years. Reversible hematologic toxicity was
the predominant adverse effect. Although
the immune response to
131
I-hMN-14 was
not measured, there was no change in the
pharmacokinetics of the radioimmunocon-
jugates (reflective of immune response) in
five initially responding patients who were
re-treated at 816 months. Four of these pa-
tients achieved PR, and in one the disease
was re-stabilized.
An analogous application in which RIT
has yielded encouraging results is in the
treatment of MRD in brain malignancies
[52]. Glioblastoma multiforme (GBM) is
one of the most deadly and rapidly grow-
ing tumors, and patients generally have a
life expectancy of less than 1 year from
the time of diagnosis. Tumor recurrences
occur following surgical resection due to
the infiltrative nature of the disease, and
patients frequently die within 46 months
of relapse. RIT offers a unique opportunity
to eradicate MRD in GBMs in order to
prolong survival. RIT has focused mainly
on targeting tenascin, a glycoprotein found
in the extracellular matrix of >90% of
brain tumors. Zalutsky et al. [52] treated
more than 300 GBM patients with
131
I-
81C6 anti-tenascin mAb administered by
direct injection into the surgical cavity. In
two Phase I studies, the maximum toler-
ated dose (MTD) of
131
I-81C6 was 3.7
4.4 GBq (100120 mCi). The predominant
adverse effects were neurologic toxicity
and reversible myelosuppression. Patients
in these early dose-finding studies survived
for 1420 months. In a follow-up Phase II
study in 33 patients with GBM treated
with 4.4 GBq (120 mCi) of
131
I-81C6, the
5.6 RIT of Solid Tumors: Encouraging Results in Minimal Residual Disease 509
Fig. 5.2 Radioimmunotherapy (RIT) with
2220 MBq m
2
(60 mCi m
2
) of
131
I-anti CEA hu-
manized monoclonal antibody hMN14 reduced
the size of a 3-cm diameter hepatic lesion (larger
arrow) by more than 50%, whereas a smaller,
1-cm lesion (smaller arrow) disappeared com-
pletely. The top set of CT images were acquired
prior to RIT, and the bottom set 3 months after
treatment. [Reprinted from Behr T.M., et al., Can-
cer 2002; 94: 13731381.]
median survival was 2022 months [52].
Riva et al. [53] similarly treated 111 pa-
tients with various brain malignancies
with
131
I-BC-2 or BC-4 anti-tenascin mAbs
administered through a catheter into the
surgical cavity. The MTD from Phase I
trials was 2.6 GBq (70 mCi), based on the
incidence of brain edema. HAMA oc-
curred in 59% of patients. In Phase II
trials utilizing 1.32.8 GBq of
131
I-BC-2 or
BC-4, the median survival of patients
ranged from 19 months in GBM (Fig. 5.3)
to 3146 months in oligodendrogliomas
(ODG). Most notably, three of seven ODG
patients survived for 7 years, and two were
still alive at 9 years. In addition, seven of
74 GBM patients survived for 2 years, and
two were still alive at 4 years. The im-
proved survival of patients with brain ma-
lignancies receiving RIT suggests a prom-
ising future role for the strategy in manag-
ing these tumors, which have a poor prog-
nosis and are difficult to treat.
In order to improve the response of sol-
id tumors to RIT, recent trials have fo-
cused on combining RIT with cytotoxic
agents that radiosensitize tumors or bio-
logic agents that up-regulate the target epi-
tope. In one study, 21 patients with meta-
static colorectal cancer were treated with
0.61 GBq m
2
(16.6 mCi m
2
) or
0.76 GBq m
2
(20.6 mCi m
2
) of
90
Y-
cT84.66 anti-CEA mAb combined with in-
creasing doses (7001000 mg m
2
) of 5-
fluorouracil (5-FU) [54]. 5-FU is an effec-
tive cytotoxic agent for colorectal cancer,
and is also a known radiosensitizer [55].
No objective responses were observed, but
11 patients with progressive disease were
stabilized for 38 months. In another
study, patients with ovarian cancer were
administered 0.52 GBq m
2
(14 mCi m
2
)
to 0.89 GBq m
2
(24.2 mCi m
2
) i.p. of
90
Y-
CC49 anti-tumor-associated glycoprotein-72
(TAG-72) antigen mAbs combined with in-
terferon a2b (IFN-a2b; four s.c. doses of
310
6
units) and paclitaxel (100 mg m
2
,
i.p.). IFN-a2b has been shown to up-regu-
late the expression of TAG-72 by tumor
cells and improve the tumor uptake of
anti-TAG-72 radioimmunoconjugates in
patients [56]. There were two PRs of 2 and
4 months duration in nine patients with
measurable disease, and in 11 patients
with non-measurable disease, the median
time to disease progression being 6
months. Four patients had no evidence of
disease at 9 to 23 months following RIT.
Long-term survival has also been reported
for ovarian cancer patients who received
RIT more than 1015 years earlier [57].
The survival for patients treated with
0.67 GBq m
2
(18 mCi m
2
) of
90
Y-HMFG1
anti-mucin mAbs following standard cis-
platinum chemotherapy at 10 years was
78%, much higher than a similar cohort
of patients who received only conventional
chemotherapy (42% at 5 years). Palliative re-
sponses to RIT in prostate cancer (i.e., de-
creased pain) [58] and significant de-
creases/stabilization in prostate-specific
antigen (PSA) associated with objective tu-
mor responses [59] have also been achieved.
Notwithstanding the promising role of RIT
in eradicating MRD in patients with solid
tumors, or its potential to improve patient
survival or to provide disease palliation,
the majority of investigators now believe
that novel strategies to improve tumor up-
take and penetration of radiotherapeutic
agents, as well as diminish their toxicity to-
wards normal tissues (especially the bone
marrow) will be required to make a signifi-
cant impact on these malignancies. These
strategies, which include the use of pre-tar-
geting techniques, peptides instead of anti-
bodies as targeting vehicles, and short-
range radionuclides such as a-emitters or
Auger electron-emitters are discussed in
the next sections.
5.7
Pre-Targeting Strategies:
Improving the Therapeutic Index of RIT
Pre-targeted RIT is a strategy, whereby the
administration of anti-tumor mAbs is sep-
arated temporally from the administration
of a radionuclide, in order to improve the
delivery of radioactivity to tumors and pro-
mote its elimination from the blood, there-
by improving the therapeutic index [34].
The majority of studies of pre-targeted RIT
have employed the streptavidinbiotin sys-
tem, which takes advantage of the very
high affinity (K
a
=10
15
M
1
) and tetrava-
lency of streptavidin (SA), a 65-kDa pro-
tein produced by Streptomyces avidinii [60],
or avidin, a 66-kDa protein found in egg
Fig. 5.3 A longer median survival (19 months)
was achieved in 70 patients with glioblastomas
(GBMs) treated with local injection of 1295
2775 MBq (3575 mCi) of
131
I-BC-2 or BC-4 anti-
tenascin monoclonal antibodies (B) following sur-
gery and radiation compared to a previously re-
ported similar control group (A) treated by sur-
gery and radiation alone (12 months). [Reprinted
from Riva P., et al. Acta Oncologica 1999; 38: 351
359.]
whites [61], for biotin, a low molecular-
weight (M
r
244 Da), water-soluble B-vit-
amin. Pre-targeting strategies were initially
conceived for tumor imaging [62], but have
since been extended to targeted in situ
radiotherapy [63].
NeoRx Inc. (Seattle, WA, USA) has been
the main proponent of pre-targeted RIT of
cancer using their three-step PreTarget
TM
approach (Fig. 5.4) [64]. In the first step,
SA-immunoconjugates of anti-tumor mAbs
are administered in a non-radioactive form
to pre-target the tumor. In the second
step, a clearing agent (CA; biotin-galac-
tose-human serum albumin) is adminis-
tered 48 h later to significantly reduce the
level of circulating, non-tumor-bound SA-
immunoconjugates. The complexes
formed between the CA and SA-mAbs in
the blood are rapidly sequestered and in-
ternalized by hepatocytes, rendering them
unavailable for binding
90
Y-1,4,7,10-tetra-
azacyclododecane-1,4,7,10-tetraacetic acid
(DOTA)-biotin in the third step.
90
Y-DOTA-
biotin is administered 24 h after adminis-
tration of the CA. Since
90
Y-DOTA-biotin is
a small molecule (M
r
~900 Da), it pene-
trates readily into tumors, where it recog-
nizes and specifically binds to SA-immu-
noconjugates pre-targeted to the surface of
cancer cells. Therefore, only anti-tumor
mAbs that are not internalized by cancer
cells are suitable for pre-targeting
approaches. Non-tumor-bound
90
Y-DOTA-
biotin is rapidly eliminated by the kidneys
into the urine, significantly decreasing the
residence time of radioactivity in the blood
and the risk for non-specific bone marrow
toxicity compared to direct RIT.
Fig. 5.4 Pre-targeted radioimmunotherapy (Pre-
Target
TM
) is a strategy which temporally separates
the delivery of a monoclonal antibody and the
radionuclide in order to improve the therapeutic
index for treating tumors. In Step 1, a streptavidin
(SA)-immunoconjugate is administered to pre-tar-
get the tumor. In Step 2, a clearing agent (biotin-
galactose-human serum albumin) is administered
to clear circulating SA-immunoconjugates from
the blood. In Step 3, yttrium-90 conjugated to bio-
tin (
90
Y-DOTA-biotin) is administered.
90
Y-DOTA-
biotin binds tetravalently with high affinity (K
a
10
15
L mol
1
) to SA-immunoconjugates bound to
tumor cells, while excess
90
Y-DOTA-biotin is rapid-
ly eliminated by renal excretion. [Reprinted from
Weiden P.L. and Breitz H.B. Crit. Rev. Oncol. He-
matol. 2001; 40: 3751.]
5.7.1
Pre-Targeted RIT of Solid Tumors
Striking preclinical results were achieved
by the NeoRx group in treating solid tu-
mors using the PreTarget
TM
approach [65].
Athymic mice implanted with s.c. LS180
or SW1222 human colon carcinoma,
MDA-MB-468 human breast carcinoma or
SHT-1 small cell lung cancer xenografts
were administered 400 lg i.v. of NR-LU-
10-SA immunoconjugates directed against
the 40-kDa Ep-CAM epithelial antigen.
Twenty-four hours later, 200 lg of CA was
administered, which decreased the level of
circulating NR-LU-10-SA immunoconju-
gates by >90% within 2 h. Finally,
90
Y-
DOTA-biotin was administered at doses
ranging from 7.4 MBq (200 lCi) to
22.2 MBq (600 lCi). Strong anti-tumor ef-
fects were achieved with dose-dependent
complete regressions and cures (defined as
no tumor recurrence for >1 year) in 10/10
mice with lung or colon cancer xenografts,
and in 8/10 mice with breast cancer xeno-
grafts, at single doses of 22.229.6 MBq
(600800 lCi). The pre-targeting approach
decreased bone marrow toxicity, allowing
up to 22.2 MBq (800 lCi) to be safely ad-
ministered compared to an MTD of
7.4 MBq (200 lCi) in mice receiving direct
RIT with
90
Y-DOTA-NR-LU-10.
Due to the highly promising results
achieved preclinically with pre-targeted
RIT, a Phase I clinical trial was initiated to
evaluate the safety of the strategy in hu-
mans [66]. Forty patients with colon, ovar-
ian, prostate, breast or other solid tumors
were administered 350 mg of NR-LU-10-
SA immunoconjugates, followed 48 h later
by 360 mg of CA. Escalating doses
(37 MBq m
2
; 1 mCi m
2
) to 5180 MBq
m
2
(140 mCi m
2
) of
90
Y-DOTA-biotin
(0.5 mg) were given 24 h after the CA. The
most common adverse effects were ele-
vated liver function enzymes, reversible
thrombocytopenia and neutropenia, and
severe and dose-limiting gastrointestinal
toxicity (nausea, vomiting and diarrhea).
Bone marrow toxicity was not dose-limit-
ing. Two PRs and four MRs were achieved.
The dose selected for Phase II testing in
order to minimize gastrointestinal toxicity
was 4070 MBq m
2
(110 mCi m
2
). This
dose of
90
Y was more than 6-fold higher
than the usual maximum dose safely toler-
ated in patients receiving direct RIT with
90
Y-conjugated mAbs (total dose of
1110 MBq corresponding to 650
740 MBq m
2
) [6].
A Phase II trial of pre-targeted RIT
using NR-LU-10-SA and
90
Y-DOTA-biotin
was conducted in 25 patients with meta-
static colon cancer [67]. Patients were ad-
ministered a patient-specific dose of NR-
LU-10-SA sufficient to achieve an initial
plasma concentration of approximately
125 lg mL
1
. The CA was infused 48 h
later at 1.04 times the dose of NR-LU-10-
SA.
90
Y-DOTA-biotin (0.5 mg) was admi-
nistered 24 h after the CA at a dose of
4070 MBq m
2
(110 mCi m
2
). Disappoint-
ingly, there were no CRs and only two
PRs, for an overall objective response rate
of 8%. There were an additional four pa-
tients who demonstrated stabilization of
disease. The most serious toxicity was diar-
rhea, which was Grade 3 or 4 in about
30% of cases, and may have even contri-
buted to the death of one patient, who had
a cardiac arrest due to underlying heart
disease aggravated by diarrhea-induced de-
hydration and hypokalemia. Elevated liver
function enzymes were observed in 30%
of patients, and two patients developed ele-
vated serum creatinine. Hematologic toxic-
ity was less severe and reversible. All pa-
tients developed HAMA, human anti-
streptavidin antibodies (HASA) and hu-
man anti-immunoconjugate antibodies.
The gastrointestinal toxicity observed in
patients receiving pre-targeted RIT was
speculated to be due to cross-reactivity of
the NR-LU-10 mAb with normal bowel,
and not because of hepatobiliary excretion
of
90
Y-DOTA-biotin. Cross-reactivity of NR-
LU-10 with renal tubules was similarly
thought to explain the increase in serum
creatinine in two patients. The authors of
the study concluded that pre-targeted RIT
of malignancies in patients was feasible,
but that the NR-LU-10 mAb was not an
appropriate pre-targeting vehicle, due to its
unfavorable normal tissue cross-reactivity.
5.7.2
Other Studies of Pre-targeted RIT
of Solid Tumors
The NeoRx group have continued to ex-
plore pre-targeted RIT of solid tumors pre-
clinically using the B3 murine IgG
1j
mAb
directed against the Lewis
y
antigen. In one
study, athymic mice bearing s.c. A431 epi-
dermoid carcinoma xenografts were in-
fused with 400 lg of B3-SA followed by
100 lg of CA, then with 9.2 to 37 MBq
(250 lCi to 1 mCi) of
90
Y-DOTA-biotin [68].
The median survival was extended to 163
days at the highest dose compared to 8
days in untreated mice, and 7/10 treated
mice achieved a cure. Reversible hemato-
logic toxicity was observed, but no gastro-
intestinal toxicity. A subsequent RIT study
in mice with A431 tumors employing the
same B3-SA pre-targeting vehicle and CA,
but using DOTA-biotin conjugated to the
a-emitter,
213
Bi was recently reported [69].
A dose of 74 MBq (2 mCi) of
213
Bi-DOTA-
biotin was acutely lethal to all mice, with
morphological evidence pointing to renal
toxicity as the cause. However, doses of be-
tween 9.25 MBq (250 lCi) and 37 MBq
(1 mCi) were relatively safe and caused
complete tumor regressions in a high pro-
portion of mice. Bone marrow hypoplasia
and hepatic necrosis were noted in some
mice treated with 37 MBq of
213
Bi-DOTA-
biotin, but not at the lower doses.
Our group constructed SA-immunoconju-
gates of the second-generation, high-affinity
TAG-72 mAb CC49 for pre-targeted imaging
and RIT of colorectal cancer [70]. TAG-72 is
present on the surface of >94% of colorectal
cancers, as well as on ovarian cancer, breast
cancer and several other solid tumors [71].
In a pre-targeted RIT study, athymic mice
bearing s.c. LS174T tumors were injected
i.p. with 250 lg of CC49-SA, and then
40 h later received 40 lg (33.3 MBq;
900 lCi) of
90
Y-DOTA-biotin [72]. No com-
plete tumor regressions were achieved, but
there was a significant 1.5-fold decrease in
the growth rate of LS174T xenografts in
treated versus untreated mice. Pre-targeted
RIT was well-tolerated, with no changes in
body weight observed (a reflection of gener-
alized or gastrointestinal toxicity) and no de-
crease in leukocyte counts detected com-
pared to untreated animals.
Dr. Giovanni Paganelli and colleagues at
the European Institute of Oncology in Mi-
lan have investigated pre-targeted RIT in
patients with GBMs [73, 74]. Patients with
GBMs treated by surgery and radiotherapy
were administered 35 mg m
2
i.v. of bioti-
nylated anti-tenascin mAbs, followed 24
36 h later by 30 mg avidin and 50 mg SA
to clear circulating biotinylated antibodies
and create strept(avidin) binding sites for
90
Y-DOTA-biotin on tumor-bound biotiny-
lated mAbs. Finally,
90
Y-DOTA-biotin was
administered i.v. 1618 h later, at a dose of
2200 to 2960 MBq m
2
(59 to 80 mCi m
2
).
In 48 patients with documented residual/
recurrent GBMs, a tumor mass reduction
of 25100% was achieved in 12 cases, with
a duration of up to 12 months [74]. There
were four CRs, two PRs and two MRs,
and four patients achieved disease stabili-
zation. Dose-related and reversible throm-
bocytopenia was observed, but it was mild-
moderate in patients treated with up to
2590 MBq m
2
(70 mCi m
2
). In a subse-
quent trial in an adjuvant setting after
complete surgical resection and radiation
treatment of GBMs, the median survival
was 33.5 months for 37 patients receiving
pre-targeted RIT, compared to only 8
months in a non-randomized control
group [73].
To summarize, pre-targeting strategies
for RIT of solid tumors have clearly dimin-
ished the risk for bone marrow toxicity
and allowed dose escalation compared to
direct RIT. Dose escalation provided strong
anti-tumor responses in preclinical mouse
tumor xenograft models. Nevertheless,
there remain significant challenges to
overcome, particularly toxicity towards
non-hematopoietic normal tissues (e.g., liv-
er, kidney and in some cases gastrointesti-
nal tract) as well as the immune response
to SA, in order for this technique to be
more successful in eradicating solid tu-
mors in patients. Due to the success in
treating hematological malignancies with
direct RIT and the lower bone marrow tox-
icity associated with pre-targeted RIT, re-
search focus has recently shifted towards
exploring the potential for pre-targeted
RIT of lymphomas and leukemias.
5.7.3
Pre-Targeted RIT of Lymphomas
and Leukemias
The NeoRx group produced novel recombi-
nant targeting vehicles consisting of a sin-
gle chain variable fragment (scFv) of the
anti-CD20 murine mAb B9E9 or anti-
CD25 (IL-2R-a receptor) murine mAb anti-
Tac, fused to the full-length SA molecule
for pre-targeted RIT of B-cell or T-cell lym-
phomas, respectively [75, 76]. In one re-
port, athymic mice bearing s.c. SUDHL-1
large cell lymphoma xenografts expressing
CD25 were pre-targeted using anti-Tac
scFv-SA followed by 29.6 MBq (800 lCi) of
90
Y-DOTA-biotin [76]. Complete tumor re-
gressions were achieved in all mice
(n=10). Using the same anti-Tac scFv-SA
fusion protein and
213
Bi-DOTA-biotin in
the MET-1 acute T-cell leukemia model, a
2-fold prolonged survival was noted com-
pared to control mice (51 versus 24 days),
and there were significant decreases in the
serum levels of the human b2l tumor
marker. No information on toxicity was
provided. In a second report, pre-targeted
RIT was compared with direct RIT in athy-
mic mice implanted with Ramos human
B-cell lymphoma xenografts [77]. Mice
bearing s.c. Ramos tumors received pre-
targeted RIT with SA-conjugated anti-
CD20 murine IgG
2a
mAb 1F5 followed by
14.8 MBq (400 lCi) or 29.6 MBq (800 lCi)
of
90
Y-DOTA-biotin, or were treated directly
with 7.414.8 MBq (200400 lCi) of
90
Y-
DOTA-1F5. Complete tumor regressions
and cures (no tumor recurrence for >140
days) were achieved in eight of nine mice
receiving pre-targeted RIT, with only mini-
mal toxicity (slight decrease in body
weight). In contrast, high doses
(14.8 MBq) of
90
Y-DOTA-1F5 were required
to obtain major and durable tumor re-
sponses, and at these dose levels, all of the
mice died from severe bone marrow sup-
pression and infection at 10 days post-
treatment.
Pre-targeted RIT has been investigated
clinically in patients with NHL using SA-
conjugated anti-CD20 mAb rituximab (Ri-
tuxan
; Roche Pharmaceuticals) and

90
Y-
DOTA-biotin [78]. Seven patients were pre-
targeted with rituximab-SA, then treated
with 1110 MBq m
2
(30 mCi m
2
) or
1850 MBq (50 mCi m
2
) of
90
Y-DOTA-bio-
tin. Six patients exhibited an objective re-
sponse, with three CRs and one PR of 6 to
12 months duration. Dosimetry estimates
(obtained by imaging using
111
In-DOTA-
biotin) projected that the dose to tumor de-
posits from pre-targeted RIT would be 25
30 cGy/37 MBq (rads mCi
1
) compared to
0.200.24 cGy/37 MBq to the bone mar-
row. Only grade 1 or 2 reversible hemato-
logic toxicity was observed, but a high pro-
portion of patients developed HASA.
These studies suggested that pre-targeted
RIT may have a promising future role in
the treatment of NHL, especially in reduc-
ing bone marrow toxicity compared to di-
rect RIT using CD20 radioimmunoconju-
gates (e.g.,
131
I-tositumomab or
90
Y-ibritu-
momab tiuxetan). The reduced risk for
bone marrow toxicity may allow dose esca-
lation to achieve response rates in NHL
similar to those with myeloablative RIT
but not requiring ASC rescue.
5.8
Peptide-Directed In situ Radiotherapy:
Targeting Somatostatin Receptors
Peptides offer an attractive alternative to
mAbs as targeting vehicles for in situ
radiotherapy of cancer, since their much
lower molecular weight (M
r
16 kDa) re-
sults in rapid elimination from the blood
by renal excretion, and facilitates their
penetration into solid tumors [24]. The de-
creased residence time of radioactivity in
the blood diminishes bone marrow toxicity
and permits dose escalation to potentially
therapeutic levels. Peptide-directed radio-
therapy (PDRT) of cancer has focused
mainly on treatment of malignancies ex-
pressing somatostatin receptors (SSR) with
90
Y-DOTA-D-Phe
1
-Tyr
3
-octreotide (
90
Y-DO-
TATOC), a synthetic octapeptide analogue
of somatostatin [79, 80]. However, many
endogenous growth factors are peptides,
and their receptors, when overexpressed
by cancer cells, also present attractive tar-
gets for radiotherapy [81]. SSR are ex-
pressed mainly by neuroendocrine tumors,
but have similarly been found on breast
cancer, NHL, melanoma and brain malig-
nancies [82]. There are five subtypes of
SSR, but
90
Y-DOTATOC has the highest af-
finity for SSR subtype 2.
In the Phase I trial of
90
Y-DOTATOC, 30
patients with various SSR-positive tumors
were administered 1.11 GBq (30 lg) to
2.59 GBq (70 lg) per cycle for a total of
three cycles over 6 months [83]. There
were no major (Grade 3 or 4) hematologic
toxicities, although most patients exhibited
reversible Grades 01 myelosuppression
up to total doses of 5.55 GBq of
90
Y-DOTA-
TOC, and 40% of patients exhibited Grade
2 toxicity at doses of 6.77.8 GBq. In par-
ticular, lymphocytopenia occurred, possibly
in part due to expression of SSR by lym-
phocytes. The most common acute adverse
effects were nausea and vomiting, but no-
tably, delayed renal toxicity at total doses
of 3.37.8 GBq. The relatively high renal
uptake of
90
Y-DOTATOC resulted in an es-
timated radiation absorbed dose of
330 cGy/GBq to the kidneys. The maxi-
mum radiation exposure tolerated by the
kidneys based on external radiation is
23002500 cGy [84], suggesting that the
maximum safe total dose for
90
Y-DOTA-
TOC may be 6.97.6 GBq. Objective re-
sponses were obtained in 23% of patients,
with two CRs, four PRs and one MR.
Kidney uptake of radiolabeled peptides
is thought to be due to charge interactions
between cationic amino acids in the pep-
tides and the negatively charged cell mem-
brane of renal tubular cells, and can be in-
hibited by co-administration of D- or L-ly-
sine [85]. Therefore, subsequent Phase II
trials of
90
Y-DOTATOC incorporated infu-
sion of 8% amino acid parenteral solu-
tions into the protocol in an attempt to
minimize renal accumulation. In one
study [86], 41 patients with neuroendo-
crine malignancies were administered a to-
tal of four doses of 6 GBq m
2
(162 mCi m
2
) of
90
Y-DOTATOC at inter-
vals of 6 weeks, with amino acid protec-
tion. The overall response rate was 24%,
with one CR, nine PRs and 5 MRs (Fig.
5.5). Stabilization of disease was achieved
in an additional 20 patients. An identical
response rate (23%) was observed in a sec-
ond Phase II trial of 39 patients with neu-
roendocrine tumors treated with four infu-
sions of 7.4 GBq
90
Y-DOTATOC [87]. Sig-
nificant clinical benefit was obtained in
two-thirds of the patients by decreasing
cancer-related diarrhea, flushing, wheezing
or pellagra. The most common side effects
were nausea and vomiting, and mild-mod-
erate bone marrow toxicity, including re-
versible lymphocytopenia. A persistently
elevated serum creatinine was observed in
one patient [87].
5.8.11
Renal Toxicity from PDRT of Malignancies
A single case report of end-stage renal dis-
ease after a total dose of 5.6 GBq m
2
(151 mCi m
2
) of
90
Y-DOTATOC [88]
sparked intense controversy and discus-
sion among investigators in the field about
the safety of PDRT of malignancies [89
92]. However, renal protection using pa-
renteral amino acid solutions was used for
only one of four single doses of
90
Y-DOTA-
TOC administered to this patient. Never-
theless, in a related study by the same
group, five of 29 patients administered
>7.4 GBq m
2
(200 mCi m
2
) of
90
Y-DOTA-
TOC without renal protection similarly de-
veloped chronic renal failure, which was
associated with thrombotic microangiopa-
thies identical to those produced by exter-
Fig. 5.5 Liver metastases in a patient with pan-
creatic insulinoma demonstrate a major decrease
in size at 8 months following treatment with two
cycles of 3.33 GBq (89 mCi) of
90
Y-DOTATOC at
an interval of 2 months. Renal protection with pa-
renteral amino acids was provided. Pre-treatment
CT overlayed with a nuclear image obtained with
111
In-DTPA-D-Phe
1
-octreotide is shown in (A). The
pre-treatment CT without the nuclear image over-
lay is shown in (B) and the post-treatment CT in
(C). [Reprinted from Bodei L. et al., Eur. J. Nucl.
Med. 2003; 30: 207216.]
nal radiation [93]. Patients administered
<7.4 GBq m
2
of
90
Y-DOTATOC did not ex-
perience renal toxicity.
The fervent discussion surrounding the
issues of renal toxicity from
90
Y-DOTATOC
raised several key points. First, the thera-
peutic benefits of PDRT often outweigh
the potential risks for serious adverse ef-
fects, including renal toxicity. Second, not-
withstanding the potential therapeutic ben-
efits, the risk for renal damage from
90
Y-
DOTATOC must be minimized through
administration of parenteral amino acid
solutions. Finally, the safe dose of
90
Y-DO-
TATOC when combined with amino acid
administration had not been adequately
determined in Phase I trials, most of
which did not employ renal protection.
Subsequently, two Phase I trials of
90
Y-DO-
TATOC in patients in whom the kidneys
were protected by administration of paren-
teral amino acid solutions were conducted
to address this last point [94, 95]. In the
first study, 40 patients with SSR-positive
tumors were treated with two cycles of
2.96 GBq (80 lg; 80 mCi) to 5.18 GBq
(150 lg; 140 mCi) of
90
Y-DOTATOC [94].
Parenteral solutions of lysine with/without
arginine were administered before and
after treatment with
90
Y-DOTATOC. He-
matologic and gastrointestinal toxicity oc-
curred, but no patients developed chronic
renal toxicity at single doses up to
5.18 GBq (total dose 10.36 GBq; 280 mCi).
Two patients exhibited mild increases in
serum creatinine and blood urea nitrogen
(BUN), which returned to normal. In the
second study, PET was used to evaluate
the effects of amino acid administration
on renal uptake of
86
Y-DOTATOC and pro-
ject the kidney dosimetry for
90
Y-DOTA-
TOC [95]. Solutions of lysine or arginine
alone, in combination (mixed amino
acids), or as a dipeptide (Lys-Arg) were
studied in 24 patients with SSR-positive
tumors administered
86
Y-DOTATOC. Ami-
no acids decreased the renal uptake of
86
Y-
DOTATOC by 21%; decreased the radia-
tion absorbed dose to the kidneys by 27%;
and increased the projected MTD for
90
Y-
DOTATOC (based on a maximum radia-
tion absorbed dose of 2300 cGy to the kid-
neys) from 2.9 GBq m
2
(79 mCi m
2
) to
4.2 GBq m
2
(113 mCi m
2
).
5.8.2
Other Studies of PDRT
of Cancer using Somatostatin Analogues
PDRT has also been studied for local treat-
ment of low-grade gliomas by the adminis-
tration of 0.55 GBq (15 mCi) to 7.0 GBq
(190 mCi) of
90
Y-DOTATOC into the surgi-
cal cavity or directly into residual tumor
[96, 97]. Disease stabilization for 1045
months was achieved in about 50% of the
patients. The toxicities observed were brain
edema and seizures. No hematologic, gas-
trointestinal or renal toxicity were ob-
served, likely due to the finding that only
50% of the locally deposited radiotherapeu-
tic agent was absorbed into the systemic
circulation [96]. Other somatostatin ana-
logues conjugated to b-emitters are under
investigation for treatment of SSR-express-
ing malignancies, including
90
Y-DOTA-lan-
reotide (
90
Y-DOTALAN) and
90
Y-DOTA-
Tyr3-octreotate (
90
Y-DOTATATE), as well as
177
Lu-DOTATATE [79].
90
Y-DOTALAN is
more lipophilic than
90
Y-DOTATOC, and
has a higher affinity for SSR subtype 5.
90
Y-DOTATATE and
177
Lu-DOTATATE have
the carboxy-terminal threoninol replaced
by threonine, and exhibit a 15-fold higher
affinity for SSR subtype 2 than
90
Y-DOTA-
TOC.
5.9
Auger Electron Radiotherapy:
Anti-tumor Effects at the Single Cell Level
One of the causes of normal tissue dam-
age, including bone marrow toxicity, from
targeted in situ radiotherapy of malignan-
cies is the cross-fire effect, in which non-
targeted normal cells are killed by long-
range (2- to 10-mm) b-particles emitted by
radiotherapeutic agents targeted to nearby
tumor cells, or circulating in the blood.
The cross-fire effect may be minimized by
decreasing the residence time of radioac-
tivity in the blood, either through pre-tar-
geting strategies or by the use of rapidly
clearing peptides. However, the effect can
be virtually eliminated by selecting radio-
nuclides such as
111
In or
125
I, that emit
low-energy Auger electrons with nano-
meter to micrometer ranges in tissues;
thus, these are theoretically only capable
of killing single targeted cancer cells,
while sparing non-targeted cells [98]. How-
ever, it has recently been shown that Au-
ger electron-emitting radiotherapeutic
agents also have a bystander effect,
whereby tumor cells experiencing DNA
damage cause the death of neighboring vi-
able non-targeted (presumably tumor)
cells, through the release of promoters of
apoptosis [99]. The extremely short range
of Auger electrons requires internalization
of the radiotherapeutic agents into the cy-
toplasm and translocation to the cell nu-
cleus in order to have the greatest cytotoxic
effect [100]. The field of Auger electron
radiotherapy of malignancies was founded
by Drs James Adelstein and Amin Kassis
at Harvard University, who showed that
the thymidine analogue,
125
I-iododeoxyuri-
dine (
125
I-IUdR) was incorporated into
DNA and was highly toxic to cells through
the emission of DNA-damaging Auger
electrons from
125
I [101].
125
I-IUdR was
not specific for malignancies however, and
recent research has focused on the identifi-
cation of biomolecules that are able selec-
tively to insert Auger electron-emitters into
the cytoplasm and nucleus of tumor cells.
5.9.1
Auger Electron Radiotherapy
with
111
In-DTPA-D-Phe
1
-octreotide
111
In-DTPA-D-Phe
1
-octreotide (Octreo-
scan
; Mallinckrodt Medical Inc.) is an oc-

tapeptide analogue of somatostatin conju-
gated to diethylenetriamine-pentaacetic
acid (DTPA) and labeled with
111
In.
111
In-
DTPA-D-Phe
1
-octreotide is routinely em-
ployed for imaging SSR-positive tumors by
taking advantage of its c-emissions (Ec 171
and 245 keV) [102], but the radiopharma-
ceutical is internalized and translocated to
the cell nucleus following binding to
membrane receptors [103105]; this sug-
gests that it could also function as a radio-
therapeutic agent for tumors expressing
SSR by exploiting the Auger electron emis-
sions from
111
In. Indeed, in clonogenic as-
says,
111
In-DTPA-D-Phe
1
-octreotide was
highly toxic to SSR-positive CA20948 rat
pancreatic tumor cells [106]. The adminis-
tration of two doses of 370 MBq (10 mCi;
0.5 lg) i.v. to rats significantly decreased,
and in some cases completely prevented,
the establishment of hepatic metastases
after inoculation of CA20948 cells into the
vena porta [107].
Based on these promising preclinical re-
sults, several clinical trials have been con-
ducted to investigate the potential of
111
In-
DTPA-D-Phe
1
-octreotide for targeted Auger
electron radiotherapy of SSR-expressing
malignancies [108110]. In one trial [110],
50 patients with SSR-positive tumors were
treated with multiple doses (67 GBq;
160190 mCi; 50100 lg] of
111
In-DTPA-D-
Phe
1
-octreotide administered every 2
weeks up to a total amount of 20 GBq
(540 mCi) to 160 GBq (4324 mCi). Antitu-
mor effects were achieved in 21 of 40 eval-
uable patients with one PR, six MRs and
stabilization of disease in 14 patients. Mild
and reversible bone marrow toxicity (i.e.,
thrombocytopenia and leukopenia) was the
most common adverse effect, as well as
profound lymphocytopenia which reached
grade 3 or 4. Myelodysplastic syndrome
and leukemia occurred in three of six pa-
tients receiving total doses >100 GBq
(2700 mCi) at 1.4 to 3 years after treat-
ment. SSR are present on testicular cells,
and
111
In-DTPA-D-Phe
1
-octreotide treat-
ment caused a major decrease in the levels
of inhibin B, a hormone associated with
spermatogenesis. Surprisingly, unlike the
prior experience with
90
Y-DOTATOC, no
patients developed renal toxicity despite
the administration of very high doses (in
some cases without renal protection) that
deposited >4500 cGy (rads) in the kidneys.
These dose estimates were well beyond the
23002500 cGy thought to cause irreversi-
ble radiation damage to the kidneys [84].
The reason for a lack of renal toxicity from
high doses of
111
In-DTPA-D-Phe
1
-octreo-
tide is not known. However, the Medical
Internal Radiation Dose (MIRD) schema
used for calculating radiation absorbed
doses to tissues is generally not believed to
be accurate for short-range Auger electron-
emitting radionuclides, especially in situa-
tions of heterogeneous tissue distribution.
In a second trial [108], 27 patients with
SSR-positive gastroenteropancreatic malig-
nancies who had failed all other therapies
were administered two single doses of
6.7 GBq (180 mCi) of
111
In-DTPA-D-Phe
1
-
octreotide. Clinical benefit was achieved in
62% of patients, and tumor-associated pan-
creastatin levels decreased significantly in
81%. Objective PRs were obtained in two
of 26 (8%) evaluable patients, and com-
puted tomography (CT) scans showed im-
provement in seven additional patients.
Decreased platelets, leukocytes and hemo-
globin were observed. The median survival
of patients treated with
111
In-DTPA-D-
Phe
1
-octreotide was 18 months (range 3 to
54 months) compared to an expected sur-
vival of 36 months in a historical control
group. Finally, in a third smaller trial, five
patients with carcinoid or pancreatic tu-
mors were treated with doses of 6 GBq
(160 mCi; 40 lg) of
111
In-DTPA-D-Phe
1
-oc-
treotide administered every 3 weeks for
three cycles [109]. Decreased tumor-asso-
ciated hormone levels were obtained, but
no changes in tumor size were detected.
Again, platelet and leukocyte counts and
hemoglobin concentrations decreased tran-
siently. There was no evidence of liver or
renal toxicity.
5.9.2
Auger Electron Radiotherapy
with
111
In-DTPA-hEGF
Our group has been exploring a type of Tro-
jan Horse targeted radiotherapeutic strat-
egy for breast cancer that exploits the inter-
nalization and nuclear translocation path-
way of human epidermal growth factor
(hEGF) following binding to its cell surface
receptor (EGFR) [111]. EGFR are overex-
pressed up to 100-fold compared to most
normal epithelial tissues in almost all estro-
gen receptor (ER)-negative, hormone-resis-
tant and poor-prognosis breast cancers
[112]. Nuclear translocation of EGF and/or
EGFR as well as direct binding to chromatin
have been reported for malignant [113] and
some dividing normal cells [114, 115], but
the precise function is not well understood,
since EGF has traditionally been thought to
regulate gene expression through EGFR-
mediated activation of intracellular signal-
ing cascades. It was recently shown that
the EGF/EGFR complex may have a novel
role as a nuclear transcription factor for
the cyclin D1 gene, particularly in rapidly di-
viding cells (e.g., cancer cells) [116]. In nor-
mal cells, EGF acts mainly by activating sig-
naling cascades, and is internalized into cy-
toplasmic vesicles and degraded. In rapidly
dividing cells, it appears that a proportion
of internalized EGF molecules circumvent
the normal lysosomal degradation pathway
and translocate to the cell nucleus, possibly
mediated by a nuclear-localizing sequence
(RRRHIVRKRTLRR) found in the EGFR
at residues 645657 [117]. The possibility
of differential nuclear versus cytoplasmic
distribution of EGF/EGFR in malignant ver-
sus normal cells has profound implications
for Auger electron-emitting radiotherapeu-
tic agents targeted at EGFR expression,
since the radiation absorbed dose deposited
in the nucleus is about 15 times higher
when
111
In or
125
I decays in the nucleus
compared to decay in the cytoplasm, and
30 times higher than when the decay occurs
on the cell surface [118]. The differential nu-
clear versus cytoplasmic uptake could pro-
vide a second level of selectivity in protect-
ing EGFR-positive normal cells, in addition
to their lower level of EGFR expression com-
pared to malignant cells.
We found that hEGF conjugated to
DTPA and labeled with
111
In (
111
In-DTPA-
hEGF) was rapidly and efficiently internal-
ized by EGFR-overexpressing MDA-MB-
468 human breast cancer cells (12 10
6
receptors per cell) [119]. About 715% of
internalized
111
In-DTPA-hEGF molecules
were translocated to the nucleus (Fig. 5.6),
and 10% of internalized
111
In-DTPA-hEGF
was bound by chromatin. The emission of
Auger electrons by
111
In-DTPA-hEGF in
close proximity to DNA reduced the sur-
viving fraction of MDA-MB-468 cells in
clonogenic assays to <5% at only 111
148 mBq per cell (34 pCi per cell) [119].
In contrast, MCF-7 cells which displayed a
100-fold lower level of EGFR on their sur-
face were unaffected by exposure to
Fig. 5.6 Fluorescence microscopy illustrating bind-
ing of fluorescein-conjugated human epidermal
growth factor (hEGF) to the cell membrane of a
single MDA-MB-468 human breast cancer cell, fol-
lowed by internalization into cytoplasmic vesicles,
and finally importation of a proportion of interna-
lized hEGF molecules into the nucleus. Nuclear
uptake is thought to be mediated by a nuclear lo-
calizing sequence (shown) found at residues 645
657 in the cytoplasmic domain of the EGFR.
133 mBq per cell (3.6 pCi per cell) of
111
In-DTPA-hEGF (Fig. 5.7). In a subse-
quent study [120], we found that
111
In-
DTPA-hEGF was 85- to 300-fold more toxic
on a molar concentration basis to MDA-
MB-468 cells (IC
50
<70 pM; 11 kBq mL
1
)
than selected chemotherapeutic agents
commonly used for breast cancer such as
paclitaxel, methotrexate or doxorubicin
(IC
50
6 nM, 15 nM and 20 nM, respec-
tively) and several logarithms more
growth-inhibitory than 5-FU (IC
50
4 lM).
Picomolar concentrations of
111
In-DTPA-
hEGF provided cytotoxic effects equivalent
to those of 4 Gy of external c-radiation.
In order to amplify the toxicity of
111
In-
DTPA-hEGF further, we conjugated hEGF
to human serum albumin (HSA) which al-
lowed substitution of multiple DTPA che-
lators for
111
In onto the HSA moiety [121].
HSA conjugation diminished the EGFR-
binding affinity about 15-fold compared to
111
In-DTPA-hEGF (K
a
5.110
7
versus
7.510
8
L mol
1
, respectively) but there
were no further decreases in affinity as the
number of DTPA groups was increased
from 12 to as many as 23 per molecule.
111
In-DTPA-HSA-hEGF was internalized
and translocated to the nucleus of MDA-
MB-468 cells similarly to
111
In-DTPA-
hEGF.
111
In-DTPA-HSA-hEGF substituted
with nine DTPA groups had a specific ac-
tivity which was 10-fold higher than
111
In-
DTPA-hEGF, and was 4-fold more cyto-
toxic to MDA-MB-468 cells (IC
50
15 versus
60 pM, respectively). One of the main ad-
vantages of
111
In-DTPA-hEGF compared to
hEGF labeled with the Auger electron-
emitter,
125
I, is that
111
In is retained in tu-
mor cells following receptor-mediated in-
ternalization and lysosomal degradation of
hEGF, whereas
125
I catabolites are ex-
ported from the cells, thus maximizing
the radiation dose from
111
In-DTPA-hEGF
to the cell nucleus [122].
In mice implanted with s.c. MDA-MB-
468 breast cancer xenografts, treatment
with five weekly doses of
111
In-DTPA-
Fig. 5.7 Clonogenic survival of MDA-MB-468 hu-
man breast cancer cells overexpressing EGFR
(1210
6
receptors/cell) or MCF-7 breast cancer
cells with a 100-fold lower level of EGFR expres-
sion (110
4
receptors/cell) exposed to increasing
amounts of
111
In-DTPA-hEGF. l, survival of MDA-
MB-468 cells based on the total amount of radio-
activity bound to the cells; ^, survival of
MDA-MB-468 cells based on the total amount of
radioactivity internalized into the cytoplasm of the
cells; n, survival of MDA-MB-468 cells based on
the total amount of radioactivity imported into the
cell nucleus; `, survival of MCF-7 cells based on
the total amount of radioactivity targeted to the
cells. [Reprinted from Reilly R.M., et al. J. Nucl.
Med. 2000; 41: 429438.]
hEGF (total 92.5 MBq; 2.5 mCi; 17 lg)
slowed the growth of established tumors
3-fold and caused regression of small,
non-established tumors (Fig. 5.8) [123].
There was a modest 1.4- to 2-fold decrease
in leukocyte and platelet counts, but there
was no toxicity towards the liver or kidneys
tissues which have moderately high
EGFR expression (10
5
EGFR per cell).
Based on these promising preclinical re-
sults, we have commenced a Phase I clini-
cal trial of
111
In-DTPA-hEGF in patients
with EGFR-positive metastatic breast can-
cer refractory to chemotherapy at Princess
Margaret Hospital in Toronto [111]. A kit
formulation for preparation of pharmaceu-
tical quality
111
In-DTPA-hEGF suitable for
human administration was recently re-
ported by our group [124]. The clinical trial
design involves administration of escalat-
ing single doses of
111
In-DTPA-hEGF
ranging from 370 MBq (10 mCi; 250 lg) to
2960 MBq (80 mCi; 500 lg) to cohorts of
three to six patients. Adverse effects and
toxicity will be monitored by clinical as-
sessment and hematological and clinical
biochemical testing. Although not the fo-
cus of the Phase I trial, anti-tumor effects
will be assessed by radiological imaging
(CT, X-ray and MRI). Tissue distribution
and macrodosimetry (MIRD schema) of
111
In-DTPA-hEGF will be evaluated by
Fig. 5.8 Tumor growth index (post-treatment vol-
ume divided by initial volume) for athymic mice
implanted with 1415 mm
3
established s.c. MDA-
MB-468 human breast cancer xenografts (A) or
10 mm
3
non-established s.c. MDA-MB-468 breast
cancer xenografts (B) at selected times (days) fol-
lowing treatment with five weekly s.c. doses of
111
In-DTPA-hEGF [cumulative dose 92.5 MBq
(2.5 mCi); 17 lg] or normal saline vehicle. Treat-
ments were started on day 0. The site of
111
In-
DTPA-hEGF injection was remote from that of tu-
mor implantation. [Reprinted from Chen P., et al.
J. Nucl. Med. 2003; 44: 14691478.]
quantitative imaging.
111
In-DTPA-hEGF is
intended for treatment of EGFR-positive
breast cancer which is almost exclusively
ER-negative, but Auger electron-emitting
agents such as 16a-[
125
I]-iodo-estradiol
[125127] or
125
I-iodotamoxifen [128] have
been proposed as treatments for ER-posi-
tive breast cancer.
111
In-DTPA-hEGF [129]
and related analogues [130, 131] are also
being investigated for treatment of GBMs
that overexpress EGFR either by local or
systemic administration.
5.9.3
Targeted Radiotherapy with Antibodies
Conjugated to Auger Electron Emitters
Auger electron-emitters are also under
study as therapeutic radionuclides conju-
gated to mAbs that specifically bind to tu-
mor-associated antigens and internalize
into cancer cells. In fact, it has been shown
that CO17-1A mAbs conjugated to Auger
electron-emitters such as
125
I or
111
In were
more effective than the same mAbs conju-
gated to the b-emitters,
131
I or
90
Y at inhib-
iting the growth of GW-39 human colon
cancer xenografts in athymic mice when ad-
ministered at equitoxic doses [132, 133].
Furthermore, CO17-1A conjugated to Auger
electron-emitters or to b-emitters was dose-
limited by bone marrow toxicity, but the
MTD was 10-fold higher for
125
I- than for
131
I-CO17-1A (11.1 MBq; 300 lCi versus
111 MBq; 3 mCi, respectively) [133] and 22
times higher for
111
In- than for
90
Y-CO17-
1A (4 MBq; 108 mCi versus 85 MBq;
2.3 mCi, respectively) [132]. Since CO17-
1A does not cross-react with bone marrow
stem cells, it was speculated that the c-emis-
sions or longer range conversion electrons
emitted by
125
I or
111
In may be responsible
for myelosuppression.
The CD74 (anti-MHC II invariant chain)
mAb LL1 conjugated to various Auger elec-
tron-emitters has been studied for killing
human B-cell lymphoma cells in vitro
[134137], and for inhibiting the growth of
B-cell lymphomas in vivo in mice [138].
Rapid recycling of the CD74 epitope al-
lowed internalization of as many as 10
7
LL1 antibody molecules per day into lym-
phoma cells. Complete killing of Raji B-
cell lymphoma cells was achieved by expo-
sure to increasing concentrations of mAb
LL1 conjugated to the Auger electron-emit-
ters,
111
In,
125
I or
99m
Tc [135]. A compari-
son of the toxicity against Raji cells of
mAb LL1 conjugated to
111
In or
125
I or to
the b-emitters,
90
Y or
131
I showed that
non-specific cytotoxicity was much higher
for the b-emitters, as expected due to the
substantial cross-fire effect [134]. In terms
of cell killing potency,
131
I and
67
Ga were
the most effective;
90
Y was intermediate in
potency; and
111
In and
125
I were the least
potent, although still able to achieve 100%
killing of Raji cells. An identical order of
cytotoxic potency was observed for Raji
cells using the anti-CD20 mAb 1F5 which
does not internalize into lymphoma cells
conjugated to
131
I,
67
Ga,
125
I or
111
In, sug-
gesting that internalization may not always
be required for cell killing with Auger elec-
tron-emitters, possibly due to the longer
range effects of concurrent conversion
electron emissions. Treatment of athymic
mice inoculated i.v. with Raji cells 35
days later with
111
In- or
67
Ga-conjugated
LL1 (8.9 MBq; 240 lCi to 12.9 MBq;
350 lCi) significantly prolonged the surviv-
al of the mice compared to controls [138].
Furthermore, the doses of
111
In- or
67
Ga-
LL1 studied did not cause significant toxic-
ity, whereas the MTD for mice injected
with
90
Y-LL1 was <1 MBq (25 lCi).
The only reported human study to ex-
amine targeted Auger electron radiother-
apy using mAbs was a Phase I/II trial of
125
I-labeled anti-EGFR mAb 425 adminis-
tered i.v. or intra-arterially to 180 patients
with high-grade gliomas [139]. Patients re-
ceived multiple single doses of 1850 MBq
(50 mCi) of
125
I-mAb 425 up to a total cu-
mulative amount of 5180 MBq (140 mCi)
in an adjuvant setting following complete
surgical excision and external radiotherapy.
The survival ranged from 4 to 150 months
for patients with GBM, and from 4 to 270
months for patients with astrocytomas.
Anti-tumor effects have also been observed
in athymic mice implanted s.c. with
U87MG human glioblastoma xenografts
treated with two i.v. doses of 18.5 MBq
(500 lCi) of
111
In-anti-human integrin a3
mAb GA17 [140].
5.9.4
Gene-targeted Auger Electron Radiotherapy
The highly localized (in some cases nano-
meter) deposition of energy associated
with Auger electrons offers the opportu-
nity to surgically cleave selected sequences
in DNA using triplex-forming oligonucleo-
tides (TFOs) conjugated to
111
In or
125
I
(antigene radiotherapy) [141, 142]. TFOs
hybridize in the major groove of the DNA
duplex by Hoogsteen non-canonical base
pairing. Dr Ronald Neumann and collea-
gues at the US National Institutes of
Health [141] showed that 14- to 31-base
pair (bp) TFOs labeled with
111
In through
a DTPA group introduced into a 5' or 3'
amine-containing flexible linker, caused
double-strand breaks in the complemen-
tary DNA sequences within 10 bases of
the
111
In decay site (Fig. 5.9). The cleavage
of DNA caused by the Auger electrons
from
111
In were in fact so precise that
skewing of breaks towards the 5' end was
observed which corresponded to the dis-
placement of the linker carrying the
111
In
along the DNA duplex backbone.
123
I-,
111
In-, or
125
I-labeled TFOs yielded 0.03,
0.38 and 0.66 DNA breaks per decay, re-
spectively [143]. These TFOs were not di-
rected against a gene sequence relevant in
cancer, but more recent investigations by
the same group has focused on antigene
radiotherapy of the MDR1 gene responsi-
ble for multidrug resistance [142, 144].
DNA breaks in the MDR1 sequence were
achieved in plasmid and purified genomic
DNA isolated from KB-V1 vinblastine-re-
sistant cancer cells with MDR1 gene am-
plification, using
125
I-TFOs. The yield of
DNA breaks in isolated plasmid or geno-
mic DNA hybridized with
125
I-TFOs was
0.50.7 breaks per decay. However, in iso-
lated nuclei or digitonin-permeabilized
cells incubated with
125
I-TFOs, the yield of
DNA breaks was decreased 10-fold (0.03
breaks per decay), suggesting that the nu-
cleosomal structure of DNA in vivo may
provide substantial protection. Neverthe-
less, an important proof-of-principle was
demonstrated in that
125
I-TFOs were able
to find their gene target within cancer cells
and cause DNA cleavage at a specific site.
Although there are challenges to overcome
in the delivery of DNA into cells [145],
these studies raise the intriguing idea that
it may be possible in the future selectively
to target and destroy specific oncogene se-
quences responsible for the malignant
phenotype through the use of TFOs emit-
ting ultrashort range Auger electrons.
5.10
a-Particle RIT: Anti-tumor Effects
at the Multi-cell Level
While Auger electron-emitters target and
kill single cancer cells, radionuclides emit-
ting a-particles can kill small clusters of ma-
lignant cells due to their 50- to 100-lm
range (510 cell diameters) [45]. RIT of leu-
kemia using
213
Bi- and
225
Ac-HuM195 has
been described, but other applications of
mAbs labeled with a-emitters are similarly
under study.
213
Bi-labeled anti-CD74 mAb
LL1 has been investigated for treatment of
disseminated Raji B-cell lymphoma xeno-
grafts in SCID mice [146]. Protection
against paralysis due to infiltration of the
spinal canal with lymphoma cells was
achieved at doses of 4.65.2 MBq (125
141 lCi). Only a slight decrease in body
weight of the mice due to normal tissue tox-
icity was noted at the highest doses. The
anti-CD20 mAb, rituximab (Rituxan
) con-
jugated to the a-emitters,
211
At [147] or
149
Tb
[148] has also been studied for RIT of B-cell
lymphoma.
211
At-rituximab selectively
killed two B-lymphoma cell lines (RAEL
and K422) while partially sparing bone mar-
row stem cells from healthy human volun-
teers.
149
Tb-rituximab administered in
doses of 5.5 MBq (150 lCi) to SCID mice in-
oculated with Daudi B-cell lymphoma cells
yielded tumor-free survival for 120 days,
whereas all mice in a control, untreated
group developed B-cell lymphoma and were
sacrificed at 37 days for humane reasons.
225
Ac conjugated to HER-2/neu mAb trastu-
zumab (Herceptin
) (Part 1, Chapter 5) pro-

longed the survival of mice implanted i.p.
with SK-OV-3 ovarian carcinoma xenografts
at very small doses of 8.116.6 mBq (220
450 nCi), but deaths from toxicity occurred
at the highest dose levels [149].
5.11
Conclusion
The application of biomolecules as target-
ing vehicles for in situ radiotherapy of ma-
lignancies, although now conceptually
more than 20 years old, is still rapidly
Fig. 5.9 Strand breaks in a
32
P-labeled oligonu-
cleotide (ODN) hybridized with a complementary
125
I- or
111
In-conjugated ODN detected by poly-
acrylamide gel electrophoresis (A; lane 4) and
quantified by gel densitometry (B). The open bars
indicate breaks produced by
125
I-ODNs, and the
dark bars indicate breaks produced by
111
In-
ODNs. The nanometer range of Auger electrons
emitted by
125
I or
111
In results in cleavage of DNA
within 10 bases of the decay site (indicated by
Y). [Reprinted with permission from Karamychev
V. N., et al., J. Nucl. Med. 2000; 41: 10931101.]
evolving. There have been recent outstand-
ing successes in treating hematological
malignancies such as NHL and leukemia
with RIT, and encouraging results have
been achieved in eradicating MRD in solid
tumors, prolonging patient survival. The
use of peptides as targeting vehicles and
innovative strategies such as pre-targeted
RIT offer the promise of improved results
in treating larger volume solid tumors in
the future, and perhaps even greater effi-
cacy in treating NHL. The incorporation of
short-range a-emitters or Auger electron-
emitters into the design of radiotherapeu-
tic agents presents the opportunity to se-
lectively kill single cancer cells or clusters
of cells, in order to eradicate micrometas-
tases and prevent tumor recurrence, while
sparing normal tissues. The demonstration
of surgical cleavage of specific gene se-
quences in the DNA of cancer cells using
TFOs labeled with Auger electron-emitters
exemplifies the great excitement that this
field holds for the future.
Acknowledgments
The author acknowledges the support of
the US Department of Defense Breast
Cancer Research Program, the Canadian
Institutes of Health Research, the Canada
Foundation for Innovation, the National
Cancer Institute of Canada and the Susan
G. Komen Breast Cancer Foundation in
providing funds for research described in
this chapter originating in his laboratory at
the University of Toronto.
References
1 Ferlay J, Bray F, Pisani P, Parkin DM. Cancer
incidence, prevalence and mortality world-
wide. Lyon, France: IARC Press, 2000.
2 Miller TP, Vhase EM, Dalton WS, Grogan
TM. The phenomenon of multidrug resistance
in non-Hodgkins lymphoma. Cancer Treat-
ment & Res 1996; 85:107117.
3 Zubercova O, Babusikova O. The multidrug
resistance in human leukemias. Minireview.
Neoplasma 1998; 45:5359.
4 Biological markers and changes induced in
their profiles following primary chemotherapy.
In: Howell A, Dowsett M, eds. Primary medi-
cal therapy for breast cancer. Amsterdam:
Elsevier, 1999.
5 Pilat MJ, Kamradt JM, Pienta KJ. Hormone
resistance in prostate cancer. Cancer Metasta-
sis Rev 1999; 17:373381.
6 Goldenberg DM. Targeted therapy of cancer
with radiolabeled antibodies. J Nucl Med
2002; 43:693713.
7 Tepmongkol S, Heyman S.
131
I MIBG therapy
in neuroblastoma: mechanisms, rationale, and
current status. Med Pediatr Oncol 1999;
32:427431.
8 Moadel RM, Nguyen AV, Lin E, Lu P, Mani J,
Blaufox MD, et al. Positron emission tomogra-
phy agent 2-deoxy-2-[
18
F]fluoro-D-glucose has
a therapeutic potential in breast cancer. Breast
Cancer Res 2003; 5:R199R205.
9 Reilly RM. Radioimmunotherapy of malignan-
cies. Clin Pharm 1991; 10:359375.
10 ODonoghue JA, Bardies M, Wheldon TE. Re-
lationship between tumor size and curability
for targeted radionuclide therapy with beta-
emitting radionuclides. J Nucl Med 1995;
36:19021909.
11 Auger P. Sur les rayons b secondaires produit
dans un gaz par des rayons X. Comp Rend
1925; 180:6568.
12 Stashenko P, Nadler LM, Hardy R, Schloss-
man SF. Characterization of a human B lym-
phocyte-specific antigen. J Immunol 1980;
125:16781685.
13 Press OW, Eary JF, Appelbaum FR, et al.
Phase II trial of
131
I-B1 (anti-CD20) antibody
therapy with autologous stem cell transplanta-
tion for relapsed B cell lymphomas. Lancet
1995; 346:336340.
14 Wahl RL, Zasadny KR, MacFarlane D, Francis
IR, Ross CW. Iodine-131 anti-B1 antibody for
B-cell lymphoma: an update on the Michigan
Phase I experience. J Nucl Med 1998; 39:21S
27S.
15 Liu SY, Eary JF, Petersdorf SH, Martin PJ,
Maloney DG, Appelbaum FR, et al. Follow-up
References 527
of relapsed B-cell lymphoma patients treated
with iodine-131-labeled anti-CD20 antibody
and autologous stem-cell rescue. J Clin Oncol
1998; 16:32703278.
16 Kaminski MS, Zasadny KR, Francis IR, Fen-
ner MC, Ross CW, Milik AW, et al. Iodine-
131-anti-B1 radioimmunotherapy for B-cell
lymphoma. J Clin Oncol 1996; 14:19741981.
17 Kaminski MS, Estes J, Zasadny KR, Francis
IR, Ross CW, Tuck M, et al. Radioimmuno-
therapy with iodine 131 tositumomab for re-
lapsed or refractory B-cell non-Hodgkin lym-
phoma: updated results and long-term follow-
up of the University of Michigan experience.
Blood 2000; 96:12591266.
18 Cheson BD. Radioimmunotherapy of non-
Hodgkin lymphomas. Blood 2003; 15:391398.
19 Vose JM, Wahl RL, Saleh M, Rohatiner AZ,
Knox SJ, Radford JA, et al. Multicenter phase
II study of iodine-131 tositumomab for che-
motherapy-relapsed/refractory low-grade and
transformed low-grade B-cell non-Hodgkins
lymphomas. J Clin Oncol 2000; 18:13161323.
20 Sgouros G, Squeri S, Ballangrud AM, Kolbert
KS, Teitcher JB, Panageas KS, et al. Patient-
specific, 3-dimensional dosimetry in non-
Hodgkins lymphoma patients treated with
131
I-anti-B1 antibody: Assessment of tumor
dose-response. J Nucl Med 2003; 44:260268.
21 Press OW, Eary JF, Gooley TA, Gopal AK, Liu
S, Rajendran JG, et al. A phase I/II trial of io-
dine-131-tositumomab (anti-CD20), etoposide,
cyclophosphamide, and autologous stem cell
transplantation for relapsed B-cell lymphomas.
Blood 2000; 96:29342942.
22 Gopal AK, Gooley TA, Maloney DG, Peters-
dorf SH, Eary JF, Rajendran JG, et al. High-
dose radioimmunotherapy versus conventional
high-dose therapy and autologous hemato-
poietic stem cell transplantation for relapsed
follicular non-Hodgkin lymphoma: a multi-
variate cohort analysis. Blood 2003; 102:2351
2357.
23 Harwood SJ, Gibbons LK, Goldner PJ, Web-
ster WB, Carroll RG. Outpatient radioimmu-
notherapy with Bexxar. Cancer 2002; 94:1358
1362.
24 Witzig TE, White CA, Wiseman GA, Gordon
LI, Emmanouilides C, Raubitschek A, et al.
Phase I/II trial of IDEC-Y2B8 radioimmuno-
therapy for treatment of relapsed or refractory
CD20+ B-cell non-Hodgkins lymphoma. J
Clin Oncol 1999; 17:37933803.
25 Witzig TE, White CA, Gordon LI, Wiseman
GA, Emmanouilides C, Murray JL, et al.
Safety of yttrium-90 ibritumomab tiuxetan
radioimmunotherapy for relapsed low-grade,
follicular, or transformed non-Hodgkins lym-
phoma. J Clin Oncol 2003; 21:12631270.
26 Wiseman GA, Gordon LI, Multani PS, Witzig
TE, Spies S, Bartlett NL, et al. Ibritumomab
tiuxetan radioimmunotherapy for patients
with relapsed or refractory non-Hodgkin
lymphoma and mild thrombocytopenia:
a phase II multicenter trial. Blood 2002;
99:43364342.
27 Witzig TE, Gordon LI, Cabanillas F, Czucz-
man MS, Emmanouilides C, Joyce R, et al.
Randomized controlled trial of yttrium-90-la-
beled ibritumomab tiuxetan radioimmunother-
apy versus rituximab immunotherapy for pa-
tients with relapsed or refractory low-grade,
follicular, or transformed B-cell non-Hodgkins
lymphoma. J Clin Oncol 2002; 20:24532463.
28 McLaughlin P, Grillo-Lopez AJ, Link BK, Levy
R, Czuczman MS, Williams ME, et al. Rituxi-
mab chimeric anti-CD20 monoclonal antibody
therapy for relapsed indolent lymphoma: half
of patients respond to a four-dose treatment
program. J Clin Oncol 1998; 16:28252833.
29 Witzig TE, Flinn IW, Gordon LI, Emmanoui-
lides C, Czuczman MS, Saleh MN, et al.
Treatment with ibritumomab tiuxetan radio-
immunotherapy in patients with rituximab-re-
fractory follicular non-Hodgkins lymphoma. J
Clin Oncol 2002; 20:32623269.
30 Juweid ME, Stadmauer E, Hajjar G, Sharkey
RM, Suleiman S, Luger S, et al. Pharmacoki-
netics, dosimetry, and initial therapeutic re-
sults with
131
I- and
111
In-/
90
Y-labeled human-
ized LL2 anti-CD22 monoclonal antibody in
patients with relapsed, refractory non-Hodg-
kins lymphoma. Clin Cancer Res 1999;
5:3292s3303s.
31 DeNardo GL, DeNardo SJ, Goldstein DS, Kro-
ger LA, Lamborn KR, Levy NB, et al. Maxi-
mum-tolerated dose, toxicity, and efficacy of
(131)I-Lym-1 antibody for fractionated radio-
immunotherapy of non-Hodgkins lymphoma.
J Clin Oncol 1998; 16:32463256.
32 Buchsbaum DJ, LoBuglio AF. Targeting of
125
I-labeled B-lymphocyte stimulator. J Nucl
Med 2003; 44:434435.
33 Riccobene TA, Miceli RC, Lincoln C, Knight
Y, Meadows J, Stabin MG, et al. Rapid and
specific targeting of
125
I-labeled B lymphocyte
stimulator to lymphoid tissues and B cell tu-
mors in mice. J Nucl Med 2003; 44:422433.
34 Reilly RM, Sandhu J, Alvarez-Diez TM, Gal-
linger S, Kirsh J, Stern H. Problems of deliv-
ery of monoclonal antibodies. Pharmaceutical
and pharmacokinetic solutions. Clin Pharma-
cokinet 1995; 28(2):126142.
35 Schwartz MA, Lovett DR, Redner A, Finn RD,
Graham MC, Divgi CR, et al. Dose-escalation
trial of M195 labeled with iodine 131 for cyto-
reduction and marrow ablation in relapsed or
refractory myeloid leukemias. J Clin Oncol
1993; 11:294303.
36 Tanimoto M, Scheinberg DA, Cordon-Cardo
C, Huie D, Clarkson BD, Old LJ. Restricted
expression of an early myeloid and monocytic
cell surface antigen defined by monoclonal
antibody M195. Leukemia 1989; 3:339348.
37 Freeman SD, Kelm S, Barber EK, Crocker PR.
Characterization of CD33 as a new member
of the sialoadhesin family of cellular interac-
tion molecules. Blood 1995; 85:20052012.
38 Burke JM, Jurcic JG, Scheinberg DA. Radio-
immunotherapy for acute leukemia. Cancer
Control 2002; 9:106113.
39 Jurcic JG, Caron PC, Nikula TK, Papadopou-
los EB, Finn RD, Gansow OA, et al. Radiola-
beled anti-CD33 monoclonal antibody M195
for myeloid leukemias. Cancer Res 1995;
55:5908s5910s.
40 Caron PC, Schwartz MA, Co MS, Queen C,
Finn RD, Graham MC, et al. Murine and hu-
manized constructs of monoclonal antibody
M195 (anti-CD33) for the therapy of acute
myelogenous leukemia. Cancer 1994;
73:10491056.
41 Co MS, Avdalovic NM, Caron PC, Avdalovic
MV, Scheinberg DA, Queen C. Chimeric and
humanized antibodies with specificity for the
CD33 antigen. J Immunol 1992; 148:1149
1154.
42 Pagel JM, Matthews DC, Appelbaum FR,
Bernstein ID, Press OW. The use of radioim-
munoconjugates in stem cell transplantation.
Bone Marrow Transplant 2002; 29:807816.
43 Jurcic JG, Larson SM, Sgouros G, McDevitt
MR, Finn RD, Divgi CR, et al. Targeted a par-
ticle immunotherapy for myeloid leukemia.
Blood 2002; 100:12331239.
44 Sgouros G, Ballangrud AM, Jurcic JG, McDe-
vitt MR, Humm JL, Erdi Y, et al. Pharmacoki-
netics and dosimetry of an a-particle emitter
labeled antibody:
213
Bi-HuM195 (anti-CD33) in
patients with leukemia. J Nucl Med 1999;
40:19351946.
45 McDevitt MR, Sgouros G, Finn RD, Humm
JL, Jurcic JG, Larson SM, et al. Radioimmu-
notherapy with alpha-emitting radionuclides.
Eur J Nucl Med 1998; 25:13411351.
46 McDevitt MR, Dangshe M, Lai L, Simon J,
Borchardt P, Frank KR, et al. Tumor therapy
with targeted atomic nanogenerators. Science
2001; 294:15371540.
47 Miederer M, McDevitt MR, Sgouros G, Kra-
mer K, Cheung N-KV, Scheinberg DA. Phar-
macokinetics, dosimetry, and toxicity of the
targetable atomic generator,
225
Ac-HuM195, in
nonhuman primates. J Nucl Med 2004;
45:129137.
48 Jordan CT. Unique molecular and cellular fea-
tures of acute myelogenous leukemia stem
cells. Leukemia 2002; 16:559562.
49 Dillehay LE, Mayer R, Zhang YG, Shao Y,
Song S-Y, Mackensen DG, et al. Prediction of
tumor response to experimental radioim-
munotherapy with
90
Y in nude mice. Int J Ra-
diat Oncol Biol Phys 1995; 33:417427.
50 Williams LE, Duda RB, Proffitt RT, Beatty BG,
Beatty JD, Wong JYC, et al. Tumor uptake as
a function of tumor mass: A mathematical
model. J Nucl Med 1988; 29:103109.
51 Behr TM, Liersch T, Greiner-Bechert L, Grie-
singer F, Behe M, Markus PM, et al. Radioim-
munotherapy of small-volume disease of me-
tastatic colorectal cancer. Cancer 2002;
94:13731381.
52 Zalutsky MR. Targeted radiotherapy of brain
tumors. Br J Cancer 2004; 90:14691473.
53 Riva P, Franceschi G, Frattarelli M, Riva N,
Guiducci G, Cremonini AM, et al.
131
I radio-
conjugated antibodies for the locoregional
radioimmunotherapy of high-grade malignant
glioma. Acta Oncologica 1999; 38:351359.
54 Wong JYC, Shibata S, Williams LE, Kwok CS,
Liu A, Chu DZ, et al. A phase I trial of
90
Y-
anti-carcinoembryonic antigen chimeric
T84.66 radioimmunotherapy with 5-fluoroura-
cil in patients with metastatic colorectal can-
cer. Clin Cancer Res 2003; 9:58425852.
55 Smalley SR, Kimler BF, Evans RG. 5-Fluoro-
uracil modulation of radiosensitivity in cul-
tured human carcinoma cells. Int J Radiat On-
col Biol Phys 1991; 20:207211.
56 Macey DJ, Grant EJ, Kasi L, Rosenblum MG,
Zhang HZ, Katz RL, et al. Effect of recombi-
nant alpha-interferon on pharmacokinetics,
References 529
biodistribution, toxicity, and efficacy of
131
I-la-
beled monoclonal antibody CC49 in breast
cancer: a phase II trial. Clin Cancer Res 1997;
3(9):15471555.
57 Epenetos AA, Hird V, Lambert H, Mason P,
Coulter C. Long term survival of patients with
advanced ovarian cancer treated with intra-
peritoneal radioimmunotherapy. Int J Gynecol
Cancer 2000; 10(Suppl 1):4446.
58 ODonnell RT, DeNardo SJ, Yuan A, Shen S,
Richman CM, Lara PN, et al. Radioimmuno-
therapy with
111
In/
90
Y-2-IT-BAD-m170 for me-
tastatic prostate cancer. Clin Cancer Res 2001;
7:15611568.
59 Milowsky M, Nanus DM, Kostakoglu L, Val-
labhajosula S, Goldsmith SJ, Bander NH.
Phase I trial of yttrium-90-labeled anti-pros-
tate-specific membrane antigen monoclonal
antibody J591 for androgen-independent pros-
tate cancer. J Clin Oncol 2004; 22:25222531.
60 Chaiet L, Wolf FJ. The properties of streptavi-
din, a biotin-binding protein produced by
streptomycetes. Arch Biochem Biophys 1964;
106:15.
61 Green NM. Avidin. Adv Protein Chem 1975;
29:85133.
62 Hnatowich DJ, Virzi F, Rusckowski M. Inves-
tigations of avidin and biotin for imaging ap-
plications. J Nucl Med 1987; 28:12941302.
63 Chinol M, Grana C, Gennari R, Cremonesi M,
Geraghty JG, Paganelli G. Pretargeted radio-
immunotherapy of cancer. In: Abrams PG,
Fritzberg AR, eds. Radioimmunotherapy of can-
cer. New York: Marcel Dekker, Inc., 2000: 169
193.
64 Theodore LJ, Fritzberg AR, Schultz JE, Ax-
worthy DB. Evolution of a pretarget radioim-
munotherapeutic regimen. In: Abrams PG,
Fritzberg AR, eds. Radioimmunotherapy of can-
cer. New York: Marcel Dekker, Inc., 2000: 195
221.
65 Axworthy DB, Reno JM, Hylarides MD, Mal-
lett RW, Theodore LJ, Gustavson LM, et al.
Cure of human carcinoma xenografts by a sin-
gle dose of pretargeted yttrium-90 with negli-
gible toxicity. Proc Natl Acad Sci USA 2000;
97:18021807.
66 Breitz H, Knox S, Weiden P, Goris M, Murtha
A, Bryan J, et al. Pretargeted radioimmuno-
therapy with antibody-streptavidin and Y-90
DOTA-biotin (Avicidin): Result of a dose esca-
lation study. J Nucl Med 1998; 39:71P.
67 Knox SJ, Goris ML, Tempero M, Weiden PL,
Gentner L, Breitz H, et al. Phase II trial of yt-
trium-90-DOTA-biotin pretargeted by NR-LU-
10 antibody/streptavidin in patients with me-
tastatic colon cancer. Clin Cancer Res 2000;
6:406414.
68 Yao Z, Zhang M, Axworthy DB, Wong KJ,
Garmestani K, Park L, et al. Radioimmuno-
therapy of A431 xenografted mice with pretar-
geted B3 antibody-streptavidin and
90
Y-labeled
1,4,7,10-tetraazacyclododecane-N,N',N'',N'''-tet-
raacetic acid (DOTA)-biotin. Cancer Res 2002;
62:57555760.
69 Yao Z, Zhang M, Garmestani K, Axworthy
DB, Mallett RW, Fritzberg AR, et al. Pretar-
geted a emitting radioimmunotherapy using
213
Bi 1,4,7,10-tetraazacyclododecane-
N,N',N'',N''-tetraacetic acid-biotin. Clin Cancer
Res 2004; 10:31373146.
70 Ngai WM, Reilly RM, Polihronis J, Shpitz B.
In-vitro and in-vivo evaluation of streptavidin
immunoconjugates of the second-generation
TAG-72 monoclonal antibody CC49. Nucl Med
Biol 1994; 22:7786.
71 Thor A, Ohuchi N, Szpak CA, Johnston WW,
Schlom J. Distribution of oncofetal antigen tu-
mor-associated glycoprotein-72 defined by
monoclonal antibody B72.3. Cancer Res 1986;
46:31183124.
72 Domingo R, Reilly RM. Pre-targeted radioim-
munotherapy of human colon cancer xeno-
grafts in athymic mice using streptavidin-
CC49 monoclonal antibody and
90
Y-DOTA-bio-
tin. Nucl Med Commun 2000; 8996.
73 Grana C, Chinol M, Robertson C, Mazzetta C,
Bartolomei M, De Cicco C, et al. Pretargeted
adjuvant radioimmunotherapy with yttrium-
90-biotin in malignant glioma patients: A pilot
study. Br J Cancer 2002; 86:207212.
74 Paganelli G, Grana C, Chinol M, Cremonesi
M, De Cicco C, De Braud F, et al. Antibody-
guided three-step therapy for high-grade glio-
ma with yttrium-90 biotin. Eur J Nucl Med
1999; 26:348357.
75 Schultz J, Lin Y, Sanderson J, Zuo Y, Stone D,
Mallett R, et al. A tetravalent single-chain anti-
body-streptavidin fusion protein for pretar-
geted lymphoma therapy. Cancer Res 2000;
60:66636669.
76 Zhang M, Zhang Z, Garmestani K, Schultz J,
Axworthy DB, Goldman CK, et al. Pretarget
radiotherapy with an anti-CD25 antibody-
streptavidin fusion protein was effective in
therapy of leukemia/lymphoma xenografts.
Proc Natl Acad Sci USA 2003; 100:18911895.
77 Press OW, Corcoran M, Subbiah K, Hamlin
DK, Wilbur DS, Johnson T, et al. A compara-
tive evaluation of conventional and pretargeted
radioimmunotherapy of CD20-expressing lym-
phoma xenografts. Blood 2001; 98:25352543.
78 Weiden PL, Breitz HB. Pretargeted radioim-
munotherapy (PRITTM) for treatment of non-
Hodgkins lymphoma (NHL). Crit Rev Oncol
Hematol 2001; 40:3751.
79 Jong M, Kwekkeboom D, Valkema R, Kren-
ning EP. Radiolabelled peptides for tumor
therapy: Current status and future directions.
Eur J Nucl Med 2003; 30:463469.
80 Bodei L, Cremonesi M, Grana C, Rocca P,
Bartolomei M, Chinol M, et al. Receptor radio-
nuclide therapy with
90
Y-[DOTA]-D-Tyr3-oc-
treotide (
90
Y-DOTATOC) in neuroendocrine
tumors. Eur J Nucl Med 2004; 31:10381046.
81 Weiner RE, Thakur ML. Radiolabeled peptides
in the diagnosis and therapy of oncological
diseases. Appl Radiat Isot 2002; 57:749763.
82 Kwekkeboom DJ, Krenning EP. Somatostatin
receptor imaging. Semin Nucl Med 2002;
32:8491.
83 Paganelli G, Zoboli S, Cremonesi M, Bodei L,
Ferrari M, Grana C, et al. Receptor-mediated
radiotherapy with
90
Y-DOTA-D-Phe1-Tyr3-oc-
treotide. Eur J Nucl Med 2001; 28:426434.
84 Cassidy JR. Clinical radiation nephropathy. Int
J Radiat Oncol Biol Phys 1995; 31:12491256.
85 Behr TM, Goldenberg DM, Becker W. Reduc-
ing the renal uptake of radiolabeled antibody
fragments and peptides for diagnosis and
therapy: present status, future prospects and
limitations. Eur J Nucl Med 1998; 25:201212.
86 Waldherr C, Pless M, Maecke HR, Halde-
mann A, Mueller-Brand J. The clinical value
of [
90
Y-DOTA]-D-Phe1-Tyr3-octreotide (
90
Y-DO-
TATOC) in the treatment of neuroendocrine
tumors: A clinical phase II study. Ann Oncol
2001; 12:941945.
87 Waldherr C, Pless M, Maecke HR, Schuma-
cher T, Crazzolara A, Nitzsche EU, et al. Tu-
mor response and clinical benefit in neuroen-
docrine tumors after 7.4 GBq
90
Y-DOTATOC.
J Nucl Med 2002; 43:610616.
88 Cybulla M, Weiner SM, Otte A. End-stage re-
nal disease after treatment with
90
Y-DOTA-
TOC. Eur J Nucl Med 2001; 28:15521554.
89 Behr TM, Behe M, Kluge G, Gotthardt M,
Schipper ML, Gratz S. Nephrotoxicity versus
anti-tumor efficacy in radiopeptide therapy:
facts and myths about the Scylla and Charyb-
dis. Eur J Nucl Med 2002; 29:277279.
90 Boerman OC, Oyen WJ, Corstens FHM. Be-
tween the Scylla and Charybdis of peptide
radionuclide therapy: hitting the tumor and
saving the kidney. Eur J Nucl Med 2001;
28:14471449.
91 Otte A, Cybulla M, Weiner SM.
90
Y-DOTATOC
and nephrotoxicity. Eur J Nucl Med 2002;
29:1543.
92 Schumacher T, Waldherr C, Mueller-Brand J,
Maecke H. Kidney failure after treatment
with
90
Y-DOTATOC. Eur J Nucl Med 2002;
29:435.
93 Moll S, Nickeleit V, Mueller-Brand J, Brunner
FP, Maecke H, Mihatsch J. A new cause of re-
nal thrombotic microangiopathy: yttrium 90-
DOTATOC internal radiotherapy. Am J Kidney
Dis 2001; 37:847851.
94 Bodei L, Cremonesi M, Zoboli S, Grana C,
Bartolomei M, Rocca P, et al. Receptor-
mediated radionuclide therapy with
90
Y-DOTA-
TOC in association with amino acid infusion:
a phase I study. Eur J Nucl Med 2003; 30:207
216.
95 Jamar F, Barone R, Mathieu I, Walrand S,
Labar D, Carlier P, et al.
86
Y-DOTA0-D-Phe1-
Tyr3-octreotide (SMT487). A phase 1 clinical
study: pharmacokinetics, biodistribution and
renal protective effect of different regimens of
amino acid co-infusion. Eur J Nucl Med 2003;
30:510518.
96 Merlo A, Hausmann O, Wasner M, Steiner P,
Otte A, Jermann E, et al. Locoregional regula-
tory peptide receptor targeting with the diffus-
able somatostatin analogue
90
Y-labeled
DOTA0-D-Phe1-Tyr3-octreotide (DOTATOC):
A pilot study in human gliomas. Clin Cancer
Res 1999; 5:10251033.
97 Schumacher T, Hofer S, Eichhorn K, Wasner
M, Zimmerer S, Freitag P, et al. Local injec-
tion of the
90
Y-labelled peptidic vector DOTA-
TOC to control gliomas of WHO grades II
and III: an extended pilot study. Eur J Nucl
Med 2002; 29:486493.
98 Kassis AI. Cancer therapy with Auger elec-
trons: Are we almost there? J Nucl Med 2003;
44:14791481.
99 Xue LY, Butler NJ, Makrigiorgos GM, Adel-
stein SJ, Kassis AI. Bystander effect produced
by radiolabeled tumor cells in vivo.
References 531
Proc Natl Acad Sci USA 2002; 99:13765
13770.
100 Hofer KG, Harris CR, Smith MJ. Radiotoxic-
ity of intracellular Ga-67, I-125 and H-3. Nu-
clear versus cytoplasmic radiation effects in
murine L1210 leukemia cells. Int J Radiat
Biol Rel Stud Phys Chem Med 1975;
28:225241.
101 Adelstein SJ. The Auger process: A thera-
peutic promise. Am J Radiol 1993; 160:707
713.
102 Krenning EP, Kwekkeboom DJ, Bakker WH,
et al. Somatostatin receptor scintigraphy
with [
111
In-DTPA-D-Phe1]- and [
123
I-Tyr3]-oc-
treotide: the Rotterdam experience with
more than 1000 patients. Eur J Nucl Med
1993; 20:716731.
103 Janson ET, Westlin J-E, Ohrvall U, Oberg K,
Lukinius A. Nuclear localization of
111
In
after intravenous injection of [
111
In-DTPA-D-
Phe1]-octreotide in patients with neuroendo-
crine tumors. J Nucl Med 2000; 41:1514
1518.
104 Likinius A, Ohrvall U, Westlin J-E, Oberg K,
Janson ET. In vivo cellular distribution and
endocytosis of the somatostatin receptorli-
gand complex. Acta Oncol 1999; 38:383387.
105 Andersson P, Forssell-Aronsson E, Johanson
V, Wangberg B, Nilsson O, Fjalling M, et al.
Internalization of indium-111 into human
neuroendocrine tumor cells after incubation
with indium-111-DTPA-D-Phe1-octreotide.
J Nucl Med 1996; 37:20022006.
106 Capello A, Krenning EP, Breeman WAP,
Bernard BF, de Jong M. Peptide receptor
radionuclide therapy in vitro using [
111
In-
DTPA0]octreotide. J Nucl Med 2003; 44:98
104.
107 Slooter GD, Breeman WAP, Marquet RL,
Krenning EP, Van Eijck CHJ. Anti-prolifera-
tive effect of radiolabelled octreotide in a
metastasis model in rat liver. Int J Cancer
1999; 81:767771.
108 Anthony LB, Woltering EA, Espenan GD,
Cronin MD, Maloney TJ, McCarthy KE. In-
dium-111-pentetreotide prolongs survival in
gastroenteropancreatic malignancies. Semin
Nucl Med 2002; 32:123132.
109 Janson ET, Eriksson B, Oberg K, Skogseid B,
Ohrvall U, Nilsson S, et al. Treatment with
high-dose [
111
In-DTPA-D-Phe1]-octreotide in
patients with neuroendocrine tumors. Acta
Oncol 1999; 38:373377.
110 Valkema R, de Jong M, Bakker WH, Bree-
man WA, Kooij PP, Lugtenburg PJ, et al.
Phase I study of peptide receptor radionu-
clide therapy with [In-DTPA]octreotide: the
Rotterdam experience. Semin Nucl Med
2002; 32:110122.
111 Larkin M. Trojan horse strategy enters
phase I trial for breast cancer. Lancet Oncol-
ogy 2003; 4:650.
112 Klijn JGM, Berns PMJJ, Schmitz PIM, Foe-
kens JA. The clinical significance of epider-
mal growth factor receptor (EGF-R) in hu-
man breast cancer: a review on 5232 pa-
tients. Endocr Rev 1992; 13:317.
113 Rakowicz-Szulczynska EM, Rodeck U, Her-
lyn M, Koprowski H. Chromatin binding of
epidermal growth factor, nerve growth factor,
and platelet-derived growth factor in cells
bearing the appropriate surface receptors.
Proc Natl Acad Sci USA 1986; 83:3728
3732.
114 Burwen SJ, Jones AL. The association of
polypeptide hormones and growth factors
with the nuclei of target cells. Trends Bio-
chem Sci 1987; 12:159162.
115 Schausberger E, Eferl R, Parzefall W, Chabi-
kovsky M, Breit P, Wagner EF, et al. Induc-
tion of DNA synthesis in primary mouse he-
patocytes is associated with nuclear pro-
transforming growth factor a and erbb-1 and
is independent of c-jun. Carcinogenesis
2003; 24:835841.
116 Lin S-Y, Makino K, Xia W, Matin A, Wen Y,
Kwong KY, et al. Nuclear localization of EGF
receptor and its potential new role as a tran-
scription factor. Nature Cell Biol 2001;
3:802808.
117 Laduron PM. Genomic pharmacology: more
intracellular sites for drug action. Biochem
Pharmacol 1992; 44:12331242.
118 Goddu SM, Howell RW, Rao DV. Cellular
dosimetry: Absorbed fractions for monoener-
getic electron and alpha particle sources and
S-values for radionuclides uniformly distrib-
uted in different cell compartments. J Nucl
Med 1994; 35:303316.
119 Reilly RM, Kiarash R, Cameron R, Porlier
N, Sandhu J, Hill RP, et al. Indium-111 la-
belled EGF is selectively radiotoxic to human
breast cancer cells overexpressing EGFR. J
Nucl Med 2000; 41:429438.
120 Chen P, Mrkobrada M, Vallis KA, Sandhu J,
Cameron R, Hendler A, et al. Comparative
antiproliferative effects of novel targeted Au-
ger electron radiotherapy, chemotherapy and
external c-radiation on EGFR-overexpressing
breast cancer cells. Nucl Med Biol 2002;
29:693699.
121 Wang J, Chen P, Su Z-F, Sandhu J, Vallis K,
Cameron R, et al. Amplified delivery of in-
dium-111 to EGFR-positive breast cancer
cells. Nucl Med Biol 2001; 28:895902.
122 Orlova A, Briskin A, Sjstrm A, Lundqvist
H, Gedda L, Tolmachev V. Cellular processing
of
125
I- and
111
In-labeled epidermal growth
factor (EGF) bound to cultured A431 tumor
cells. Nucl Med Biol 2000; 27:827835.
123 Chen P, Cameron R, Wang J, Vallis KA,
Reilly RM. Antitumor effects and normal tis-
sue toxicity of
111
In-labeled epidermal
growth factor administered to athymic mice
bearing epidermal growth factor receptor-
positive human breast cancer xenografts. J
Nucl Med 2003; 44:14691478.
124 Reilly RM, Scollard DA, Wang J, Mondal H,
Chen P, Henderson LA, et al. A kit formu-
lated under good manufacturing practices
for labeling human epidermal growth factor
with
111
In for radiotherapeutic applications.
J Nucl Med 2004; 45:701708.
125 Bronzert DA, Hochberg RB, Lippman ME.
Specific cytotoxicity of 16-a-[
125
I]iodoestradiol
for estrogen receptor-containing breast can-
cer cells. Endocrinology 1982; 110:2177
2179.
126 DeSombre ER, Hughes A, Hanson RN,
Kearney T. Therapy of estrogen receptor-pos-
itive micrometastases in the peritoneal cavity
with Auger electron-emitting estrogens. Acta
Oncol 2000; 39:659666.
127 McLaughlin WH, Milius RA, Pillai KMR,
Edasery JP, Blumenthal RD, Bloomer WD.
Cytotoxicity of receptor-mediated 16a-
[
125
I]iodo-estradiol in cultured MCF-7 hu-
man breast cancer cells. J Natl Cancer Inst
1989; 81:437440.
128 Bloomer WD, McLaughlin WH, Weichsel-
baum RR, Tonnesen GL, Hellman S, Seitz
DE, et al. Iodine-125-labeled tamoxifen is
differentially cytotoxic to cells containing
oestrogen receptors. Int J Radiat Biol 1980;
38:197202.
129 Sundberg L, Orlova A, Bruskin A, Gedda
L, Carlsson J, Blomquist E, et al. [
111
In]Bz-
DTPA-hEGF: Preparation and in vitro char-
acterization of a potential anti-glioblastoma
targeting agent. Cancer Biother Radiopharm
2004; 18:643654.
130 Carlsson J, Blomquist E, Gedda L, Liljegren
, Malmstrm P-U, Sjstrm A, et al. Con-
jugate chemistry and cellular processing of
EGF-dextran. Acta Oncol 1999; 38:313321.
131 Kurihara A, Deguchi Y, Pardridge WM. Epi-
dermal growth factor radiopharmaceuticals:
111
In chelation, conjugation to a blood-brain
barrier delivery vector via a biotin-polyethy-
lene linker, pharmacokinetics, and in vivo
imaging of experimental brain tumors. Bio-
conjug Chem 1999; 10:502511.
132 Behr TM, Behe M, Lohr M, Sgouros G, An-
gerstein C, Wehrmann E, et al. Therapeutic
advantages of Auger electron- over b-emit-
ting radiometals or radioiodine when conju-
gated to internalizing antibodies. Eur J Nucl
Med 2000; 27:753765.
133 Behr TM, Sgouros G, Vougioukas V, Mem-
tsoudis S, Gratz S, Schmidberger H, et al.
Therapeutic efficacy and dose-limiting toxici-
ty of Auger-electron vs. beta emitters in
radioimmunotherapy with internalizing anti-
bodies: evaluation of
125
I- vs.
131
I-labeled
CO17-1A in a human colorectal cancer mod-
el. Int J Cancer 1998; 76:738748.
134 Govindan SV, Goldenberg DM, Elsamra SE,
Griffiths GL, Ong GL, Brechbiel MW, et al.
Radionuclides linked to a CD74 antibody as
therapeutic agents for B-cell lymphoma:
comparison of Auger electron emitters with
b-particle emitters. J Nucl Med 2000;
41:20892097.
135 Griffiths GL, Govindan SV, Sgouros G, Ong
GL, Goldenberg DM, Mattes MJ. Cytotoxicity
with Auger electron-emitting radionuclides
delivered by antibodies. Int J Cancer 1999;
81:985992.
136 Mattes MJ. Radionuclide-antibody conjugates
for single-cell cytotoxicity. Cancer 2002;
94:12151223.
137 Ong GL, Elsamra SE, Goldenberg DM,
Mattes MJ. Single-cell cytotoxicity with radio-
labeled antibodies. Clin Cancer Res 2001;
7:192201.
138 Ochakovskaya R, Osorio L, Goldenberg DM,
Mattes MJ. Therapy of disseminated B-cell
lymphoma xenografts in severe combined
immunodeficient mice with an anti-CD74
antibody conjugated with
111
Indium, 67Gal-
lium, or
90
Yttrium. Clin Cancer Res 2001;
7:15051510.
References 533
139 Quang TS, Brady L. Radioimmunotherapy
as a novel treatment regimen:
125
I-labeled
monoclonal antibody 425 in the treatment of
high-grade brain gliomas. Int J Radiat Oncol
Biol Phys 2004; 58:972975.
140 Saga T, Sakahara H, Nakamoto Y, Sato N,
Zhao S, Aoki T, et al. Radioimmunotherapy
of human glioma xenografts in nude mice
by indium-111 labelled internalizing mono-
clonal antibody. Eur J Cancer 1999; 35:1281
1285.
141 Karamychev VN, Panyutin IG, Kim M-K,
Le N, Paik CH, Carrasquillo JA, et al. DNA
cleavage by
111
In-labeled oligodeoxyribonu-
cleotides. J Nucl Med 2000; 41:10931101.
142 Panyutin IG, Sedelnikova OA, Karamychev
VN, Neumann RD. Antigene radiotherapy.
Ann NY Acad Sci 2003; 1002:134140.
143 Karamychev VN, Reed MW, Neumann RD,
Panyutin IG. Distribution of DNA strand
breaks produced by iodine-123 and indium-
111 in synthetic oligodeoxynucleotides. Acta
Oncol 2000; 39:687692.
144 Sedelnikova OA, Luu AN, Karamychev VN,
Panyutin IG, Neumann RD. Development of
DNA-based radiopharmaceuticals carrying
Auger-electron emitters for antigene radio-
therapy. Int J Radiat Oncol Biol Phys 2001;
49:391396.
145 Dachs GU, Dougherty GJ, Stratford IJ, Cha-
plin DJ. Targeting gene therapy to cancer: A
review. Oncology Res 1997; 9:313325.
146 Michel RB, Rosario AV, Brechbiel MW, Jack-
son TJ, Goldenberg DM, Mattes MJ. Experi-
mental therapy of disseminated B-cell lym-
phoma xenografts with
213
Bi-labeled anti-
CD74. Nucl Med Biol 2003; 30:715723.
147 Aurlien E, Larsen RH, Kvalheim G, Bruland
OS. Demonstration of highly specific toxicity
of the a-emitting radioimmunoconjugate
211
At-rituximab against non-Hodgkins lym-
phoma cells. Br J Cancer 2000; 83:1375
1379.
148 Beyer GJ, Miederer M, Vranjes-Duric S, Co-
mor JJ, Kunzi G, Hartley O, et al. Targeted
alpha therapy in vivo: direct evidence for sin-
gle cancer cell kill using
149
Tb-rituximab.
Eur J Nucl Med 2004; 31:547554.
149 Borchardt PE, Yuan RR, Miederer M, McDe-
vitt MR, Scheinberg DA. Targeted actinium-
225 in vivo generators for therapy of ovarian
cancer. Cancer Res 2003; 63:50845090.
150 Davies AJ, Rohatiner AZS, Howell S, Britton
KE, Owens SE, Micallef IN, et al. Tositumo-
mab and iodine I 131 tositumomab for re-
current indolent and transformed B-cell
non-Hodgkins lymphoma. J Clin Oncol
2004; 22:14691479.
151 Tempero M, Leichner P, Dalrymple G, Har-
rison K, Augustine S, Schlom J, et al. High-
dose therapy with iodine-131-labeled mono-
clonal antibody CC49 in patients with gas-
trointestinal cancers: A phase I trial. J Clin
Oncol 1997; 15:15181528.
152 Wong JYC, Chu DZ, Yamauchi DM, Wil-
liams LE, Liu A, Wilczynski S, et al. A phase
I radioimmunotherapy trial evaluating
90
Yt-
trium-labeled anti-carcinoembryonic antigen
(CEA) chimeric T84.66 in patients with me-
tastatic CEA-producing malignancies. Clin
Cancer Res 2000; 6:38553863.
153 Macey DJ, Grant EJ, Kasi L, Rosenblum
MG, Zhang HZ, Katz RL, et al. Effect of re-
combinant alpha-interferon on pharmacoki-
netics, biodistribution, toxicity, and efficacy
of
131
I-labeled monoclonal antibody CC49 in
breast cancer: a phase II trial. Clin Cancer
Res 1997; 3:15471555.
154 DeNardo SJ, Kramer EL, ODonnell RT,
Richman CM, Salako QA, Shen S, et al.
Radioimmunotherapy for breast cancer
using indium-111/yttrium-90 BrE-3: results
of a phase I clinical trial. J Nucl Med 1997;
38:11801185.
155 Meredith RF, Partridge EE, Alvarez RD,
Khazaeli MB, Plott G, Russell CD, et al. In-
traperitoneal radioimmunotherapy of ovarian
cancer with lutetium-177-CC49. J Nucl Med
1996; 37:14911496.
156 Juweid M, Swayne LC, Sharkey RM, Dunn
R, Rubin AD, Herskovic T, et al. Prospects
of radioimmunotherapy in epithelial ovarian
cancer: results with iodine-131-labeled mur-
ine and humanized MN-14 anti-carcinoem-
bryonic antigen monoclonal antibodies. Gy-
necol Oncol 1997; 67:259271.
157 Rosenblum MG, Verschraegen CF, Murray
JL, Kudelka AP, Gano J, Cheung L, et al.
Phase I study of
90
Y-labeled B72.3 intraperi-
toneal administration in patients with ovar-
ian cancer: Effect of dose and EDTA coadmi-
nistration on pharmacokinetics and toxicity.
Clin Cancer Res 1999; 5:953961.
158 Alvarez RD, Huh WK, Khazaeli MB, Mere-
dith RF, Partridge EE, Kilgore LC, et al. A
phase I study of combined modality
90
Yt-
trium-CC49 intraperitoneal radioimmuno-
therapy for ovarian cancer. Clin Cancer Res
2002; 8:28062811.
159 Deb N, Goris M, Trisler K, Fowler S, Saal J,
Ning S, et al. Treatment of hormone-refrac-
tory prostate cancer with
90
Y-CYT-356 mono-
clonal antibody. Clin Cancer Res 1996;
2:12891297.
References 535
Abstract
Medical Enzymes core technology and
products are based on a proven approach
using amino acid-depleting enzymes as
therapeutic agents. This technique capital-
izes on the dependence of cancer cells on
particular amino acids relative to normal
cells. The use of amino acid-depleting en-
zymes provides a way to starve diseased
cells by efficiently reducing the concentra-
tion of the selected amino acid. The amino
acid-depleting strategy was demonstrated
with asparaginase (Elspar
, Merck), a che-
motherapeutic agent for leukemia. Joseph
Roberts pioneered the development of as-
paraginase, his team isolating the asparagi-
nase-producing bacterial organism and de-
veloped economically feasible methods for
producing and purifying the enzyme. These
accomplishments related to advancing the
field of therapeutic enzymes form the basis
of Medical Enzymes core technology.
Abbreviations
ALL acute lymphatic leukemia
DON 6-diazo-5-oxo-l-norleucine
GA glutamine-asparaginase
PEG-PGA PEGylated Pseudomonas 7A
Glutaminase-Asparaginase
PGA Pseudomonas 7A Glutami-
nase-Asparaginase
6.1
Rationale for GlutaDON
Therapy
GlutaDON is a combination therapy for
the treatment of cancer, and consists of
PEGylated glutaminase enzyme and the
glutamine analogue 6-diazo-5-oxo-l-norleu-
cine (DON). The rationale for using gluta-
mine antagonists in combination with the
enzyme glutaminase is based on the pre-
mise that the effectiveness of the antago-
nist will be enhanced when the available
pool of glutamine is depleted by the en-
zyme. Glutamine, the most abundant cir-
culating amino acid in the body, is the ma-
jor respiratory fuel for tumor cells [1, 2].
Tumor cells are avid glutamine consumers
due to their decreased expression of gluta-
mine synthetase and the need for gluta-
mine as a substrate in nucleotide and pro-
tein biosynthesis, for energy production,
and for the generation of key metabolic in-
termediates [3, 4]. In order to cope with
the demand for glutamine, tumor cells ex-
press highly efficient transporters to en-
sure that substrate availability does not be-
come rate-limiting [5]. Human hepatoma
cells, for example, transport glutamine at a
537
6
New Directions in Tumor Therapy
Amino Acid Depletion with GlutaDON
as Treatment for Cancer

ISBN: 3-527-31184-X
rate 10- to 20-fold faster than do normal
hepatocytes [6].
DON is an anti-tumor antibiotic isolated
from cultures of Streptomyces [7]. As a struc-
tural analogue of l-glutamine, DON func-
tions as a glutamine antagonist, interfering
with the utilization of that amino acid in
several key biochemical reactions [8].
Through this activity, DON inhibits DNA
replication and protein synthesis, resulting
in the inhibition of tumor growth [911].
DON has been shown to possess promising
anti-neoplastic activity against a variety of
animal tumors and human tumor xeno-
grafts (in nude mice), including colon,
breast, and lung carcinomas [12, 13].
DON, however, has limited potential when
used as a single agent in the treatment of
cancer in humans because of severe toxicity
that prevents dose escalation into the re-
quired therapeutic range [1416].
DON uptake by tumor cells is greatly re-
duced by the presence of even very low
concentrations of glutamine, indicating
that glutamine and DON are transported
into the cells through the same transport
system [17]. By depleting glutamine in the
bloodstream through the activity of a glu-
tamine-depleting enzyme, DON should be
more rapidly taken up by cells. Tumor
cells, which possess an enhanced transport
mechanism for glutamine, are expected to
transport DON to a greater extent than
normal cells. This increased efficiency of
DON uptake by tumor cells, allows lower
levels of the analogue to be used, resulting
in preferential inhibition of tumor growth
[18]. (For an example of this, refer to the
data in Table 6.3, where a four-fold lower
dose of DON when combined with gluta-
minase, was more effective in inhibiting
the growth of sarcoma 180 tumor with less
toxicity to the mice [18].)
Prior to the studies by Roberts and col-
leagues, the available forms of glutami-
nase were not suitable for use as therapeu-
tic agents. The Roberts laboratory isolated
new forms of glutaminase from soil bacte-
ria with the desired kinetic and molecular
characteristics for therapeutic value [20
22]. The most therapeutically effective of
these is PEGylated Pseudomonas 7A Gluta-
minase-Asparaginase (PEG-PGA). PEGylat-
ing this glutaminase increased the plasma
half-life in animals compared to the un-
PEGylated form, and shielded the enzyme
from host-mediated inactivation (see Sec-
tion 6.3; see also Part VI, Chapter 2).
Using PEG-PGH in combination with
DON, it is hoped that a better therapeutic
response will be achieved in patients with
cancer that is refractory to standard treat-
ment.
In preclinical studies utilizing human
lung, breast, colorectal, and ovarian tumor
xenografts growing in athymic nude mice,
this combination treatment showed prom-
ising anti-cancer activity, with minimal
host toxicity. In addition, numerous stud-
ies have shown that tumors do not develop
resistance to glutaminase treatment as
they do to most anti-cancer therapies. Pre-
clinical toxicology studies in animals were
completed in 2001, and a clinical dose-
finding multicenter study Phase I/IIa is
currently under way in Germany, compris-
ing three different trial steps. GlutaDON
targets lung, breast, ovarian, colorectal,

and prostate cancers, the predominant
cancers in the Western World.
6 New Directions in Tumor Therapy Amino Acid Depletion with GlutaDON

6.2
Preclinical Studies
6.2.1
Investigations of Anti-tumor,
In vitro Effects of PGA and DON
6.2.1.1 DON
Extensive studies have been performed on
the activity of DON, as a single agent,
against human tumor cell lines. These
data have been compiled by the National
Cancer Institute, and can be accessed via
their website at www.nci.nih.com under
the code NSC 7365 (Table 6.1).
6.2.1.2 PGA+DON
The ability of PGA to inhibit the growth of
ovarian cancer cells was examined in vitro,
in the presence and absence of DON. The
MA148 cell line is a primary ovarian can-
cer cell line established from a human pa-
tient. As documented in Table 6.2, this cell
line is highly sensitive to PGA alone, even
at low concentrations, but when PGA was
combined with DON the growth inhibition
was almost complete. Other established
ovarian cancer lines such as OVCAR3
were also found to be sensitive both to
PGA alone and PGA+DON therapy (Table
6.2; Fig. 6.1).
Table 6.1 Tumor cell line sensitivity to DON (National Cancer Institute. Code NSC 7365. www.nci.nih.gov).
Cell line Total tested Moderate sensitivity More sensitive Not sensitive
Leukemia 6 2 (33) 4 (66) 0
NSCLC 10 1 (10) 3 (30) 6 (60)
SCLC 2 1 (50) 1 (50) 0
Colon 9 2 (22) 7 (78) 0
CNS 8 4 (50) 4 (50) 0
Melanoma 9 2 (22) 4 (45) 3 (33)
Ovarian 6 3 (50) 1 (17) 2 (33)
Renal 9 4 (45) 2 (22) 3 (33)
Prostate 2 0 2 (100)
Breast 8 0 3 (38) 5 (62)
Values in parentheses are percentages.
Table 6.2 Efficacy of PGA and DON treatment towards ovarian
cells in tissue culture [19] (Sethuraman and Roberts, unpublished
data).
Ovarian cell line Treatment group Growth inhibition [%]
MA 148 PGA (0.001 IU mL
1
)
PGA (0.003 IU mL
1
)
PGA (0.009 IU mL
1
)
51
82
93
OVCAR3 PGA (0.01 IU mL
1
)
DON (1.5 lg mL
1
)
PGA+DON
83
52
97
MA 148 PGA (0.01 IU mL
1
)
DON (1.5 lg mL
1
)
PGA+DON
97
52
100
The ability of PGA to inhibit the growth
of prostate cancer cells was examined in vi-
tro in the presence and absence of DON.
Fig. 6.2 illustrates the synergistic effect of
PGA and DON, against prostate cancer
cells grown in tissue culture.
6.2.2
In vivo Effects of Glutaminase-Asparaginase
(GA) and DON Treatment
Sarcoma 180 tumor cells were injected into
BDF1 mice to study the effects of DON at
varying concentrations, compared with a
combination therapy of DON and GA. Sub-
sequently, a four-fold lower dose of DON,
when combined with GA, was more effec-
tive in inhibiting tumor growth and was
less toxic towards the mice [18] (Table 6.3).
Four murine tumors were tested in vivo
for sensitivity to treatment with GA and
DON, both alone and in combination. The
results indicated that combination therapy
is far more effective than either drug alone
[18] (Table 6.4).
Seven human tumors (LX-1 lung, MX-1
breast, and Redmond, CX-1, CX-5, CX-2,
and SK-CO-1 colon) were tested in vivo for
sensitivity to DON treatment alone, and in
combination with GA. The human tumors
were grown in athymic mice, and efficacy
of the therapy was determined by measur-
ing tumor diameter and recording animal
survival time [18] (Table 6.5).
Additional studies were conducted with
LX-1 tumors, the cell line being injected
into athymic mice. When the tumors were
palpable, the following treatment regimen
was started. Group 1 (control) received no
treatment, while group 2 received PEG-
PGA, group 3 received DON, and group 4
received PEG-PGA+DON. As seen in

Fig. 6.1 Efficacy of PGA and
DON treatment on OVCAR3
cells in tissue culture [19].
Fig. 6.2 Effectiveness of PGA
and DON treatment on PC-3
prostate cancer cells in tissue
culture [19].
Fig. 3, treatment with PEG-PGA+DON
was seen to be considerably more effective
at inhibiting tumor growth over a pro-
longed period than DON alone (Fig. 6.3)
[19].
In studies with MX-1 mammary adeno-
carcinoma, the cell line was injected into
athymic mice, and the following treatment
regimen was started when the tumors
were palpable. Group 1 (control) received
no treatment, while group 2 received
DON, group 3 received GA, and group 4
received GA+DON. The results showed
treatment with GA+DON to be more ef-
fective at reducing tumor growth than
DON alone (Fig. 6.4) [18].
6.3
PEGylation and Protection from Inactivation
As the biotechnology industry comes of
age, and protein-based therapeutics be-
come a practical reality, the importance of
protein-stabilizing technologies comes in-
creasingly to the fore. The therapeutic effi-
cacy of many of these biopharmaceuticals
is impeded by the hosts natural immune
defense system [23]. When a host encoun-
ters a foreign protein in its circulation, the
hosts immune system initiates an im-
mune response, which results in the pro-
duction of protein-inactivating antibodies
that clear the protein from the circulation.
When the immune system develops an
ability to inactivate the protein, the therapy
becomes ineffective. Hence, to counteract
Table 6.3 Inhibition of sarcoma 180 tumor cell growth
a)
by GA+DON [18].
Treatment
b)
Average weight change
at day 14 [g]
Inhibition of tumor growth
at day 21 [%]
Control +1.2 0
GA +0.8 0
DON (6 mg kg
1
) +0.4 0
DON (12 mg kg
1
) 1.9 38
DON (24 mg kg
1
) 4.3 51
GA+DON (6 mg kg
1
) 0.6 66
a) S180 cells injected subcutaneously into BDF mice. DON was
injected at 4 h after GA administration.
b) Treatment administered twice weekly for 2 weeks.
Table 6.4 Combination treatment of mouse tumors with GA and DON [18]
Tumor Increase in median survival time [%]
GA DON GA+DON
B16 Melanoma 10 10 94
L1210 Leukemia 0 64 220
P388 Leukemia 0 45 127
C1498 Myelogenous leukemia 19 56 112

Table 6.5 Effect of glutaminase-asparaginase (GA) and DON on human tumors in vivo [18].
Tumor Treatment group Mean tumor diameter [cm] Median survival time
[days]
Start of treatment 6 weeks
MX-1 Breast Control 0.34 2.7 49
GA+DON 0.33 0.16 189
GA 0.26 2.3 56
DON 0.34 0.71 140
LX-1 Lung Control 0.57 2.3 56
GA+DON 0.51 0.33 105
GA 0.53 1.6 45
DON 0.65 1.5 76
Redmond colon Control 0.41 3.3 41
GA+DON 0.53 3.2 46
GA 0.45 3.4 39
DON 0.33 2.9 47
CX-1 colon Control 0.81 3.3 45
GA+DON 0.76 3.2 52
GA 0.82 3.1 47
DON 0.81 3.1 47
CX-5 colon Control 0.49 1.9 NA
GA+DON 0.47 0.9 NA
DON 0.48 1.0 NA
CX-2 colon Control 4.6 20.6 67
GA+DON 5.0 8.3 127
GA 5.4 27.8 47
DON 4.7 18.8 73
SK-CO-1 Control 0.52 2.09 NA
Colon GA+DON 0.52 1.02 NA
NA, data not available.
Fig. 6.3 Effect of PEG-PGA+DON treatment on human LX-1 lung tumors in vivo [19].
antibody formation and host-induced inac-
tivation, the specific protein must some-
how be protected.
One strategy is to modify the proteins
with agents such as carbohydrates or bio-
compatible polymers (see Part VI, Chap-
ters 1 and 5), an example being that of
polyethylene glycol (PEGylation) (see Part
VI, Chapter 2). Previously, PEGylation has
been used successfully to shield a number
of protein drugs from the hosts immune
system [24, 25]. PEGylation, ideally, should
confer stealth properties, thereby making
therapeutic proteins undetectable to the
hosts immune system, as well as provid-
ing increased stability to proteins [24].
In practice, however, PEGylation often re-
duces the biological activity of the protein.
For each therapeutic protein it is necessary
to develop a specific PEGylation strategy to
shield it effectively from host-mediated in-
activation. Medical Enzymes AG has devel-
oped a strategy to monitor and evaluate
the effectiveness of PEGylation, and this
can be used to determine the optimal mod-
ifying agent and extent of modification for
individual biopharmaceuticals. The PEGyla-
tion strategy developed by Medical Enzymes
AG for PGA produces only a minimal loss
of enzyme activity.
An example of the clinical benefits of
PEGylation can be seen with asparaginase.
L-Asparaginase has shown success against
the infantile form of acute lymphatic leu-
kemia (ALL), increasing the overall cure
rate by 1030%. Unfortunately, this useful-
ness is limited by the development of hy-
persensitivity reactions, it having been re-
ported that 73% of newly diagnosed ALL
patients developed a hypersensitivity reac-
tion to E. coli L-asparaginase during main-
tenance therapy. The plasma half-life of L-
asparaginase is approximately 1 day;
hence, three injections are required each
week to maintain a sufficient depletion of
asparagine, but this increases both the risk
of developing hypersensitivity reactions
and the cost of treatment. Oncospar
TM
(PEGylated L-asparaginase) has an in-
creased half-life, such that the injection
frequency is reduced to 14 days, and the
incidence of hypersensitivity reactions is
also significantly reduced. Thus, the bene-
fits of PEGylation are clearly indicated [26].
Fig. 6.4 Treatment of MX-1 mammary adenocarcinoma with GA+DON [18].
In the case of unmodified Pseudomonas
7A Glutaminase-asparaginase (PGA), the
plasma half-life in normal BDF1 mice is
13 hours [22], but this is extended by PEG-
ylation to approximately 5 days in mice,
rats, and rabbits (Figs. 6.56.8). Thus, the
PEGylated drug requires much less-fre-
quent administration to the tumor-bearing
host than the unPEGylated drug.
To demonstrate the efficacy of PEGylation
in preventing enzyme inactivation, mice
were treated with twice-weekly doses of
PEG-PGA at 5 IU kg
1
or 15 IU kg
1
. Blood
samples were taken from the animals twice

Fig. 6.5 Half-life of unmodified PGA in mice.
Fig. 6.6 Half-life of PEG-PGA in Balb-C mice [19].
Fig. 6.7 Half-life of PEG-PGA in Sprague-Dawley rats [19].
each week, at 4 hours after the injection.
The enzyme recovery remained good after
more than 40 days of therapy, indicating
that the enzyme was protected against
host-mediated inactivation (Fig. 6.9).
6.4
Toxicology
Toxicology studies of PEG-PGA and DON
were performed under contract by the Labo-
ratory of Pharmacology and Toxicology
(LPT) in Hamburg, Germany. The studies
were carried out using current production
lots of PEG-PGA and DON, and included
acute toxicity studies of PEG-PGA alone,
DON alone, and a combination of both in
Sprague-Dawley rats. PEG-PGA studies in
rats and rabbits were also conducted to deter-
mine the dose which would provide an opti-
mal depletion of glutamine. In addition, six-
week subchronic studies were conducted in
rats and rabbits using a combination of PEG-
PGA (at the optimal dose) and DON. Further
toxicology and pharmacokinetics studies
were performed in mice, rats, and rabbits
at the University of South Carolina, USA.
6.5
Clinical Trial
A multicenter, Phase IIIa trial is currently
under way in Germany (principal investi-
gator Prof. Dr. Clemens Unger, Clinic for
Tumor Biology in Freiburg, Germany).
The trial comprises three steps in patients
with advanced solid tumors that are refrac-
tory to standard treatment:
Fig. 6.8 Half-life of PEG-PGA in New Zealand White rabbits [19].
Fig. 6.9 Comparison of plasma enzyme activity in mice
treated twice weekly with various doses of PEG-PGA.
1. Dose-escalation of PEG-PGA
2. Dose-escalation of DON with optimized
PEG-PGA-dose
3. Treatment with optimized PEG-PGA-
and DON-doses
Currently, Step 1 has been completed,
with a PEG-PGA dose level of 120 IU m
2
showing a sufficient level of glutamine de-
pletion in all patients. However, this study
is ongoing, with the second part using
PEG-PGA at 120 IU m
2
combined with
DON. The first positive results relating to
efficacy and tolerability have proved en-
couraging [27].
6.6
PGA and DON have complementary
mechanisms of action against tumor cells,
with both affecting the ability of the cells
to replicate by interfering with glutamine
usage. PGA depletes glutamine from the
bloodstream, while DON, as an antimeta-
bolite of glutamine, inhibits its use in re-
actions. This synergistic effect has been
studied through in vitro and in vivo studies
of tumor cell lines. As monotherapy, both
drugs were found to reduce tumor volume
in several cancers; however, based on find-
ings in animal models, a combination
therapy is not only safer but also more ef-
fective in long-term treatment.
References
1 Bergstrom J, Furst P, Noree LO, Vinnars E.
Intracellular free amino acid concentration in
human muscle tissue. J. Appl. Physiol. 36:
693699, 1974.
2 Kovacevic Z, Morris HP. The role of gluta-
mine in the oxidative metabolism of malig-
nant cells. Cancer Res. 32: 326335, 1972.
3 Matsuno T, Hirai H. Glutamine synthetase
and glutaminase activities in various hepato-
ma cells. Biochemistry Int. 19(2): 219225,
1989.
4 Souba WW. Glutamine and cancer. Ann. Surg.
218(6): 715728, 1993.
5 Medina MA, Sanchez-Jimenez F, Marquez J,
Rodriguez Quesada A, Nunez de Castro I.
Relevance of glutamine metabolism to tumor
cell growth. Mol. Cell. Biochem. 113: 115,
1992.
6 Bode BP, Kaminski DL, Souba WW, Li AP.
Glutamine transport in isolated human hepa-
tocytes and transformed liver cells. Hepatology
21: 511520, 1995.
7 Dion HW, Fusari SA, Jakubowski ZL, Zora
JG, Bartz QR. 6-diazo-5-oxo-L-norleucine, a
new tumor inhibitory substance I. Isolation
and characterization. J. Am. Chem. Soc. 78:
30753077, 1956.
8 Kisner DL, Catane R, Muggia FM. The redis-
covery of DON (6-diazo-5-oxo-L-norleucine).
Recent results. Cancer Res 74: 258263, 1980.
9 Levenberg B, Melnick O, Buchanon JM. Bio-
synthesis of the purines. XV. The effect of
aza-L-serine and 6-diazo-5-oxo-L-norleucine on
the inosinic acid biosynthesis de novo. J. Biol.
Chem. 225: 163176, 1957.
10 Eidinoff ML, Knoll JE, Marano B, Cheong L.
Pyrimidine Studies. I: Effect of DON (6-diazo-
5-oxo-L-norleucine) on incorporation of precur-
sors into nucleic acid pyrimidines. Cancer Res.
18: 105109, 1958.
11 Rosenbluth RJ, Cooney DA, Jayaram HN,
Milman HA, Homan ER. DON, CONV, and
DONV II. Inhibition of L-asparaginase
synthetase in vivo. Biochem. Pharmacol. 25:
18511858, 1976.
12 Ovejera AA, Houchens DP, Catane R, Sheri-
dan MA, Muggia FM. Efficacy of 6-diazo-5-
oxo-L-norleucine and N-[N-gamma-glutamyl-6-
diazo-5-oxo-norleucinyl}-6-diazo-5-oxo-norleu-
cine against experimental tumors in conven-
tional and nude mice. Cancer Res. 39(8):
32203224, 1979.
13 Houchens DP, Ovejera AA, Sheridan MA,
Johnson RK, Bogden AE, Neil GL. Therapy
for mouse tumors and human tumor xeno-
grafts with the antitumor antibiotic AT-125.
Cancer Treat. Rep. 63(3): 473476, 1979.
14 Sklaroff R, Casper E, Magill G, Young C.
Phase I study of 6-diazo-5-oxo-L-norleucine

(DON). Cancer Treat. Rep. 64(12): 12471251,
1980.
15 Weiss GR, McGovern JP, Schade D, Kufe DW.
Phase I and pharmacological study of acivicin
by 24-hour continuous infusion. Cancer Res.
42(9): 38923895, 1982.
16 Le Page GA, Loo TL. In: Cancer Medicine
(JF Holland, E Frei, III, Eds.), Lea & Febiger,
Philadelphia, pp. 754756, 1973.
17 Huber KR, Rosenfeld H, Roberts J. Uptake of
glutamine antimetabolites 6-diazo-5-oxo-L-nor-
leucine (DON) and acivicin in sensitive and
resistant tumor cell lines. Int. J. Cancer 41:
752755, 1988.
18 Roberts J (unpublished data).
19 Sethuraman N, Roberts J (unpublished re-
sults).
20 Roberts J, Holcenberg JS, Dolowy WC. Anti-
neoplastic activity of highly purified bacterial
glutaminases. Nature 227(263): 11361137,
1970.
21 Roberts J, Holcenberg JS, Dolowy WC. Isola-
tion, crystallization, and properties of Achro-
mobacteriaceae glutaminase-asparaginase with
antitumor activity. J. Biol. Chem. 247(1): 84
90, 1972.
22 Roberts J. Purification and properties of a
highly potent antitumor glutaminase-asparagi-
nase from Pseudomonas 7A. J. Biol. Chem. 251:
21192123, 1976.
23 Roberts J, McGregor WG. Inhibition of mouse
retroviral disease by bioactive glutaminase as-
paraginase. J. Gen. Virol. 72: 299305, 1991.
24 Nucci ML, Short R, Abuchowski A. Advanced
Drug Delivery Rev. 6: 133151, 1991.
25 Zalipsky S. Advanced Drug Deliv. Rev. 16: 157
182, 1995.
26 Enzon. Summary Basis of Approval 4/15/93.
27 Unger C, Baas F, Wiessner S et al. Phase I
dose escalating study of PEG-PGA and DON:
A new amino acid depleting anticancer drug
approach. ASCO 2003.
References 547
Abstract
The human immunodeficiency virus type 1
(HIV-1) is the causative agent of AIDS. HIV-
1 is a member of the retrovirus family, and
belongs to the lentivirus genus. Due to the
complexity of the HIV-1 infection and the
difficulty of finding a safe and efficient ther-
apy, a major effort has been made during
the past two decades to develop gene thera-
py approaches aimed at inhibiting viral re-
plication and making cells resistant to the
virus, or eliminating the infected cells.
Many different approaches have been made
that can be classified into RNA-based strate-
gies (using ribozymes, RNA decoys, anti-
sense mRNAs or siRNAs), or protein-based
strategies (using dominant negative viral
proteins, intracellular antibodies or suicide
genes). At present, virus replication is quite
efficiently blocked by conventional therapy
(HAART). However, the existence of long-
lasting, latently HIV-1 infected cells in the
patient does not allow eradication of the
virus. In fact, the viral reservoirs of latently
infected cells are not affected by HAART,
and have become the most problematic area
in HIV-1 therapy. For this reason, we have
developed a gene therapy strategy that is
not aimed at inhibiting viral replication,
but is designed to destroy the viral reser-
voirs. The latently infected resting T cells
are probably the largest reservoir, and the
one that can persist for the whole life of
the patient. We have constructed a lentiviral
vector designed to destroy latently infected
cells with silent viruses. The vector has been
designed in such a way that any alteration
of normal cell functions is prevented, with
the objective that uninfected transduced
cells remain unaffected and fully func-
tional. In summary, the genetic construc-
tion contains an externally inducible system
that promotes the expression of any latent
HIV-1 provirus without affecting the cell cy-
cle state by the expression of the potent viral
transactivation TATprotein. A second genet-
ic system included in the vector allows the
549
ISBN: 3-527-31184-X
7
AIDS Gene Therapy:
A Vector Able to Selectively Destroy Latently HIV-1-infected Cells
Mundus Vult Decipi
High Mutation Rates of HIV and New Paradigms for Treatment
expression of a suicide gene in response to
the presence of the essential viral REV pro-
tein, which in turn induces a quick death of
the cell by apoptosis. This second system
permits the destruction of any HIV-1 in-
fected cell. The vector can be packaged into
HIV-1 viral particles to deliver the genetic
construction in any cell susceptible to
HIV-1. At present, the most important lim-
itation of this approach is that, despite re-
cent improvements in lentiviral packaging
systems, it remains difficult to obtain titers
of packaged vectors high enough to carry
out an efficient in vivo gene therapy.
Abbreviations
AIDS acquired immunodeficiency
syndrome
AP-1 activator protein 1
CCR3 chemokine receptor type 3
CMV cytomegalovirus
CNS central nervous system
crHIV conditionally replicating HIV-1
vectors
CRM1 host nuclear export protein re-
quired for REV activity. Nuclear
export of late HIV-1 mRNAs oc-
curs via a cellular protein export
pathway
env HIV-1 gene coding for the En-
velope surface and transmem-
brane proteins
Ets-1 ETS-1 transcription factor
FIV feline immunodeficiency virus
gag HIV-1 gene coding for the cap-
sid and matrix structural pro-
teins
HAART highly active antiretroviral ther-
apy
HSC hematopoietic stem cells
IPTG isopropyl-b-d-thiogalactopyrano-
side
LBP-1 leader binding protein 1
LEF lymphoid enhancer binding fac-
tor
LTR long-terminal-repeat
nef negative factor. HIV-1 gene with
pleiotroc effects. Down-regu-
lates CD4 and major histocom-
patibility class I molecules. It
stimulates viral infectivity and
modulates cellular activation
pathways
NF-AT nuclear factor of activated T-
cells
NFjB nuclear factor jB
PBS primer binding site
pol HIV-1 gene coding for the re-
verse transcriptase, integrase
and protease enzymes
REV HIV-1 gene involved in the reg-
ulation of the viral expression
RRE REV response element
RSV Rous sarcoma virus
SA splicing acceptor sequence
SD splicing donor sequence
SIV simian immunodeficiency virus
SP1 SP1 transcription factor
TAR TAT response element
TAT HIV-1 gene coding for the
transactivator
USF-1 upstream stimulatory factor 1
vif viral infectivity factor. HIV-1
gene that increases viral trans-
mission, it helps in virion as-
sembly
vpr viral protein R. HIV-1 gene that
arrest cell cycle at G2 phase
and facilitates the transport to
the nucleus of the viral pre-inte-
gration complex
vpu viral protein U. HIV-1 gene that
enhances the release of viral
particles and promotes degrada-
tion of CD4
VSV-G vesicular stomatitis virus G-pro-
tein
7 AIDS Gene Therapy: A Vector Able to Selectively Destroy Latently HIV-1-infected Cells 550
7.1
The Genes and Life Cycle of HIV-1
The human immunodeficiency virus type
1 (HIV-1) is the causative agent of AIDS.
This virus is currently spread worldwide
and affects millions of individuals; the
present situation is especially critical in
sub-Saharan countries, where the virus af-
fects more than 30 million people. The
impact of this virus in health and social
relationships has been vast in many coun-
tries worldwide since it was discovered in
the early 1980s. HIV-1 is a member of the
retrovirus family, and belongs to the lenti-
virus genus, which shows remarkable dif-
ferences from other retroviruses such as
the oncoviruses. The HIV-1 genome is
comprised of nine genes (Fig. 7.1); three
of them are common to all retroviruses
(gag, pol, and env), and six are specific to
HIV-1 and related lentiviruses (tat, rev,
nef, vif, vpr, and vpu). Viral particles con-
tain two genomic single-stranded RNAs
that are reverse-transcribed shortly after in-
fection of susceptible cells. The double-
stranded cDNA once inside the nucleus of
the cell is integrated into chromosomal
DNA. At this stage, it becomes similar to
a cellular gene and is called a provirus.
The provirus is expressed in response to
host factors and normally leads to the pro-
duction of new viral particles and to cell
death (Fig. 7.2).
The gag and env genes encode for the
viral structural proteins. The gag gene en-
codes for a precursor protein Pr55, which
is further processed by proteolytic cleavage
to give the p17 matrix protein, p24 capsid
protein and the p9 and p7 nucleocapsid
proteins. The processed products of this
gene form the viral nuclear particles where
the genomic RNA is packaged. The env
gene encodes for a precursor (Pr160) of
the envelope glycoproteins gp120 and
gp40. The pol gene encodes for the pro-
tease, reverse transcriptase and integrase
enzymes. The pol gene is expressed in a
gag-pol fusion product made by ribosomal
frameshifting, giving a long precursor pro-
tein that is processed by the protease activ-
ity into the final protein products. The rate
at which the frameshifting occurs is a type
of regulation that ensures the correct rate
between the enzymatic Pol proteins and
the structural Gag proteins. The vif, vpr,
vpu, and nef genes encode for proteins that
have been termed accessory or auxili-
Fig. 7.1 The genomic structure of the HIV-1 pro-
virus comprises six accessory genes (tat, rev, nef,
vpu, vpr and vif ) and three genes common to all
retroviruses gag, pol, and env that produce the
proteins. MA=matrix; CA=capsid; NC=nucleo-
capsid; PR=protease; RT=reverse transcriptase;
IN=integrase; SU=surface glycoprotein;
TM=transmembrane glycoprotein.
ary, to reflect the fact that they are needed
for virus replication in some cellular types.
However, these genes are essential to pro-
duce an effective infection in the patient.
Especially relevant for the regulation of
transcription of the proviral genes are the
5'-long-terminal-repeat (5'-LTR) promoter
and the regulatory proteins Tat and Rev
(see Fig. 7.2). The viral genes are ex-
pressed from the 5'-LTR promoter, which
contains binding sites for a number of cel-
lular transcription factors involved in T-cell
activity such as SP1, NFjB, AP-1 and NF-
AT [15] (Fig. 7.3). For this reason, the pro-
virus is expressed in response to cellular
activation [68]. The integrated provirus is
expressed in a full-length RNA that con-
tains several splicing donors and acceptor
sites. Fully spliced RNA encodes three pro-
teins that accumulate in the cell nucleus,
namely Tat, Rev, and Nef. Meanwhile, Nef
is involved in the increase of T-cell activa-
tion state [913], Tat is needed to enhance
the transcription rate by promoting com-
pletion of initiated transcriptional activity.
Tat together with host factors binds a 5'
stem-loop of the nascent RNA (TAR) that
stabilizes the RNA polymerase II complex
by hyperphosphorylation [14, 15]. Rev pro-
tein is accumulated until a threshold is
reached, then multiple Rev proteins bind a
sequence present inside the env gene
(RRE) and at the same time associates
with the host nuclear export protein
CRM1 [1618]. Rev is therefore involved in
nuclear export of unspliced or single
spliced transcripts, and in this way it al-
lows a change in the pattern of gene ex-
pression from early (tat, rev, and nef) to
late genes (gag, pol, env, vif, vpu, and
Fig. 7.2 The integrated provirus is expressed in a
full-length RNA from the 5'-LTR promoter that
contains several splicing donor and acceptor sites.
Fully spliced RNA encodes three proteins that ac-
cumulate in the cell nucleus, Tat, Rev, and Nef. A:
Tat enhances the transcription rate; B: once the
Rev protein is accumulated, multiple Rev proteins
bind the RRE sequence; C: Rev activity produces
the nuclear export of unspliced or single spliced
transcripts, and the gag, pol, env, vif, vpu and vpr
proteins are synthesized.
vpr) [1921]. There is also much evidence
showing that HIV-1 can be transcribed
when it is present in the cell in a non-inte-
grated reverse-transcribed form; in this
case, only Tat, Rev, and Nef proteins have
been detected [8, 2228].
Most retroviruses can only infect cells
that are dividing. However, lentiviruses
can infect non-dividing cells, probably be-
cause they have the capacity to transport
the reverse-transcribed, double-stranded
DNA to the nucleus. However, resting cells
do not allow the integration of the viral
DNA, which will remain in the nucleus in
a label form for a variable time, expressing
only early genes, as mentioned previously.
7.2
Gene Therapy of AIDS
7.2.1
Main Strategies
Retroviral or lentiviral vectors are preferen-
tially used for HIV-1 gene therapy studies.
Two main strategies for gene therapy of
HIV-1 infection have been used:
Elimination of cells infected by the virus.
Making the cells resistant to the virus or
to its pathogenic effects.
A third approach is aimed at making the
infection tolerable for the patient.
7.2.1.1 Elimination of Cells Infected
by the Virus
This approach is based on the use of suicide
vectors that express a toxic gene in response
to viral regulatory proteins such as Tat and/
Fig. 7.3 The 5' long terminal repeat (5'-LTR) promoter has cis elements for Tat binding
(TAR), and for some host transcription factors (USF-1, Ets-1, LEF, NF-jB, SP1, TATA
box, LBP and AP1) that causes the virus expression to be activated in response to cellu-
lar transcription factors.
or Rev [2933]. This strategy has the advan-
tage that the target is the infected cell, and
not the virus. In this way, the problem of
the virus high mutation rate can be over-
come, which in turn is responsible for the
emergence of resistant forms found in ther-
apy schemes designed against viral targets.
However, this strategy presents two limita-
tions. One limitation is the fact that trans-
duced cells are eliminated if infected by
the HIV-1 virus, and their number will de-
crease after some time, and only non-trans-
duced cells will remain in the patient.
Therefore, this approach would be most in-
teresting if high in vivo transduction effi-
ciency was obtained and a huge number
of cells were transduced. In this case, the
virus will infect with high frequency trans-
duced cells, and it will be eliminated with-
out replicating. In addition, the fact that
about 99% of the target cells are in a resting
G
0
state means that in most cells the vector
will not be able to become integrated in a
host chromosome, and as a consequence
will be eliminated in a few days. The second
limitation that the vector that contains a
lethal gene should not compromise cell vi-
ability and function at all.
7.2.1.2 Production of Cells Resistant to the
Virus or to its Pathogenic Effects
This approach has received more attention
because the protected cells have a selective
advantage over the non-transduced ones.
Therefore, the requirement for a very high
efficient transduction rate is less critical.
Protected T cells should be used to repopu-
late the immune system with virus-resistant
cells, and gene transfer to hematopoietic
stem cells (HSC) is a promising approach
to achieve this goal [34] (see also Part II,
Chapter 8). However, it is still unclear
whether the repopulated immune system
would be able to generate a diversity of anti-
gen-binding receptors high enough to con-
fer a real effective immunity to the patient.
7.2.1.3 Reduction of HIV-1 Virulence
A third interesting approach consists of
the use of conditionally replicating HIV-1
vectors (crHIV-1) [35, 36]. The crHIV-1 vec-
tors will act as a parasite for the wild-type
HIV-1 virus. These vectors contain all the
cis-acting sequences necessary for packa-
ging, but do not have the trans-acting ele-
ments. Instead, they carry an antiviral
gene that inhibits any or various HIV-1
functions. In fact, the inhibition can affect
not only a viral function, but also a cellular
function required for viral replication, and
in this way the crHIV-1 is less susceptible
to mutational escape by HIV-1. Transduced
cells, when infected with a wild HIV-1, will
produce a low HIV-1 progeny, and many
of the viral particles will contain the
crHIV-1 vector instead of the wild HIV-1.
In this way, the replication rate of the
virus is strongly diminished and the vector
is spread. The aim of this approach is not
to eliminate the virus but to reduce its rep-
lication in order to make it less aggressive
and more tolerable by the host. A theoreti-
cal study shows the feasibility of this strat-
egy [37], but there are still some possible
in vivo limitations to this approach that
need to be addressed for example, the
fact that recombination and mutational
events may modify the crHIV-1 vector to
an inactive form; in contrast, aggressive
forms of the vector might be generated.
7.2.2
Anti-viral Genes
According to the main strategies of anti-
HIV gene therapy, two types of anti-viral
genes are used. First, toxic genes that pro-
duce a lethal effect on infected cells are
used to destroy the provirus population.
Second, genes that inhibit the viral life cy-
cle are used to protect the cells from the
infection.
7.2.2.1 Toxic Genes
The elimination of the infected cells is per-
formed using suicide vectors that express
a toxic gene in response to viral regulatory
proteins such as Tat or Rev. Virus spread
inhibition and selective killing of infected
cells was obtained in vitro with the herpes
simplex virus type 1 thymidine kinase
gene, which induces acyclovir and ganci-
clovir sensitivity [38, 39], and with the
diphtheria toxin A chain gene [2933], the
interferon a2 gene [31] and the cytosine
deaminase gene [31]. Induction of apopto-
sis by the over-expression of P53 in re-
sponse to the presence of HIV-1 provirus
produces a very efficient and specific de-
struction of infected cells (R. Oya, L. Sani-
ger and F. Luque, submitted). Although
this type of genes and vectors has received
little attention in recent years, it should be
considered that if all infected cells were to
be eliminated, the patient would immedi-
ately be cured. For that reason and
although we are still far from that possibil-
ity, due to technical limitations it is
worth making an effort in that direction.
7.2.2.2 Genes that Inhibit the Viral Cycle
Any step of the viral life cycle can be the tar-
get for intervention using anti-viral genes.
The discovery of co-receptors that are essen-
tial for the viral entry in the cell, but not for
cell function, has opened the possibility of
preventing the viral entry into the cell.
These are ideal targets because they prevent
the infection of modified cells. CCR5 is the
co-receptor that is used by most viral strains
[40, 41], while CXCR4 is another co-receptor
used by some viral strains [42]. Using the
expression of ribozymes, intracellular single
chain antibodies, intrakines and siRNA, the
expression of both co-receptors has been
successfully blocked [4349] (see also Part
II, Chapter 8). One concern with down-reg-
ulating a co-receptor is that escape mutants
might switch the virus tropism to a differ-
ent co-receptor such as CCR3, and subse-
quently infect the cells freely. Viral entry
has been also achieved by the expression
of modified peptides derived from the Env
protein gp41, which interferes with the
membrane [50]. Once the virus has en-
tered the cell, the viral RNA is reverse-
transcribed and the cDNA integrated into
a cellular chromosome. Both steps re-
verse transcription and integration have
been blocked by the use of intracellular
monoclonal antibodies [5153]. If the pro-
virus is already present in an integrated
form it can be also the target of anti-viral
genes. The provirus gene expression has
been the focus of a number of different ex-
perimental approaches to avoid the effec-
tive expression of the inserted provirus.
The Tat function is the preferred target to
inhibit the provirus gene expression be-
cause the Tat protein produces multiple
adverse effects [5463]. To obtain inhibi-
tion of the Tat activity, different anti-viral
genes have been used against the Tat pro-
tein itself, or directed to suppress its inter-
action with the TAR element. Anti-Tat
genes used are transdominant negative Tat
mutants [64, 65], anti-Tat intracellular anti-
bodies [66, 67], and anti-hCyclinT1 intra-
bodies that prevent the Tat interaction with
the positive transcription elongation factor
b [68] essential for the Tat activity. To sup-
press the Tat interaction with the TAR ele-
ment RNA decoys [6972], TAR antisenses
[73, 74] or ribozymes-anti-TAR [75] have
been used. Ribozymes have also been di-
rected against the 5' of the transcripts cor-
responding to the LTR U5 region to de-
grade the viral transcripts, and in this way
to suppress transcription of the provirus
[35, 7678]. The Rev activity has also been
inhibited to prevent the nuclear export of
unspliced or partially spliced transcripts.
Thus, expression of the genes that are pro-
cessed in the multiple spliced mRNA is
prevented, and many of their toxic effects
avoided, although there remains the prob-
lem of side effects produced by the Tat
protein. To obtain an efficient inhibition of
the Rev function, transdominant negative
Rev mutants have been broadly used [74,
7986] as well as RRE decoys RNAs [83],
and siRNA [34, 8688] (see also Part I,
Chapter 9 and Part II, Chapter 8). Finally,
structural genes have also been the target
to prevent virion formation. In this way,
transdominant Gag mutants [89, 90] and
intrabodies anti-Env [67] have been used,
as well as the a1 antitrypsin gene which
inhibits cellular serine proteases as well as
the viral protease, which produces a potent
inhibition of HIV-1 replication in trans-
duced lymphocytes [91]. The use of genes
that interfere with the latest steps in the
viral cycle has the disadvantage that the cy-
totoxic effects of the other viral proteins
are not prevented. However, the main lim-
itation of the anti-viral genes is the emer-
gence of resistant mutant strains. The de-
velopment of vectors that contain two anti-
viral genes is perhaps a good way of delay-
ing the onset of resistant strains; thus, a
combination of a transdominant negative
Rev and an antisense env can inhibit HIV-
1 replication quite efficiently [80].
7.2.3
Vectors Used in HIV-1 Gene Therapy
Viral vectors are preferred to deliver anti-
viral genes into target cells in the HIV-1
gene therapy, and among these, lentiviral
vectors are preferentially used. Retro-
viruses are in general quite good vectors
for the stable transduction of target cells,
as lentiviral vectors have the advantage of
maintaining the ability of lentiviruses to
infect both dividing and non-dividing cells,
while oncoretroviruses can only infect di-
viding cells [92]. Feline immunodeficiency
virus (FIV), HIV-1, HIV-2 and simian im-
munodeficiency virus (SIV) -derived vec-
tors also have the advantage that they can
be directed very specifically towards the
same target cells as the wild-type virus as
T cells and dendritic cells, as well as pre-
cursor cells HSC. In order to make the
use of lentiviral vectors safer, self-inactivat-
ing vectors have been designed to avoid
the uncontrolled spread of the vector and
the 3'-LTR-driven expression of chromo-
somal genes downstream of the insertion
[9399]. The safety of the self-inactivating
vectors can be improved by mutation of
the primer binding site (PBS) and primer
complementation in the packaging cells
[100]. However, conditionally replicating
vectors that can be mobilized by the wild
HIV-1 virus are also used. In this case, the
vector is designed to carry anti-viral genes
that protect the cell from the virus. In ad-
dition, the vector RNA can compete with
the viral RNA during the packaging. In
this way, a stronger inhibition of the viral
spread is obtained, while new cells can be
simultaneously transduced by the vector
and protected against the virus infection.
One difficulty in the therapy of HIV-1
comes from the fact that some viral strains
show a tropism towards central nervous
system (CNS) cells. Most anti-retrovirals
cannot reach CNS cells, which represent a
viral reservoir that escapes therapy. In or-
der to provide an efficient anti-HIV-1 gene
therapy strategy, the delivery system
should be able to reach the CNS cells, and
this is one reason why the vesicular stoma-
titis virus G-protein (VSV-G) is used to
pseudotype the vectors. This envelope pro-
tein broadens the spectrum of target cells
(including CNS cells), and also yields
higher vector titers, as well as providing
greater stability to the packaged vector vir-
al particles [101]. VSV-G has been success-
fully used to obtain the delivery of genes
by pseudotyped lentiviral vectors into the
CNS [102107].
SV40-based vectors are of great interest
because they can be very efficiently pack-
aged and can transduce non-dividing cells,
including CNS cells [108110]. This type
of vector has been used to confer efficient
protection of neurons from HIV-1 with the
delivery of anti-HIV-1 transgenes [79].
However, SV40-based vectors can trans-
duce and modify genetically many cells
that are not HIV-1 targets, which it would
be preferable to avoid. Lentiviral vectors
are still less efficiently packaged than
other viral systems, and need to be im-
proved before being able to proceed with
an in vivo gene therapy. It should be noted
that some relevant improvements have re-
cently been made, and an efficient large-
scale production and concentration of
HIV-1 derived lentiviral vectors has been
obtained [111]. Thus, it is conceivable that
future technical improvements will render
these vectors ideal for efficient gene thera-
py both in AIDS and other illnesses.
Gene therapy of an illness such as HIV-
1/AIDS must deal with the problem repre-
sented by the existence of millions of cells
spread throughout different organs and
tissues of the patient, all of which must be
protected from the virus. The reconstruc-
tion of an efficient immune function from
a subpopulation of cells is a task that is
likely to be possible only if the protected
cells are HSC. Attempts are currently
being made to create an in vitro genetic
modification of a subpopulation of HSC
that might later engraft in the bone mar-
row to produce a wide progeny of immu-
nity cells, and this experimental approach
deserves special mention. Lentiviral de-
rived vectors have been shown able to
transduce HSC, and such modified HSC
are capable of engrafting and differentiat-
ing into multiple hematopoietic lineages
[112115]. Transgene silencing tends to oc-
cur in undifferentiated cells that pass
through a long differentiation process.
This phenomenon is commonly observed
in HSC, where transgenes are frequently
shut off during the lineage differentiation
[116120]. However, lentiviral vectors ap-
pear not to be affected by silencing [121],
and efficient gene transfer into rhesus re-
populating HSC has been achieved using
a SIV-based lentiviral vector [122]. This
lack of gene silencing of transgenes deliv-
ered with lentiviral vectors might be re-
lated to the finding that HIV-1 seems to
integrate preferentially into active genes
[123]. Whether this is the reason, or not,
lentiviral vectors are currently the vectors
of choice for delivering genes into undif-
ferentiated embryonic or adult cells.
7.3
Viral Latency: the Real Challenge
Viral replication is often quite efficiently
reduced by HAART to undetectable levels
[124132], and can greatly prolong the
time to progression to AIDS [133]. How-
ever, viral sanctuaries of replication-compe-
tent HIV-1 persist in the patient, prevent-
ing eradication of the virus [124, 128, 129].
Latently infected resting CD4
+
T cells re-
present the most important sanctuary be-
cause they can persist for the whole life of
the patient [127, 130, 134, 135]. In fact, in
most treated patients and despite several
years of efficient HAART treatment the
latently infected resting cells sanctuary
does not decline, or does so very slowly
[127, 130, 131]. Intensification of HAART
therapy has recently permitted an increase
in the rate of decay of the latent reservoir,
but still does not eliminate ongoing virus
replication [135].
The latently infected resting CD4
+
lym-
phocytes contain an integrated provirus
that is silent, but that can become produc-
tive at any moment if the cell is activated
by interaction with a specific antigen (Ag).
Activated lymphocytes form a small part of
the lymphocyte pool at any moment; thus,
most lymphocytes are in a resting state.
About half of the resting lymphocytes are
nave cells, and the other half are resting
memory cells. The resting CD4
+
lympho-
cytes have a long life span, and in addition
the nave population is constantly replen-
ished with newly generated T cells [136],
while memory cells are maintained by pro-
liferative renewal that is Ag-independent.
It is thought that the HIV-1 infection of
an activated cell when it is in the process
of becoming a resting memory cell may, at
some stages, allow provirus integration,
but not viral gene expression at levels high
enough to proceed with the viral life cycle
[137, 138]. In this way, the infected lym-
phocyte would become a latently infected
resting cell, which could be maintained for
decades by the cells long life span and the
Ag-independent proliferative renewal.
It is unclear as to whether the latently
infected resting reservoir is important in
the strong viral load re-emergence that fol-
lows the cessation of HAART therapy.
However, it is clear that the lifetime of
these infected resting cells may prevent
the eradication of the HIV-1 from the pa-
tient [127]. At the same time, the efficacy
of prolonged HAART therapy is compro-
mised by the side effects of the drugs, and
by the emergence of resistant viruses.
Two forms of HIV-1 latency can be
found in resting cells: a labile preintegra-
tion form; and a stable, post-integration
form [125, 126, 132, 139142]. Whilst the
labile form decays rapidly after initiation
of the HAART therapy, the stable form re-
mains almost unchanged [132, 139]. One
important question to be addressed here is
whether the stable integrated latent pro-
viruses are able to produce infectious
viruses or, in contrast, whether they can-
not be activated to produce replication-
competent viruses since they are defective
or are silenced by epigenetic mechanisms
such as CpG methylation affecting the
promoter [143] or local chromatin struc-
ture [144, 145]. The activation of highly
purified populations of resting CD4
+
lym-
phocytes from patients undergoing long-
term HAART therapy reveals that about
1% of integrated proviruses can produce
high levels of mRNA after cellular activa-
tion, and that a subset of them can pro-
duce infectious viruses [146]. This small
proportion of replication-competent latent
proviruses is enough to allow the recovery
of replication-competent viruses from all
patients after cellular activation of resting
CD4
+
lymphocytes [146].
As cellular activation induces viral ex-
pression, and causes the former latent
viruses to become exposed to HAART ther-
apy, several efforts have been made to acti-
vate in vivo the silent proviruses by activat-
ing the resting CD4
+
T cells. However, the
adverse effects and poor clinical benefit
found indicate that this approach is not of
value [147, 148]. Prostratine and DPP are
not tumor-promoting phorbol esters, but
are under examination because they can
activate the provirus without complete acti-
vation of the cell [149, 150]. However, the
elimination of cells infected with latent
proviruses is not yet sufficiently efficient,
and much needs to be done to address the
possible side effects of these agents.
Hence, this approach is still in its very
early stages for evaluating the real possibil-
ity of eliminating latently infected cells.
Although latently infected cells are
highly relevant in achieving eradication of
the virus from the patient, no gene thera-
py approaches have previously been tar-
geted against these cells. Gene therapy
approaches are generally performed inde-
pendently of the achievements and limita-
tions of the current HAART therapies. In
the following section we describe a prelim-
inary scheme that aims to complement
HAART therapies with a gene therapy
strategy designed to destroy the viral reser-
voir of latently infected resting cells.
7.4
A Vector Able Selectively to Destroy Latently
Infected Cells
As mentioned above, current gene therapy
strategies run in parallel with conventional
anti-retroviral drug treatments, mainly be-
cause HAART therapies cannot eradicate
the virus from the patient, and hence an
alternative is required. The other possibili-
ty is to exploit the successes of HAART in
inhibiting virus replication, and to try to
overcome the limitations of that approach.
Following this rationale, a HIV-1-derived
lentiviral vector was designed to destroy
the viral sanctuaries, and especially the la-
tently infected cell reservoir. This vector is
capable of destroying any cell infected by
HIV-1, and even those that contain a silent
provirus. The vector was designed to effi-
ciently destroy the infected cells, but at the
same time preventing any change in nor-
mal cell functions, the objective being that
uninfected transduced cells remain unaf-
fected and fully functional.
7.4.1
Vector Design
The vector design should provide special
attention to the following points:
To prevent the emergence of viral-resis-
tant mutants, the target should be the
infected cell rather than the virus. In-
fected cells are not functional, so their
destruction does not have any negative
effect in the patient.
Transduced uninfected cells should be
preserved from any damage and remain
fully functional; therefore, the vector
should be silent in those cells.
The easiest way of detecting a latent
virus in a cell is first to induce the virus
expression; however, for reasons of pa-
tient safety the cell cycle state of the
transduced cells must not be modified.
Once the provirus is being expressed, it
should be detected, and a lethal gene
then expressed.
The lethal gene that destroys the cell
should be completely repressed in unin-
fected cells. In any case, it is better that,
if a weak expression of the lethal gene
occurs, it has no adverse effects for the
cell.
The vector also contains the cis elements
needed for packaging in HIV-1 particles,
and with two genetic systems to obtain an
effective and selective elimination of the
latently infected cells, without altering the
normal function of the transduced unin-
fected cells. In summary, the genetic con-
struction contains an externally inducible
system that promotes the expression of
any latent HIV-1 provirus without affecting
the cell cycle state, whilst a second genetic
system included in the vector allows the
detection of an essential viral protein and
expression of a suicide gene in response to
the viral protein (Fig. 7.4).
7.4.2
How the Vector Works
The vector has been designed to remain si-
lent in non-infected cells, and to be innoc-
uous to them. To achieve this, the first ge-
netic system consists of the HIV-1 transac-
tivator tat gene driven by a promoter con-
trolled exogenously by isopropyl-b-d-
thiogalactopyranoside (IPTG) (Fig. 7.5). An
alternative to IPTG is to use tetracycline as
inducer, which is currently approved for
use in humans. Thus, a tet-based vector is
being made to replace the IPTG. This sys-
tem will activate the expression of the si-
lent provirus without affecting the cell cy-
cle state. This point is of major impor-
tance in order to avoid any harm being
caused to the uninfected cells. Thus, in
resting cells Tat protein is produced transi-
ently only when IPTG is present, and as a
consequence the provirus is transcribed
and their proteins produced, including the
provirus Tat protein. Once the inducer is
removed, the vector returns to the silent
state if no provirus is present in the cell.
However, if the cell is infected by a now
activated provirus, the viral Tat protein will
keep the transcriptional activity of the LTR
promoter of the vector (see Fig. 7.5).
The second system induces a rapid death
of the cell by apoptosis if viral proteins are
being synthesized in that cell. This second
system permits the destruction of any
HIV-1 infected cell, no matter whether the
provirus was formerly active or silent. The
active provirus is detected by the presence
of Rev active proteins. As this protein is in-
volved in the rapid transport of the tran-
scripts to the cytoplasm, avoiding or reduc-
ing the rate of splicing of the introns, its ac-
tivity is essential for a correct viral gene ex-
pression, so no replication-competent
viruses can escape from its detection. An
apoptotic gene, p53, is inside an intron that
can only be over-expressed through the 5'-
LTR promoter of the vector and the pres-
ence of the Rev protein; otherwise, it is
Fig. 7.4 The HIV-1 lentiviral-based vector contains
the 5' long terminal repeat (5'-LTR), transactiva-
tion response region (TAR), packaging sequence
(), primer binding site (pbs), splice donor (SD)
and acceptor (SA) sites, Rev response element
(RRE), the 3' long terminal repeat (3'-LTR) and the
polypurine tract (ppt) from HIV-1. It contains a
cassette with the tat cDNA placed under the LacS-
witch II inducible mammalian expression system
with the following elements: RSV LTR=Rous sar-
coma virus long terminal repeat promoter; op=o-
perator; tk poly A=thymidine kinase poly A;
CMV=cytomegalovirus promoter; Lac I =Lac in-
hibitor; NLS=nuclear localization signal. It also
contains a second cassette with a lethal gene
(p53 cDNA) placed between SD and SA sites that
is expressed in response to Tat and Rev proteins.
Fig. 7.5 (a) The vector is si-
lent in uninfected cells be-
cause the transactivator ctat
gene is repressed. Addition of
the inducer (isopropyl-b-d-
thiogalactopyranoside [IPTG]
or tetracycline) abolishes the
repression of ctat (1) and Tat
protein is produced (2). The
Tat protein produces the tran-
scription driven by the 5'-LTR
of the vector, which includes
the p53 gene (3). P53 protein
is not synthesized because
the intron including p53 is
spliced (4). No apoptosis is
produced, and when the indu-
cer is removed there is no
longer any production of Tat.
(b) If a latent provirus is in
the cell, the addition of the in-
ducer (1) allows production
of the vector Tat (2). The Tat
protein produces the tran-
scription driven by the 5'-LTRs
of the vector and provirus (3),
and Rev protein accumulates
(4). The Rev activity causes
transcripts to be exported to
the cytoplasm before the spli-
cing of the introns occurs (5),
P53 is produced at a high rate
(6), and the cell dies by apop-
tosis (7). (c) When the cell is
infected with an active pro-
virus, the Tat (1) and Rev (2)
viral proteins cause the vector
p53 gene to be transcribed
and not spliced (3), P53 pro-
tein is highly produced (4),
and the cell dies by apoptosis
(5).
spliced from the transcript. Sustained high
expression of p53 induces apoptosis of the
cell, and therefore the death of cells contain-
ing an active provirus. The presence in the
vector of both genetic systems allows first
the activation of the silent provirus and sec-
ond the death of the cells that contain the
formerly silent, but now activated provirus
(see Fig. 7.5).
The use of a gene such as p53 to induce
apoptosis of the cell is another feature
aimed at avoiding unwanted harm to unin-
fected cells. This is because if leaky expres-
sion of p53 were able to exist in the ab-
sence of Rev, this physiological protein
would be expressed at low levels in normal
cells that do not compromise the cell life,
unless sustained high levels of expression
were to be achieved. In fact, it might have
a protective role against cancer for the
widely known tumour suppression activity
of this gene.
The behavior of the vector under different
experimental situations is summarized in
Fig. 7.5ac. When the vector is introduced
into a latently infected cell, it expresses
the tat gene in response to the inducer
added externally. The Tat protein transacti-
vates the 5'-LTR of the HIV-1 provirus and
the viral proteins begin to accumulate.
Among the first viral proteins to accumu-
late are Tat and Rev. Tat protein causes the
5'-LTR of the vector to be highly activated,
and many transcripts are produced. The
lethal gene is inside an intron, and initially
is spliced, but when the Rev protein reaches
a threshold level it allows transport of the
transcripts to the cytoplasm before the in-
tron is processed. In this way, the Rev pro-
tein of the HIV-1 provirus allows over-ex-
pression of the lethal gene and the cell dies
by apoptosis (R. Oya, L. Saniger, and F. Lu-
que, submitted) (Fig. 7.6).
If the vector enters a cell which is in-
fected with an active provirus, then no in-
ducer is needed and the vector is ex-
pressed in response to the viral Tat and
Rev proteins. However, if the cell is unin-
fected there is no expression of the genetic
systems of the vector. In these uninfected
cells, the tat gene is repressed and the
lethal gene is not transcribed, or is tran-
scribed at a very low level. However, the
transcripts are processed and the lethal
gene removed as an intron, such that no
lethal protein is synthesized (R. Oya, L. Sa-
niger, and F. Luque, submitted).
As stated previously, most lymphocytes
are in a resting G
0
state. This applies also
for HIV-1 infected patients, especially if
they are taking HAART therapy, and there-
fore have a reasonably healthy state. A lenti-
viral vector that enters a resting cell will be
reverse-transcribed, but will remain in a la-
bile, non-integrated form for a few days be-
fore being degraded. Therefore, if this vec-
tor is used in vivo, most cells will contain
a non-integrated vector for only some days,
and eventually will be degraded. If the rest-
ing cell is infected with a latent provirus, is
expected that the vector will be expressed in
response to the inducer, and as a conse-
quence the cell would die by apoptosis.
Only a minority of the cells will integrate
the vector permanently into a chromosome.
A strategy that modifies the genome of the
lowest number of cells is preferable to a
huge number of genetically manipulated
cells. At the same time, no harm should
be done to the permanently transduced
cells, in which the vector was designed to
be silent. In addition, the vector accepts
new modifications to render it even safer,
such as the introduction of the self-inacti-
vating technology in the LTRs.
7.4.3
Future Perspectives and Possible
Therapeutic Use
The treatment of HIV-1 infection is such a
complex task that, over the past 20 years,
conventional therapies have been unable
to provide a cure, and indeed a cure is not
envisaged in the near future. Given this sce-
nario, gene therapy represents a hope in the
search for efficient treatments. Most efforts
are directed to obtaining a subpopulation of
cells which are resistant to the virus, and to
reconstituting an effective immune re-
sponse with those resistant cells. Although
many technical limitations remain, the ge-
netic modification of bone marrow stem
cells represents the most promising
Fig. 7.6 Apoptosis observed by fluorescence mi-
croscopy. Apoptotic cells appear labeled with
green or red fluorescence if very late apoptosis
was achieved (Annexin V and propidium iodide
staining). Necrotic cells appear labeled with red
fluorescence. Left column: images obtained with
Nomarsky; right column: images obtained with
fluorescence staining. Almost no apoptosis is de-
tected if cells are treated with IPTG and trans-
fected with an empty vector (p53-deficient) (A and
B); transfected with the vector alone (C and D); or
with a provirus alone (E and F). However, a mas-
sive apoptosis is observed in cells treated with
IPTG and co-transfected with the vector plus an
active provirus (G and H); or with the vector and
a silent (tat-defective) provirus (I and J).
approach to achieve this goal. However,
even if limitations such as efficient gene
transfer and expression of the transgene
and especially a highly efficient engrafting
of the modified cells were overcome, and
the patients could tolerate the virus infec-
tion, eradication of the virus would not be
achieved. The only hope for a real cure
and virus eradication can, in theory, be the
destruction of all cells infected with replica-
tion-competent viruses. Although at present
this approach appears unaffordable because
the number of cells that need to be trans-
duced is vast, it is mainly a problem of the
efficiency of packaged vector production. It
is true that we are far from the simple pro-
duction of billions of packaged lentiviral
vectors. However, recent technical improve-
ments have shown great promise, and there
is no reason why, in the years to come, that
important advances will not be made. The
application of massive retroviral vectors will
need studies to determine the correct condi-
tions for efficient in vivo transduction.
Moreover, it is likely that short-term immu-
nosuppression of the patients will be
needed in order to avoid an effective im-
mune response against the vector. Efficient
in vivo gene therapy might be most useful
in other approaches such as stem cell genet-
ic modification, as no later engrafting is re-
quired. The targeting of lentiviral vectors
against metastasis tumor cells would also
benefit from efficient in vivo gene therapy.
Vectors targeted against any infected cell,
including latently infected resting cells, are
required for an HIV-1 gene therapy
approach, since no virus eradication can
be achieved if replication-competent latent
proviruses survive the therapy. This strategy
also allows a synergistic effect between
HAART therapy and gene therapy. While
HAART is able to block viral replication to
very low (undetectable) levels, a short-term
arrest of reverse transcriptase inhibitor up-
take would permit the in vivo delivery of
massive amounts of packaged vectors to
eliminate not only the latently infected cells
but also virus-producing cells that avoid the
HAART drugs. Although the technical diffi-
culties of this approach have been identi-
fied, a selective in vitro destruction of in-
fected cells has already been shown possi-
ble, even if the provirus is silent. It is hoped
that the other technical challenges will also
be overcome in the future to yield in a new
and efficient class of biopharmaceuticals.
References
1 Garcia, J. A., Wu, F. K., Mitsuyasu, R. and Gay-
nor, R. B. (1987) EMBO J 6, 37613770.
2 Kawakami, K., Scheidereit, C. and Roeder,
R. G. (1988) Proc Natl Acad Sci USA 85,
47004704.
3 Nabel, G. and Baltimore, D. (1987) Nature
326, 711713.
4 Li, C., Lai, C. F., Sigman, D. S. and Gaynor,
R. B. (1991) Proc Natl Acad Sci USA 88,
77397743.
5 Ross, E. K., Buckler-White, A. J., Rabson, A. B.,
Englund, G. and Martin, M. A. (1991) J Virol
65, 43504358.
6 Siekevitz, M., Josephs, S. F., Dukovich, M.,
Peffer, N., Wong-Staal, F. and Greene, W.C.
(1987) Science 238, 15751578.
7 Tong-Starksen, S. E., Luciw, P. A. and Peterlin,
B. M. (1987) Proc Natl Acad Sci USA 84,
68456849.
8 Stevenson, M., Stanwick, T. L., Dempsey, M. P.
and Lamonica, C. A. (1990) EMBO J 9, 1551
1560.
9 Wang, J. K., Kiyokawa, E., Verdin, E. and Tro-
no, D. (2000) Proc Natl Acad Sci USA 97,
394399.
10 Baur, A. S., Sawai, E. T., Dazin, P., Fantl, W. J.,
Cheng-Mayer, C. and Peterlin, B. M. (1994)
Immunity 1, 373384.
11 Rhee, S. S. and Marsh, J. W. (1994) J Immunol
152, 51285134.
12 Alexander, L., Du, Z., Rosenzweig, M., Jung,
J. U. and Desrosiers, R. C. (1997) J Virol 71,
60946099.
13 Schrager, J. A. and Marsh, J. W. (1999) Proc
Natl Acad Sci USA 96, 81678172.
14 Herrmann, C. H. and Rice, A. P. (1995) J Virol
69, 16121620.
15 Berkhout, B., Silverman, R. H. and Jeang, K. T.
(1989) Cell 59, 273282.
16 Fornerod, M., Ohno, M., Yoshida, M. and Mat-
taj, I. W. (1997) Cell 90, 10511060.
17 Henderson, B. R. and Percipalle, P. (1997) J
Mol Biol 274, 693707.
18 Neville, M., Stutz, F., Lee, L., Davis, L. I. and
Rosbash, M. (1997) Curr Biol 7, 767775.
19 Hammarskjold, M. L., Heimer, J., Hammar-
skjold, B., Sangwan, I., Albert, L. and Rekosh,
D. (1989) J Virol 63, 19591966.
20 Felber, B. K., Hadzopoulou-Cladaras, M., Cla-
daras, C., Copeland, T. and Pavlakis, G. N.
(1989) Proc Natl Acad Sci USA 86, 1495
1499.
21 Malim, M. H., Hauber, J., Le, S. Y., Maizel, J. V.
and Cullen, B. R. (1989) Nature 338, 254257.
22 Wu, Y. and Marsh, J.W. (2001) Science 293,
15031506.
23 Stevenson, M., Haggerty, S., Lamonica, C. A.,
Meier, C. M., Welch, S. K. and Wasiak, A. J.
(1990) J Virol 64, 24212425.
24 Cara, A., Guarnaccia, F., Reitz, M. S., Gallo,
R. C. and Lori, F. (1995) Virology 208, 2428.
25 Engelman, A., Englund, G., Orenstein, J. M.,
Martin, M. A. and Craigie, R. (1995) J Virol
69, 27292736.
26 Wiskerchen, M. and Muesing, M. A. (1995) J
Virol 69, 376386.
27 Ansari-Lari, M. A., Donehower, L. A. and
Gibbs, R. A. (1995) Virology 211, 332335.
28 Spina, C. A., Guatelli, J. C. and Richman, D. D.
(1995) J Virol 69, 29772988.
29 Harrison, G. S., Long, C. J., Curiel, T. J., Max-
well, F. and Maxwell, I. H. (1992) Hum Gene
Ther 3, 461469.
30 Harrison, G. S., Long, C. J., Maxwell, F.,
Glode, L. M. and Maxwell, I. H. (1992) AIDS
Res Hum Retroviruses 8, 3945.
31 Ragheb, J. A., Couture, L., Mullen, C., Ridg-
way, A. and Morgan, R. A. (1999) Hum Gene
Ther 10, 103112.
32 Dinges, M. M., Cook, D. R., King, J., Curiel,
T. J., Zhang, X. Q. and Harrison, G. S. (1995)
Hum Gene Ther 6, 14371445.
33 Curiel, T. J., Cook, D. R., Wang, Y., Hahn,
B. H., Ghosh, S. K. and Harrison, G. S. (1993)
34 Banerjea, A., Li, M. J., Bauer, G., Remling, L.,
Lee, N.S., Rossi, J. and Akkina, R. (2003) Mol
Ther 8, 6271.
35 Dropulic, B., Hermankova, M. and Pitha,
P. M. (1996) Proc Natl Acad Sci USA 93,
1110311108.
36 Dropulic, B. (2001) Mol Ther 4, 511512.
37 Weinberger, L. S., Schaffer, D. V. and Arkin,
A. P. (2003) J Virol 77, 1002810036.
38 Brady, H. J., Miles, C. G., Pennington, D. J.
and Dzierzak, E. A. (1994) Proc Natl Acad Sci
USA 91, 365369.
39 Caruso, M. and Klatzmann, D. (1992) Proc
40 Ross, T. M., Bieniasz, P. D. and Cullen, B. R.
(1999) Adv Virus Res 52, 233267.
41 Choe, H., Farzan, M., Sun, Y., Sullivan, N.,
Rollins, B., Ponath, P. D., Wu, L., Mackay,
C. R., LaRosa, G., Newman, W., Gerard, N.,
Gerard, C. and Sodroski, J. (1996) Cell 85,
11351148.
42 Bleul, C. C., Farzan, M., Choe, H., Parolin, C.,
Clark-Lewis, I., Sodroski, J. and Springer, T. A.
(1996) Nature 382, 829833.
43 Lee, M. T., Coburn, G. A., McClure, M. O. and
Cullen, B.R. (2003) J Virol 77, 1196411972.
44 Anderson, J., Banerjea, A., Planelles, V. and
Akkina, R. (2003) AIDS Res Hum Retro-
viruses 19, 699706.
45 BouHamdan, M., Strayer, D. S., Wei, D.,
Mukhtar, M., Duan, L. X., Hoxie, J. and Po-
merantz, R.J. (2001) Gene Ther 8, 408418.
46 Feng, Y., Leavitt, M., Tritz, R., Duarte, E.,
Kang, D., Mamounas, M., Gilles, P., Wong-
Staal, F., Kennedy, S., Merson, J., Yu, M. and
Barber, J. R. (2000) Virology 276, 271278.
47 Cagnon, L. and Rossi, J. J. (2000) Antisense
Nucleic Acid Drug Dev 10, 251261.
48 Bai, J., Gorantla, S., Banda, N., Cagnon, L.,
Rossi, J. and Akkina, R. (2000) Mol Ther 1,
244254.
49 Bai, X., Chen, J. D., Yang, A. G., Torti, F. and
Chen, S. Y. (1998) Gene Ther 5, 984994.
50 Egelhofer, M., Brandenburg, G., Martinius,
H., Schult-Dietrich, P., Melikyan, G., Kunert,
R., Baum, C., Choi, I., Alexandrov, A. and von
Laer, D. (2004) J Virol 78, 568575.
51 Shaheen, F., Duan, L., Zhu, M., Bagasra, O.
and Pomerantz, R. J. (1996) J Virol 70, 3392
3400.
52 BouHamdan, M., Duan, L. X., Pomerantz, R. J.
and Strayer, D.S. (1999) Gene Ther 6, 660
666.
53 Levy-Mintz, P., Duan, L., Zhang, H., Hu, B.,
Dornadula, G., Zhu, M., Kulkosky, J., Bizub-
References 565
Bender, D., Skalka, A. M. and Pomerantz, R. J.
(1996) J Virol 70, 88218832.
54 Frankel, A. D. and Pabo, C. O. (1988) Cell 55,
11891193.
55 Zauli, G., La Placa, M., Vignoli, M., Re, M. C.,
Gibellini, D., Furlini, G., Milani, D., Marchi-
sio, M., Mazzoni, M. and Capitani, S. (1995) J
Acquir Immune Defic Syndr Hum Retrovirol
10, 306316.
56 Ensoli, B., Barillari, G., Salahuddin, S. Z., Gal-
lo, R. C. and Wong-Staal, F. (1990) Nature
345, 8486.
57 Gallo, R. C. (1999) Proc Natl Acad Sci USA
96, 83248326.
58 Zagury, D., Lachgar, A., Chams, V., Fall, L. S.,
Bernard, J., Zagury, J. F., Bizzini, B., Gringeri,
A., Santagostino, E., Rappaport, J., Feldman,
M., Burny, A. and Gallo, R. C. (1998) Proc Natl
Acad Sci USA 95, 38513856.
59 Li, C. J., Friedman, D. J., Wang, C., Metelev, V.
and Pardee, A. B. (1995) Science 268, 429431.
60 Huang, L., Bosch, I., Hofmann, W., Sodroski,
J. and Pardee, A. B. (1998) J Virol 72, 8952
8960.
61 Secchiero, P., Zella, D., Capitani, S., Gallo,
R. C. and Zauli, G. (1999) J Immunol 162,
24272431.
62 Zocchi, M. R., Rubartelli, A., Morgavi, P. and
Poggi, A. (1998) J Immunol 161, 29382943.
63 Ott, M., Emiliani, S., Van Lint, C., Herbein,
G., Lovett, J., Chirmule, N., McCloskey, T.,
Pahwa, S. and Verdin, E. (1997) Science 275,
14811485.
64 Fraisier, C., Abraham, D. A., van Oijen, M.,
Cunliffe, V., Irvine, A., Craig, R. and Dzierzak,
E. A. (1998) Gene Ther 5, 946954.
65 Rossi, C., Balboni, P. G., Betti, M., Marconi,
P. C., Bozzini, R., Grossi, M. P., Barbanti-Bro-
dano, G. and Caputo, A. (1997) Gene Ther 4,
12611269.
66 Poznansky, M. C., La Vecchio, J., Silva-Arietta,
S., Porter-Brooks, J., Brody, K., Olszak, I. T.,
Adams, G. B., Ramstedt, U., Marasco, W. A.
and Scadden, D. T. (1999) Hum Gene Ther 10,
25052514.
67 Poznansky, M. C., Foxall, R., Mhashilkar, A.,
Coker, R., Jones, S., Ramstedt, U. and Maras-
co, W. (1998) Hum Gene Ther 9, 487496.
68 Bai, J., Sui, J., Zhu, R. Y., Tallarico, A. S., Gen-
nari, F., Zhang, D. and Marasco, W. A. (2003)
J Biol Chem 278, 14331442.
69 Browning, C. M., Cagnon, L., Good, P. D., Ros-
si, J., Engelke, D. R. and Markovitz, D. M.
(1999) J Virol 73, 51915195.
70 Fraisier, C., Irvine, A., Wrighton, C., Craig, R.
and Dzierzak, E. (1998) Gene Ther 5, 1665
1676.
71 Lisziewicz, J., Sun, D., Smythe, J., Lusso, P.,
Lori, F., Louie, A., Markham, P., Rossi, J.,
Reitz, M. and Gallo, R.C. (1993) Proc Natl
Acad Sci USA 90, 80008004.
72 Lisziewicz, J., Rappaport, J. and Dhar, R.
(1991) New Biol 3, 8289.
73 Chadwick, D. R. and Lever, A. M. (2000) Gene
Ther 7, 13621368.
74 Vandendriessche, T., Chuah, M. K., Chiang, L.,
Chang, H. K., Ensoli, B. and Morgan, R. A.
(1995) J Virol 69, 40454052.
75 Sun, L. Q., Ely, J. A., Gerlach, W. and Sy-
monds, G. (1997) Mol Biotechnol 7, 241251.
76 Gervaix, A., Li, X., Kraus, G. and Wong-Staal,
F. (1997) J Virol 71, 30483053.
77 Yu, M., Leavitt, M. C., Maruyama, M., Yamada,
O., Young, D., Ho, A. D. and Wong-Staal, F.
(1995) Proc Natl Acad Sci USA 92, 699703.
78 Leavitt, M. C., Yu, M., Yamada, O., Kraus, G.,
Looney, D., Poeschla, E. and Wong-Staal, F.
(1994) Hum Gene Ther 5, 11151120.
79 Cordelier, P., Van Bockstaele, E., Calarota,
S. A. and Strayer, D.S. (2003) Mol Ther 7,
801810.
80 Mautino, M. R. and Morgan, R. A. (2002) Gene
Ther 9, 421431.
81 Mautino, M. R., Keiser, N. and Morgan, R. A.
(2001) J Virol 75, 35903599.
82 Bonyhadi, M. L., Moss, K., Voytovich, A., Au-
ten, J., Kalfoglou, C., Plavec, I., Forestell, S.,
Su, L., Bohnlein, E. and Kaneshima, H. (1997)
J Virol 71, 47074716.
83 Caputo, A., Rossi, C., Bozzini, R., Betti, M.,
Grossi, M. P., Barbanti-Brodano, G. and Balbo-
ni, P. G. (1997) Gene Ther 4, 288295.
84 Plavec, I., Agarwal, M., Ho, K. E., Pineda, M.,
Auten, J., Baker, J., Matsuzaki, H., Escaich, S.,
Bonyhadi, M. and Bohnlein, E. (1997) Gene
Ther 4, 128139.
85 Bahner, I., Zhou, C., Yu, X. J., Hao, Q. L., Gua-
telli, J. C. and Kohn, D. B. (1993) J Virol 67,
31993207.
86 Bauer, G., Valdez, P., Kearns, K., Bahner, I.,
Wen, S. F., Zaia, J. A. and Kohn, D. B. (1997)
Blood 89, 22592267.
87 Bahner, I., Kearns, K., Hao, Q. L., Smogor-
zewska, E. M. and Kohn, D. B. (1996) J Virol
70, 43524360.
88 Bahner, I., Kearns, K., Coutinho, S., Leo-
nard, E. H. and Kohn, D. B. (1997) Blood 90,
17871798.
89 Trono, D., Feinberg, M. B. and Baltimore, D.
(1989) Cell 59, 113120.
90 Tewari, D., Notkins, A. L. and Zhou, P.
(2003) J Gene Med 5, 182189.
91 Cordelier, P., Zern, M. A. and Strayer, D.S.
(2003) Gene Ther 10, 467477.
92 Lewis, P. F. and Emerman, M. (1994) J Virol
68, 510516.
93 Yu, S. F., von Ruden, T., Kantoff, P. W., Gar-
ber, C., Seiberg, M., Ruther, U., Anderson,
W. F., Wagner, E. F. and Gilboa, E. (1986)
Proc Natl Acad Sci USA 83, 31943198.
94 Yee, J. K., Moores, J. C., Jolly, D. J., Wolff,
J. A., Respess, J. G. and Friedmann, T. (1987)
95 Hawley, R. G., Covarrubias, L., Hawley, T.
and Mintz, B. (1987) Proc Natl Acad Sci
USA 84, 24062410.
96 Olson, P., Nelson, S. and Dornburg, R.
(1994) J Virol 68, 70607066.
97 Zufferey, R., Dull, T., Mandel, R. J., Bukovs-
ky, A., Quiroz, D., Naldini, L. and Trono, D.
(1998) J Virol 72, 98739880.
98 Miyoshi, H., Blomer, U., Takahashi, M.,
Gage, F. H. and Verma, I. M. (1998) J Virol
72, 81508157.
99 Xu, K., Ma, H., McCown, T. J., Verma, I. M.
and Kafri, T. (2001) Mol Ther 3, 97104.
100 Grunwald, T., Pedersen, F. S., Wagner, R.
and Uberla, K. (2004) J Gene Med 6, 147
154.
101 Yee, J. K., Miyanohara, A., LaPorte, P., Bouic,
K., Burns, J. C. and Friedmann, T. (1994)
102 Blomer, U., Naldini, L., Kafri, T., Trono, D.,
Verma, I. M. and Gage, F. H. (1997) J Virol
71, 66416649.
103 Naldini, L., Blomer, U., Gallay, P., Ory, D.,
Mulligan, R., Gage, F. H., Verma, I. M. and
Trono, D. (1996) Science 272, 263267.
104 Kyrkanides, S., Miller, J. H. and Federoff, H. J.
(2003) Brain Res Mol Brain Res 119, 19.
105 Jakobsson, J., Ericson, C., Jansson, M.,
Bjork, E. and Lundberg, C. (2003) J Neurosci
Res 73, 876885.
106 Zhao, C., Strappe, P. M., Lever, A. M. and
Franklin, R. J. (2003) Glia 42, 5967.
107 Kordower, J. H., Emborg, M. E., Bloch, J.,
Ma, S. Y., Chu, Y., Leventhal, L., McBride, J.,
Chen, E. Y., Palfi, S., Roitberg, B. Z., Brown,
W. D., Holden, J. E., Pyzalski, R., Taylor,
M. D., Carvey, P., Ling, Z., Trono, D., Han-
traye, P., Deglon, N. and Aebischer, P.
(2000) Science 290, 767773.
108 Strayer, D. S. (1996) J Biol Chem 271,
2474124746.
109 Strayer, D. S. and Milano, J. (1996) Gene
Ther 3, 581587.
110 Strayer, D. S., Kondo, R., Milano, J. and
Duan, L. X. (1997) Gene Ther 4, 219225.
111 Coleman, J. E., Huentelman, M. J., Kasparov,
S., Metcalfe, B. L., Paton, J. F., Katovich,
M. J., Semple-Rowland, S. L. and Raizada,
M. K. (2003) Physiol Genomics 12, 221228.
112 Sutton, R. E., Wu, H. T., Rigg, R., Bohnlein,
E. and Brown, P. O. (1998) J Virol 72, 5781
5788.
113 Miyoshi, H., Smith, K. A., Mosier, D. E., Ver-
ma, I. M. and Torbett, B. E. (1999) Science
283, 682686.
114 Roesler, J., Brenner, S., Bukovsky, A. A.,
Whiting-Theobald, N., Dull, T., Kelly, M., Ci-
vin, C. I. and Malech, H.L. (2002) Blood 100,
43814390.
115 Davis, B. M., Humeau, L. and Dropulic, B.
(2004) Mol Ther 9, 160172.
116 Neff, T., Shotkoski, F. and Stamatoyanno-
poulos, G. (1997) Stem Cells 15 Suppl 1,
265271.
117 Zentilin, L., Qin, G., Tafuro, S., Dinauer,
M. C., Baum, C. and Giacca, M. (2000) Gene
Ther 7, 153166.
118 Klug, C. A., Cheshier, S. and Weissman, I. L.
(2000) Blood 96, 894901.
119 Pannell, D., Osborne, C. S., Yao, S., Sukon-
nik, T., Pasceri, P., Karaiskakis, A., Okano,
M., Li, E., Lipshitz, H. D. and Ellis, J. (2000)
EMBO J 19, 58845894.
120 Osborne, C. S., Pasceri, P., Singal, R., Sukon-
nik, T., Ginder, G. D. and Ellis, J. (1999) J
Virol 73, 54905496.
121 Pfeifer, A., Ikawa, M., Dayn, Y. and Verma,
I. M. (2002) Proc Natl Acad Sci USA 99,
21402145.
122 Hanawa, H., Hematti, P., Keyvanfar, K.,
Metzger, M.E., Krouse, A., Donahue, R. E.,
Kepes, S., Gray, J., Dunbar, C. E., Persons,
D. A. and Nienhuis, A. W. (2004) Blood.
References 567
123 Schroder, A. R., Shinn, P., Chen, H., Berry,
C., Ecker, J. R. and Bushman, F. (2002) Cell
110, 521529.
124 Chun, T. W., Stuyver, L., Mizell, S. B., Ehler,
L. A., Mican, J. A., Baseler, M., Lloyd, A.L.,
Nowak, M. A. and Fauci, A. S. (1997) Proc
125 Chun, T. W., Carruth, L., Finzi, D., Shen, X.,
DiGiuseppe, J. A., Taylor, H., Hermankova,
M., Chadwick, K., Margolick, J., Quinn,
T. C., Kuo, Y. H., Brookmeyer, R., Zeiger,
M. A., Barditch-Crovo, P. and Siliciano, R. F.
(1997) Nature 387, 183188.
126 Chun, T. W., Finzi, D., Margolick, J., Chad-
wick, K., Schwartz, D. and Siliciano, R. F.
(1995) Nat Med 1, 12841290.
127 Finzi, D., Blankson, J., Siliciano, J. D., Mar-
golick, J. B., Chadwick, K., Pierson, T.,
Smith, K., Lisziewicz, J., Lori, F., Flexner,
C., Quinn, T. C., Chaisson, R. E., Rosenberg,
E., Walker, B., Gange, S., Gallant, J. and Sili-
ciano, R. F. (1999) Nat Med 5, 512517.
128 Finzi, D., Hermankova, M., Pierson, T., Car-
ruth, L. M., Buck, C., Chaisson, R. E., Quinn,
T. C., Chadwick, K., Margolick, J., Brook-
meyer, R., Gallant, J., Markowitz, M., Ho,
D. D., Richman, D. D. and Siliciano, R. F.
(1997) Science 278, 12951300.
129 Wong, J. K., Hezareh, M., Gunthard, H. F.,
Havlir, D. V., Ignacio, C. C., Spina, C. A. and
Richman, D. D. (1997) Science 278, 1291
1295.
130 Ramratnam, B., Mittler, J. E., Zhang, L.,
Boden, D., Hurley, A., Fang, F., Macken,
C. A., Perelson, A. S., Markowitz, M. and Ho,
D. D. (2000) Nat Med 6, 8285.
131 Zhang, L., Ramratnam, B., Tenner-Racz, K.,
He, Y., Vesanen, M., Lewin, S., Talal, A.,
Racz, P., Perelson, A. S., Korber, B. T., Mar-
kowitz, M. and Ho, D. D. (1999) N Engl J
Med 340, 16051613.
132 Persaud, D., Pierson, T., Ruff, C., Finzi, D.,
Chadwick, K. R., Margolick, J. B., Ruff, A.,
Hutton, N., Ray, S. and Siliciano, R.F. (2000)
J Clin Invest 105, 9951003.
133 Powderly, W. G., Landay, A. and Lederman,
M. M. (1998) JAMA 280, 7277.
134 Siliciano, J. D., Kajdas, J., Finzi, D., Quinn,
T. C., Chadwick, K., Margolick, J. B., Kovacs,
C., Gange, S. J. and Siliciano, R. F. (2003)
Nat Med 9, 727728.
135 Ramratnam, B., Ribeiro, R., He, T., Chung,
C., Simon, V., Vanderhoeven, J., Hurley, A.,
Zhang, L., Perelson, A. S., Ho, D. D. and
Markowitz, M. (2004) J Acquir Immune De-
fic Syndr 35, 3337.
136 Douek, D. C., McFarland, R. D., Keiser, P. H.,
Gage, E. A., Massey, J. M., Haynes, B. F., Po-
lis, M. A., Haase, A. T., Feinberg, M. B., Sulli-
van, J. L., Jamieson, B. D., Zack, J. A., Picker,
L. J. and Koup, R. A. (1998) Nature 396, 690
695.
137 Stevenson, M. (1997) AIDS 11 Suppl A,
S25S33.
138 Persaud, D., Zhou, Y., Siliciano, J. M. and Si-
liciano, R. F. (2003) J Virol 77, 16591665.
139 Blankson, J. N., Finzi, D., Pierson, T. C., Sa-
bundayo, B. P., Chadwick, K., Margolick,
J. B., Quinn, T. C. and Siliciano, R. F. (2000)
J Infect Dis 182, 16361642.
140 Bukrinsky, M. I., Stanwick, T. L., Dempsey,
M. P. and Stevenson, M. (1991) Science 254,
423427.
141 Spina, C. A., Guatelli, J. C. and Richman,
D. D. (1995) J Virol 69, 29772988.
142 Zack, J. A., Arrigo, S. J., Weitsman, S. R., Go,
A. S., Haislip, A. and Chen, I. S. (1990) Cell
61, 213222.
143 Bednarik, D. P., Cook, J. A. and Pitha, P. M.
(1990) EMBO J 9, 11571164.
144 Jordan, A., Defechereux, P. and Verdin, E.
(2001) EMBO J 20, 17261738.
145 Winslow, B. J., Pomerantz, R. J., Bagasra, O.
and Trono, D. (1993) Virology 196, 849854.
146 Hermankova, M., Siliciano, J. D., Zhou, Y.,
Monie, D., Chadwick, K., Margolick, J. B.,
Quinn, T. C. and Siliciano, R. F. (2003) J Vir-
ol 77, 73837392.
147 Stellbrink, H. J., Hufert, F. T., Tenner-Racz,
K., Lauer, J., Schneider, C., Albrecht, H.,
Racz, P. and van Lunzen, J. (1998) Antivir
Ther 3, 209214.
148 Dybul, M., Hidalgo, B., Chun, T. W., Belson,
M., Migueles, S. A., Justement, J. S., Herpin,
B., Perry, C., Hallahan, C. W., Davey, R. T.,
Metcalf, J. A., Connors, M. and Fauci, A. S.
(2002) J Infect Dis 185, 6168.
149 Bocklandt, S., Blumberg, P. M. and Hamer,
D. H. (2003) Antiviral Res 59, 8998.
150 Korin, Y. D., Brooks, D. G., Brown, S., Korot-
zer, A. and Zack, J. A. (2002) J Virol 76,
81188123.
Abstract
Lentiviral vector delivery of RNA-derived
modalities offers a valid and potentially effi-
cacious gene therapy-based approach to aug-
ment current anti-HIV-1 therapeutics. Many
lentiviral vector systems have been exten-
sively studied including those based on
FIV, HIV-1 and HIV-2/SIVas well as replica-
tion incompetent, self-inactivating (sin) ver-
sus conditionally replicating (mobilizable)
vectors. However, a major limitation in uti-
lizing these lentiviral vectors and a gene
therapy-based approach in treating HIV-1
infection has been the relative lack of an ef-
ficacious therapeutic modality. Recently, the
emerging technology of RNA interference
the natural mechanism by which small in-
terfering RNAs (siRNAs) operate to specifi-
cally and potently downregulate the expres-
sion of a target gene has been described,
and shown to potently and specifically sup-
press HIV-1. Moreover, since siRNAs are a
small nucleic acid reagent they are unlikely
to elicit an immune response, making them
a theoretically good future therapeutic to be
used to treat HIV-1 infection. This chapter
will focus on the development, delivery
and potential therapeutic use of RNA-based
antiviral modalities including ribozymes
and siRNAs to employ in conjunction with
current anti-HIV-1 therapies.
Abbreviations
crHIV conditionally replicating HIV
HAART highly active antiretroviral therapy
MoMLV Moloney murine leukemia virus
PTGS post-transcriptional gene silencing
RISC RNA-induced silencing complex
shRNA short hairpin RNA
TGS transcriptional gene silencing
8.1
Introduction
HIV-1 has been a major cause of death
and morbidity in the human population
for at least 20 years now, with roughly
70% of the world cases in sub-Saharan
Africa [1] and with little hope for an effica-
cious vaccine in the near future [2]. Unde-
niably, much progress has been made in
understanding and treating individuals in-
fected with HIV-1. Importantly, highly ac-
tive antiretroviral therapy (HAART), con-
sisting of compounds that specifically
block HIV-1 reverse transcriptase and pro-
tease, has proven invaluable in extending
the lives of those infected with HIV-1.
HAART, however, does have its limita-
tions, including the emergence of drug-re-
569
8
Combinatorial RNA-based Therapies for HIV-1
ISBN: 3-527-31184-X
sistance mutants [3], patient noncompli-
ance [4] and the overall cost of multidrug
combination therapy. Therefore, alternative
strategies to inhibit HIV-1 viral replication
either alone or in combination with the
currently administered therapies are ur-
gently needed (see also Part II, Chapter 7).
One alternative strategy involves a gene-
therapy-based approach and the use of
virus-based vectors to deliver anti-HIV-1
genes to target cells. This chapter will fo-
cus specifically on anti-HIV-1 genetic strat-
egies, such as ribozymes, small interfering
RNAs (siRNAs) and the use of lentiviral
vectors in delivering these antiviral genes
as well as functioning as competitive inhi-
bitors to HIV-1 virion packaging. This
chapter assumes some knowledge of the
HIV-1 lifecycle.
8.2
RNA-based Antiviral Agents
8.2.1
Ribozymes
In the past RNA has been considered pri-
marily as a form of information storage
for the transfer of genetic information from
gene to protein. This traditional view was
drastically altered when RNA was demon-
strated to function catalytically [5, 6]. These
enzymatic forms of RNA have been termed
as ribozymes. Trans-acting ribozymes (func-
tioning on other RNA molecules) can be
easily generated in vitro, and function by
specifically base pairing with and cleaving
mRNA phosphodiester bonds catalytically
in a sequence-specific manner [7, 8]. Ribo-
zymes can act in cis (on the same tran-
script) or trans to target specific RNA se-
quences [9, 10]. In theory, only a single
RNA species will be targeted and degraded
by the ribozyme [9]. Essentially two classes
of ribozyme motifs are currently being uti-
lized to combat HIV-1 infection: hammer-
head [11] and hairpin ([12, 13]; reviewed in
[14, 15]) (see also the Introduction to this
book and Part III, Chapter 3). Ribozymes
specific for HIV-1 Integrase, the U5 region
of the long terminal repeat, Tat, Pol and
Rev have all demonstrated successful sup-
pression of HIV-1 gene expression [1620].
Ribozymes can be delivered and expressed
endogenously from either PolII and PolIII
promoters in both plasmid and viral vector
constructs [8, 9]. Ribozyme transcription
can be achieved from either a stably trans-
fected or transduced cell line, i.e., where in-
tegration of the ribozyme gene into the ge-
nome has occurred or from cells which have
been transiently transfected with a plasmid
carrying the ribozyme gene [8]. The endog-
enous ribozyme expression levels rely on a
suitable promoter expressing the ribozyme,
and in vitro a free Mg
2+
concentration of
around 10 mM [8]. It is generally assumed
that the Mg
2+
concentration in the typical
cell is around 5 mM. Exogenous delivery
of net negatively charged ribozymes is be-
lieved to face many barriers such as hydro-
phobic cell membranes, sequestering in en-
dosomes or lysosomes, thereby limiting
their biologic availability (see also Part VI,
Chapters 68). Therefore, if ribozymes are
to become a bona fide anti-HIV-1 therapeu-
tic, a delivery system capable of more effec-
tively localizing these molecules with the
target mRNA will be required.
8.2.2
RNA Interference (RNAi)
RNAi, first described in plants and termed
cosuppression (reviewed in [21]), is a pro-
cess in which double-stranded RNA in-
duces homology-dependent degradation of
mRNA [2224]. RNAi is a process involv-
ing small interfering double-stranded
RNAs (siRNAs) 2122 bp in length with 2-
base 3' overhanging ends that can induce
a homology-dependent degradation of cog-
nate mRNA [23]. The generation of siRNA
is the result of a multistep process that in-
volves the action of an RNase III endonu-
clease Dicer [25] (Fig. 8.1). The approxi-
mately 22-bp siRNAs that are processed by
Dicer provide much of the specificity in
the silencing process. The necessity for an
exact sequence match in the sense strand
of siRNA duplexes has, however, been
questioned, as single-stranded antisense
siRNAs can guide target RNA cleavage [26]
and as many as five mismatches in the
sense strand RNA may be tolerated [27].
In contrast, a single base pair mismatch
relative to the target RNA on the antisense
strand has been shown to significantly re-
duce siRNA mediated message degrada-
tion [28] (see also Part I, Chapter 10). Fol-
lowing the action of dicer the approxi-
mately 21-bp siRNAs are incorporated into
the RNA-induced silencing complex
(RISC), which identifies and silences by
slicing the mRNAs complementary to the
21-bp siRNA through interactions with Ar-
gonaute 2 [29] (Fig. 8.1). The specificity
juxtaposed with potent suppression of tar-
get genes by siRNA has truly made RNAi
a standard methodology for gene-specific
silencing in mammalian cells.
Mechanistically RNAi can suppress gene
expression via two distinct pathways: tran-
scriptional gene silencing (TGS) and post-
transcriptional gene silencing (PTGS) [30,
31]. PTGS involves siRNAs targeting either
mRNA or pre-mRNA, including intronic se-
quences in Caenorhabditis elegans and yeast
[32]. TGS was first observed in plants that
were infected with a virus and simulta-
neously contained viral promoters express-
ing integrated transgenes. Interestingly,
these promoters became methylated at sites
matching the small double-stranded viral
RNAs, and transcription was suppressed
as a result of these homologous viral RNAs
entering the nucleus and inducing TGS [33,
34], i.e., RNA-specific targeted suppression
of gene expression at the level of the promo-
ter. In human cells, gene silencing induced
by RNAi was initially thought to be re-
stricted to action on cytoplasmic mRNA or
RNA at the nuclear pore [35], similar to
most reports in C. elegans and Trypanosoma
brucei [22, 36, 37]. Until recently, TGS was
only found to occur in plants, Drosophila
and in Schizosaccharomyces pombe in centro-
meric regulation [38]. However, TGS has re-
cently been reported to operate in mamma-
lian cells and appeared to rely on the deliv-
ery of the siRNA to the nucleus ([39, 40]; re-
viewed in [41]). However, the strict
requirements of nuclear delivery may not
be necessary if temporal factors are in-
cluded in the analysis [39]. Consequently,
siRNAs can be used to not only silence tar-
geted transcripts (PTGS-based silencing),
but also targeted promoters (TGS-based si-
lencing), subsequently increasing the diver-
sity of siRNA targets. Indeed, the field of
RNAi, and the fact that siRNAs can silence
a target gene not only at the mRNA level
but also interact, modify and suppress gene
expression at the chromatin level, has sig-
nificantly shifted the paradigm and chal-
lenged the current dogma about the func-
tion(s) of RNA in the cell, similar to pre-
vious observations with ribozymes (see also
Part I, Chapters 1 and 7).
8.3
RNAi: Diversity of Viral Targets
Targeted suppression of HIV-1 has been
achieved through siRNAs directed against
HIV-1 tat and rev [4345], reverse tran-
scriptase [4547], TAR and the 3' untrans-
lated region [48], Vif [48], as well as gag,
Fig. 8.1 RNAi pathways in mammalian cells. RNAi
can operate at the transcriptional and post-tran-
scriptional level or possibly a combination of
both, and is based on the specific targeting of siR-
NAs to a mRNA or promoter. A cell can be stably
transduced with a lentiviral vector that expresses
siRNAs either from two independent promoters
(U6 PolIII) or a single promoter driving the ex-
pression of a hairpin shRNA targeting a particular
gene of interest (1). The expressed shRNAs are
probably bound by Exportin 5 (2) [85, 86], get
shuttled out of the nucleus and handed off to Di-
cer, which then cleaves the loop from the hairpin
(3), producing the siRNA that is then loaded into
RISC, ultimately leading to slicing of the target
mRNA (4) and essentially driving PTGS. Alterna-
tively, siRNAs can function in a TGS-based man-
ner following expression from the lentiviral vector
(A), the siRNAs get bound by DNA methyltrans-
ferases (B), which can interact with histone deace-
tylase and the histone methyltransferase
(SUV39H1) (C), to essentially replace the acetate
group of histone 3 lysine 9 with a methyl group,
subsequently silencing the targeted promoter in a
chromatin modifying-based pathway which may or
may not result in robust methylation of the tar-
geted promoter. Finally, it is possible that the pro-
moter-directed siRNAs might be expressed and
exported out of the nucleus, loaded into RISC in
the cytoplasm, and shuttled back to the nucleus
where they can function to suppress gene expres-
sion.
and the HIV-1 coreceptor CD4 [44] and
coreceptor CCR5 ([49]; reviewed extensively
in [50]). However, despite the excitement
and the early proof-of-principle in siRNA
targeting of HIV-1, there are some signifi-
cant issues and concerns about therapeutic
application of this technology, including
difficulties with efficient delivery, uncer-
tainty about potential toxicity and the
emergence of siRNA-resistant viruses. In
particular, certain viruses encode proteins
that block one or more steps in the RNAi
pathway [5156]. Indeed, resistance to si-
RNA occurs rather rapidly and is only con-
tingent on a single nucleotide substitution
[57], and recently HIV-1 demonstrated an
ability to elude siRNA targeting by the evo-
lution of alternative splice variants for the
siRNA-targeted transcripts [58]. Thus, a
priori routes to circumventing such a con-
clusion in siRNA-mediated therapies for
HIV-1 could be to (1) design siRNAs to
best fit targets from an extensive database
of the variants in the particular target virus
[41, 46, 47] and/or (2) to incorporate these
best fit siRNAs into a multiple antiviral
siRNA expressing transgene vector. Unde-
niably, the multiplexing of several different
siRNAs targeting dissimilar sites in the
HIV genome along with non-essential cel-
lular targets such as CCR5 should be uti-
lized to harness the full potential of this
mechanism in treating HIV-1 infection
with siRNA technology. Interestingly, ribo-
zymes can also be used to increase the
breadth of targets by their incorporation as
the hairpin loop in small hairpin RNAs
(shRNAs). Recent evidence indicates that
the expression of a ribozyme in cis from a
shRNA (i.e., the ribozymes acts as the loop
while the siRNAs function as the stem in
these stemloop shRNAs) is efficiently lo-
calized as well as functional in suppres-
sing target mRNAs expression [59]. Alter-
natively, designing siRNAs to more con-
served regions such as viral intron/exon
splice junctions might also prove effective
in reducing the emergence of viral vari-
ants.
8.4
Delivery of siRNAs to Target Cells
In the context of an experimental setting,
direct delivery of particular siRNAs is re-
quired for an initial assessment of siRNA-
mediated suppressive effects, either PTGS
or TGS based. Cationic lipid-based com-
plexes have proven remarkably useful for
this purpose, especially when determining
the efficacious nature of a particular si-
RNA directed to a target mRNA (i.e., a
PTGS-based mode of suppression). More-
over, cationic lipid complexes have been
shown in mice to systemically deliver si-
RNAs, suggesting such a methodology
could prove useful in using siRNAs in hu-
mans to aid or augment the immune re-
sponse during times of duress [60, 61] (see
also Part VI, Chapter 7).
Peptide-based siRNA delivery systems
have also been shown to be important in
the nuclear delivery of siRNAs designed to
specifically target gene promoters [46, 47]
(see also Part VI, Chapter 1). Two particu-
lar nuclear-specific peptides have proven
useful in nuclear delivery of siRNAs and
induction of transcriptional gene silencing:
MPG [62] and NLSV404 [63]. Others have
found oligofectamine useful in promoter-
specific siRNA targeting [41]. Indeed the
initial determination of a particular pro-
moter susceptibility to siRNA-mediated
TGS will generally rely on transient-based
transfection procedures.
Once a siRNA targeting a particular viral
RNA or promoter has been designed,
tested and shown to be effective in vitro, it
will probably be necessary to stably express
the siRNA or a multiple siRNA cassette.
The introduction of siRNAs into mamma-
lian cells can be achieved through varieties
of standard commercially available trans-
fection methods. The strength and dura-
tion of the silencing response delivered in
the context of such transfection methods,
however, is determined and/or limited by
several factors. On a population basis, the
overall efficiency of transfection is a major
determinant, which must be addressed by
optimizing conditions. In each individual
cell, silencing depends upon a combina-
tion of the amount of siRNA that is deliv-
ered and upon the potential of that indi-
vidual siRNA to suppress its target (the po-
tency). Even a relatively poor siRNA can si-
lence its target provided that sufficient
quantities are delivered. However, over-
loading the system with a high concentra-
tion of siRNAs is likely to lead to unde-
sired effects, including off-target suppres-
sion as well as the induction of a PKR re-
sponse [6466].
There are severable methodologies for
expressing siRNAs. One method for stable
and controlled expression of siRNAs from
the context of the cell is via the use of len-
tiviral vectors. Lentiviruses, unlike retro-
viruses such as Moloney murine leukemia
virus (MoMLV), tend to preferentially inte-
grate downstream of active promoters
within the active transcriptional unit, po-
tentially limiting their overall oncogenicity
[67]. Moreover, lentiviral-based vectors are
capable of transducing nondividing cells
[68] and specifically targeting the nucleus
[69] (see also Part I, Chapters 6 and 7).
Previous gene therapies targeting HIV-1
with stable integrating virus systems have
been predominantly based upon vectors
constructed from the retrovirus MoMLV
and lentiviruses such as HIV-1 (reviewed
by [69, 70]). An effective gene therapy for
HIV-1 will require a vector system that can
(1) target and transduce the same cell
types infected by HIV-1, (2) express the
anti-HIV genes at levels capable of inhibit-
ing the virus or burden the virus infected
cell in such a manner as to cause its
demise, and/or (3) specifically disrupt the
provirus to be unable to produce infectious
virions. While both MoMLV- and HIV-
based vectors can stably transduce cell
populations, the lentiviral-based vectors of-
fer significant advantages. Lentiviruses,
unlike retroviruses such as MoMLV, tend
to preferentially integrate downstream of
active promoters within the active tran-
scriptional unit, potentially limiting their
overall oncogenicity [67]. Another advan-
tage to the use of lentiviral vectors such as
HIV-1 or HIV-2 as an adjunctive therapy
for HIV-1 infection is found in their ability
to compete with the wild-type virus for
packaging as well as their ability to be mo-
bilized and spread to the same target cells
infected by the virion (Fig. 8.2). Conse-
quently, lentiviral-based vectors are one of
the best candidates to meet the above crite-
ria.
Lentiviral vectors such as HIV-1, HIV-2/
SIV or FIV are generally produced as de-
picted in Fig. 8.3, and are capable of stably
transducing many cell types, including he-
matopoietic stem cells [71, 72], and inte-
grating and expressing desired transgenes
[7275]. Recently, lentiviruses have been
shown to cross-package one another [76
78]. This observation has been carried over
experimentally with HIV-2 vectors being
cross-packaged by FIV, and capable of sta-
bly transducing and protecting human pri-
mary blood mononuclear cells [46] from
HIV-1 infection. The cross-packaging of
lentiviral vectors offers a unique and possi-
bly safer method for delivering antiviral
vectors to target cells in HIV-1-infected in-
dividuals. For instance, FIV packaged HIV-
1 or HIV-2 vectors reduce the likelihood of
Fig. 8.2 Vector mobilization in an HIV-1-infected
cell. Initial stages of HIV-1 infection involve an in-
teraction of the virion with the target cell CD4 and
CCR5 coreceptor (A), followed by adsorption of
the virion, reverse transcription of the viral RNAs
(B), pre-integration complex and integration (C).
The conditionally replicating vector (crHIV vector)
also undergoes similar cell surface binding,
although non-specifically with VSV-G pseudotyped
envelope, reverse transcription and integration
(1). Once the virus and vector are integrated, tran-
scription occurs and is facilitated by HIV-1 Tat
and Rev (1). Both full-length viral and crHIV vec-
tor mRNAs are made. Any antiviral moieties such
as siRNAs would also be expressed at this stage
(2). Next, vector and viral RNA are either trans-
lated or processed and packaged (5) into the
forming virion (6). Finally, three possible species
of virions would be expected to be produced from
this vector and virus-containing cell: (i) either con-
taining two viral copies (HIV), (ii) one crHIV vec-
tor and one viral RNA (Vector/HIV) or (iii) two
crHIV vector copies (Vector).
immune recognition, or seroconversion,
due to exposure to structural proteins.
Moreover, in certain cell types such as
stem cells, HIV-based vectors are markedly
enhanced in marker gene expression rela-
tive to FIV vectors [74]. Consequently, FIV
packaged HIV-1 vectors could be used to
both reduce seroconversion or immune re-
cognition while simultaneously maintain-
ing the highest level of transgene expres-
sion within the transduced HSC. Finally,
lentiviral vectors can be pseudotyped to tar-
get specific cell types [79, 80] or alterna-
tively designed with a receptorligand
bridge to target specific cell types [81].
Fig. 8.3 Production of lentiviral vectors. Lentiviral
vectors are produced by (1) transfecting 293 pro-
ducer cells with the lentiviral vector, lentiviral
packaging plasmid and the envelope plasmid.
Next, the transfected cell transcribes the respec-
tive plasmids (2 and 3) subsequently producing
the packaging cofactors (4) and vector RNA which
is then packaged into the budding particles (4).
Between 48 and 72 h later the culture superna-
tants are collected and vector concentration deter-
mined by tittering on target cells.
8.5
Challenges for RNA-based Therapies
A major advantage to using ribozymes
and/or siRNAs to treat HIV-1 is the rela-
tive ease of design, construction and test-
ing. The emerging field of RNAi (siRNAs,
in particular) provides a potentially cost-ef-
fective and relatively quick methodology to
treat or augment current drug regimens
for HIV-1 infection. There are, however,
some important issues which remain to be
addressed before the use of RNA-based
therapy can act as a realized therapeutic
for HIV-1. These constraints are the avoid-
ance of off-target effects, delivery of the
siRNA to the target cell and targeting in a
manner that avoids the emergence of viral-
resistant strains.
8.6
Summary and Conclusion
Steady progress has been made with re-
gard to gene-therapy-based delivery sys-
tems such as the lentiviral vector-based
system, but much more work will be re-
quired prior to moving into a clinical set-
ting. Regarding off-target effects, the use
of siRNAs to target specific cellular or viral
transcripts relies essentially on hijacking
the endogenous RNAi machinery, of which
we know very little, i.e., what is the poten-
tial for saturating the RNAi pathway? In-
deed there is evidence that RISC can be
saturated at least in the context of cultured
cells [82]. Consequently, endogenous RNAi
pathways appear to be susceptible to high
concentrations of exogenous siRNA sug-
gesting that it will probably be imperative
to not only quantitate siRNA-mediated si-
lencing, but to also monitor other genes
for untoward off-target effects (see Part III,
Chapter 3). Finally, to avoid the emergence
of HIV-1 resistance [3], a multitargeting
approach should be taken. Importantly,
anti-HIV-1 ribozymes either alone or as
the shRNA loop [59] could be incorporated
along with multiple siRNAs targeting the
most conserved regions of HIV-1 as well
as splice junctions. Moreover, Tat or Rev
decoys [83] (see also Part I, Chapter 9)
could also be incorporated as well as siR-
NAs targeting the viral coreceptor CCR5
[49]. Indeed lentiviral vectors can accom-
modate a roughly 6.5-kb payload [84] offer-
ing a promising delivery vehicle for RNA-
based modalities such as siRNAs and ribo-
zymes to be used as biopharmaceuticals in
the treatment HIV-1 infections.
References
1 Idemyor, V. 2003. 20 years since human im-
munodeficiency virus discovery: considerations
for the next decade. Pharmacotherapy 23, 384
387.
2 Ho, D. D., Y. Huang. 2002. The HIV-1 vaccine
race. Cell 110, 135138.
3 DAquila, R. T., J. M. Schapiro, et al. 2003.
Drug resistance mutations in HIV-1. Topics
HIV Med 11 (MayJune), 9296.
4 Maggiolo, F., D. Ripamonti, et al. 2003. Once-
a-day HAART: dream or reality? HIV Clin
Trials 4, 193201.
5 Kruger, K., P. J. Grabowski, et al. 1982. Self-
splicing RNA: autoexcision and autocycliza-
tion of the ribosomal RNA intervening se-
quence of Tetrahymena. Cell 31, 147157.
6 Guerrier-Takada, C., K. Gardiner, et al. 1983.
The RNA moiety of ribonuclease P is the cata-
lytic subunit of the enzyme. Cell 35, 849857.
7 Cech, T. R. 1990. Self-splicing of group I in-
trons. Annu Rev Biochem 59, 543568.
8 Marschall, P., J. B. Thomson, et al. 1994. Inhi-
bition of gene expression with ribozymes. Cell
Mol Neurobiol 14, 523538.
9 Amarzguioui, M., H. Prydz. 1998. Hammer-
head ribozyme design and application. Cell
Mol Life Sci 54, 11751202.
10 James, H. A., I. Gibson. 1998. The therapeutic
potential of ribozymes. Blood 91, 371382.
References 577
11 Zhou, C., I. C. Bahner, et al. 1994. Inhibition
of HIV-1 human T-lymphocytes by retrovirally
transduced anti-tat and rev hammerhead ribo-
zymes. Gene 149, 3339.
12 Yamada, O., M. Yu, et al. 1994. Intracellular
immunization of human T cells with a hair-
pin Ribozyme against HIV-1. Gene Ther 1, 38.
13 Akhtar, S., H. James, et al. 1995. Molecular
DIY with hairpins and hammerheads. Nat
Med 1, 300302.
14 Macpherson, J. L., J. A. Ely, et al. 1999. Ribo-
zymes in gene therapy of HIV-1. Front Biosci
4, 497505.
15 Rossi, J. J. 2000. Ribozyme therapy for HIV in-
fection. Adv Drug Deliv Rev 44, 7178.
16 Sioud, M., K. Drlica. 1991. Prevention of HIV-1
integrase expression in Escherichia coli by a ri-
bozyme. Proc Natl Acad Sci USA 88, 73037307.
17 Dropulic, B., N. H. Lin, et al. 1992. Functional
characterization of a U5 ribozyme: intracellu-
lar suppression of HIV-1 expression. J Virol
66, 14321441.
18 Leavitt, M. C., M. Yu, et al. 1994. Transfer of an
anti-HIV-1 ribozyme gene into primary human
lymphocytes. Human Gene Ther 5, 11151120.
19 Sun, L. Q., L. Wang, et al. 1995. Target se-
quence-specific inhibition of HIV-1 replication
by ribozymes directed to tat RNA. Nucleic
Acids Res 23, 29092913.
20 Michienzi, A., S. Li, et al. 2002. A nucleolar
TAR decoy inhibitor of HIV-1 replication. Proc
21 Tijsterman, M., R. F. Ketting, et al. 2002. The
genetics of RNA silencing. Annu Rev Genet
36, 489519.
22 Montgomery, M. K., S. Xu, et al. 1998. RNA as
a target of double-stranded RNA-mediated ge-
netic interference in Caenorhabditis elegans.
23 Nishikura, K. 2001. A short primer on RNAi:
RNA-directed RNA polymerase acts as a key
catalyst. Cell 107, 415418.
24 Sharp, P. A. 2001. RNA interference. Genes
Dev 15, 485490.
25 Sui, G., C. Soohoo, et al. 2002. A DNA vector-
based RNAi technology to suppress gene ex-
pression in mammalian cells. Proc Natl Acad
Sci USA 99, 55155520.
26 Martinez, J., A. Patkaniowska, et al. 2002. Sin-
gle-stranded antisense siRNAs guide target
RNA cleavage in RNAi. Cell 110, 563574.
27 Sumimoto, H., M. Miyagishi, et al. 2003. De-
velopment of an efficient small interfering
RNA (siRNA) expression system with a lenti-
viral vector. Mol Ther 7, S34.
28 Hamada, M., T. Ohtsuka, et al. 2002. Effects
on RNA interference in gene expression
(RNAi) in cultured mammalian cells of mis-
matches and the introduction of chemical
modifications at the 30162-ends of siRNAs.
Antisense Nucleic Acid Drug Dev 12, 301309.
29 Liu, J., M. A. Carmell, et al. 2004. Argonaute2
is the catalytic engine of mammalian RNAi.
Science 305, 14371441.
30 Sijen, T., I. Vign, et al. 2001. Transcriptional
and posttranscriptional gene silencing are me-
chanistically related. Curr Biol 11, 436440.
31 Pal-Bhadra, M., U. Bhadra, et al. 2002. RNAi
related mechanisms affect both transcriptional
and posttranscriptional transgene silencing in
Drosophila. Mol Cell 9, 315327.
32 Bosher, J. M., P. Dugourcq, et al. 1999. RNA
interference can target pre-mRNA: conse-
quences for gene expression in a Caenorhabdi-
tis elegans operon. Genetics 153, 12451256.
33 Wassenegger, M. 2000. RNA-directed DNA
methylation. Plant Mol Biol 43, 203220.
34 Wassenegger, M., M. W. Graham, et al. 1994.
RNA-directed de novo methylation of genomic
sequences in plants. Cell 76, 567576.
35 Zeng, Y., B. R. Cullen. 2002. RNA interference
in human cells is restricted to the cytoplasm.
RNA 8, 855860.
36 Fire, A., S. Xu, et al. 1998. Potent and specific
genetic interference by double stranded RNA
in Caenorhabditis elegans. Nature 391, 806811.
37 Ngo, H., C. Tschudi, et al. 1998. Double-
stranded RNA induces mRNA degradation in
Trypanosoma brucei. Proc Natl Acad Sci USA
95, 1468714692.
38 Volpe, T. A., C. Kidner, et al. 2002. Regulation
of heterochromatic silencing and histone H3
lysine-9 methylation by RNAi. Science 297,
18331837.
39 Kawasaki, H., K. Taira. 2004. Induction of
DNA methylation and gene silencing by short
interfering RNAs in human cells. Nature 431,
211217.
40 Morris, K. V., C. Chung, et al. 2004. Inhibition
of HIV-1 replication by siRNA targeting con-
served regions of gag/pol. RNA Biol 1, 114
117.
41 Kawasaki, H., K. Taira, et al. 2005. siRNA in-
duced transcriptional gene silencing in mam-
malian cells. Cell Cycle 4(3), in press.
42 Coburn, G. A., B. R. Cullen. 2002. Potent and
specific inhibition of human immunodefi-
ciency virus type 1 replication by RNA inter-
ference. J Virol 76, 92259231.
43 Lee, N. S., T. Dohjima, et al. 2002. Expression
of small interfering RNAs targeted against
HIV-1 rev transcripts in human cells. Nat Bio-
technol 19, 500505.
44 Novina, C. D., M. F. Murray, et al. 2002. si-
RNA-directed inhibition of HIV-1 infection.
Nat Med 8, 681686.
45 Surabhi, R. M., R. B. Gaynor. 2002. RNA inter-
ference directed against viral and cellular tar-
gets inhibits human immunodeficiency virus
type 1 replication. J Virol 76, 1296312973.
46 Morris, K. V., J. Gilbert, et al. 2004. Transduc-
tion of cell lines and primary cells by FIV-
packaged HIV vectors. Mol Ther 10, 181190.
47 Morris, K. V., S. W. Chan, et al. 2004. Small in-
terfering RNA-induced transcriptional gene si-
lencing in human cells. Science 305, 5688,
12891292.
48 Jacque, J., K. Triques, et al. 2002. Modulation
of HIV-1 replication by RNA interference. Na-
ture 26, 14.
49 Qin, X., D. An, et al. 2002. Inhibiting HIV-1
infection in human T cells by lentiviral
mediated delivery of small interfering RNA
against CCR5. Proc Natl Acad Sci USA 100,
183188.
50 Lee, N. S., J. J. Rossi. 2004. Control of HIV-1
replication by RNA interference. Virus Res
102, 5358.
51 Llave, C., K. D. Kasschau, et al. 2000. Virus-en-
coded suppressor of posttranscriptional gene
silencing targets a maintenance step in the si-
lencing pathway. Proc Natl Acad Sci USA 97,
1340113406.
52 Johansen, L. K., J. C. Carrington. 2001. Silenc-
ing on the spot. Induction and suppression of
RNA silencing in the Agrobacterium-mediated
transient expression system. Plant Physiol 126,
930938.
53 Mallory, A. C., L. Ely, et al. 2001. HC-Pro sup-
pression of transgene silencing eliminates the
small RNAs but not transgene methylation or
the mobile signal. Plant Cell 13, 571583.
54 Hamilton, A., O. Voinnet, et al. 2002. Two
classes of short interfering RNA in RNA silen-
cing. EMBO J 21, 46714679.
55 Li, H., W. X. Li, et al. 2002. Induction and sup-
pression of RNA silencing by an animal virus.
Science 296, 13191321.
56 Mallory, A. C., B. J. Reinhart, et al. 2002. A
viral suppressor of RNA silencing differen-
tially regulates the accumulation of short in-
terfering RNAs and micro-RNAs in tobacco.
57 Gitlin, L., S. Karelsky, et al. 2002. Short inter-
fering RNA confers intracellular antiviral im-
munity in human cells. Nature 26, 15.
58 Westerhout, E. M., M. Ooms, et al. 2005. HIV-
1 can escape from RNA interference by evolv-
ing an alternative structure in its RNA ge-
nome. Nucleic Acids Res 33, 796804.
59 Nahreini, P., A. J. Hanson, et al. 2004. Alter-
ing cellular signaling pathways enhance gene
silencing activity of shRNA, shRNA, ribo-
zyme, and shRNA antisense in neuroblastoma
cells. Cell Mol Neurobiol 24, 781792.
60 Sioud, M., D. R. Sorensen. 2003. Cationic lipo-
some-mediated delivery of siRNAs in adult
mice. Biochem Biophys Res Commun 312,
12201225.
61 Soutschek, J., A. Akinc, et al. 2004. Therapeu-
tic silencing of an endogenous gene by sys-
temic administration of modified siRNAs. Na-
ture 432, 7014, 173178.
62 Morris, M. C., P. Vidal, et al. 1997. A new pep-
tide vector for efficient delivery of oligonucleo-
tides into mammalian cells. Nucleic Acids Res
25, 27302736.
63 Ritter, W., C. Plank, et al. 2003. A novel trans-
fecting peptide comprising a tetrameric nuclear
localization sequence. J Mol Med 81, 708717.
64 Sledz, C. A., M. Holko, et al. 2003. Activation
of the interferon system by short-interfering
RNAs. Nat Cell Biol 5, 834839.
65 Pebernard, S., R. D. Iggo. 2004. Determinants
of interferon-stimulated gene induction by
RNAi vectors. Differentiation 72, 103111.
66 Persengiev, S. P., X. Zhu, et al. 2004. Nonspe-
cific, concentration-dependent stimulation and
repression of mammalian gene expression by
small interfering RNAs (siRNAs). RNA 10,
1218.
67 Wu, X., Y. Li, B. Crise, et al. 2003. Transcrip-
tion start regions in the human genome are
favored targets for MLV integration. Science
300, 17491751.
68 Buchschacher, G. L., F. Wong-Staal. 2000. De-
velopment of lentiviral vectors for gene thera-
py for human diseases. Blood 95, 24992504.
69 Greber, U. F., A. Fassati. 2003. Nuclear import
of viral DNA genomes. Traffic 4, 136143.
References 579
70 Duzgunes, N., S. Simoes, et al. 2001. Delivery
of novel macromolecular drugs against HIV-1.
Expert Opin Biol Ther 6, 949970.
71 Gervaix, A., L. Schwarz, et al. 1997. Gene ther-
apy targeting peripheral blood CD34
+
hemato-
poietic stem cells of HIV-infected individuals.
72 Yam, P. Y., S. Li, et al. 2002. Design of HIV
vectors for efficient gene delivery into human
hematopoietic cells. Mol Ther 5, 479484.
73 Poeschla, E., P. Corbeau, et al. 1996. Develop-
ment of HIV vectors for anti-HIV gene thera-
py. Proc Natl Acad Sci USA 93, 1139511399.
74 Price, M. A., S. S. Case, et al. 2002. Expression
from second-generation feline immunodefi-
ciency virus vectors is impaired in human he-
matopoietic cells. Mol Ther 6, 645652.
75 Quinonez, R., R. E. Sutton. 2002. Lentiviral
vectors for gene delivery into cells. DNA Cell
Biol 12, 937951.
76 White, S. M., et al. 1999. Lentivirus vectors
using human and simian immunodeficiency
virus elements. J Virol 73, 28322840.
77 Browning, M. T., R. D. Schmidt, et al. 2001.
Primate and feline lentivirus vector RNA
packaging and propagation by heterologous
lentivirus virions. J Virol 75, 51295140.
78 Goujon, C., L. Jarrosson-Wuilleme, et al. 2003.
Heterologous human immunodeficiency virus
type 1 lentiviral vectors packaging a simian
immunodeficiency virus-derived genome dis-
play a specific postentry transduction defect in
dendritic cells. J Virol 787, 92959304.
79 Kobinger, G. P., D. J. Weiner, et al. 2001. Filo-
virus-pseudotyped lentiviral vector can effi-
ciently and stably transduce airway epithelia
in vivo. Nat Biotechnol 19, 225230.
80 Sandrin, V., Russell, S. J., et al. 2003. Target-
ing retroviral and lentiviral vectors. Curr Topics
Microbiol Immunol 281, 137178.
81 Boerger, A. L., S. Snitkovsky, et al. 1999. Retro-
viral vectors preloaded with a viral receptorli-
gand bridge protein are targeted to specific
cell types. Proc Natl Acad Sci USA 96, 9867
9872.
82 Pasquinelli, A. E. 2002. MicroRNAs: deviants
no longer. Trends Genet 18, 171173.
83 Michienzi, A., S. Prislei, et al. 1996. U1 small
nuclear RNA chimeric ribozymes with sub-
strate specificity for Rev pre-mRNA of HIV.
84 Wolkowicz, R., G. P. Nolan, et al. 2004. Lenti-
viral vectors for the delivery of DNA into
mammalian cells. Methods Mol Biol 246, 391
411.
85 Lee, Y., C. Ahn, et al. 2003. The nuclear
RNase III Drosha initiates microRNA process-
ing. Nature 425, 415419.
86 Lund, E., S. Guttinger, et al. 2004. Nuclear ex-
port of microRNA precursors. Science 303, 95
98.
ISBN: 3-527-31184-X
Part III
Improving the Development of Biopharmaceuticals
Abstract
The development of modern biopharmaceu-
ticals increasingly relies on protein engi-
neering to tailor pharmacological character-
istics. Such approaches are, however, limited
by the enormous complexity inherent in
protein libraries. Current directed evolution
approaches therefore tackle both sides of the
problem: the throughput in the screening of
protein variants while maintaining pharma-
cologically relevant conditions and the de-
sign of protein libraries to enhance the
chances of finding improved variants. Prop-
er design of molecular diversity allows effi-
cient exploration of protein sequence space
through the selective application of random
or selective mutagenesis techniques with
homologous recombination. The technical
capabilities that are required to screen large
libraries under pharmacologically relevant
conditions include the ability to handle large
numbers of liquid samples in the submicro-
liter range and to provide the infrastructure
to process the data resulting from such
miniaturized screens, in order to reliably
rank protein phenotypes. The integration
of these fundamental capabilities into a cy-
clic process enables the engineering of pro-
teins even for ambitious tasks, e.g., the en-
gineering of the substrate specificity of a
protease to hydrolyse a pharmacologically
relevant target sequence.
Abbreviations
BAR barrier proteinase (Saccharomyces
cerevisiae)
DSPA Desmodus (rotundus) salivary plas-
minogen activator
ETM enhanced thematic mapper
RCR recombination chain reaction
SCR selective combinatorial randomi-
zation
TEM-1 baterial enzyme b-lactamase
TNK tenecteplase
t-PA tissue plasminogen activator
uHTS ultra-high-throughput screening
583
1
Design of Modern Biopharmaceuticals
by Ultra-high-throughput Screening and Directed Evolution
Markus Rarbach, Wayne M. Coco, Andre Koltermann, Ulrich Kettling, and Manfred Eigen
ISBN: 3-527-31184-X
Citius, Altius, Fortius Acceleration by High Throughput and Ultra-HT
1.1
Biopharmaceuticals suffer from a wide-
spread imbalance between perception and
facts. Some perceive biopharmaceuticals as
new, exotic and not well understood, de-
spite the fact that the first recombinant
therapeutic protein was approved by the
FDA more than two decades ago (Eli Lillys
insulin Humulin, approved in 1982) and
approximately 100 other approved biophar-
maceutical products have since followed
[1] (see also the Introduction to this vol-
ume and Part VII, Chapter 4). Indeed, sev-
eral biologics have now been released as
second- and even third-generation prod-
ucts that exhibit improved efficacy, fewer
side-effects and better production effi-
ciency (see also Part II, Chapter 3 and Part
IV, Chapter 13). Moreover, the first bio-
pharmaceutical products with established
applications and markets are coming off
patent, and are now being challenged in
their respective markets by follow-on bio-
logics and biogenerics (see also Part VII,
Chapter 5 and Part VIII, Chapter 3). With
the above in mind, we should rather think
of biopharmaceuticals as a classical class
of drugs with established risk profiles and
proven, often remarkable efficacies. In
short, biologics have all the characteristics
of a mature pharmaceutical platform.
So what makes modern biopharmaceuti-
cals modern? What has actually changed
since the inception of recombinant protein
drugs? The answer has to do with the
growing understanding of how biologics
work on the molecular level. This under-
standing has led to the development of a
broad set of powerful tools that enable pro-
tein engineers to generate tailor-made bio-
pharmaceuticals (see also Part V, Chapter
2). In the past, protein drugs were cloned
and manufactured more or less as nature
designed them. A well-known example of
this early biopharmaceutical approach is
the above-mentioned insulin. Other exam-
ples include erythropoietin, growth factors
and the interferons. This limitation, how-
ever, has now been widely overcome. Mod-
ern biopharmaceuticals are now custo-
mized to provide particular mechanisms
of action; tools have been developed that
allow biochemical diversification of recom-
binant proteins in a manner that is much
more versatile than is possible with combi-
natorial chemistry or any other class of
pharmaceuticals. The difference between
early and modern biopharmaceuticals is
profound we are no longer forced to dis-
cover a compound with a particular activity,
but are instead able to intentionally develop
it. The hunter becomes a craftsman.
The first examples of this paradigm
change, i.e., therapeutic monoclonal anti-
bodies, have already altered modern medi-
cine. Monoclonal antibodies used as thera-
peutics became a multibillion dollar busi-
ness in recent years and proved their po-
tential to efficiently treat diseases for
which no effective treatment option had
yet existed (see also Part I, Chapter 5, Part
IV, Chapter 16 and Part V, Chapter 1). The
reason why monoclonal antibodies pro-
vided the first platform technology of tai-
lor-made biopharmaceuticals lies in the
fact that nature provides a tool for their
engineering the immune system. Anti-
bodies arose early in the evolution of verte-
brates as one of natures ingenious plat-
form technologies to fight diseases of all
kinds. In addition to their obvious impor-
tance in infectious diseases, it has since
become clear that antibodies are also im-
portant in the defense against disease
states in which the human body is con-
fronted with abnormal self-derived struc-
tures as in the case of cancer. After
learning about the fascinating nature of
1 Design of Modern Biopharmaceuticals by Ultra-high-throughput Screening and Directed Evolution 584
the immune system and the structure of
antibodies, we were able to produce mono-
clonal antibodies in large quantities and of
high purity. The antibodies themselves,
however, were, and in many cases still are,
developed by an animals immune system.
More recently, we also learned to modify
and artificially develop monoclonal anti-
bodies without the help of the natural im-
mune system (see also Part V, Chapter 2).
By the use of screening technologies and
combinatorial mutagenesis, we are now
able to efficiently modify antibody affinity,
cross-reactivity or serum half-life.
While antibody technology relies on one
particular molecular scaffold (the immuno-
globulin fold) and its specific functional-
ities, more and more protein engineering
is also done with non-immunoglobulin
structures. Together, these represent the
future modern biopharmaceuticals: man-
made therapeutic proteins with tailored
functionalities. Such functionalities range
from high binding affinities to highly spe-
cific catalytic activities (see also Part III,
Chapter 6). The protein engineering tool
box comprises discovery and engineering
tools to develop new molecular scaffolds,
tools to analyze and exploit structurefunc-
tion relationships, directed evolution tools
to mimic natures way of generating highly
complex libraries, and high-throughput
and high-content screening tools (see also
Part III, Chapter 2). In particular, directed
evolution strategies combined with high-
throughput and high-content screening
technologies provide the basis for generat-
ing proteins that will be the future block-
buster drugs.
1.2
Directed Evolution Fundamentals
The term directed evolution denotes the lab-
oratory reproduction of the fundamental
processes of natural evolution at the mo-
lecular level, i.e., mutation, selection and
amplification. Directed refers to the appli-
cation of molecular evolution to direct the
characteristics of biomolecules towards in-
tended characteristics. In natural evolu-
tion, a myriad of biomolecular species
have been selected to perform specified
tasks in an intricately coupled and con-
trolled interplay of molecular interactions.
(Biomolecules or, from a different perspec-
tive, their respective genes, may be consid-
ered as the evolving species and their host
organisms only as their vehicles; or, in Sol
Spiegelmans famous words . . . one might
have wondered why DNA invented man. It
is now evident that man was invented to
provide DNA with the opportunity to ex-
plore also extraterrestrial possibilities for
replication [2].) While nature selected so-
lutions for natures challenges, it will not
necessarily provide optimal solutions for
man-made tasks. Nevertheless, using na-
tures strategies to solve problems is still
good advice. Using natures strategies in
directed evolution means establishing li-
braries of molecular variants of a given
biomolecule, producing and amplifying
them in the test tube, and selecting vari-
ants according to their fitness to perform a
given task. In principle, the above scheme
is not restricted to a certain chemical class.
However, proteins have proven to be the
most versatile class: first of all because in-
formation stored at the DNA level can be
both relatively simply replicated and trans-
lated into proteins, and, secondly, because
the large set and chemical diversity of
building blocks (i.e., the 20 natural amino
acid residues) provide an unimaginably
large combinatorial diversity of functional
groups that can be precisely oriented in
three-dimensional structures to carry out
highly diverse functions such as activating
a cell-surface receptor or preferentially cat-
alyzing a specific chemical reaction. Final-
ly, using natures tools to tailor a biomole-
cule does not require extensive prior
knowledge of the molecules structure and
functions. Such knowledge can certainly
be helpful, particularly in the design and
implementation of particular libraries, but
is by no means a prerequisite.
The establishment of a productive evolu-
tionary cycle necessitates a combination of
techniques from several scientific disci-
plines, the most important of which are
molecular biology, biochemistry, physics,
engineering sciences and informatics. Mo-
lecular biology provides the tools to gener-
ate libraries of protein variants by intro-
ducing variations on the DNA level. It also
provides the necessary expression systems
and the tools needed to determine and to
leverage the created genotypic alterations
in successive rounds of library creation.
The methods employed to generate these
libraries are carefully chosen to maximize
the chances of finding optimal variants for
the particular protein of interest. Biochem-
istry provides the capabilities to design
customized assays for a wide range of pro-
tein targets. Such assays ultimately provide
the linkage between a variants molecular
phenotype (i.e., the sum of its properties)
and a measurable signal or physical modi-
fication that allows for the selection of op-
timized variants. Engineering sciences es-
tablish the technical means to screen large
numbers of phenotypes under application-
relevant conditions in order to select the
right variants for the given task. While
automation is a prerequisite for an effi-
cient screening process, modern laboratory
automation can additionally be employed
for other tasks, such as amplifying, purify-
ing and modifying DNA, in order to gain
a higher degree of reproducibility and to
speed up the process. Finally, state-of-the-
art informatics and modern computer
hardware allow for the storage and proces-
sing of enormous amounts of data gener-
ated in screening-based directed evolution.
Although any of these disciplines may
allow accomplishment of the correspond-
ing individual tasks, the integration of
each step in a productive evolutionary opti-
mization process requires the logistics to
coordinate all these tasks. The combina-
tion of all these elements in a cyclic pro-
cess enables the generation of optimized
protein variants in a relatively short time,
even for ambitious tasks.
1.3
Generation of Protein Diversity
Our understanding that mutations can im-
prove an organisms phenotype gained in-
creasing acceptance in the decades follow-
ing the work of Gregor Mendel. However,
in vitro mutagenesis of proteins (through
their respective genes) was only made pos-
sible after the structure of DNA was dis-
covered, the genetic code was determined
and basic DNA manipulation technologies
became available. Methods for oligonucleo-
tide-mediated site-directed mutagenesis
were established in the late 1970s [3, 4], as
were methods for random mutagenesis of
limited regions in a template or plasmid
of interest [5, 6]. The term protein engi-
neering was soon coined (see, e.g., [7])
and by 1986 a journal was founded specifi-
cally dedicated to this by then already im-
portant technology [8]. Since then, a re-
markable number of methods have been
developed for the in vitro diversification of
proteins, each with their individual advan-
tages and limitations.
Expression systems are a prerequisite
for the generation and exploitation of pro-
tein diversity. This is because we still lack
the ability to copy peptide sequences in
the same way as we are able to copy nu-
cleic acid molecules. For this reason, any
changes in a proteins structure are first
made at the DNA level and are then trans-
lated into the corresponding polypeptide.
Only upon synthesis can the linear poly-
peptide fold into the active three-dimen-
sional protein structure which can be
tested for a particular phenotype. The
choice of expression system for a given
protein is determined by many factors.
First, the protein needs to be produced in
its native conformation at an expression
level that makes reliable measurement of
the proteins phenotype technically feasi-
ble. In addition to in vitro systems, expres-
sion hosts capable of secreting proteins to
the growth medium or displaying them on
their surface are generally preferred be-
cause this markedly reduces the technical
complexity of the selection or screening
processes. Some proteins require post-
translational modifications, which limits
the range of suitable expression systems
further. Finally, where expression host or-
ganisms are used, suitable transformation
protocols are required that generate 10
6
or
more transformants, in order to establish
libraries large enough for efficient screen-
ing and/or selection processes (see also
Part V, Chapter 2). Frequently used expres-
sion systems satisfying the above require-
ments include bacteria such as Escherichia
coli, Pseudomonas sp. or Bacillus sp., yeasts
such as Saccharomyces cerevisiae (see also
Part IV, Chapters 12 and 13), insect cells
(see also Part IV, Chapter 14), and, particu-
larly when post-translational modifications
are critical, mammalian cell lines (see also
Part IV, Chapters 14). In addition, phages
or viruses may be used in combination
with several of these hosts, the most com-
mon combination being E. coli filamen-
tous phages such as M13 or insect cell
viruses like baculovirus.
The choice of the library strategy is one
of the most crucial decision points in any
directed evolution project. The library
strategy determines the composition of
protein variants present in the libraries as
well as the number of possible mutants
and mutations. The number of possible
mutants is termed the complexity of a li-
brary. Table 1.1 summarizes the complexi-
ties of protein libraries designed by differ-
ent strategies. As is obvious from these
numbers, the complexities of simple pro-
tein libraries easily surpass by orders of
magnitudes the number of variants that
can be generated and technically evaluated.
This imbalance is known as the complex-
ity problem in protein evolution. A simple
calculation demonstrates the problem. The
number of all possible variants of a 200
amino acid protein is approximately 10
260
.
The enormous dimension of this theoreti-
cal number can be better grasped when
comparing it with real figures. For exam-
ple, if the mass of the entire (currently
known) universe were to consist of such
proteins, this would amount to only ap-
proximately 10
75
molecules. It is obvious
that evolution on earth, even in 4 billion
years (which is only approximately 10
17
s),
could not have evaluated all possible pro-
tein variants to come up with those mil-
lions of proteins that form the basis of to-
days living cells. Natural evolution works
by selecting variants with improved fitness
from smaller populations of closely related
variants within a species. Selection pres-
sure favors a subset of variants of such a
quasispecies over competing variants and
confers a higher replication rate to these
variants, which eventually leads to a shift
of the quasispecies towards a distinct phe-
notype (for a detailed description of the
quasispecies concept, see [9]).
Common strategies for the generation of
protein libraries, when no useful informa-
tion about the proteins structure and func-
tion is available, are based on the introduc-
tion of random mutations at random posi-
tions in the gene. Nowadays, the technical
protocols for such random mutagenesis
methods have become rather simple and a
large number of protocols are available.
This includes, among others, chemical
mutagenesis [10], polymerase chain reac-
tion (PCR) mutagenesis with or without
nucleotide analogs [11, 12], ensemble mu-
tagenesis [13], spiked oligonucleotide mu-
tagenesis [14], complete randomization of
contiguous codons [15] as well as muta-
genesis in mutator strains [1618]. How-
ever, for all these protocols the most chal-
lenging question remains: what is the opti-
mal mutation frequency? That is, on aver-
age, how many mutations per sequence
length should be introduced?
In order to answer this question, it is
again helpful to review natures strategies
for efficient evolutionary adaptation. In
natural evolution, the entities with the
fastest adaptation to changing environ-
mental conditions are viruses. Presumably,
viruses developed this extreme adaptability
to escape eradication by the often quickly
changing defense mechanisms of their
hosts. Looking at the molecular mecha-
nisms that confer this ability, it becomes
apparent that replication of the viral ge-
nome operates at very high mutation rates
[19]. The high variability within viral qua-
sispecies increases the probability of find-
ing a mutated variant that will escape
elimination by the response of the host or-
ganism. Viral infection can be considered
a competition between co-evolving sys-
tems, i.e., the virus and the collective or
individual repertoire of the hosts defense
mechanisms. In order to adapt as quickly
as possible, viruses operate at the highest
feasible mutation rate. There is, however,
an upper limit for the mutation rate.
Above this upper limit, the information
stored in the variants genotype is lost
among the too many errors introduced in
each generation; inheritance of informa-
tion from parents to offspring becomes
impossible at mutation rates larger than
this species-specific error threshold. The
most efficient random mutagenesis proto-
cols imitate this strategy and operate also
Table 1.1 Combinatorial complexity in protein diversity
Complexity
Protein library
(200 amino acids)
no. of alanine scan mutants 210
2
no. of all mutants with Hamming distance 1 410
3
7
10
no. of all possible mutants 210
260
Reference no. of bacterial cells in a 1.5-ml tube 110
9
no. of proteins in a 1.5-ml tube (no water left) 810
19
no. of proteins equal to mass of the Earth 210
47
no. of proteins equal to mass of the universe 210
75
at error rates close to the error threshold
[9, 20]. As with naturally evolving species,
the error threshold of protein libraries is
defined as the error rate above which the
functionality of the protein is destroyed
over consecutive directed evolution rounds.
Therefore, the error threshold is not a
fixed rate, but is dependent on the se-
quence length and on the fitness land-
scape, which in turn is dependent on the
selection conditions and the nature of each
individual protein. Practically, the error
threshold is determined by generating pro-
tein libraries at different error rates and,
by screening the library phenotypes, deter-
mining the error rate above which the
functionality of the protein is lost.
There are numerous examples demon-
strating the most efficient evolutionary
adaptation at the highest mutagenesis
rates. Daugherty et al., for example, ob-
served in a model experiment for directed
evolution of antibodies that the antibody
variants with highest affinity were found
under the highest mutation frequency
used in the study [21]. At this error rate
the proportion of functional clones in the
library was well below 1%. The authors
conclude that exploring a larger region in
sequence space is more efficient in the
evolution of a protein than evaluating a
small portion of sequence space with a
larger number of mutants. This conclu-
sion is consistent with the error threshold
strategy. Additionally, Daugherty et al. re-
port the rather common finding in direct-
ed evolution experiments that a large pro-
portion of beneficial mutations are distant
from the antigen binding site of the anti-
body, suggesting a complex and poorly un-
derstood influence of these sites on the
function of the protein.
Zaccolo et al. [22] followed a similar
approach in their study of high-frequency
random mutagenesis of TEM-1 b-lacta-
mase and selection of variants from com-
parably small pools of less than 10
5
geno-
types. Clones with improved resistance to
the b-lactam antibiotic cefotaxime were se-
lected by an amplification-coupled selec-
tion strategy (see below). Variants confer-
ring highest resistance (20000-fold im-
provement compared to wild-type) were
found under the highest mutation rates.
The total improvement found was compar-
able to that found in a study by Stemmer
involving random mutagenesis combined
with DNA shuffling (see also Part VIII,
Chapter 5) of mutated positions [23]. Anal-
ysis of variants by Zaccolo et al. showed
that the best clones exhibited synergistic
mutations. These data lead to the same
conclusion: under high mutagenesis rates,
portions of the sequence space are ex-
plored that are inaccessible by other meth-
ods even if only a small fraction of the
theoretical complexity is screened.
Randomization, however, does not al-
ways imply random mutation positions.
Where there is prior knowledge of the pro-
teins structure or specific functional resi-
dues (e.g., the active site of an enzyme or
the ligand-interaction site of a receptor),
this information can be used for identify-
ing sites for selective randomization (see
also Part III, Chapter 6). Several protocols
have been established that allow the rando-
mization of such sites or regions. Most
protocols are based on oligonucleotides
bearing the respective randomized posi-
tions. For an application of such protocols,
see elsewhere (e.g., [24]) (see also Part V,
Chapter 2).
In order to further increase efficiency of
evolutionary optimization projects, random
mutagenesis is typically combined with
two alternative strategies: (1) homologous
recombination of mutated positions and
(2) selective randomization of previously
mutated positions (see Fig. 1.1 for illustra-
tion). From a quasispecies perspective, ho-
mologous recombination allows muta-
tional short-cuts between different vari-
ants, thereby increasing the connectivity
within a quasispecies distribution. Selec-
tive randomization, on the other hand, ac-
celerates relocation of a quasispecies distri-
bution by artificially increasing the error
rate and thus the potential fitness at pre-
selected hot spots in a given sequence.
The first hints that homologous recom-
bination can beneficially be employed for
protein engineering date back into the
1980s, when several groups identified pro-
tein variants with improved activities from
small sets of in vitro recombined chimeric
genes containing one to three crossovers
[25, 26]. In addition to the development by
Pompon and Nicolas of in vivo family
shuffling in 1989 [27], beginning in 1989
and continued in the 1990s, several PCR
methods were implemented that used
gene fragments as templates, thereby caus-
ing in vitro recombination events leading
to shuffled clones [23, 2830], a strategy
that is now commonly referred to as DNA
shuffling (see also Part VIII, Chapter 5).
Other DNA shuffling methods employ
shortened elongation times to pause or ter-
minate polymerization before elongation
reaches the end of a PCR template (e.g.,
StEP) [31], the annealing and flap trim-
ming of gene fragments on linear tem-
plates (RACHITT) [32], oligonucleotide-
mediated polymorphism recombination
(e.g., DHR) [33], in vivo repair of in vitro
generated heteroduplexes [34] or exonu-
clease digestion and extension reactions
on heteroduplexed genes: recombination
chain reaction (RCR) (see also Fig. 1.2)
Fig. 1.1 Generation of protein diversity. Starting
from a parent genotype, mutations are introduced
by random mutagenesis (ETM) followed by
screening of phenotypically improved variants.
Sets of these mutations can be randomly recom-
bined by homologous recombination, e.g., by the
RCR, to identify variants with combinations of
beneficial mutations. Alternatively, positions at
which beneficial mutations were introduced by
random mutagenesis can be selectively random-
ized, thereby allowing identification of the optimal
amino acid residue at each position. SCR enables
such site-specific randomization without requiring
time-consuming sequence analysis and oligonu-
cleotide synthesis.
[35]. While all these protocols lead to ran-
dom recombination of a set of parental
genes, the protocols vary in their ability to
work on DNA fragments of different
lengths, in their homology requirements,
and how accurately they can be tuned in
terms of recombination frequency and re-
combination sites within a gene. The RCR
protocol, for example, allows the practi-
tioner due to its cyclic nature to adjust
both the number of recombination events
and the recombination sites (see Fig. 1.2
for details). The experimental choice of re-
combination sites and events becomes val-
uable when, for example, sets of synergis-
tic mutations have already been identified
in the gene, or when elementary protein
domains should not be disrupted to pre-
serve protein function. Disruption of sche-
mata by random recombination protocols
has been found to be deleterious to li-
braries as for example shown by Voigt et
al. [36]. In terms of evolutionary theory,
RCR can be experimentally tuned to en-
rich successful schemata in the library of
protein variants. In addition to the recom-
bination of variants carrying different mu-
tations, recombination can also be em-
ployed for backcrossing of protein variants
with wild-type genes. This allows the iden-
tification of optimal subsets of mutations
in a variant without the need for sequence
Fig. 1.2 Homologous recombination by RCR. (a)
Random recombination protocol. Initially, the gen-
otypes of all parental variants are pooled, and het-
eroduplexes are formed by melting and subse-
quent re-annealing. In a second step, double-
stranded heteroduplexes are partially degraded
from the 3'-ends by a double-strand-specific exo-
nuclease like E. coli ExoIII. The resulting single-
stranded termini then serve as templates for the
re-synthesis of the degraded strands by a polymer-
ase, yielding a double-stranded gene containing a
recombination site. By iteration of these process
steps, a defined number of recombination events
can be introduced into the DNA pool. (b) Regio
selective recombination protocol. In an alternative
setup, the position of recombination sites within
a gene can also be controlled. By controlling the
exonucleolytic rate of ExoIII, the extent of DNA de-
gradation in the second step can be limited. The
introduction of site specific recombination events
within a gene is enabled by protecting the 3-termi-
nus of one strand of the parental gene is pro-
tected against exonucleolytic degradation (e.g., by
using modified PCR primers). This second option
is in especially valuable when intending to mini-
mize gene inactivation by disruption of vital sche-
mata within a protein structure.
analysis and extensive site-directed muta-
genesis.
The second commonly used strategy to
accelerate evolutionary adaptation and to
maximize progress in directed evolution
projects is the selective randomization of
positions within a gene that were pre-
viously found to be mutated in variants se-
lected from random mutagenesis libraries
(see Fig. 1.1). This strategy is especially
important when the optimization of pro-
tein properties is required using the mini-
mum number of residue substitutions,
e.g., where the potential for immunogenic-
ity is a concern. Selective randomization is
beneficial for two reasons. First, it over-
comes the mutational limitations of ex-
changing single nucleotides instead of co-
dons. (The overwhelming majority of mu-
tations created by random mutagenesis
methods like error-prone PCR include a
single nucleotide change per mutated co-
don. This, however, enables substitution of
amino acid residues with on average only
5.6 other amino acids instead of the 19
possible amino acid residues that are pos-
sible with multiple changes per codon.)
Second, selective randomization enables
one to quickly identify the best among sev-
eral alternative beneficial mutations at a
specific position. The latter effect is in par-
ticular important when operating at high
mutation rates (see above, error threshold
mutagenesis). Library complexities in-
crease exponentially with the average num-
ber of mutations per gene. Therefore, vari-
ants selected from highly mutated libraries
usually exhibit beneficial, but frequently
not the optimal, amino acid residue substi-
tution. Selective randomization at these
positions overcomes the necessity to
further optimize this position by random
mutagenesis. The utility of selective ran-
domization has been demonstrated in sev-
eral studies. Reetz et al. [24], for example,
used random mutagenesis to find impro-
vable residues in a lipase that enhanced
enantioselectivity. Further significant im-
provements, however, were found when
previously identified mutated positions
were selectively randomized.
Practically, selective randomization is
commonly carried out by sequence analy-
sis of variants selected from mutagenesis
libraries, followed by oligonucleotide-based
randomization of the respective codons in
which mutations were found. Such a se-
quence-based approach, however, is intrin-
sically slow. Nevertheless, it is well suited
to the analysis of the effects of specific
amino acid substitutions, and thereby to
deepen our understanding of the proteins
structure and function. Efficient evolution-
ary adaptation, however, requires protocols
that are fast and sequence analysis-inde-
pendent. The selective combinatorial ran-
domization (SCR) method described by
Koltermann et al. is such a protocol [37].
Instead of identifying the relevant posi-
tions by sequence analysis, such positions
are recognized by mismatch-recognizing
enzymes after generation of heteroduplex
molecules and then, without further analy-
sis of the positions, selectively random-
ized. By running the protocol cyclically,
multiple positions from distinct parent
molecules can be randomized in a single
gene. In this fashion, randomization is
achieved in a combinatorial manner, en-
abling in a single step both the recombina-
tion of mutated positions and the identifi-
cation of the best amino acid residue at
each position (see Fig. 1.1 for illustration).
In conclusion, there is an extensive and
still growing arsenal of methods and proto-
cols available to generate diversity in protein
libraries. These technologies introduce en-
ormous flexibility and thus allow the exploi-
tation of the tremendous potential of phar-
maceutically relevant proteins. The problem
is thus not how to modify the protein struc-
ture, but how to modify it in the right way
and how to select the best variant. As we
will see below, this requires not only the
right mixture of library generation meth-
ods, but also their efficient combination
with appropriate selection strategies.
1.4
Selection Strategies
The selection of improved variants from a li-
brary of genotypes is crucial because it
guides the development of a protein toward
a certain functionality. In other words, the
precise definition and implementation of
the selection criteria comprise the most im-
portant steps in any directed evolution pro-
ject. When, for example, improper library
generation techniques are chosen, a direc-
ted evolution project will run slower and
less efficiently. In the extreme, there will
be no progress at all. However, if improper
selection processes are chosen, the directed
evolution project will run in the wrong di-
rection generating useless protein pheno-
types. Careful consideration and evaluation
of available selection processes is therefore
key to successful directed evolution projects.
In its simplest form, the fitness of a
variant can be directly associated with the
amplification of the target molecule itself,
or can confer a growth advantage to the
host organism. Such an amplification-
coupled selection strategy is most easily im-
plemented if the target protein is directly
connected to the carbon or energy metabo-
lism of the expression host as has been
demonstrated by Hall et al. [38]. In a mod-
ified approach, Dube et al. [15] utilized the
change in the death rate of the host organ-
ism to select variants of an antibiotic resis-
tance protein. As a variation of the same
principle, Duenas et al. [39] and Krebber
et al. [40] linked the binding affinity of an
antibody to the infectivity of a phage and
selected antibody variants with improved
affinity. The most significant drawback of
such selection strategies, however, is their
limited applicability. In relatively few ex-
amples can the phenotype of a protein be
directly coupled to the growth or survival
of the expression host. In addition, when
put under amplification-coupled selection
pressure, hosts often find short-cuts,
adapting to the selection pressure without
generating the intended protein pheno-
type. In conclusion, amplification-coupled
selection is technically simple but limited
to a rather small subset of projects.
In an artificial system, such as a direct-
ed evolution project, the amplification of
variants can be fully separated from the se-
lection step. If the selection pressure is ap-
plied to the pool comprising the entire li-
brary of protein variants, we call this strat-
egy selection by physical enrichment. This
approach is most straightforward if the im-
proved fitness of variants confers higher
binding affinities to a target molecule.
Such target molecules are usually immobi-
lized on a matrix (e.g., the bottom of a Pet-
ri dish, the surface of micro beads or a
chromatography matrix) and target-specific
binders from a library are bound to this
matrix and selectively eluted by changing
the buffer conditions. If applied to the se-
lection of proteins, the approach is compli-
cated by the necessity to provide a physical
coupling between genotype and pheno-
type. A variety of display techniques have
addressed this issue, the most common of
which are phage display, bacterial display,
yeast cell surface display, ribosome display
and DNA-protein hybrid display (see also
Part V, Chapter 2) [41]. Selection by physi-
cal separation is mainly applied in the se-
lection of improved binders such as recep-
tor ligands and antibodies, but has also
been used in a limited number of exam-
ples on enzyme libraries.
Selection by screening, the third alternative,
relies on the separation of variants into sin-
gle compartments on a sample carrier. This
compartmentalization of variants exhibits
several advantages. First, the compartment
provides the coupling of genotype and phe-
notype without the necessity and the biolog-
ical and technical limitations of display
techniques. This, in turn, increases the
range of available expression hosts for a giv-
en protein. In many cases, screening-based
directed evolution projects are directly run
on the basis of expression systems that are
later used as production hosts for large-
scale industrial protein production. This re-
duces the risk of generating proteins that
are highly functional but difficult to pro-
duce. It also enables the screening for high-
er expression yields as one of several opti-
mization parameters. Second, and more im-
portantly, single variants can be tested for
their phenotype even when the biochemis-
try to detect the phenotype is complex and
includes extreme conditions. Diverse pro-
cesses such as extensive liquid handling op-
erations, incubations under a broad range
of conditions or solid-phase separations
can be performed to measure phenotypic
characteristics of the protein variants under
application-relevant conditions. Moreover,
such manipulations do not have to be com-
patible with biological systems. For exam-
ple, incubations at high temperatures or as-
says with toxic or denaturing components
are easily performed with aliquots of sam-
ples from the compartments, leaving the
cells (i.e., the genotype) unaffected by the
assay components.
Aside from its advantages, screening-
based approaches are marked by a high
degree of technical complexity. In practice,
however, the flexibility of the screening
platform and the potentially much higher
information content by far outweigh the
challenges of its technical complexity.
Many protein targets of economic impor-
tance are not addressable by other selec-
tion strategies and it is foreseeable that
the future success of directed evolution
approaches will continue to heavily rely on
the implementation of suitable, efficient
and broadly applicable screening plat-
forms.
1.5
High-throughput and High-content Screening
of Protein Libraries
The purpose of screening protein libraries
is to provide a quantitative measure of
each variants fitness under defined reac-
tion conditions, to link this quantity to
each variants genotype and to enable the
physical isolation of the genotypes of those
variants whose fitness values match prede-
fined criteria. Although this definition is
rather simple, the options for the technical
implementation of screening systems are
numerous.
Selection by screening has been done
for decades in almost every molecular biol-
ogy laboratory throughout the world.
Whenever the blue/white marker en-
zyme, b-galactosidase, is used as an indica-
tor for correct cloning of a DNA fragment
and a white colony is isolated from an
agar plate, all essential features of a
screening process are found: genotypes
and phenotypes are physically linked with-
in each bacterial colony, the phenotype of
each variant is determined, and variants
that fulfill the predefined criteria are
picked from the plate. The system is tech-
nically simple and the throughput can
reach the capacity of much more complex,
modest-throughput robotics-based screen-
ing systems. The major limitation, how-
ever, is obvious: selection is based solely
on the activity of the marker enzyme. The
next level, i.e., the quantification of a vari-
ants fitness via colony-based screens, has
been implemented for several enzyme
classes, particularly for oxygenases and hy-
drolyzing enzymes such as proteases, li-
pases or amylases. For example, a library
of bacterial clones secreting protease vari-
ants can be spread on agar plates contain-
ing skim milk and variants producing an
active protease that hydrolyzes the protein
in the agar can be ranked according to the
halo sizes around the colonies. Such deter-
mination of halo sizes can be done either
visually or with the help of image-process-
ing devices. The throughput of these
screens is already significantly impaired
compared to binary sorting such as blue/
white screening because the density of the
colonies has to be adapted to allow for
identification and measurement of halos.
The most severe restriction, however, is
their limitation to relatively few protein
functions and, even more importantly, to
usually highly artificial reaction condi-
tions.
A screen that is predictive for complex ap-
plication conditions generally requires that
individual variants are grown in closed com-
partments that allow application relevant as-
says. Usually, single clones from a library
are inoculated into liquid medium in the
wells of a microtiter plate. Secondary pro-
cessing steps like induction of protein ex-
pression by addition of an inducing agent
or taking an aliquot of the cell suspension
or supernatant for storage or parallel pro-
cessing are easily implemented, and liquid
samples from the wells can be analyzed by
a large variety of homogeneous and hetero-
geneous assay formats. This variety of assay
principles and analytical methods, and the
flexibility of choice of reaction conditions
is key to the success of screening-based di-
rected evolution projects. The selection of
the right variant for the intended purpose
critically depends on the choice of the right
selection pressure and hence the optimal re-
action conditions.
Spectroscopic methods such as mass
spectrometry, IR spectroscopy, UV/visual
spectroscopy or fluorescence spectroscopy
are standard laboratory techniques and
most of them have also been adapted to
high-throughput screening of protein li-
braries. For screening of biocatalysts,
Schrader et al., for example, established a
mass spectroscopy-based high-throughput
screening system that is able to measure
up to 10
4
variants per day [42]. The sophis-
ticated technical setup is based on multi-
plexed high-performance liquid chromatog-
raphy/electrospray ionization mass spectro-
metry and is especially suited for measure-
ment of low-molecular-weight reaction
products, even against a background of
other contaminating substances in the sam-
ple. In an alternative approach, Tielmann et
al. used Fourier transform IR spectroscopy
to quantify the enantioselectivity of biocata-
lysts [43]. The authors reported a through-
put of up to 10
4
variants per day. While
the measurement technique is technically
less complex, it requires higher purity of
the samples to limit interference from other
substances. A further major limitation is
that larger substrates, like proteins, are
rather difficult to quantify. The most crucial
drawback of both approaches, however, is
the limited throughput.
UV/visual spectroscopy based on colori-
metric assays is regularly used in low-
throughput screens based on microtiter
plates with a low density of wells (usually
96 wells per plate). Further increases in
throughput by using high-density sample
carriers has proven difficult for UV/visual
spectroscopy detection systems. For most
such applications, sample volume and thus
the thickness of the absorbing layer become
too small to yield satisfactory discrimination.
Fluorescence techniques are not subject
to such restrictions and are therefore stan-
dard in ultra-high-throughput screening
(uHTS). They provide the flexibility to im-
plement a multitude of assays under var-
ious experimental conditions. A wide
range of instruments is available featuring
different optical designs. The instruments
allow for screening throughputs of up to
several 10
6
samples per day, which is unri-
valled by any other technique. Pope et al.
[44] provide a detailed overview of fluores-
cence-based screening techniques and
their applications. The range of fluores-
cence-based techniques includes fluores-
cence intensity, fluorescence polarization,
fluorescent resonance energy transfer,
time-resolved fluorescence energy transfer
and confocal fluorimetry techniques such
as fluorescence correlation spectroscopy.
Confocal fluorimetry significantly extends
classical fluorescence spectroscopic tech-
niques (see also Part V, Chapters 4 and 5).
By confining the detection volume to a
tiny portion within the sample (usually in
the femtoliter range), confocal fluorimetry
is characterized by a surprising insensitiv-
ity to miniaturization. Signal-to-noise ra-
tios are for the most part independent of
the sample volume, whether this is several
milliliters or in the submicroliter range.
The other powerful feature of confocal
fluorimetry is the ability to analyze molec-
ular populations on the single-molecule
level, rather than in ensembles. In its var-
ious implementations, this approach en-
ables the measurement of identity, concen-
tration, size, diffusion properties, bright-
ness and the interaction of different fluo-
rescent molecules. A broad overview on
confocal fluorimetry can be found in Ze-
manova et al. [45]. A particularly advanta-
geous modification of confocal methods is
the extension to dual-color techniques.
This optical setup allows for simultaneous
excitation and detection of two spectrally
distinct fluorophores, and thereby even
higher flexibility in the assay design. A
survey of dual-color confocal techniques
and their application to enzyme-catalyzed
reactions is provided by Rarbach et al. [46].
More recently, confocal fluorimetry itself
has been impressively extended. In partic-
ular, the implementation of multi-photon
excitation opened the potential to excite
different fluorescent labels by a single la-
ser line [47]. This considerably simplified
the optical setup of confocal instruments.
For example, Heinze et al. [48] described a
setup for two-photon excitation confocal
fluorimetry where three molecular species
were quantified simultaneously using a
single laser. When included in screening
systems, these spectroscopic advancements
enable the quantification of enzymatic re-
action rates on several substrates in paral-
lel or, when applied for peptide or protein
ligands, the simultaneous measurement of
binding affinities on different target recep-
tors. In this way, biopharmaceuticals can
be selected on the basis of their specificity
and selectivity. As a consequence, unde-
sired side activities can be controlled very
early in the hit identification process.
The technical implementation of uHTS
systems also made significant progress in
recent years. In practically every larger
pharmaceutical company, the growing
chemical libraries as well as the introduc-
tion of combinatorial chemistry led to the
investment and build-up of appropriate
uHTS systems. As might be expected, sev-
eral of the accompanying technical devel-
opments for chemical compound screens
are also useful for protein screens. How-
ever, although the processes may at first
glance seem rather similar, there are a
number of critical differences.
First, the number of substances in com-
mon chemical libraries is small compared
to the complexity of combinatorial protein
libraries. The total number of compounds
in a typical pharma company library is
usually in the range of several 10
5
and the
actual subset that is screened on a particular
target or for a particular indication is in
many cases significantly smaller. In con-
trast, the complexities of protein libraries
frequently reach 10
9
variants, limited only
by technical restrictions such as the expres-
sion system. Therefore, throughput and
productivity requirements of directed evolu-
tion projects usually exceed the require-
ments of compound screens by several or-
ders of magnitude. As a consequence, min-
iaturization in order to increase well density
and to lower operating costs is of even high-
er relevance for protein screens.
Furthermore, the strategies and boundary
conditions for library storage and handling
are notably different. Chemical libraries
are usually static libraries that are stored
in huge archives with a tremendous de-
mand on logistics. The identity (chemical
structure) of the compounds is usually
known and the compounds are stored over
decades. Replicas are made from the library
stock for every screen and, as a conse-
quence, every screen consumes a portion
of the substances which eventually need to
be re-synthesized. Protein libraries in direc-
ted evolution projects, on the other hand, do
not require costly storage techniques, but
are newly and often uniquely generated
for each screening campaign. The genera-
tion of DNA libraries and the production
of protein variants is therefore an integral
part of the overall uHTS process itself.
The most striking difference between
chemical compound screens and protein
screens, however, lies in the demands on
data processing and hit isolation proce-
dures. The result from chemical screens is
the information on potentially active com-
pounds and this information can be drawn
from the screening data at any time after
the screen is completed. Manual follow-up
after chemical screens is thus common-
place. Compounds in a protein library are
protein variants and their coding DNA
fragments which due to the enormous
complexity of protein libraries usually ex-
ist exclusively in a single compartment of
a sample carrier. The result from a protein
screen is therefore the compound itself
(i.e., the coding DNA) and this compound
has to be identified and isolated in real
time, i.e., while the screen is in progress.
Additionally, the information on the geno-
type can be extremely unstable and can be
destroyed by hydrolysing enzymes. As a
consequence, powerful real-time data pro-
cessing and automated, stable and reliable
online hit picking strategies have to be im-
plemented into the screening process.
Fig. 1.3 shows an example of the technical
implementation of a uHTS system espe-
cially adapted to the screening of protein
libraries.
In addition to the automated screening
process, other processes of the evolution-
ary optimization cycle are amenable to
process automation. Candidates for auto-
mation include techniques that are labor
intensive, are needed regularly in the pro-
cess, are performed on a large number of
samples or which require complex and de-
licate handling when performed manually.
Unit operations in molecular biology com-
monly fulfill these criteria. Automated liq-
uid handling workstations will in most
cases be the instrument of choice to ad-
dress processes automation issues here.
This includes isolation of plasmid or geno-
mic DNA from various hosts, PCR setup
and purification, setup of DNA-sequencing
reactions, and the implementation of var-
ious library generation protocols.
1.6
Directed Evolution of Biopharmaceuticals
Screening-based directed evolution requires
the integration of techniques for the genera-
tion of protein diversity and high-through-
put protein screening technology into a
stable and productive lab-scale evolution
process (see Fig. 1.4 for a graphical repre-
sentation). After a library is generated on
the DNA level, the genotypes are distributed
into single compartments of a sample car-
rier. In each compartment, a clone of a sin-
gle genotype is amplified and the respective
protein variant is expressed. The fitness of
each variant is then determined by means
of an appropriate assay and, finally, geno-
types of the best performing variants are
isolated for the next optimization round. Lo-
gistic integration and automation of the cy-
clic process is the key challenge. The most
important factors limiting the progress of
a directed evolution project are the time
needed by a particular expression host to
produce a consistent and sufficient amount
of protein when growing from a single cell,
the complexity of protocols for library gen-
eration, and, for some projects, the assay in-
cubation time. Efficient screening logistics
and maximum process stability are further
key factors. Quality control in all process
steps is of utmost importance, since subpro-
cesses are technically intricate and interde-
pendencies between them are complex.
Directed evolution optimization of pro-
teins has proven its potential in an enor-
mous number of studies targeting various
proteins, including enzymes, antibodies,
peptide hormones and cytokines, to name
a few, and aiming for a broad variety of
optimization goals such as binding affinity,
catalytic activity, thermostability, pH stabil-
ity, expression yield and many others. In
terms of biopharmaceuticals, directed evo-
lution is mainly employed to improve the
characteristics of protein drugs that are
under development, for the engineering of
marketed drugs, i.e., for the generation of
second- and third-generation products, or
for the engineering of follow-on biologics,
i.e., unrelated proteins that have the same
functionality as marketed protein drugs.
Fig. 1.3 uHTS protein screening. Automated high-
throughput and high-content protein screening
technology as implemented in the authors labs.
The technology is based on different reader sys-
tems with a focus on the latest confocal fluorime-
try techniques, including multi-photon excitation
and multi-color detection. It enables the auto-
mated screening of protein libraries with complex-
ities of more than 10
6
variants per 24 h, and fea-
tures unattended over-night and over-weekend
runs. In contrast to typical chemical compound
screening systems which are usually based on
low-density sample carriers (below 1000 wells per
plate; usually 96-well plates), the system is based
on proprietary, high-density sample carriers (above
1000 wells per plate). Sample volumes in these
sample carriers are in the range of 50 nl to a few
microliters. This places high demands on liquid
handling equipment in terms of accuracy, speed
and flexibility. In addition to dispensing and pipet-
ting, other technical features include reproducible
and quick mixing of reagents, controlled incuba-
tion under defined conditions, flexible read-out pa-
rameters, real-time data processing and online
hit-picking to recover improved genotypes. The
technology has been validated in a broad series of
directed evolution projects, including extreme pro-
cess steps like handling of viscous and aggressive
components, handling of solids or suspensions,
high-temperature incubations, use of multiple
read-outs in a single screen, etc.
The development of such a follow-on
protein by directed evolution was carried
out in the authors labs using the example
of plasminogen activators. The application
of protein drugs to dissolve blood clots
during myocardial infarction, ischemic
stroke and embolism was one of the ear-
liest applications of recombinant therapeu-
tic proteins. This is impressively shown in
a video animation on the supplementary
CD-ROM. Stroke results in more than 4.5
million deaths per year worldwide and sur-
viving patients often suffer severe, irrever-
sible disabilities resulting from prolonged
oxygen deprivation [49]. Biopharmaceuti-
cals have been developed to accelerate the
dissolution of clots and the reperfusion of
tissue, eventually reducing tissue damage
and limiting the irreversible effects of
stroke. A first-generation biopharmaceuti-
cal for this indication was streptokinase, a
bacterial protein that activates plasmino-
gen through a specific structural change
induced by binding to streptokinase [50]
(see also Part II, Chapter 1). Streptokinase
was approved by the FDA for human use
in 1987. A second generation agent was re-
combinant human tissue-type plasmino-
gen activator (t-PA) produced in Chinese
hamster ovary cells (Alteplase; Genentech)
[51]. Third generation products are modi-
fied proteins derived from human t-PA (r-
PA, TNK, n-PA), as well as proteins from
alternative sources (e.g., rDSPA) [49, 51].
(Note: the supplementary Volume V of
Modern Biopharmaceuticals will contain a
case study on DSPA from Oliver Kops
from Paion.) Human t-PA is a 70-kDa ser-
ine-type protease that produces active plas-
min in vivo from its inactive precursor,
plasminogen, by cleavage of a single pep-
tide bond (Arg561Val562). Plasmin is it-
self a protease that specifically degrades fi-
brin, a major constituent of blood clots
(see also Part III, Chapter 6).
In a directed evolution approach pursued
in our laboratories, the sequence specificity
of a microbial protease was systematically
altered away from its natural substrate to
match the activation site of plasminogen
Fig. 1.4 Screening-based directed evolution. The process involves the cyclic iteration
of mutation, selection and amplification of improved variants in a stable laboratory-
scale process.
(see Fig. 1.5 for experimental details). This
strategy allowed the development of a t-
PA-like protease on the basis of a protein
scaffold that is significantly more active
and production cost-efficient when com-
pared to human t-PA or alternative en-
zymes. The resulting engineered protease
is based on the BAR protease from S. cerevi-
siae and can be readily produced in this host
at comparably high yields. In order to effi-
ciently screen for proteases with altered
specificity, the authors also devised a specif-
ic screening strategy simulating a co-evolu-
tion between enzyme and substrate [53].
As illustrated in Fig. 1.6, the amino acid se-
quence of the substrate was altered in a
stepwise fashion, starting with a peptide
substrate resembling the natural substrate
and finally employing a peptide substrate
containing the target sequence. By using
this strategy, a constant selection pressure
can be exerted on the evolving protein,
and new subpopulations can be selected
and amplified.
Substrate specificities of variants se-
lected by this strategy over six consecutive
rounds of directed evolution are summa-
rized in Fig. 1.5a. The example shows that
Fig. 1.5 Experimental results. Engineering of a
new substrate specificity into the BAR protease
from S. cerevisiae. (a) Evolution of proteolytic ac-
tivities on three different substrates over six
rounds of mutagenesis and screening. (b) Three-
dimensional ribbon model of the protease indicat-
ing the positions of mutated residues in the final
enzyme from round six. Experimental details. The
gene encoding the aspartic protease, BAR, from S.
cerevisiae was used as the starting point to engi-
neer a protease that efficiently activates plasmino-
gen at its precise activation site. BAR is a highly
specific, 587-amino-acid enzyme from the chymo-
sin fold protease family [52]. Protein libraries were
generated by random mutagenesis, homologous
recombination and SCR, and transformed into S.
cerevisiae as the expression host. Screening was
performed on an automated platform using confo-
cal fluorimetry read-outs as described in Fig. 1.3.
The screening strategy involved the consecutive
use of intermediate peptide substrates (see text).
Substrate 1 contains the original cleavage se-
quence for the BAR protease, while substrate 3
besides two cysteines which for technical reasons
were replaced by two serines contains the plas-
minogen activation sequence. Oligopeptides were
fluorescently labeled and used as screening sub-
strates. Read-out parameters were fluorescence in-
tensity, molecular rotational diffusion and molecu-
lar translational diffusion. A weighted combination
of these parameters was used to rank perfor-
mance of variants in the library. Various nontarget
control peptides labeled with spectroscopically
distinguishable fluorophores were used to simul-
taneously select for sequence specificity. The wild-
type enzyme showed high catalytic activity on sub-
strate 1 and very low activity on intermediate sub-
strate 2, while activity on substrate 3 was below
the detection limit. After four rounds of directed
evolution, the catalytic rate on intermediate sub-
strate 2 was improved by a factor of 44, while the
catalytic activity on substrate 1 dropped to ap-
proximately 40%. After two further rounds the ac-
tivity on substrate 3 was increased 6-fold, while
activity on the two other substrates dropped to
approximately 10%. Activity was normalized to the
maximum catalytic activity on each substrate.
the sequence specificity of a protease can
be fundamentally changed in only a few
cycles of directed evolution to a predeter-
mined target sequence by modifying the
screening substrate in a stepwise fashion
towards the intended substrate, with si-
multaneous adaptation of the enzyme
specificity. Using this approach, a novel
and completely unrelated specificity was
generated in a protease.
The putative structure of the BAR pro-
tease can be inferred from its sequence
homology to chymosin (see Fig. 1.5b; Pro-
tein Data Bank entry 1CMS). When the mu-
tations that were found in the final engi-
neered BAR variant are mapped onto this
structure, approximately 50% of the muta-
tions lie in or close to the substrate binding
site of the protease, while the other 50% are
found at distant sites in the protein. This
distribution confirms that directed evolu-
tion processes efficiently generate muta-
tions at relevant positions.
1.7
Conclusions
In the study described above, the sequence
specificity of a protease was changed to
specifically address a new, pharmacologi-
cally relevant target sequence. This is only
one example of the potential of screening-
based directed evolution in the design of
modern biopharmaceuticals. The number
of potential applications is virtually inex-
haustible. The technology, of course, is not
restricted to catalytic proteins, but can also
be employed for the selection of binding
and signaling proteins. The requirement
that the final product be functional under
physiologic conditions often imposes addi-
tional constraints on protein function that
cannot easily be reproduced by other pro-
tein engineering strategies. Such con-
straints include the stability of proteins to
proteolytic digestion by endogenous pro-
teases, inactivation by natural inhibitors or
undesired clearance of biologics via bind-
ing to serum proteins. Furthermore, bio-
logics that exhibit side-effects can be modi-
fied to target efficiently and selectively the
pharmacologically relevant target. When
applied appropriately, directed evolution is
indisputably the single most powerful pro-
tein engineering tool that is available to
date. The reality and continuing promise
of modern biopharmaceuticals is thus the
modification of natural proteins for the
generation of novel, specialized molecular
drugs that carry out predefined pharmaco-
logical functions. We are already experi-
encing an exciting expansion of the impact
of engineered biopharmaceutic of human
health and this trend will undoubtedly ac-
celerate for many years to come.
Acknowledgments
The authors wish to thank their coworkers
at Direvo Biotech AG for their indispensa-
ble contributions to the described results
and protocols.
1.7 Conclusions 601
Fig. 1.6 Co-evolution principle. See text for de-
tails.
References
1 G. Walsh, Nat. Biotechnol. 2000, 18, 831833.
2 S. Spiegelman, Q. Rev. Biophys. 1971, 4, 213
253.
3 E. Domingo, R. A. Flavell, C. Weissmann,
Gene 1976, 1, 325.
4 C. A. Hutchison, S. Phillips, M. H. Edgell, S.
Gillam, P. Jahnke, M. Smith, J. Biol. Chem.
1978, 253, 65516560.
5 D. Shortle, D. Nathans, Proc. Natl Acad. Sci.
USA 1978, 75, 21702174.
6 C. T. Chu, D. S. Parris, R. A. Dixon, F. E. Far-
ber, P. A. Schaffer, Virology 1979, 98, 168181.
7 K. M. Ulmer, Science 1983, 219, 666671.
8 R. Wetzel, Protein Eng. 1986, 1, 35.
9 M. Eigen, Naturwissenschaften 1971, 58, 465
523.
10 J. F. Sambrook, E. F. Fritsch, T. Maniatis, Mo-
lecular Cloning: A Laboratory Manual. Cold
Spring Harbor Laboratory Press, Cold Spring
Harbor, NY, 1989.
11 O. P. Kuipers, Methods Mol. Biol. 1996, 57,
351356.
12 R. C. Cadwell, G. F. Joyce, PCR Methods Appl.
1992, 2, 2833.
13 S. Delagrave, D. C. Youvan, Biotechnology 1993,
11, 15481552.
14 J. D. Hermes, S. M. Parekh, S. C. Blacklow, H.
Koster, J. R. Knowles, Gene 1989, 84, 143151.
15 D. K. Dube, L. A. Loeb, Biochemistry 1989, 28,
57035707.
16 A. Greener, M. Callahan, B. Jerpseth, Methods
Mol. Biol. 1996, 57, 375385.
17 O. Selifonova, F. Valle, V. Schellenberger,
Appl. Environ. Microbiol. 2001, 67, 36453649.
18 E. C. Cox, Annu. Rev. Genet. 1976, 10, 135
156.
19 E. Domingo, L. Menendez-Arias, M. E. Qui-
nones-Mateu, A. Holguin, M. Gutierrez-Rivas,
M. A. Martinez, J. Quer, I. S. Novella, J. J. Hol-
land, Prog. Drug Res. 1997, 48, 99128.
20 M. Eigen, K. Henco, PCT patent application
WO 92/18645, 1992.
21 P. S. Daugherty, G. Chen, B. L. Iverson, G.
Georgiou, Proc. Natl Acad. Sci. USA 2000, 97,
20292034.
22 M. Zaccolo, E. Gherardi, J. Mol. Biol. 1999,
285, 775783.
23 W. P. Stemmer, Nature 1994, 370, 389391.
24 M. T. Reetz, Tetrahedron 2002, 58, 65956602.
25 M. Streuli, A. Hall, W. Boll, W. E. Stewart, S.
Nagata, C. Weissmann, Proc. Natl Acad. Sci.
USA 1981, 78, 28482852.
26 A. Meister, G. Uze, K. E. Mogensen, I. Gres-
ser, M. G. Tovey, M. Grutter, F. Meyer, J. Gen.
Virol. 1986, 67, 16331643.
27 D. Pompon, A. Nicolas, Gene 1989, 83, 1524.
28 S. Paabo, D. M. Irwin, A. C. Wilson, J. Biol.
Chem. 1990, 265, 47184721.
29 A. Meyerhans, J. P. Vartanian, S. Wain-Hob-
son, Nucleic Acids Res. 1990, 18, 16871691.
30 W. P. Stemmer, Proc. Natl Acad. Sci. USA
1994, 91, 1074710751.
31 H. M. Zhao, L. Giver, Z. X. Shao, J. A. Affhol-
ter, F. H. Arnold, Nat. Biotechnol. 1998, 16,
258261.
32 W. M. Coco, W. E. Levinson, M. J. Crist, H. J.
Hektor, A. Darzins, P. T. Pienkos, C. H.
Squires, D. J. Monticello, Nat. Biotechnol.
2001, 19, 354359.
33 W. M. Coco, L. P. Encell, W. E. Levinson, M. J.
Crist, A. K. Loomis, L. L. Licato, J. J. Arensdorf,
N. Sica, P. T. Pienkos, D. J. Monticello, Nat.
Biotechnol. 2002, 20, 12461250.
34 A. A. Volkov, Z. Shao, F. H. Arnold, Nucleic
Acids Res. 1999, 27, e18ie18vi.
35 M. Eigen, A. Koltermann, U. Kettling, PCT
patent application WO 01/34835, 2000.
36 C. A. Voigt, C. Martinez, Z. G. Wang, S. L.
Mayo, F. H. Arnold, Nat. Struct. Biol. 2002, 9,
553558.
37 A. Koltermann, U. Kettling, J. Pilling, O.
Spangenberg, PCT patent application WO 04/
018674, 2003.
38 B. G. Hall, T. Zuzel, Proc. Natl Acad. Sci. USA
1980, 77, 35293533.
39 M. Duenas, C. A. Borrebaeck, Biotechnology
1994, 12, 9991002.
40 C. Krebber, S. Spada, D. Desplancq, A.
Pluckthun, FEBS Lett. 1995, 377, 227231.
41 D. Ma, M. Li, J. Cell. Biochem. Suppl. 2001, 37,
3441.
42 W. Schrader, A. Eipper, D. J. Pugh, M. T. Reetz,
Can. J. Chem. 2002, 80, 626623.
43 P. Tielmann, M. Boese, M. Luft, M. T. Reetz,
Chem. Eur. J. 2003, 9, 38823887.
44 A. J. Pope, U. M. Haupts, K. J. Moore, Drug
Discov. 1999, 4, 350362.
45 L. Zemanova, A. Schenk, M. J. Valler, G. U.
Nienhaus and R. Heilker, Drug Discov. Today
2003, 8, 10851093.
46 M. Rarbach, U. Kettling, A. Koltermann, M.
Eigen, Methods 2001, 24, 104116.
47 K. G. Heinze, A. Koltermann, P. Schwille,
Proc. Natl Acad. Sci. USA 2000, 97, 10377
10382.
48 K. G. Heinze, M. Jahnz, P. Schwille, Biophys.
J. 2004, 86, 506516.
49 P. A. Lapchak, Expert Opin. Invest. Drugs 2002,
11, 16231632.
50 J. A. Marcum, D. L. Kline, Comp. Biochem.
Physiol. B Comp. Biochem. 1983, 75, 389394.
51 A. R. Xavier, A. M. Siddiqui, J. F. Kirmani,
R. A. Hanel, A. M. Yahia, A. I. Qureshi, CNS
Drugs 2003, 17, 213224.
52 V. L. MacKay, J. Armstrong, C. Yip, S. Welch,
K. Walker, S. Osborn, P. Sheppard, J. For-
strom, Adv. Exp. Med. Biol. 1991, 306, 161
172.
53 A. Koltermann, U. Kettling, U. Haupts, J.
Tebbe, P. Scholz, J. Pilling, S. Werner, M. Rar-
bach, PCT patent application WO 03/095670,
2003.
References 603
Abstract
The Gateway
recombinational cloning
system represents a new paradigm in mo-
lecular biology by improving speed and ef-
ficiency beyond traditional restriction clon-
ing methods. By harnessing the recombi-
nation mechanism of bacteriophage lamb-
da, it provides an efficient method for
manipulating DNA elements and acceler-
ates the discovery process and the develop-
ment of biopharmaceuticals. It does this
by allowing efficient, high-throughput, and
automatable cloning and transfer of DNA
elements to analytical platforms. This
chapter will discuss the evolution of DNA
cloning which led to the development of
the Gateway system, the recombination
mechanism which leads to its high effi-
ciency, and several examples of novel ex-
perimental approaches which it enables.
Abbreviations
attB bacterial attachment site
CFP cyan fluorescence protein
ECFP enhanced cyan fluorescence protein
EYFP enhanced yellow fluorescence pro-
tein
HTP high-throughput
IHF integration host factor
Int integrase
Xis exisionase
YFP yellow fluorescence protein
2.1
Introduction
As with most information whether it is
encoded as bits and bites or as deoxyribo-
nucleic acids the challenge is funneling
it to a useful analytical platform. Until re-
cently, the volume of information in the
form of sequenced genes has been relative-
ly small such that it could be managed
(cloned and expressed) well enough to
avoid major bottlenecks. With high-
throughput (HTP) sequencing efforts and
sequencing of the human genome [1] (as
well as several other genomes), we are in-
undated with a vast number of identified
genes and their sequence information.
This leads to one of the new bottlenecks
in proteomics, the isolation, and transfer
of open reading frames (ORFs) and other
DNA elements to suitable systems for ex-
pression and analysis. In the following sec-
tions, we will explore the Gateway system
which is a genetic engineering tool based
on a well-evolved viral recombination sys-
605
2
Learning from Viruses: High-throughput Cloning using the
Gateway
System to Transfer Genes without Restriction Enzymes

Jonathan D. Chesnut
ISBN: 3-527-31184-X
tem and allows HTP gene isolation and
transfer to and from essentially any vector-
based genetic/proteomic platform.
The Gateway system enables HTP clon-
ing and transfer of genetic material for ap-
plications such as high-level protein pro-
duction for structural analysis, generation
of antibodies or other detection reagents,
HTP screening of chemical compounds
for pharmaceuticals, functional analysis
studies such as protein interactions, sub-
cellular localization, post-translational
modification, and RNA interference of
translation. In addition to HTP applica-
tions, Gateway offers utility for researchers
who need to optimize protein expression,
purification, and detection by use of var-
ious epitope tags, promoters, and host sys-
tems (Fig. 2.1).
2.2
Background
The age of recombinant DNA technology
was born with early descriptions of DNA
cloning and plasmid development in the
1970s [24]. The development of these
technologies revolutionized almost every
facet of biology, biochemistry and genetics
by giving us tools to cut, join, and repli-
cate DNA. During the early 1980s, empha-
sis moved toward defining the boundaries
of genes, cloning ORFs, and expressing re-
combinant protein in bacteria. Many or-
ganisms have been used as hosts for re-
combinant protein expression since then.
The ability to manipulate and express re-
combinant protein is still a valuable skill
that has evolved into more of a necessity
than a novelty.
System to Transfer Genes 606

Fig. 2.1 The Gateway recombinational cloning system.
In order to clone genes by standard re-
striction enzyme and ligase methods, a re-
searcher must choose the correct restric-
tion enzymes to allow isolation of the frag-
ments, cut and purify the vector and in-
sert, and assemble them in a ligation
reaction in the proper ratios. Each gene
fragment must be considered on a case-by-
case basis, given that the particular restric-
tion enzymes needed to isolate the frag-
ment may not be compatible with the vec-
tor for which it is intended. Also, the gene
itself might contain the same restriction
sites internally that match those needed
for subcloning. To mitigate these types of
problems, the gene is commonly amplified
using the polymerase chain reaction
(PCR), with additional restriction sites en-
coded in the primers. This approach can
be used to create a small number of ex-
pression vectors. Working at modern
scales, however, HTP cloning calls for the
transfer of hundreds, thousands, or even
millions of genes from platform to plat-
form. These types of operations pose a
major challenge to even the best cloning
wizard contemplating using standard re-
striction enzyme and ligase techniques.
Over the past decade, there has been sig-
nificant evolution in the field of recombi-
national cloning towards the development
of an efficient system to handle large
amounts of genetic information for analy-
sis. The Gateway system represents the lat-
est in this evolution, as the most versatile
and efficient method available for manipu-
lating nucleic acid fragments. It facilitates
efficient gene transfer via a re-engineered
phage recombinase system, and represents
an enabling technology for the field of
HTP gene cloning and expression. The
Gateway system offers an efficient mecha-
nism to streamline cloning without con-
cern for the specific sequence of an ORF.
Since it does not use restriction enzymes,
ORFs are transferred to an expression vec-
tor without the requirement of amplifica-
tion by PCR or subsequent concern for
PCR-induced mutations. Gateway is an ex-
ample of a new genre of recombination-
based cloning systems that facilitate sub-
cloning of genetic material by site-specific
recombination and enable transfer and as-
sembly of large numbers of DNA elements
using a standard mechanism.
Variations of cloning by recombination
and site-specific transposition have been
demonstrated previously in E. coli and
yeast. Many of these systems relied on
using the host machinery to integrate
DNA in vivo [511], while others used Cre
recombinase-based (Cre/lox) systems re-
constituted in vitro to clone and transfer
DNA [1214]. Until the late 1990s, the
most effective systems for vector transfer
were those based on the Cre/lox system
mentioned above. The plasmids in those
systems contain lox sites and are fused, or
segments transferred, upon addition of
Cre recombinase. A notable member of
this group is the Creator System (BD-Clon-
tech). This system is similar to Gateway in
that a gene is first cloned into a vector
then transferred in a second reaction to a
second (expression) vector for use. Major
differences with the Gateway system are
that gene transfer by Cre/lox recombina-
tion is relatively inefficient and, more im-
portantly, unidirectional. The multiple spe-
cificities only available with the Gateway
system also allow for reversible assembly
of multiple DNA fragments in a single re-
action (as will be discussed below).
During the mid-1980s, James Hartley
and Michael Brasch recognized the on-
coming rush of genetic information that
would require efficient manipulation. They
set out to develop a recombinational clon-
ing system to handle this flood of informa-
tion that would facilitate the manipulation
2.2 Background 607
of DNA without restriction enzyme and li-
gase cloning [15]. Their ideal system
would have maximum compatibility and
flexibility. It would be useful with a mini-
mum amount of planning, and it would
maintain the orientation and reading
frame of the ORFs transferred within the
system. Moreover, it would be rapid, need-
ing no restriction enzymes, no gel purifi-
cation of DNA fragments, and no lengthy
ligation steps. Ultimately, the Gateway sys-
tem would be ideal for HTP cloning
though this approach had barely been con-
sidered at that time.
After evaluating several potential sys-
tems, Hartley and Brasch set their sights
on the integration and excision activity of
bacteriophage lambda. With help from Ho-
ward Nash, who is considered a pioneer of
the study of lambda biology [1620], they
set up an in vitro system for transferring
DNA fragments from one vector to an-
other. Lambda, a double-stranded DNA
bacteriophage, uses the bacterial host cel-
lular machinery to propagate itself. To ac-
complish this, it first infects the cell and
inserts its entire genome into the host
chromosome. In this conservative and site-
specific reaction, a 25-base pair region in
the bacterial chromosome (termed the attB
site for bacterial attachment) is recognized
by the phage recombination complex and
aligns with a 243-base pair sequence in
the phage genome (attP for phage attach-
ment). In the presence of a dimeric bacte-
rial protein, Integration Host Factor (IHF)
and phage-encoded Integrase (Int), the
DNA align and exchange strands, effec-
tively inserting the phage genome into the
bacterial chromosome at the attB site. The
BP reaction that integrates the phage DNA
into the bacterial genome creates two new
sites that now flank the phage DNA
termed attL (for left) and attR (for right)
(Fig. 2.2, Panel A). The phage DNA is rep-
licated along with the host DNA and can
remain integrated in the host genome for
an extended period of time. At some point
possibly after some sort of stress or
other insult to the bacterium the phage
DNA is excised from the host chromo-
some and is re-encapsulated, thereby al-
lowing it to move on to infect another bac-
terium.
The excision reaction is essentially the
reverse of the integration reaction de-
scribed above. Int and IHF are again in-
volved, but are joined by an additional
phage-encoded protein, Excisionase (Xis).
These enzymes functionally join the attL
and attR sites leading to strand cleavage
and exchange, and creation of attB and
attP sites on the two respective circular
chromosomes.
Within each att site, no matter what type
(attL, attR, attB, or attP), there is a 7-base
pair core sequence, or overlap region,
where strand exchange occurs. It is at
these sites where the reaction specificity is
determined (Fig. 2.2, Panel B). There is no
known specific sequence requirement for
a functional core region, only that both re-
combination partners (attL and attR or
attB and attP) have homologous cores.
This property of lambda recombination al-
lowed the development of various core se-
quences that recombine specifically and ef-
ficiently. For standard Gateway reactions,
the wild-type att site core region described
in Fig. 2.2 (Panel B) was mutated to att1
and att2. As will be discussed below, more
complex reactions can be accomplished
using att sites 1 through 6 (see Table 2.2).
The attP site is the largest and most
complex of all the att site sequences, as it
contains elements of all the other sites.
There is the P arm which consists of the
majority of the attR site, and the P' arm,
which consists essentially of attL sequence.
A complete attB sequence, containing the

specificity-determining overlap region,
separates these two sites. In each arm,
there are several specific protein binding
sites separated by spacer regions. This is
where Int, IHF, and Xis bind and bend
the DNA to form a large proteinDNA
complex which aligns the core sequence of
the attP site with that of the bacterial attB
site. The core sequence in the attP and
attB sites also have Int binding sites where
cleavage and strand exchange is catalyzed.
The result of the integration reaction (BP
recombination) is the generation of attL
and attR sites on the bacterial chromo-
some. In nature, these sites are cis-acting,
although the reaction can occur in trans.
In the Gateway system, the attP and attB
sites remain on their own respective (do-
nor and expression) plasmids, while the
attL and attR sites exist on different (entry
and destination) plasmids respectively (see
Fig. 2.4).
2.3
Engineering the Lambda System
to Create Gateway
There are two key aspects of lambda re-
combination that are important to the
Gateway system:
1. Multiple specificities can be generated
by changing one or more bases in the
overlap region of the att sites.
2. The att sites can be used in pairs and
flipped in orientation to allow transfer
DNA cassettes between vectors.
This important concept is shown in Fig.
2.3, where the attP1 site is designated as
L-B1-R and the attP2 site as R-B2-L.
This designation denotes the different
arms or regions of the attP sites. It is im-
portant to remember that all the specificity
lies within the core of the attB sites. Dur-
ing a BP recombination reaction, the attP1
site recombines only with an attB1 se-
quence, and the attP2 site only with an
attB2 sequence. This allows separate
strand exchange at these two sites that ef-
fectively splits the attP sites, thereby creat-
Fig. 2.2 Site-specific recombi-
nation in bacteriophage lamb-
da.
ing attL (L-B1 and B2-L) sites in the entry
vector and attR (B1-R and R-B2) sites in
the destination vector.
2.4
The Gateway Reactions
The heart of the Gateway system is the en-
try clone. In this plasmid, attL1 and attL2
sites flank the DNA fragment of interest
(Fig. 2.4). Once a gene, ORF, or other
DNA element has been inserted between
the attL sites to create the entry clone, it
can be transferred to any compatible desti-
nation vector. This transfer maintains the
orientation and reading frame (if the DNA
is an ORF) and is accomplished regardless
of the sequence of the DNA. Since there is
no specific sequence requirement that the
element must have (besides containing
flanking att sites), entire pools or highly
diverse libraries of DNA fragments can be
captured and transferred with the same
high efficiency, and in a single reaction.
To transfer the DNA fragment from an
entry vector to another plasmid, the entry
vector is mixed with a destination vector
(that carries attR1 and attR2 sites), after
which LR clonase (Int, IHF, and XIS) is
added to catalyze the recombination be-
tween specific attL and attR sites. The
products of this reaction consist of a donor
vector that contains the ccdB gene flanked
by attP sites and the expression clone that
carries the DNA of interest flanked by attB
sites (i.e., an ORF now downstream of a
promoter sequence). Given the appropriate
amounts of reactants, the efficiency of the
recombination reaction (the amount of en-
try clone converted to expression clone) is
approximately 70%.
An aliquot of the LR reaction mix is
used to transform competent E. coli, and
the desired expression clone is selected by
two means. First, any cell carrying either a
donor vector or an unreacted destination
vector is selected against by the presence
of the ccdB gene [19]. The product of this
gene is extremely toxic to most strains of
E. coli, and therefore any cell containing
this gene is efficiently killed. Second, the
transformation mixture is plated on the
appropriate selective antibiotic (ampicillin
in the example shown in Fig. 2.4) that ef-
fectively selects against cells transformed
with an unreacted entry vector. The recom-
binational specificity and the stringency of

Fig. 2.3 Gateway bacterioph-
age (BP) and lambda recogni-
tion (LR) recombination reac-
tions.
the positive and negative selection leads to
an overall efficiency (the proper clones ex-
isting as colonies on the agar plate) of
greater than 95%. This high cloning effi-
ciency is a key attribute of the Gateway
system. The strong selection system often
makes it unnecessary to screen multiple
colonies in order to select the desired
clone, thereby facilitating many HTP gene
cloning and expression applications.
As mentioned above, the recombination
reaction can be reversed depending on the
presence of the Xis protein in the enzyme
mix. To transfer an ORF from an expres-
sion clone back into an entry vector, the
expression clone (with the DNA fragment
flanked by attB1 and attB2 sites) is com-
bined with a donor vector (containing
attP1 and attP2 sites flanking the ccdB
gene) and BP Clonase (Int and IHF). The
BP reaction, like the LR, is specific and
unidirectional and results in a similar
cloning efficiency.
2.5
Creating Gateway Entry Clones
There are several ways in which to create
entry clones. The easiest method is direct
ligation of a PCR product with a TOPO-
charged entry vector (Fig. 2.5). A TOPO
cloning vector is linearized and covalently
attached to Vaccinia topoisomerase I at the
3' terminus, or TOPO-charged. The PCR
product is simply added to the TOPO-
charged molecule, incubated for 5 minutes
at room temperature, and used to trans-
form E. coli. This reaction is efficient, such
that generally greater than 90% of the re-
sulting colonies contain the correct clone.
Currently, there are four TOPO entry vec-
tors available. This method, whilst simple
to perform, is not suited for some HTP
cloning applications because it has an opti-
mal window of insert size and insert-to-
vector ratios. While average sized inserts
(12 kb) clone with very high efficiency,
overall colony numbers decrease when the
insert is larger than 34 kb. If the insert is
in more than 10-fold molar excess of the
Fig. 2.4 The Gateway vector
system.
TOPO vector, then ligation of separate in-
serts at each TOPO site is thought to be-
come more prevalent, leading to linear
products that are unable to transform bac-
teria.
The efficiency of the BP reaction is less
affected by insert size and concentration
than TOPO cloning, which makes it more
amenable to HTP cloning and cloning
large inserts. With this strategy, forward
and reverse PCR primers are designed that
contain the 24-base pair attB1 and attB2
sequences. The fragment of interest is
then amplified and the product, which
now has an attB1 or an attB2 site on
either end, is recombined with a donor
vector. The high efficiency and specificity
of the recombination reaction makes this
process robust and allows it to support
cloning of fragments equal to or greater
than 10 kb (Table 2.1). The entire process
is easily automated, and can be used to
amplify and clone large sets of genes in
parallel [20, 21].
Creating entry vectors allows the re-
searcher easily to isolate and archive DNA
fragments for use in any of a multitude of
applications. Once the entry clone is se-
quenced, it serves as a defined resource
that can be shared within the research
community and easily transferred to vari-
ous expression systems using the Gateway
mechanism. Elements from single ORFs
to complex cDNA libraries can be intro-
duced into a destination vector in a single
reaction with high efficiency. A single
ORF or set of ORFs can also be cloned
into a set of different destination vectors
using the same reaction conditions and
carried out in parallel. These types of ma-
nipulations are not always feasible using
restriction enzyme and ligase systems.

Fig. 2.5 Gateway entry vector systems.
2.6
Gateway Destination Vectors
The Gateway system has opened new ave-
nues in proteomic and genomic research
by enabling parallel, HTP transfer of ge-
netic information into various expression
and analytical systems. These Gateway-
compatible plasmid systems are termed
destination vectors. To be considered a
destination vector, a plasmid need only
contain a destination cassette, an origin of
replication, and an antibiotic-selectable
marker. In most cases, the plasmid con-
tains the elements needed for gene expres-
sion in a particular host (promoter, poly-
adenylation signal, epitope tag, etc.). The
destination cassette, which is used to con-
vert standard vectors, consists of attR sites
to match with the attL sites of the entry
vector(s) along with a ccdB expression cas-
sette. In the case of standard Gateway,
these are attR1 and attR2 (things get more
complicated with Multisite Gateway de-
scribed below). The attR sites flank the
ccdB gene cassette and usually another se-
lectable marker (generally the chlor-
amphenicol resistance gene). Once posi-
tioned just downstream of a promoter se-
quence, the destination cassette is poised
to receive an ORF or other genetic element
from an entry clone via the LR reaction.
Using this configuration, it is easy to
convert any vector to a destination vector
using commercially available Gateway con-
version cassettes. These are supplied as
blunt-ended DNA fragments that can be li-
gated with a properly prepared plasmid
vector. The cassette is ligated with a blunt
vector and used to transform ccdB-resis-
tant E. coli (available from Invitrogen). The
correct clones are selected by plating the
transformation reaction on agar plates con-
taining chloramphenicol. The plasmid
Table 2.1 Efficiency of cloning PCR fragments by bacteriophage
recombination
Size
(kb)
PCR DNA
(fmol)
PCR DNA
(ng)
Colonies/ml
Transformation
b)
Correct Clones/
Total Clones Examined
0.26 15 3 1223 10/10
a)
38 7.5 2815
1.0 15 10 507 49/50
38 25 1447
1.4 15 14 271 48/50
38 35 683
3.4 15 34 478 9/10
a)
38 85 976
4.6 15 46 190 10/10
a)
38 115 195
6.9 15 69 30 (235)
b)
47/50
38 173 54 (463)
b)
10.1 7.5 50.5 16 (112)
b)
15/16
37.5 252.5 42 (201)
b)
a) DNA mini-prep analysis
b) 1 pUC=10
8
CFU/ml
c) After overnight incubation
DNA from the resulting colonies is then
screened for the correct cassette orienta-
tion (the ligation is bi-directional) by re-
striction enzymes or PCR. Once the prop-
erly oriented clone is isolated, a destina-
tion vector is born.
For a complete list of commercially avail-
able destination vectors, please visit www.in-
vitrogen.com. These cover a broad array of
applications including protein expression
in E. coli, yeast, insect, and mammalian
cells, yeast two hybrid protein interaction,
and gene knock-down RNAi vectors.
2.7
Applications Enabled by Gateway Cloning
Gateway cloning clearly facilitates the
management of nucleic acid fragments,
making cloning and transfer operations
easier and more efficient. The emergence
of the Gateway system in the late 1990s
accelerated the genome cloning revolution.
Mark Vidals group was the first to isolate
and clone an entire genome using the
Gateway system. They started with a Cae-
norhabditis elegans cDNA library, amplified
each ORF via attB-PCR, and then created a
nearly complete entry clone ORFeome li-
brary [20, 22]. Vidal recognized a key fea-
ture of the Gateway system in that all the
manipulations necessary to clone, transfer,
and express a pool of ORFs can be accom-
plished essentially in-batch, without con-
cern for the composition of specific ORFs
in the pool or library. The entire process
was automated and performed using ro-
botic liquid handling systems, thereby al-
lowing completion of the entire C. elegans
ORFeome collection in two years. The
collection has subsequently been used for
various studies by the C. elegans commu-
nity, most notably, the pioneering work
which uses the ORF pool in yeast two-hy-
brid studies of the interactome [23, 24].
Interaction mapping of the entire ORF-
eome is possible because entry clones
could be easily transferred to bait or
prey destination vectors [22] (Proquest
TM
System, Invitrogen) for screening. Since
this groundbreaking work with HTP clon-
ing and interaction mapping, several other
ORF sets have been cloned using Gateway
(see Ref. [21] and references therein).
2.8
HTP Expression Analysis
in Mammalian Cells
Another example of Gateway technology
enabling new types of experimentation by
offering the ability to handle pools and li-
braries of ORFs and cDNAs in HTP fash-
ion is the study reported by Stefan Wei-
manns group [25]. In this study, Gateway
cloning was used to shuttle a set of ap-
proximately 100 novel cDNAs (no experi-
mentally confirmed function) into N-termi-
nal CFP and C-terminal YFP fusion vec-
tors. HeLa cells were transfected with
these constructs, and then visualized with
fluorescence microscopy to determine the
subcellular localization of each fusion pro-
tein (Fig. 2.6). The proteins were catego-
rized by their cellular location and com-
pared with bioinformatic predictions of
their function. This represents the first ex-
ample of a system in which HTP screen-
ing of cDNAs for their cellular location
was achieved. Further development of sim-
ilar approaches could lead (potentially) to
HTP, high-content cellular assay screening
of wild-type and mutant proteins of phar-
maceutical value.

2.9
HTP Cloning and Expression
in a Baculovirus System
Recombinational cloning, specifically Gate-
way, works with such high efficiency due
to the favorable kinetics of the strand ex-
change reaction. This allows the cloning
efficiency to be much less dependent on
insert size. The reduced size dependence
enables direct cloning into very large plas-
mids such as the baculovirus genome
(128 kb) using the Gateway-enabled bacu-
lovirus expression system, BaculoDirect
TM
(Invitrogen).
Common baculoviral expression systems
require the use of a bacterial transfer vec-
tor to transfer an ORF of interest into the
baculovirus genome by homologous re-
combination in bacteria. This requires that
the reaction be carried out in E. coli cells
that have fully functional recombination
machinery (this is disabled in standard
cloning strains). Whilst this facilitates the
homologous recombination required to
transfer the ORF of interest into the bacu-
lovirus genome, it also leads to a variety of
undesirable recombination events and the
subsequent screening of many colonies to
obtain the correct expression clone. This
inefficiency essentially precludes standard
baculovirus cloning from being used as a
HTP cloning tool.
In the BaculoDirect system, the entire
circular baculovirus genome serves as a
destination vector. The ccdB selection sys-
tem used in other E. coli-based cloning
systems has been replaced with the thymi-
dine kinase/gancyclovir system that effec-
tively kills eukaryotic cells [26]. Also, the
destination cassette contains a LacZ gene
for detection of parent virus prior to and
during the selection on gancyclovir (Fig.
2.7). This allows selection of the correct re-
combinants in insect cells (i.e., the baculo-
viral host). The efficiency of Gateway re-
Fig. 2.6 Systematic subcellular localization of novel proteins. (Reproduced from Ref. [27].)
combination, along with the strength of
the gancyclovir selection, yields the same
high cloning efficiency in insect cells as
observed in bacteria. This system lends it-
self well to HTP cloning applications,
since an ORF of interest can be trans-
ferred to the baculovirus expression sys-
tem without exposing it to mutation-induc-
ing PCR amplification or homologous re-
combination,
After transfecting the insect cell popula-
tion with the LR reaction solutions, the
cells are grown in the presence of gancy-
clovir for one to two passages that effec-
tively kills cells expressing the TK gene
(all those not containing the gene of inter-
est). After selection, viral particles can be
harvested and/or protein production as-
sayed.
The generation of soluble, active protein
is now more important than ever with the
increased emphasis on expression of pro-
teins for structural analysis. One class of
proteins that has gained importance as
druggable targets are human membrane
proteins. Currently, baculoviral expression
is one of the best avenues for the genera-
tion of soluble membrane proteins. There-
fore, easier access to baculovirus expres-
sion technology is now more important
than ever. The use of Gateway cloning to
facilitate transfer of ORFs into large plas-
mids represents the enablement of baculo-
viral expression for HTP applications.
2.10
Multisite Gateway
In this chapter, the attributes of the Gate-
way system have been discussed in terms
of its ability to transfer a single DNA frag-
ment from one vector to another. The
term standard Gateway suggests that
there is a type of Gateway recombination
that is different than what has been dis-
cussed so far. As the mechanism of att site
recombination specificities has been
worked out, it became clear the same spec-
ificity offered the system by the core att se-
quence that generated att1 and att2 would
potentially allow many more att site speci-
ficities. As mentioned previously, the over-
lap sequence within the att site that actu-
ally must have homology to its partner
consists of seven bases and very minor dif-
ferences in the sequence will knock out or
change the specificity of the reaction. Giv-

Fig. 2.7 The BaculoDirect
TM
system.
GFP
en this concept, and taking into account
that all the potential combinations of sev-
en bases equals 4
7
or 16384, one could ex-
trapolate that there are many potential att
site specificities. Clearly, one would not ex-
pect all of these combinations to show
high recombinational activity or absolute
specificity, but it allows for the develop-
ment of multiple att site specificities for
assembly of more that one fragment with
a destination vector. To date, six att sites
have been isolated that show high reactiv-
ity and absolute specificity (minimal cross-
reactivity with other att sites). The Multi-
site Gateway system was developed by
using these sites in combination with the
same enzymatic mechanism used in stan-
dard Gateway [2729]. Using the same
clonase mixes as mentioned above, the
specificity at the level of the overlap se-
quence determines which fragments re-
combine in which order. In Multisite Gate-
way, the entry vectors have different attL
and attR sites, but the destination vectors
remain the same given the correct config-
uration. By relying on the same positive
and negative selection system described
previously, up to four fragments can be ef-
ficiently assembled into a destination vec-
tor in a single reaction tube (Table 2.2).
The recombination mechanism for creat-
ing a Multisite construct is the same as
that with a single fragment Gateway re-
combination. As with standard Gateway,
planning a Multisite construction starts
with the entry clones. The difference is
the att site specificity and the configura-
tion (and orientation) of the att sites. As
seen in Fig. 2.8, entry clones are con-
structed as described above but in order
for them to have the correct configuration
in the final LR assembly reaction, there
needs to be a combination of flanking attL
and attR sites. This is facilitated by the
modular nature of the att sites (see Fig.
2.2, Panel B). For example, by reversing
the standard orientation of the attP1 site
(to an attP1r site), a BP recombination
with an attB1 site (in reverse orientation,
attB1r) results in the formation of an attR1
site instead of an attL1 site. This concept
is key to the elegance of the Multisite
Gateway system.
Table 2.2 attB sequences representing specificities 1 through 6
T S L Y K K V V
attB1 ACA AGT TTG TAC AAA AAA GCA GTG
A T F L Y K V V
attB2 CCA GCT TTC TTG TAC AAA GTG GTN
A T L L Y K V V
attB3 GCA ACT TTA TTA TAC AAA GTT GTG
A T L L Y K V V
attB4 GCA ACT TTG TAT AGA AAA GTT GTB
A T F V Y K V V
attB5 GCA ACT TTT GTA TAC AAA GTT GTG
A T L Y E K V V
attB6 GCA ACT TTG TAC GAA AAA GTT GTG
2.11
Creation of Entry Vectors and
Three-fragment Multisite Assembly Reaction
With the appropriately configured vectors,
the assembly of multiple fragments is
achieved by combining entry vectors with
an appropriate destination vector contain-
ing the correct flanking att sites to capture
the inserts. LR clonase (or Clonase +) is
then added and the reaction incubated for
between 1 and 16 hours at room tempera-
ture. Using the same selection criteria as
with a single fragment recombination, es-
sentially the only way to obtain a colony is
if the proper att sites line up and recom-
bine with each other and the destination
vector. Figure 2.8 shows an example of a
three-fragment Multisite assembly. Five-
fragment assemblies have been achieved
in the laboratory, although colony num-
bers are lower with more complex reac-
tions. Nonetheless, the cloning efficiency
remains in the 90% range. Using the six
currently available att site specificities, as-
semblies of more than 10 fragments are
possible by creating intermediate destina-
tion vectors [30].
Modular assembly of expression con-
structs including promoters, ORFs, epitope
and purification tags, can now be achieved
in HTP, without the use of restriction en-
zymes and ligase. Once a set of entry vec-
tors has been created, the elements can be
mixed and matched in an LR reaction. Ele-
ments in a multi-fragment construct can
be selectively removed via a specific BP re-
combination reaction, thereby creating an
intermediate destination vector. For exam-
ple, an enzymatic pathway can be as-
sembled using modular promoters and
ORFs to create a synthetic operon. Once
the wild-type activity is established, a li-
brary of mutants can easily replace specific

Fig. 2.8 Three-fragment Multisite Gateway cloning. Creation
of entry vectors and three-fragment Multisite assembly reac-
tion.
ORFs in the pathway, without affecting the
surrounding elements. In addition, multi-
domain proteins (e.g., membrane recep-
tors) can be assembled using a modular
approach [27]. Using the Gateway recombi-
nation technology, this can be completed
within a matter of days.
2.12
Perspective
This chapter aims at providing a broad over-
view of Gateway recombinational cloning
and its significance to modern molecular
biology and proteomics. This system repre-
sents a new paradigm in molecular biology
where procedures not previously possible
are now enabled. For example, HTP cloning
of ORFs was essentially impossible prior to
Gateway cloning, but since its introduction
over 100000 ORFs from various organisms
have been cloned as entry vectors and serve
as a valuable resource for the scientific com-
munity. Today, this and future information
archived as DNA sequence can be
handled easily and transferred either singly
or en masse to essentially any plasmid-
based analytical platform. HTP ORFeome
cloning, categorization of novel proteins
by their cellular location, and the enable-
ment of HTP baculovirus cloning and ex-
pression are only three examples of novel
experimental models that are now possible
since the advent of Gateway.
The full power of this system is realized
in the Multisite Gateway system, where
the same mechanism which is used to
transfer single DNA elements from vector
to vector is employed to assemble multiple
fragments. The future of Gateway cloning
will include HTP, controllable assembly of
multi-element plasmids, multi-domain
proteins, and entire enzymatic pathways
that could be expressed in a desired cellu-
lar background. These engineered cells
could be used for many applications from
cell-based assays to production of biophar-
maceuticals.
The Gateway system was designed to
serve as an efficient conduit to shuttle the
incredible volume of genetic information
available to the scientific community to
proteomic analytical platforms for study.
The ability efficiently to manipulate DNA
elements afforded by the Gateway cloning
system has opened (see also Part III,
Chapters 1 and 7) and will continue to
open many doors to novel experimental
protocols that will be invaluable in basic
research linked with the development of
References
1 Venter, et al. 2001. Science 291: 13041351.
2 Jackson, D. A., Symons, R. H., and Berg, P.
1972. Proc. Natl. Acad. Sci. USA 69: 2904
2909.
3 Lobban, P. and Kaiser, A. 1973. J. Mol. Biol.
78: 453471.
4 Cohen, S. N., Chang, A., Boyer, H., and Hel-
ling, R. 1973. Proc. Natl. Acad. Sci. USA 70:
32403244.
5 Bubeck, P., Winkler, M., and Bautsch, W.
1993. Nucleic Acids Res. 21: 36013602.
6 Oliner, J., Kinzler, K., and Vogelstein, B. 1993.
Nucleic Acids Res. 21: 51925197.
7 Degryse, E. 1996. Gene 170: 4550.
8 Zhang, Y., Buchholtz, F., Muyrers, J. P., and
Stewart, A. F. 1998. Nat. Genet. 20: 123128.
9 Lafontaine, D. and Tollervey, D. 1996. Nucleic
Acids Res. 24: 34693471.
10 Storck, T., Kruth, U., Kolekar, R., Sprengel, R.,
and Seeburg, P. H. 1996. Nucleic Acids Res. 24:
45944596.
11 Lucklow, V., Lee, S., Barry, G., and Olins, P.
1993. J. Virol. 67: 45664579.
12 Peakman, T., Harris, R., and Gewert, D. 1992.
Nucleic Acids Res. 20: 495500.
13 Boyd, A. 1993. Nucleic Acids Res. 21: 817821.
14 Liu, Q., Li, M., Leibham, D., Cortez, D., and
Elledge, S. 1998. Curr. Biol. 8: 13001309.
References 619
15 Hartley, J., Temple, G., and Brasch, M. 2000.
Genome Res. 10: 17881795.
16 Nash, H. 1977. Curr. Top. Microbiol. Immunol.
78: 171199.
17 Nash, H. 1981. Annu. Rev. Genet. 15: 143167.
18 Nash, H. 1990. Trends Biochem. Sci. 6: 222
227.
19 Bernard, P. and Couturier, M. 1992. J. Mol.
Biol. 226: 735745.
20 Walhout, A., Temple, G., Brasch, M., Hartly,
J., Lorson, M., van den Heuvel, S., and Vidal,
M. 2001. Methods Enzymol. 328: 575592.
21 Brasch, M. A., Hartley, J. L., and Vidal, M.
2004. Genome Res. 14: 20012009.
22 Reboul, J., et al. 2003. Nat. Genet. 34: 3541.
23 Walhout, A. and Vidal, M. 2001. Methods 24:
297306.
24 Li, S., Armstrong, C. M., Bertin, N., Ge, H.,
Milstein, S., Boxem, M., Vidalain, P. O., Han,
J. D., Chesneau, A., Hao, T., et al. 2004.
Science 303: 540543.
25 Simpson, J. C., Wellenreuther, R., Poustka, A.,
Pepperkok, R., and Wiemann, S. 2000. EMBO
Rep. 1: 287292.
26 Rubsam, L. Z., Boucher, P. D., Murphy, P. J.,
Kuruga, M., and Shewach, D.S. 1999. Cancer
Res. 59: 669675.
27 Cheo, D. L., Titus, S. A., Byrd, D. R. N., Hartley,
J. L., Temple, G. F., and Brasch, M. A. 2004.
Genome Res. 14: 21112120.
28 Sasaki, Y., Sone, T., Yoshida, S., Yahata, K.,
Hotta, J., Chesnut, J., Honda, T., and Imamo-
to, F. 2004. J. Biotech. 107: 233243.
29 Hope, I., Stevens, J., Garner, A., Hayes, J.,
Cheo, D., Brasch, M., and Vidal, M. 2004. Ge-
nome Res. 14: 20092110.
30 Sone, T., Yahata, K., Sasaki, Y., Hotta, J.,
Kishine, H., Chesnut, J., and Imamoto, F.
2005. J. Biotech. (in press).

Abstract
Scientific breakthroughs and novel tech-
niques in basic science are often rapidly
adapted by the biopharmaceutical industry,
and can also originate from the drug dis-
covery process itself. In this chapter, an
overview is given of state-of-the-art tech-
niques in molecular and cellular biology
used in the drug discovery process (espe-
cially at the early step of target validation).
The spectrum ranges from RNA/DNA-
based methods to protein/peptide-based
approaches. In addition, cellular assays are
described in which these techniques can
be employed. Some of these approaches
especially the use of RNA interference
(RNAi) and of transporter peptides are
described in more detail, because both are
relatively novel, have been proven to be
very successful, and have great potential to
be developed into new biopharmaceuticals
themselves. In particular, RNAi has made
a quantum leap from its initial discovery
in 1998 to its application as the method of
choice in targeted gene silencing (see Part
I, Chapter 10 and Part II, Chapter 8). Over
500 protein kinases have recently been de-
scribed in the human genome, and the
protein kinase family is used here as an
important example for a class of cellular
targets. Several techniques for the valida-
tion of members of the protein kinase
family are described, and two novel, com-
mercially available biopharmaceutical
drugs, which target cellular kinases, are
portrayed. We further provide here a list of
signaling pathways and a compilation of
potential molecular targets with their links
to specific diseases. We hope to give the
more experienced reader new insights to
novel developments in the biopharmaceu-
tical field and, at the same time, to excite
the less experienced reader about the drug
discovery process.
621
ISBN: 3-527-31184-X
3
Target Validation: An Important Early Step in the Development
of Novel Biopharmaceuticals in the Post-genomic Era
In Vivo Veritas Early Target Validation in Knock-out Mice and More
Abbreviations
Aa amino acid
ADME Absorption, Distribution, Meta-
bolism and Excretion
BrdUTP bromodeoxyuridine triphos-
phate
CML chronic myeloid leukemia
CMV cytomegalovirus
DN dominant negative
DNA desoxyribonucleic acid
EGFR epidermal growth factor recep-
tor
ELISA enzyme-linked immunosorbent
assay
EST expressed sequence tags
GIST gastrointestinal stromal tumors
HER human epidermal growth factor
receptor
hnRNA heterogeneous nuclear RNA
HPV human papillomavirus
KD kinase dead
MAPK mitogen-activated protein kin-
ase
miRNA microRNAs
mRNA messenger RNA
MTS methylthiazoltetrazolium deri-
vate
MTT methylthiazoltetrazolium
NHEK normal human epidermal kera-
tinocytes
NIH National Institutes of Health
NMDA N-methyl-D-aspartate
NSCLC non-small cell lung cancer
nt nucleotide
PARP poly (ADP-ribose) polymerase
PDGFR platelet-derived growth factor re-
ceptor
PDZ protein interaction domain de-
rived from the first three identi-
fied members: PSD95, DLG-1
and ZO-1
PS phosphatidylserine
RNA ribonucleic acid
rRNA ribosomal RNA
RTK receptor tyrosine kinase
SELEX systematic evolution of ligands
by exponential enrichment
shRNA small hairpin RNA
snRNA small nuclear RNA
stRNA small temporal RNA
Tat transactivator of transcription
(HIV protein)
TET tetracycline
tRNA transfer RNA
Tunel Terminal deoxynucleotide trans-
ferase dUTP nick end labeling
VEGF vascular endothelial growth fac-
tor
3.1
Introduction
Some 10 years ago, the biopharmaceutical
industry was very eager to identify novel
targets and to use them in their drug dis-
covery programs. Today, the incentive is
more to conserve money by focusing on
more validated targets. With the human
genome project finished, the identification
of targets is easier and, in principle, all of
the roughly 30,000 genes identified by the
Human Genome project could be used as
potential targets for drug discovery pro-
grams. More realistic estimates range from
a few thousand to 10,000 putative drug
discovery targets in humans. As no bio-
pharmaceutical company has the resources
to investigate more than a handful of tar-
gets at a time, assessing the quality of
these targets or in other words validating
3 Target Validation: An Important Early Step in the Development of Novel Biopharmaceuticals 622
these targets is a crucial step in the drug
discovery process (Fig. 3.1).
Several target databases have been estab-
lished and are helpful tools for future ad-
vances in drug discovery. Examples are the
NIH target database, which includes data
from the worldwide structural genomic
and proteomic project, and the therapeutic
target database of the National University
of Singapore. However, the key questions
remain: What is a target, and what makes
it a valid one? Here, companies generally
do differ in their requirements for target
validation. In general terms, a target is a
molecule (often a protein) that is instru-
mental to a disease process, although it
might not be directly involved. Target valida-
tion verifies that the target molecule is an
essential element in the disease process,
and that it constitutes a potential point of
therapeutic intervention. During the target
identification process, targets are linked to
the generation, the progression or the
symptoms of a disease. A potential target
becomes a validated target when it is con-
vincingly demonstrated that altering (in-
creasing or inhibiting) its biological activity
leads to an improvement in a respective dis-
ease model. The functional validation of a
target prior to the start of a drug discovery
program is a very critical early step, and fast
and accurate target validation technologies
are the foundation for future successful
drug discovery. Diametrical but equally
important at this early stage can be the
less pleasant target invalidation. The early
investment in target validation technologies
can clearly pay off at later stages, or at least
save much money later spent. In addition,
many different technologies including tar-
geted gene silencing, protein inhibition, cel-
lular assays, chemical genetics and combi-
natorial biology are used for the target val-
idation process. In most cases, a combina-
tion of tools, including gene/protein
modifications and model system ap-
proaches, along with expression and proteo-
mic data, are utilized to validate and priori-
tize targets of interest (see Part I, Chapter
4). Several, but not all, of the target valida-
tion technologies will be discussed here.
Some of these methods bear the potential
to turn into biopharmaceutical drugs them-
selves. A crucial aspect of the target valida-
tion process is to understand systems biol-
ogy, and to use (if possible) a good model
system for the respective disease. To help
understand the systems biology it is neces-
sary to be familiar with the cellular signal
transduction networks. This encompasses
aspects of qualitative and quantitative biol-
ogy, and involves the timing and localiza-
tion of signaling complexes, as well as their
regulation. Many intracellular signaling
pathways have been described, and many
of these are very highly conserved from
yeast, to worms, to flies, and to humans.
A good example is the epidermal growth
factor receptor (EGFR) pathway, which
Fig. 3.1 Stairway to drug discovery. The drug dis-
covery process as modeled here is not necessarily
a linear process. Novel potential biopharmaceuti-
cals (e.g., siRNAs) used in early steps of target
identification and validation can, in principle, skip
the next steps and be tested directly in ADME/tox-
icity studies (see Part VIII, Chapter 4). ADME, Ad-
sorption, Distribution, Metabolism, and Excretion;
HTS, high-throughput screening.
leads to the activation of mitogen-activated
kinases (MAPK). The EGFR pathway is con-
served from worms (Caenorhabditis elegans),
where it plays a role in vulva development,
to flies (Drosophila melanogaster) where it
is involved in the development of the eye,
to humans where it is important for cell
growth and differentiation. In order for
the pharmaceutical industry to benefit from
the huge progress made in genomics, pro-
teomics and the associated bioinformatics,
it remains crucial to understand and test
the systems biology.
Understanding systems biology can pro-
vide means for identifying pathways that
are critical to disease. It is important to
understand not only qualitative but also
quantitative aspects of systems biology. As
an example, proteinprotein interactions,
such as the interaction of a PDZ protein
with its PDZ ligand motif containing bind-
ing partner, depend on cellular localization
and also very strongly on binding affini-
ties. Both play important roles for deter-
mining the actual relevant interactions in
a biological system. Changes in these in-
teractions can be crucial for disease devel-
opment and progression.
To study signaling pathways in silico, an
excellent resource is a site of the NIH
(http://cgap.nci.nih.gov/Pathways), which
is linked to biocarta (www.biocarta.com),
and the Kyoto Encyclopedia of Genes and
Genomes (KEGG) (www.genome.ad.jp/
kegg/). Additional information on individu-
al components of these pathways is available
at various locations, and one useful site that
is also linked to the biocarta is Genecards
of the Weizmann Institute in Rehovot
(http://bioinfo.weizmann.ac.il/cards/).
However, two-dimensional models of signal-
ing networks will only partially help in an
understanding of systems biology, and bear
the potential danger of misleading and over-
simplifying systems biology properties.
Several important signaling pathways
and potential targets, together with their
implications for different diseases, are
listed in Table 3.1.
3.2
RNA- and DNA-based Techniques
for Post-transcriptional Regulation
of Molecular Targets, and their Potential
as Biopharmaceutical Drugs
Nucleic acids are increasingly and heavily
used for target validation throughout the
drug discovery process. In addition, nu-
cleic acids are also being considered for
therapeutic uses, either to interfere with
the function of specific nucleic acids or to
bind specific proteins. Three types of nu-
cleic acid drugs are discussed here: anti-
gene compounds, such as small interfer-
ing (si) double-stranded (ds) RNA, anti-
sense oligos and antisense RNA, which
lead to post-transcriptional gene silencing
(see Part I, Chapters 1 and 10); and ribo-
zymes (catalytic RNA), which bind and
cleave RNA targets and also lead to gene
silencing. The phenomenon of catalytic
RNA was discovered by Thomas R. Cech,
who was awarded the Nobel Prize for
these findings. Oligonucleotide aptamers,
which can bind and inhibit proteins direct-
ly, can also lead to functional inactivation.
3.2.1
RNA Interference (RNAi):
Silencing Genes as Tools and Therapeutics
RNAi is a very promising new method for
targeted post-transcriptional gene silencing
in plants and animals. It has major implica-
tions for the drug development process, and
can also be used for target identification as
well as target validation. Moreover, it has
great potential for the development of fu-
ture novel biopharmaceutical therapeutics.
3.2 RNA- and DNA-based Techniques for Post-transcriptional Regulation of Molecular Targets 625
Table 3.1 Important signaling pathways and potential targets, and their implications for different diseases.
Pathways Potential targets Disease/disorder
Synuclein Synuclein, Synphilin, Parkin (E3), UBAI (E1),
UBCH7,8 (E2), GPR37, SNCAIP, CDCrel1, PKA,
Reactive Oxygen Species (ROS), p38, JNK, Cas-
pases
Parkinsons
Amyloid Precursor
Protein Processing
Amyloid Precursor Protein (APP), a,b,c-Secre-
tases, Caspase3, Presenilin, Aph-1, Pen-2, IDE,
GSK3a
Alzheimers
Angiogenesis VEGF, VEGFR, FGF, FGFR, PDGF, PDGFR,
FLT3, PKC, PI3K, AKT, MAPK, MMP-2
Cancer
Anti-inflammatory IL-2, IL-2R, IL-4, IL-4R, IL-10, IL-10R Inflammation, Cancer
Apolipoprotein
metabolism
Apolipoprotein E, LDL Receptor Alzheimers
Apoptosis FASL, FAS(CD95), FAP, FADD, Caspases, TNFR,
RIP, DR, BCL-2, p53
Stroke, Cancer
Arachidonic acid
metabolism
Lipoxygenases, Prostaglandins, Thromboxanes,
Cox, Leukotrienes, LT receptors
Hypertension, Inflam-
mation
cAMP-dependent
signaling
GPCR, Adenyl cyclase (AC), PKA Polycystic kidney dis-
ease, Depression, Cancer
Cell adhesion and
motility
E-Cadherin, a,b-Catenin, FAK, Cadherin5, P-Se-
lectin, ICAMs, LIMK1,2, p130CAS, Integrins,
PECAM, Paxillin, Talin, Calpain, FAK, Vincullin,
Src, CSK, Actin stress fibers
Peripheral vascular
disease, Cancer
Cell cycle regulation Growth factors and their receptors, Cyclins,
Cyclin-dependent kinases (CDKs), MAPKs,
HSP90, CDC37, p16
Cancer, Osteoarthritis
Cell cycle G
1
/S TGF-b, ATM (ATR), Rb, E2F, Cdk4, 6/CyclinD,
DHFR, TK
Cancer
Cell cycle G
2
/M Cdc2/CyclinB, Myt1, Wee1, Cdc25, p90RSK,
Chk1,2, p53, p300, DNA-PK, ATM
Cancer
Ceramide signaling CAPK, Raf, MEK, ERK1,2, MEKK1, MEK4, JNK,
TNFR, Fan, Bax, Bad, Bcl-2
Cancer, Alzheimers,
Fabrys
Cell survival AKT, PDK1, IKK, NFjB, Bad, Casp9, FKHRL1 Cancer, Huntingtons,
Alzheimers
CFTR and a-adrenergic
receptor signaling
CFTR, PKA, Ezrin, EBP50R, b2-AR, G-protein,
AC
Cystic fibrosis
Chemokine signaling CXCR2, CCR3, CCR4, CCR5, CCR6, ERK, AKT,
p38
Cardiovascular disease,
Inflammation, Infection,
Metabolic disease, Respi-
ratory disease, Cancer
Cyclic nucleotide
metabolism
PDE2, PDE3, PDE4, PDE5 Depression, Asthma,
Cancer
DNA damage DNA-PK, ATM (ATR), Chk1,2, p53, p73, MDM2,
RAD51, GADD45, NFjB, BRCA-1, c-Abl
Cancer
Double-stranded (ds)
RNA induced gene
expression
PKR, p53, p58IPK, eIF2, IKK, NIK, IkBa, NFjB Huntingtons
Double-stranded RNA-
induced gene silencing
(RNA interference)
Dicer, Ago4, RISC (Paz, Ago1,2,3) No disease associated
EGFR transactivation
by GPCR
Angiotensin, Endothelin, LPA, GPCR, PLCy,
PKC, ADAM, MAPK pathways, NFjB, Rho
Cardiac hypertrophy,
Arteriosclerosis
EPO signaling EPOR, PLCy, STAT5, JAK2, SHP-1, SHC, GRB-2,
SOS, RAF, MEK, ERK, ELK-1
Huntingtons, Cancer,
Renal and kidney
disease, Friend disease
Glucose homeostasis Glucagon, Glucagon receptor, Glucokinase,
GFPT1, GFPT2, Glycogen phosphorylase,
Glycogen synthase
Diabetes
Growth hormone
signaling
GHR, JAK2, STAT3,5, PIAS, IRS-1, PI3K, AKT,
FAK, SRC, PYK2, PKC, CAMK
GH-deficient dwarfism,
Kidney disease, Prader-
Willi syndrome, Turner
syndrome
Hormone signaling DPP4, CRHR1, Melanocortin3 and 4 R, Neuro-
peptide Y receptor Y5, GPR24, ADCYAP1,
ADCYAP1R1
Diabetes, Depression,
Obesity, Urogenital
Hypoxia-induced
signaling
HIF1a, MDM2, p53, 28S, Proteasome, IkBb Cancer, Chronic inflam-
matory bowel disease,
Rheumatoid arthritis,
Ischemia/Reperfusion
Insulin signaling Insulin, Insulin receptor (IR), IGF1R, PTPN1,
IRS-1, GRB2, SHP-2, PI3K, GLUT4, SHC, SOS,
Ras, ERK, JNK
Diabetes, Cancer
IFN pathway IFN-a,b,c, JAK1,2, STAT1,2, TYK1 Multiple sclerosis, Anti-
viral immune response,
Cancer
Lipid homeostasis,
Cholesterol biosynthesis
HMG-CoA reductase, SCAP, SREBP, ACAT-1,
ACAT-2, ACACA, ACACB, LXR-a, DGAT1, LDLR,
LRP2, ABCA1
Hyperlipidemia, Dys-
lipidemia, Obesity, Alz-
heimers, Coronary heart
disease, Atherosclerosis
Mitogen-activated
protein kinase (MAPK)
signaling
EGFR, HER2, PDGFR, GRB-2, SOS, Shc, Ras,
Raf, MEK, Erk, Rac, Cdc42, MEK4,7, JNK, MEK6,
TAK, ASK, MLKs, MAPKAP2, p38
Inflammation, Cancer,
Neurological
(Parkinsons)
Matrix metalloproteinase
signaling
MMP2, MMP9, TIMP1,2,3,4, RECK Cancer, Metastasis
Neurokinin signaling Tachykinin receptors Depression, Urogenital
Neurotransmission Neuropeptide signaling (Orexin receptor, Canna-
binoid receptor, NAALAD2, Neurotensin recep-
tors, Serotonin receptors, GABA A receptors,
Metabotropic Glutamate receptors, Folate Hydro-
lase1)
Depression, Alzheimers,
Anxiety, Schizophrenia,
Obesity, Urogenital
Neurotransmitter
receptor regulation
Neuregulin 1, ErbB4 (HER4) Schizophrenia
NFjB signaling LPS, TLR, TRAF6, IRAK, MYD88, TNF, TNFR,
FADD, TRADD, RIP, IL-1, IL-1R, IKK, MEKK1,
NIK, IkBa
Inflammatory diseases
NO signaling NMDAR, PSD95, NOS, Calcineurin, PKA, PKC,
HSP90, Caveolin, G-Cyclase, PDE, PKG
Stroke, Cardiovascular
diseases
Nociception Opioid receptors, Vanilloid receptors Acute pain
Notch signaling Delta, Notch, Next, Presenilin, NICD, ADAM,
Furin
Cancer (T-cell leukemia),
Alzheimers
Nuclear hormone
receptor signaling
PPAR, RAR, RXR, Progesterone receptor (PR),
Androgen receptor (AR), Estrogen receptor (ER),
LXR-a, FXR
Lipid metabolism,
Inflammation, Athero-
sclerosis, Diabetes,
Cancer, Alzheimers
Phosphatidyl inositol
signaling
PI3K, AKT, PDK, PTEN, mTOR, GSK3b, SHIP,
PLC, PKC, SHC, GRB-2, SOS, Ras, Raf, Erk,
Elk1, Fos
Inflammation, Rheuma-
toid arthritis, Cancer,
Respiratory diseases
Prion signaling PRP
c
, PRP
SC
, Laminin, Laminin receptor, GFAP,
Bcl-2, NRF2, HSP40, HSPD1, HSPA5, BIP
Creutzfeldt-Jakob,
Bovine spongiform
encephalitis
Pro-inflammatory TNF-a, TNF-aR, Tace, IKK, IL-1, IL-1R, IRAK,
IL-6, IL-6R, Toll-like receptors, BlyS, CD40L, LT-b,
GM-CSF
Inflammatory diseases,
Autoimmune disorders,
Cancer
Protein acetylation HDAC family Cancer, Inflammation,
Neurodegenerative
diseases
3.2.1.1 History and Background
Nucleic acids such as desoxyribonucleic
acids (DNA) and ribonucleic acids (RNA)
are unbranched nucleotide chains, which
are joined together by phosphodiester
bonds. RNA contains ribose and the pyrimi-
dine uracil, instead of desoxyribose and thy-
midine used in DNA. There are many dif-
ferent RNA species with different func-
tions, for example transfer-RNA (tRNA), ri-
bosomal RNA (rRNA), small nuclear RNA
(snRNA), heterogeneous nuclear RNA
(hnRNA), snoRNA involved in RNA methy-
lation and RNA stability, catalytic active
Hammerhead and hairpin ribozymes, mi-
croRNAs (miRNA), small interfering RNAs
(siRNA) both of which are involved in
gene regulation and messenger RNA
(mRNA). mRNA is the intermediary be-
tween the information-containing DNA in
a cells nucleus and the production of a pro-
tein in a cells cytoplasm. In recent years,
more complex activities of RNA have been
discovered, and small temporal RNA
(stRNA) and small interfering RNA (siRNA)
involved in RNA interference (RNAi) have
been added to the list. RNAi is a natural
process that some organisms use to destroy
viruses by intercepting and destroying the
genetic message encoded by the virus and
transcribed by the cellular machinery. The
interception takes place at the level of the
mRNA, which is destroyed in the cytoplasm
or shortly before its release into the cyto-
plasm. Consequently, the genes function
is silenced, or knocked down, without in-
terfering with the gene itself. In addition
to the role of RNA interference in anti-viral
protection, some organisms use RNAi to re-
press the translation of endogenous genes.
Protein prenylation Protein farnesyltransferase a and b Cancer
Sonic Hedgehog
signaling
Shh, PTC-1, Smoothened, GSK3b, PKA, SUFU,
GLI-1,2,3, DYRK1, Cdc2/CyclinB
Cancer, Parkinsons
T- and B-cell signaling TCR, Calcineurin, IgE, FceRI, FceRII Immune response,
Allergic disorders
Tau phosphorylation Tau, GSK3b, CDK5, MARK, p35, Calpain1, PP2A Alzheimers
Telomerase signaling TERT, HSP90, TEP1, TRF1, AKT, PKCa, SP1,
Myc/Max, Estrogen, Rb, NF-Y, SP3, WT-1, Mad/
Max, HDAC, p53
Aging, Mortality
TGF-b signaling TGF bR, SMAD2,3,4,7, TAK1, TAB1, SARA,
SnoN
Cancer, Atherosclerotic
plaques
Toll-like receptor
signaling
TLR, MYD88, IRAK, TRAF6, ESCIT, MEKK1,
NFjB, TAK1, TAB1,2, JNK, p38, IL-1, IL-12,
TNFa
Infectious diseases
Ubiquitin-mediated
proteolysis
E1 (UBA), E2 (UBC), E3 (UBR), F-Box proteins,
SKP1, cdc34, APC, cdc4, Grr1, VHL, Cul2, Rbx1
Von HippelLindau,
Cancer, Parkinsons
WNT signaling Frizzled, DSH, Naked, PP2A, GSK3b, APC, b-
Catenin, Axin, bTRCP, Bcl-9, TCF/LEF
Cancer, Hair growth
The silencing phenomenon was first discov-
ered in plants and in the nematode C. ele-
gans. In 1998, Fire and Mello discovered
the gene silencing mechanism by double-
stranded (ds) RNA molecules in worms,
and it was at this time that the term RNAi
was created. In 2001, Tuschl identified 21
nucleotide (nt) -long, siRNAs as being the
mediators of this gene-silencing mecha-
nism, and demonstrated its use for gene si-
lencing in mammalian tissue culture cells.
These findings were rapidly followed by
the discovery of the cellular machinery in-
volved in RNAi, and by identification of
the enzymes dicing up the targeted mRNA.
A model for the mechanism of RNAi is
shown in Fig. 3.2.
The rapid transition from the initial find-
ings and discoveries to the widespread use
of RNAi in academic laboratories, as well
as in the pharmaceutical industry, under-
lines the potency of this new method of tar-
geted gene silencing. In addition, current
progress has been made to bring RNAi into
the clinic and turn siRNAs directly into
drugs. Before the discovery of RNAi, sin-
gle-stranded antisense RNA or antisense
DNA oligos were used for gene silencing.
It was due to the findings of Tuschl and
others that a functional RNAi pathway was
seen to be operating in mammalian cells,
whereupon RNA interference became a
powerful tool for scientists to study the
function of genes not only in organisms
(e.g., in C. elegans and D. melanogaster),
but also in mammalian cells, tissues, and
whole organisms. However, the initial appli-
cation of this process to mammalian cells
proved unworkable, as dsRNA provoked
non-sequence-specific responses through
the interferon or PKR pathways. The inter-
feron pathway is normally used to combat
Fig. 3.2 (A) The RNAi pathway: Dicer with its
RNAse III domains cleaves long double-stranded
RNA (dsRNA) twice per strand, and generates the
siRNA duplex. After cleavage, a siRNA/protein
complex (siRNP) is formed. In the next step, the
siRNA is unwound and the RNA-induced silencing
complex (RISC) becomes activated. The target
mRNA is recognized based on its complementary
sequence to the siRNA, and is cleaved by the
RISC complex. (B) A single-stranded RNA can be
expressed which forms a 19 nt-long hairpin struc-
ture with a loop and 3' terminal uridines. The
hairpin siRNA is recognized by Dicer and cleaved
to form a functional siRNA.
viral infections, and one consequence of
this response is that cells undergo pro-
grammed cell death (apoptosis) in order to
prevent infiltration of the entire organism
with the propagated virus. By using smaller,
21- to 23-nt-long siRNAs, this central limita-
tion was circumvented, and it became possi-
ble to apply RNAi to mammals, without
triggering unspecific responses. It is impor-
tant to know that high doses of siRNAs may
still lead to unspecific gene silencing ef-
fects, and it is crucial to use several control
siRNAs. In addition, some siRNAs have
been found to have off-target effects and to
silence other genes that are not related to
the initial target gene. However, this is a
rare occurrence, and as yet is not fully un-
derstood. A possible hypothesis is that these
siRNAs affect gene and/or RNA stability in
some other way, perhaps involving hetero-
chromatin formation or other gene regula-
tory functions. Nonetheless, a combination
of genomic data with RNAi-directed gene si-
lencing allows the functional determination
of almost any gene expressed in a cell or
pathway, and is one of the newest and most
successful methods in target validation.
3.2.1.2 Mechanism of Action of RNAi
RNAi is mediated by siRNAs. Starting from
longer, double-stranded precursors, these
long dsRNAs are cleaved into smaller dou-
ble-stranded segments of 2123 nt by an en-
zyme with RNAse III activity, known as Di-
cer. These siRNAs act as guides and se-
quence the recognition elements of the mul-
ticomponent RNA-induced silencing
complex (RISC), which then specifically
binds the corresponding mRNA and targets
it for degradation. RISC is thought to probe
for the target mRNA using the antisense si-
RNA strand as a template, and then med-
iates its cleavage. A model for the RNAi
pathway and the enzymes involved is shown
in Fig. 3.2 (see also a video animation on the
supplement CD-ROM). The process may
also be triggered directly by the introduction
of chemically or enzymatically synthesized
siRNAs. These siRNAs can, in principle,
suppress the expression of any gene or gene
family with no unwanted, non-specific dam-
age to the cells. The small size of the siRNAs
enables them to knock down gene expres-
sion without provoking undesirable cellular
responses, such as the interferon pathway.
3.2.1.3 Design of siRNAs
Important characteristics for siRNAs are
the short overhanging 3' nucleotides and
the ideal length, which varies somewhere
between 19 nt to 28 nt. Typically used are
19- to 22-nt-long siRNAs with 3'-terminal
dinucleotide overhangs. The initial step in
the design of siRNAs is usually to identify
the target sequences on the mRNA. The
guidelines originally derived from Tuschls
studies start by identifying a specific,
usually unique 19 nt-long target sequence
typically preceded by a dinucleotide leader
used to define the sequence composition
of the two-base 3' overhangs. These target
sites are then evaluated for GC content,
with 50% being ideal. It is now recom-
mended to compare the sequences to ex-
pressed sequence tags (EST) and other da-
tabases to assure the uniqueness and spec-
ificity of the siRNA. It is important to state
that the method of choice with several ad-
vantages is to design the siRNAs through
rational siRNA design using different algo-
rithms available from the manufacturers
of siRNAs [e.g., Ambion, Xeragon (now
Quiagen) or Dharmacon]. These more
comprehensive methods of predicting si-
lencing take more structural and sequence
characteristics into account than just the
GC content. Success rates depend on the
targeted genes, and it is recommended to
test at least two (or better still, three) si-
RNAs for their gene knock-down effects.
(For a comprehensive overview on the ra-
tional design and use of up to four siRNAs
in parallel, see Part I, Chapter 10.)
3.2.1.4 Methods for Generating siRNAs
In general, there are four methods to gener-
ate siRNA: 1) chemical synthesis; 2) in vitro
cleavage of larger dsRNA using recombi-
nant Dicer (as developed by Myers and Fer-
rell) [1]; 3) in vitro transcription of siRNAs;
and 4) expression of hairpin vectors encod-
ing small hairpin RNAs (shRNA).
Chemical synthesis RNA synthesis was
first developed in the 1980s, and synthesiz-
ing both strands of the siRNA and then an-
nealing them together allows the generation
of any size of RNA, including siRNAs and
longer sequences for shRNA. The chemical
synthesis allows for modification, is suitable
for high-throughput synthesis approaches
(e.g., for target identification), and for gen-
erating large quantities of siRNAs. A disad-
vantage is the currently relatively high cost
of chemically synthesized RNAs.
In vitro dicing of larger dsRNA siRNAs
can be generated from larger dsRNAs by
using recombinant, purified enzymes such
as Dicer. In principle, in vitro-synthesized
or in vitro-transcribed longer RNAs can be
annealed to generate full-length or larger
dsRNA fragments. These longer dsRNAs
can be cleaved in vitro into the functional si-
RNAs using recombinant Dicer the same
enzyme that is processing the RNA in vivo
and which was identified by Hannons
group. This method, which uses the gener-
ated pool of siRNAs to induce gene silen-
cing, has both advantages and disadvan-
tages. Using a pool of siRNAs circumvents
the design of a single effective siRNA, and
thus the chances are higher for silencing.
On the other hand, the chances are also
higher for off-target effects on other genes,
and to control this by blasting all putative
sequences is not feasible. Overall, this
approach adds an alternative method for
RNAi which can be used to silence more
difficult genes. The siRNA pool approach
can allow lower doses of individual siRNA
to be effective, or yield a higher degree of
gene silencing if several individual siRNAs
within the pool have additive effects.
In vitro transcription of siRNAs Commer-
cially available in vitro transcription sys-
tems are available which commonly use
T7 or SP6 to promoters for transcription.
Both strands for the respective double-
stranded siRNA can be transcribed in vitro.
The same vector can be used to generate
the sense and antisense strands, and it is
easy to anneal these in order to generate
dsRNAs. Alternatively, a single-stranded
shRNA can be transcribed in vitro. This is
a low-cost method, which is not suitable
for directed high-throughput approaches.
Expression of vector-based approaches for
RNAi For this approach, shRNAs encoded
by expression vectors are used. The DNA-
based vectors generally used have RNA poly-
merase III promoters to generate shRNAs
which then lead to the generation of siRNAs
by the cellular RNAi machinery. These hair-
pin constructs seem to be efficient in gene
silencing for almost 50% of the time, and
this vector-based expression of siRNAs is a
way of expressing siRNAs in cells, without
the need for in vitro-transcribed dsRNA or
synthetic siRNAs. This is by far the cheapest
method of using RNAi, and retroviral vec-
tors encoding shRNAs have been shown to
be successful for RNAi. Expression vectors
with selection markers for the stable
knock-down of non-essential genes have
also been shown to work in this way. The
use of a selection marker allows for long-
er-term stable expression of the shRNAs.
The non-selected hairpin vector expression
is relatively transient, and siRNAs are usual-
ly detectable for not much longer than
72 hours post transfection. Stability for the
in vitro-transcribed or the chemically synthe-
sized siRNAs seems to range from 36 days,
and is most likely comparable to that of
shRNAs. Inducible silencing, for example
through the tetracycline system (TET on/
off) or other inducible systems, has also
been proven to work. The in vivo expression
of shRNA vectors holds some promise for
the future use of RNAi in gene therapy.
3.2.1.5 Delivery of siRNAs
In general, synthesized siRNAs, in vitro-
transcribed siRNAs, pools of in vitro diced
siRNAs and plasmids for expression of the
shRNAs all need to enter the target cells.
Feeding DNA or RNA to worms (e.g., C. ele-
gans) can render the uptake of RNA and
DNA with food relatively straightforward.
The task is more difficult in mammalian
cells, however, and in the case of mamma-
lian cells transfection and infection meth-
odologies depend largely on the cell type
used (see Part VI, Chapter 7). Standard
DNA transfections such as lipofection and
calcium chloride transfection appear to be
efficient, and to function in many cell types.
Electroporation is an excellent alternative,
and works especially well for dsRNAs. Mi-
croinjection can be used especially in larger
cells and on a single cell level. For high effi-
ciencies in mammalian cells and espe-
cially in primary mammalian cells indi-
vidual optimization is necessary. Other ap-
proaches, for example using AMAXA or
the siFect system from Gene Therapy Sys-
tems, have been shown to have high effi-
ciencies, including primary cells. In addi-
tion, the retroviral transduction of hairpin
expression vectors for siRNAs may be a use-
ful alternative, and generally yields high in-
fection rates (see Part I, Chapter 7 and Part
VI, Chapters 1, 3, and 6).
3.2.1.6 siRNAs as Potential Novel
Biopharmaceutical Drugs
The rapid transition, from initial findings
and discovery to the widespread use of
RNAi in both academic laboratories and
the pharmaceutical industry, underlines
the potency of this new method of targeted
gene silencing. In addition, current pro-
gress has been made to bring RNAi into
the clinic and to convert siRNAs directly
into biopharmaceutical drugs. Today, the
use of RNAi for genetic-based therapies is
the subject of many studies, especially for
viral infections (one of the natural roles of
RNAi is anti-viral defense), for cancer, and
for inherited genetic disorders. Indeed, ini-
tial experiments in the use of siRNA to treat
influenza, hepatitis and HIV have shown
great promise, and RNAi was used success-
fully to block HIV replication in cell culture
(see Part II, Chapters 7 and 8). Further-
more, successful in vitro results using si-
RNA have been achieved for the inhibition
of liver-cell inflammation by targeting the
FAS receptor, for inhibition of K-Ras-in-
duced tumorigenesis, and for targeting the
mutated allele that causes amyotrophic lat-
eral sclerosis. Many other applications for
RNAi are currently under investigation,
and RNAi holds great promise for success-
ful new biopharmaceuticals. Despite such
promise, RNAi-based therapeutics have a
long way to go before their useful applica-
tion in the clinic. Inherited delivery prob-
lems must be solved; moreover, as yet RNAi
amplification of RNAi-based gene silencing
has been observed in worms and plants, but
not in mammalian cells. In worms, RNAi
can also be passed on to the offspring.
3.2.2
Antisense RNA and Antisense Oligos
as Useful Tools for Gene Silencing
It is interesting to look back at the history
of early antisense technology, which first
emerged in the early 1980s and originally
was aimed at interrupting the biology of
diseases in a more fundamental and pre-
sumably better way than had been
achieved with recombinant protein thera-
pies. Rather than interfering with protein
interactions and activities in the organism,
antisense drugs were intended to inhibit
gene expression and thus block the initial
production of a disease-causing protein.
These antisense drugs were short oligonu-
cleotides designed to pair with an mRNA.
In theory, such pairing should block the
translation of the mRNA into protein di-
rectly by blocking the protein synthesis
machinery, or indirectly by inducing the
enzymatic destruction of the antisense-
bound mRNA. Initial results with anti-
sense drug candidates were disappointing.
DNA and RNA have poor pharmacoki-
netics in part because they are rapidly
degraded in blood and within cells.
Furthermore, the promised specificity of
antisense technology was unrealized;
many antisense drug candidates were
found to induce a therapeutic response
simply by eliciting a non-specific stimulus
to the immune system, rather than by ex-
erting a specific effect on the targeted
gene. In recent years, the chemical struc-
tures of oligonucleotides have been opti-
mized, and the half-life of the antisense
molecules has been improved. These next-
generation compounds have recently re-
captured the interest of the pharmaceutical
industry to antisense technology. Vitravene
(fomivirsen), which is the first FDA-ap-
proved antisense drug, is used to treat a
condition called cytomegalovirus (CMV)
retinitis in people with AIDS. Isis Pharma-
ceuticals (Carlsbad, CA, USA) has devel-
oped the drug and licensed the worldwide
commercial rights to Novartis. OncoGenex-
Technologies, Inc. (Vancouver, British Co-
lumbia, Canada) and Isis Pharmaceuticals
established a drug development collabora-
tion in 2001 to develop and commercialize
OGX-011 (ISIS 112989), an anti-cancer
antisense drug which inhibits clusterin.
OGX-011 has currently entered Phase II
clinical trials for patients with prostate
cancer and other solid tumors. Another ex-
ample is LY 2181308 (ISIS 23722), a sec-
ond-generation antisense drug which was
licensed to Eli Lilly. In preclinical studies,
LY2181308 demonstrated activity in multi-
ple in vivo models of cancer, and in No-
vember 2004 Lilly initiated Phase I clinical
trials in cancer patients. LY2181308 targets
survivin, a molecule that allows the surviv-
al of cells that would normally undergo
apoptosis. When cancer cells grow, they
appear to need the help of survivin, and
this molecule is abundant in many types
of cancers, including colon, brain, lung,
skin and others, but nearly non-existent in
normal cells.
Another such example is that of Gena-
sense
, produced by Genta Inc. (Berkeley

Heights, NJ, USA). This drug inhibits the
production of a protein which is known as
Bcl-2 and is broadly expressed in most
common types of cancer. Genta Inc. is cur-
rently conducting or planning clinical
trials in a number of these cancers, includ-
ing melanoma, multiple myeloma, acute
myeloid leukemia, chronic lymphocytic
leukemia, prostate cancer, and lung cancer.
In comparison to antisense RNA, RNA
interference is just a little over 5 years old.
RNAi has been shown to inhibit gene ex-
pression post-transcriptionally via cytoplas-
mic mRNA degradation; it has also been
used successfully for tissue-specific gene
knockdown in mice, thus proving its
function in the whole animal. The effec-
tive siRNAs can be routinely and rapidly
synthesized, without any time-consuming
search for active sequences as required by
antisense approaches. Most importantly,
however, siRNA-mediated RNAi appears to
be more efficient and longer-lasting than
the inhibition achieved by antisense oligo-
nucleotides. In addition, effective concen-
trations of siRNAs are many-fold lower
than those typically used in antisense ex-
periments. The remarkable efficacy of si-
RNAs lies in the fact that small, double-
stranded RNAs seem to be relatively
stable, and target mRNA destruction is
achieved by eliciting a natural cellular pro-
cess. In contrast, cellular processes that
antagonize the effect of antisense oligonu-
cleotides have limited the effectiveness of
antisense technology over the past 15
years.
Some current limitations of RNAi
should be mentioned: in organisms such
as Xenopus laevis (the African clawed toad)
and Danio rerio (the zebrafish), RNAi has
had limited success as a routine method,
and only functions under certain circum-
stances. Here, the standard use of mor-
pholinos (chemical modified antisense oli-
gos) has routinely proven to be the meth-
od of choice, though more studies on
RNAi are necessary to establish a routine
use in cells and embryos of these verte-
brates.
3.2.3
Ribozymes as an Alternative Means
of Post-transcriptional Gene Silencing
Ribozymes are RNA molecules with a cata-
lytic activity, and bind and cleave substrate
RNAs in a sequence-specific manner. This
phenomenon was originally discovered by
Cech (Part I, Chapter 1). Hammerhead
and hairpin ribozymes form base pairs,
with their substrate RNAs using specific
binding arms, which contain conserved
structural and catalytic domains. The ham-
merhead ribozyme is the simplest form,
and was originally found in plant viroids.
It can act intermolecularly (trans), and can
be tailored to recognize different target se-
quences. The minimum structure (ca.
40 nt) that maintains catalytic activity is
composed of a conserved catalytic core and
two flanking regions complementary to
the substrate sequence. Upon dissociation,
the substrate RNA is cleaved at a specific tri-
plet sequence and its expression is
knocked-down. Thus, cleaving renders
the hydrolyzed RNA substrate biologically
inactivate. This targeted gene silencing by
ribozymes spurs the hope of developing
them into new gene therapy-based drugs.
Due to their catalytic action, ribozymes
can produce the desired inhibitory effect at
lower concentrations than are used in anti-
sense oligonucleotide-based approaches. Ri-
bozymes have already shown, both in vitro
and in vivo, to be effective for targeted gene
silencing.
In order to obtain therapeutic agents
based on synthetic ribozymes, it is neces-
sary to modify these structures chemically.
Such modifications as in the traditional
antisense strategy should confer resis-
tance to nucleases, selectivity, and proper
hybridization and uptake characteristics.
In the case of ribozymes, the design of
new modified nucleotides becomes more
complex, since correct folding of the nu-
cleic acid is needed in order to maintain
the catalytic activity. Several studies (e.g.,
X-ray structure elucidation and mapping
with modified nucleotides) have shown
that the presence of the 2'-hydroxyl group
at specific positions in the catalytic core is
essential for hydrolytic activity. Ribozymes
are currently used in larger screens as a
target identification tool, and randomized
ribozyme libraries are utilized to identify
genes involved in various pathways and
different aspects of biology. An example is
the identification and validation of Peter
Pan, a tumor suppressor gene. Knockdown
of Peter Pan by a specific ribozyme leads
to cellular transformation and anchorage-
independent growth [2]. In future, re-
search investigations will doubtless gener-
ate many more specific ribozymes which
can then be used for targeted gene silenc-
ing in target validation.
3.2.4
Aptamers: Screening to Fit Oligonucleotides
for Different Targets
Oligonucleotide aptamers are small nu-
cleotide chains, which have selected bind-
ing capabilities to their targets (Latin: ap-
tus =fit). These aptamers can be generated
by using the SELEX technique (systematic
evolution of ligands by exponential enrich-
ment), which is an oligonucleotide-based
combinatorial library approach that has
been extensively used to isolate high-affini-
ty ligands (called aptamers) for a wide vari-
ety of proteins and small molecules. The
in vitro selection of RNA and DNA ligands
against specific targets obeys the same
rules as natural selection. For this pur-
pose, a partial randomized synthetic DNA
template is constructed containing a ran-
dom inner region that is flanked on both
sides by constant sequences. The random
sequence classically consists of 15 to 75
random positions, where all four bases are
incorporated with equal probabilities. This
pool containing 1012 to 1015 different se-
quences can either be directly used for se-
lection or first transcribed to RNA using
T7 RNA polymerase. In this case, a T7
promoter site needs to be placed on the 5'
site of the DNA template. The random
DNA/RNA pool is exposed to the protein
target for screening. The final pool is
cloned into a bacterial vector and individu-
al colonies are sequenced. The previous
random regions are aligned and searched
for consensus motifs which are often lo-
cated in stem-loop structures and are
thought to mediate binding specificity. Ap-
tamers can be chemically modified for im-
proved efficiency and stability, and for use
in complex biological media. In addition
to their promising application as molecu-
lar sensors, many aptamers are also able
to interfere with the proteins biological
function. Moreover, aptamers selected in
vitro may retain their activity in vivo and
thus offer novel perspectives for gene ther-
apy and the design of new drugs. Recently
developed techniques facilitate the intracel-
lular application of aptamers and their use
as in vivo modulators of cellular physiol-
ogy. Using these approaches, one can
quickly obtain highly specific research re-
agents that act on defined intracellular tar-
gets in the context of the living cell. Thus,
aptamers are suitable for different applica-
tions based on the molecular recognition
of a target molecule for both diagnostic
and therapeutic purposes. Oligo aptamers
can be obtained for almost every target,
whether complex or small. This technology
has been applied to a wide range of tar-
gets, including various enzymes of HIV,
growth factors, and inflammation-inducing
enzymes. The first aptamer to proceed to
Phase I clinical studies is NX-1838, an an-
giogenesis inhibitor. In addition, aptamers
that recognize small molecules are increas-
ingly applied as tools in molecular biology,
from the detection of oxidative damage in
DNA to conditional gene expression and
from their use as modules for the engi-
neering of allosteric ribozymes to biosen-
sors.
3.3
Peptide and Protein-based Approaches
3.3.1
Transporter Peptides
Delivering peptide-based and non-peptide
based drugs into cells is crucial to the suc-
cess of many biopharmaceutical drugs,
which function inside the cell (see Part I,
Chapter 7 and Part VI, Chapters 1, 3, and
6). The drug has not only to enter the tar-
get cells, but also an effective dose must
be incorporated. Before facing the vast
realm of the human body and dealing with
questions such as bioavailability and tar-
geted drug delivery in the entire organism,
it becomes important to validate the target
and to prove the concept of cellular inter-
vention in cell-based disease models, and
ultimately in animal models. For this pur-
pose, it is important that the drugs enter
the cells and, in the best-case scenario, are
delivered specifically to the target cells.
The majority of small-molecule drugs
reach their targets because they are able to
diffuse passively through the cell mem-
brane. As a requisite and limitation, this
requires that the drug is both soluble in
the commonly more polar extracellular
milieu, and also in the non-polar cell
membrane environment. A wide variety of
methods are used to solve the problems of
cellular uptake and drug delivery for drugs
that are not soluble in both the extracellu-
lar milieu and/or the membrane (see Part
VIII, Chapter 4). The main approach is to
transport bioactive substances such as
small molecules, peptides, proteins or nu-
cleic acids inside the cell. Many cells par-
ticularly epithelial cells are designed spe-
cifically to present boundaries for various
substances and are, by nature, pro-
grammed to resist the cellular uptake of
unwanted substances. Here, we provide an
overview of state-of-the-art techniques and
their limitations to overcome this obstacle,
by describing the different methods of de-
livery of biopharmaceuticals into cells a
vital step in the drug discovery process. It
is also important at an early stage to vali-
date a given target in a cell-based assay or
in an animal model. The means to deliver
macromolecules are broad-based, and have
been studied widely over the past few de-
cades. The transduction of nucleic acids
into cells has been established as a stan-
dard laboratory method, though a variety
of efficient methods have been used for
different cell types. Whilst the transduc-
tion of peptides and proteins has pro-
gressed, it is at present a far from stan-
dard technique. Peptide and protein trans-
duction is often attempted with the help
of so-called transporter peptides. These are
derived from various naturally occurring
proteins, including the Tat protein from
HIV, the VP22 protein from herpes sim-
plex virus, and the antennopodia protein
from Drosophila (also known as penetra-
tin). Naturally occurring protein transduc-
tion sequences have common features. For
example, they are cationic, and many of
them are arginine-rich (see Part VI, Chap-
ter 3). Arginine has a guanidinium head
group, and can form two hydrogen bonds;
this is in contrast to lysine, which is also
cationic but can only form one hydrogen
bond. It is of note that not only poly-argi-
nine sequences but also poly-lysine, poly-
ornithine, and even poly-histidine se-
quences are capable of penetrating cell
membranes. Specific short sequences (9
30 amino acids), derived from the above,
account for these transduction abilities
and can be used to deliver a variety of
molecules by covalently linking to them. A
comparison between three different cell
types treated with a Tat transporter peptide
is shown in Fig. 3.3. The figure illustrates
that transduction and cellular localization
of the transporter peptide depends
strongly on the cellular background. Based
on these poly-arginine peptide transport-
ers, molecular transporters have been de-
signed with changes in their backbone and
side chains, and have been shown to be ef-
fective. Although most investigations into
protein transduction are still in their pre-
clinical phases, two good examples for can-
didates in clinical trials have been identi-
fied. The first is a morphine-related drug
that is conjugated to a peptide transporter
(SYN2001 Pep: trans for acute pain by
Synt : em), and this is currently under-
going Phase I clinical trials. The hope is
that the use of the transporter will reduce
morphine-related side effects and thus in-
crease the efficacy. (Synt:em was recently
acquired by Sonus Pharmaceuticals,
Bothell, WA, USA.) The second example is
a topically administered cyclosporine con-
jugate ointment that consists of a poly-ar-
ginine-based transporter coupled to cyclos-
porine. The product, known as PsorBan
(Cellgate, Sunnyvale, CA, USA), showed
positive results in a randomized Phase IIa
clinical trial for psoriasis, and is currently
in early Phase III clinical trials.
The exact mechanism of protein trans-
duction remains unclear. In the case of
the Tat peptide, the sequence can be
scrambled or the D-isomers of the amino
acids used, which argues against a specific
ligand/receptor-mediated mechanism. As
yet, an elusive physical interaction of the
peptide with the cell membrane is thought
to permit cellular entry. However, whether
the process involves endocytosis or not is
also arguable, and in some experiments
the cargo is spread throughout the cyto-
plasm, whereas in others it is concentrated
Fig. 3.3 Different human cell types were treated
with a fluorescein-labeled Tat-Transporter peptide
(FITC-Tat-Peptide). The Tat sequence is derived
from the HIV1 Tat protein, and consists of a 9
amino acid-long poly-arginine-rich sequence
(Tat
4957
RKKRRQRRR). The transported peptide is
an 8-mer peptide with the amino acid sequence
SMTASSVS. Cells were incubated with the FITC-
Tat-peptide (FITC-RKKRRQRRR-SMTASSVS) at a fi-
nal concentration of 10 lM. After incubation for
4 hours, cells were fixed and stained with DAPI to
visualize cell nuclei (shown in blue). The Tat-
Transporter peptide was FITC-labeled and is
shown in green. Both cell lines, the human em-
bryonic kidney cells (HEK 293) and the human
melanoma cells (MeWo), show localization of the
peptide exclusively in the Golgi apparatus. The
primary human foreskin fibroblast cells (HFF) in
addition show cytoplasmic staining. (Photomicro-
graphs taken using confocal microscopy; courtesy
of Jaya Besser.)
in vesicles or the nucleus (see Fig. 3.3).
The molecular transporters also function
in many different cell types, which argues
for a more general mechanism. It appears
that protein transduction is capable of de-
livering all kinds of substances, including
small molecules, oligonucleotides, pep-
tides, proteins, iron beads, and liposomes.
It is important to point out certain lim-
itations of these techniques, however. For
example, peptide delivery does not work in
all cell types with a satisfactory efficiency,
and toxicity has also been observed when
using different transporter peptides. It is
clearly dependent on the individual appli-
cation as to whether molecular transport-
ers can be used, and much optimization is
required for each application. In principle,
when considering proteinprotein interac-
tions, inhibitory peptides coupled to a
transporter and used in a cell-based assay
may provide early validation when target-
ing proteinprotein interactions. Likewise,
agonistic or antagonistic antibodies can be
used, and may be transported inside the
cell with the help of a transporter peptide.
Under certain conditions, transporter pep-
tides have been successfully used in ani-
mal models, an example being the use of
Tat transporter peptides to reduce ischemic
brain damage in rats by perturbing N-
methyl-d-aspartate (NMDA) receptor
PSD95 protein interactions [3].
3.3.2
Antibodies
Humanized antibodies have shown great
promise, and are currently in use as suc-
cessful biopharmaceuticals. Further details
on the development of antibodies and
their present and future success as drugs
are described elsewhere in this book (see
Part I, Chapter 5; Part IV, Chapter 16; Part
V, Chapters 1, 2, and 6).
3.3.3
Peptide Aptamers
Peptide aptamers are molecules that are
selected for their intracellular binding to a
specific target protein. They are useful
tools for target validation, and can block
the intracellular function of a target pro-
tein with high specificity, thereby allowing
the study of distinct physiological and
pathological processes within living cells.
In addition, peptide aptamers provide a ba-
sis for the development of novel diagnostic
and therapeutic strategies, with implica-
tions for a broad variety of different dis-
eases, including cancer, viral infections,
metabolic diseases and neurological disor-
ders.
During the past decade, significant de-
velopments have occurred in the use of
compound libraries for the discovery of
molecules with new binding or catalytic
properties, and in the discovery of protein
aptamers. The basic idea is to shorten the
time and effort associated with the search
of drug leads by developing a huge combi-
nation of molecules, which are further
tested against a certain target, or for a spe-
cific activity. Whilst high-throughput
screening (HTS) is carried out using pep-
tide pools, the main problem with this
technique is the design of a correct strat-
egy to determine the identity of the active
molecules from the original pool. To solve
this problem, either untagged or tagged
methodologies have been developed. An
excellent example is one of the most
powerful tag technologies: the phage dis-
play technology that combines peptide bio-
chemistry with molecular biology (see Part
V, Chapter 2). The principle is that the
peptide library is presented on the surface
of bacteriophages, each of them displaying
a unique peptide that is coded by its ge-
nome. This idea has been further devel-
oped using combinatorial chemical li-
braries based on alternating parallel com-
binatorial synthesis, where the sequence of
a certain peptide is tagged by a natural oli-
gonucleotide [4]. In these examples, the li-
brary diversity is developed by peptide-
based chemistry.
3.4
Protein Kinases as Targets
for Drug Development
Protein kinases play an essential role in
many signaling pathways, and therefore
have the potential to contribute to diseases
ranging from cancer and inflammation to
diabetes and cardiovascular disorders.
They are important for cell growth and
survival and many other biological func-
tions. Recent investigations reveal that
there are over 500 human kinases in the
human genome (approximately 1.7% of
the human genome). Of these 518 kinases
identified, 71 were novel kinases and 56
were extended or corrected sequences of
already known kinases. Almost half of all
kinases (244 of 518) map to known disease
loci or cancer amplicons [5]. This shows
very nicely how important protein kinases
are both for normal biological processes,
as well as for pathological processes. It is
apparent that this has led and will con-
tinue to lead to many drug discovery
projects. The identification of specific ki-
nase inhibitors and functional validation
of kinases as targets has been a challenge,
and classical as well as innovative
approaches are used to address these is-
sues. Methods used to identify novel kin-
ase inhibitors range from HTS to struc-
ture-based design approaches. ATP mi-
metic kinase inhibitors and non-ATP com-
petitive kinase inhibitors are being
developed. Several companies screen in
both directions they screen the kinome
for their specific inhibitors, and they
screen their inhibitome for specific ki-
nases. Many kinase inhibitors have been
(and are currently) in clinical trials in-
deed, one of the most successful novel
cancer drugs, Gleevec (Imatinib mesylate;
Novartis), is a kinase inhibitor which tar-
gets the Abl tyrosine kinase but also inhib-
its other tyrosine kinases such as the
PDGFR or the KIT receptor. Gleevec was
initially approved by the FDA for the treat-
ment of Philadelphia chromosome-positive
chronic myeloid leukemia (CML) in blast
crisis. In CML, a chromosomal transloca-
tion leads to the bcr/abl gene fusion (Phi-
ladelphia chromosome, named after the
city in which it was first discovered). This
translocation leads to a fusion protein
product with a constitutive activity of the
abl kinase. This constant signaling to stem
cells of the myeloid lineage leads to the
overproduction of abnormal white blood
cells in the body. Recently, the FDA has
granted approval for Gleevec for treatment
of KIT (CD117)-positive gastrointestinal
stromal tumors (GIST). Apart from the
chemical approach to target kinases, an-
other approach, mostly used for target vali-
dation, employs dominant negative (DN)
and catalytic inactive (also termed kinase
dead or KD) kinase mutants to link specif-
ic kinases to different diseases. This has
been successfully used to validate the re-
ceptor tyrosine kinase (RTK) Flk-1 (also
known as VEGFR), the receptor for vascu-
lar endothelial growth factor (VEGF). Flk-1
DN mutants have been shown to block
both vascularization and tumor growth in
nude mice [6]. Another approach is to use
siRNAs for gene silencing of specific kin-
ase genes, and many functional kinase-di-
rected siRNAs have been identified.
Furthermore, antibodies can be used to
block kinase activation; the anti-cancer
drug herceptin (Trastuzumab; Genentech/
Roche) is an example of the success of this
approach (see Part I, Chapter 5). Herceptin
is a monoclonal antibody which targets the
human epidermal growth factor receptor
family member HER2 (also known as erB-
b2 or neu). It is an FDA-approved cancer
drug for the treatment of HER2-positive
breast cancers (about one in four breast
cancers). HER2 was originally identified in
1984 by Robert Weinbergs group, it being
found later (by Slamon and Ullrich) that
the HER2/neu gene was amplified in
breast cancers. Slamon and Genentech
pursued the successful approach by target-
ing HER2 with an antibody. Herceptin is
often used in combination with other
types of chemotherapy.
Another example of an EGFR (HER1 re-
ceptor) inhibitor is gefitinib (Iressa
; As-
traZeneca Pharmaceuticals LP, Wilming-
ton, DE, USA). In May 2003, gefitinib
(ZD1839) received accelerated approval by
the US Food and Drug Administration as
monotherapy for patients with locally ad-
vanced or metastatic non-small cell lung
cancer (NSCLC), after failure of both plati-
num-based and docetaxel chemotherapies.
3.5
Cell-based Assays for In vitro Target
Validation in the Drug Discovery Process
Tissue culture cells are being used for target
identification and validation, for primary
and secondary screening, and also to pro-
duce biopharmaceutical proteins (see Part
IV, Chapter 1 and 3). To handle all currently
used cellular assays is beyond the scope of
this chapter, and we will focus here on
cell-based assays that are used as cancer
models, or which are relevant for cellular
transformation or cell migration and inva-
siveness. We will also discuss their suitabil-
ity for target validation. Cancer is a multi-
step process, which involves a facet of cellu-
lar de-regulations. Alterations of regulatory
pathways involved in proliferation and
homeostasis will lead to the transformation
of normal human cells into malignant can-
cers. In a recent publication, six essential al-
terations in cell physiology that collectively
dictate malignant growth were summarized
[7], and include: 1) self-sufficient growth
signals; 2) unresponsiveness to growth-inhi-
bitory signals (anti-growth signals); 3) es-
cape from the apoptosis program; 4) unlim-
ited replicative potential; 5) sustained angio-
genesis; and 6) tissue invasion and metasta-
sis. It was suggested that most cancers have
acquired the same set of these functional
capabilities during their development, albeit
through different mechanisms. For a long
time, cancer cell biology has taken these
hallmarks of cancerous cells into account,
and many cell biology assays have been de-
veloped reflecting aspects of these capa-
bilities.
3.5.1
Cell Growth and Viability Assays
There are several cellular assays measur-
ing cell growth and viability under differ-
ent conditions. Colorimetric assays can be
used to quantify cell survival and prolifera-
tion. These are cell-based assays, which
measure cell viability as well as an in-
crease in cell number. The original assay
uses MTT [3-(4,5-cimethylthiazol-2-yl)-2,5-
diphenyl tetrazolium bromide], which is a
pale yellow substrate that is cleaved by liv-
ing cells to yield a dark blue formazin
product. Similar assays (e.g., the MTS as-
say) have the advantage of better solubility.
Recently, more sensitive viability assays
using fluorescent read-outs are available.
However, the easiest viability assay to per-
form remains the cell count assay. These
cell growth and viability assays can be uti-
lized to measure initial transformation by
two assays:
Growth in low serum: Transformed cells
are able to grow under low serum condi-
tions (e.g., 0.5%) in contrast to untrans-
formed cells.
Saturation density assay: Untransformed
cells, which are still contact inhibited,
will grow in a monolayer, whereas trans-
formed cells will grow on top of each
other and thus can be grown to a higher
density which can be measured by the
different viability assays or by a higher
cell count.
Another possibility to quantitate prolif-
eration is by measuring DNA synthesis
and DNA content. Cell proliferation corre-
lates with a higher DNA synthesis rate.
The rate of DNA synthesis can be mea-
sured by [
3
H]thymidine incorporation into
the DNA of cells, or in a non-radioactive
manner by bromodeoxyuridine (BrdU) in-
corporation using either a BrdU antibody
in a colorimetric ELISA setting or by using
FITC-conjugated BrdU and flow cytometry.
Alternatively, or in parallel, propidium
iodide staining of cells and flow cytometry
detection measures the DNA content of
the cells and distinguishes between cells
in different cell cycle phases (e.g., S-phase
and G
1
phase). It also measures cell death,
and to some extent the appearance of ab-
normal cells can be detected.
3.5.2
Classical Transformation Assays
Soft agar assay: The soft agar assay mea-
sures the anchorage-independent growth
of cells on a soft agar layer, and visual-
izes the colonies formed.
Focus formation assay: The focus forma-
tion assay measures the loss of contact
inhibition. Normal cells grow to a mono-
layer and become contact-inhibited by
inhibitory signals from their neighbor-
ing cells. Transformed cells will grow
out of the monolayer and form colonies
on top of the monolayer. These colonies
(or foci) can be stained above back-
ground, and when counted foci are a
classic parameter for cellular transfor-
mation. These in vitro transformation as-
says have first been established in ro-
dent fibroblast cell lines such as
NIH3T3 and Rat1 fibroblasts. There is a
good (but not complete) correlation be-
tween the formation of foci and tumori-
genicity in nude mice. The same is true
for the growth in soft agar.
The disadvantages of these assays are
their relative long time frame (1021 days)
and, in case of the soft agar assay, poor
solubility in the agar of many compounds.
3.5.3
Genetic Instability and Loss
of Genomic Integrity
It has long been considered that genetic
instability is an integral component of hu-
man neoplasia. In a small fraction of tu-
mors, mismatch repair deficiency leads to
a microsatellite instability at the nucleotide
sequence level. In other tumors, an abnor-
mal chromosome number (aneuploidy)
has suggested an instability.
The importance of maintaining genomic
stability is evidenced by the fact that trans-
formed cells often contain a variety of
chromosomal abnormalities such as eu-
ploidy, translocations, and inversions.
Gene amplification is a well-characterized
hallmark of genomic instability, and is
thought to result from recombination
events following the formation of double-
strand, chromosomal breaks. Therefore,
3.5 Cell-based Assays for In vitro Target Validation in the Drug Discovery Process 641
gene amplification frequency can serve as
an indicator of genomic stability. The
PALA assay is designed to directly mea-
sure the frequency with which a specific
gene, CAD, is amplified within a cells ge-
nome.
Another manifestation of genetic insta-
bility is an abnormal number of centro-
somes. Centrosomes can be stained and
quantified in an assay using anti-centrin
antibodies. Both, genetic instability and
loss of genomic integrity, are often post-
transformation events. However, under
certain circumstances they can contribute
to cellular transformation directly. An ex-
ample is the transformation by the human
papillomavirus (HPV) E7 protein. The
HPV E7 and E6 proteins are the two HPV
oncogenes found in 99.3% of all cervical
cancers.
3.5.4
Morphological Changes as Indicators
for Cellular Transformation
For many cell types, morphological
changes are indicators associated with
transformation into cancerous cells. The
classical example are rodent fibroblasts
(e.g., Rat1, NIH3T3 or 3Y1 cells), which
dramatically change their appearance (due
to transformation) from a more roundish,
droplet-like phenotype to a more
stretched bean-like phenotype. Morpho-
logical transformation is observed in NBT-
II cells, rat bladder cells and is associated
with a more invasive phenotype. In 3Y1
rat fibroblasts, morphological transforma-
tion can be observed by swirl formation of
confluent cultures. It is important to note
that these assays are not classical transfor-
mation assays and, in their own right, are
not sufficient as all cells mentioned here
are established cell lines and already im-
mortalized. In certain cells another mor-
phological change the loss of stress fi-
bers is also associated with cellular
transformation.
3.5.5
Invasiveness and Cell Migration
Cell migration is a fundamental function
of normal cellular processes, including
embryonic development, angiogenesis,
wound healing, immune responses, and
inflammation. It is also linked to invasive-
ness and metastasis in cancer. Many differ-
ent cell migration assays have been estab-
lished. The simplest and easiest to per-
form is the in vitro-wound healing assay,
which is also called the scratch assay.
Cells are grown to a high density and the
cell monolayer is scratched (usually with a
yellow pipette tip). Migration is quanti-
tated by measuring closure of the
scratch, usually within the first 20 hours
after incision. Preferentially used cells are
NBT-II cells, rat bladder cancer cells,
which migrate depending on growth fac-
tors such as epidermal growth factor
(EGF). Chinese hamster ovary cells are
also used in this type of assay. NBT-II
cells, in addition to enhanced migration,
also show morphological changes. In the
so-called scatter assay, NBT-II cells are
treated with growth factors such as EGF,
and the morphology and dispersion of
small colonies is measured microscopically
and documented photographically. Quanti-
fication of migration is performed by
counting single cells with fibroblastoid
migration morphology compared with cells
in groups with epithelial morphology.
NBT-II cells change their morphology
from a roundish cell shape to a more
stretched and sometimes star-shaped mi-
gratory phenotype.
3.5.5.1 The Boyden Chamber Assay for
Chemotaxis and Cellular Migration
The Boyden chamber is a simple appara-
tus used to test for chemotaxis, especially
of leukocytes. It can also be used to assess
tumor cell transmigration across an endo-
thelial monolayer in vitro. It consists of
two compartments separated by a Milli-
pore filter (38 lm pore size). A chemotac-
tic factor is placed in one compartment,
and a gradient develops across the thick-
ness of the filter (ca. 150 lm). Cell move-
ment into the filter is measured after an
incubation period less than the time taken
for the gradient to decay. Cell motility can
be measured in Boyden chambers contain-
ing filters precoated with different materi-
als, for example fibronectin or fibronectin
fragments. The method, when applied to
malignant and non-malignant cell lines,
shows that the variable invasive potentials
of these cells correlate with their ability to
disrupt the endothelial cell monolayer.
3.5.5.2 Three-dimensional Collagen Matrix
Assay
In addition to the Boyden chamber, inva-
siveness can be measured in several in vi-
tro assays in which cells are grown in var-
ious three-dimensional matrices. One ex-
ample is the collagen gel assay, where the
invasive activity of cells can be monitored
in reconstituted collagen gels. Carcinoma
cell lines, for example, are added to col-
lagen type 1 gels, which have been mixed
with various fibroblasts. The prepared col-
lagen/fibroblast gels are overlaid with the
tumor cell suspension and initially cul-
tured for 1 day. The gels are then lifted
onto nylon membranes (100 lm pores),
placed on stainless steel grids, and the me-
dium is added. After 14 days culture, the
organotypic cultures are processed for his-
tology and the depth of invasion of tumor
cells assessed. These experiments can be
carried out in the absence or presence of
various inhibitors. The assay allows for the
use of transporter peptides or also neutra-
lizing antibodies.
3.5.5.3 Acinar Formation Assay
(Three-dimensional Basement
Membrane Culture Model)
In vitro three-dimensional basement mem-
brane models allow evaluation of the biolog-
ical activities of growth factors and other
genes associated with breast cancer in
events related to the initiation and progres-
sion of breast tumors. This in vitro 3D mod-
el involves the use of the immortalized hu-
man mammary epithelial cell line, MCF10A
cells. These cells undergo a program of
morphogenetic events in Matrigel basement
membrane cultures, leading to the develop-
ment of growth-arrested acini-like spheroid
structures that are composed of a single
layer of epithelial cells surrounding a hol-
low lumen. Such cultures allow the exami-
nation of the ability of growth factors and
other breast cancer-associated genes to al-
low cells to escape proliferative suppression,
to survive in the lumen, to disrupt apical po-
larity, and to break down/invade the base-
ment membrane. Infiltrating the hollow
middle of the acinar ring shows the first in-
dication of invasiveness, and may represent
the initial step of penetrating the basement
membrane and leaving the carcinoma in
situ state.
3.5.6
Escape from Differentiation
and Differentiation Signals
Differentiation of murine keratinocytes
and of normal human epidermal keratino-
cytes (NHEK) can be triggered by a variety
of stimuli. Changes in extracellular Ca
2+
3.5 Cell-based Assays for In vitro Target Validation in the Drug Discovery Process 643
or stimulation with tumor necrosis factor
will induce terminal differentiation in
these cells, which can be measured by
monitoring specific differentiation mark-
ers. Keratinocytes will also spontaneously
differentiate at high cell densities in vitro.
Already transformed keratinocytes will not
follow these signals and will not differenti-
ate. The keratinocyte differentiation assay
can be used as initial evidence for transfor-
mation in this particular cellular back-
ground. Correlation with terminal differen-
tiation indicates cell-cycle arrest, and the
amount of cell cycle-arrested cells can be
measured by flow cytometry using propi-
dium iodide, as detailed before.
3.5.7
Escape from Apoptosis
In a healthy organism, a balance between
the destructive and proliferative processes
of cells defines cellular homeostasis. Apop-
tosis is a lifelong programmed, energeti-
cally dependent and genetically regulated
cell destruction process. This cellular de-
struction process functions through a spe-
cial signaling mechanism, and removes
weak, unnecessary and damaged cells
from an organism. Each day, approxi-
mately 5% of all cells of an organism un-
dergo apoptosis. In cancer cells however,
regulation of the apoptosis program can
be disturbed and cells destined to die will
survive. A very important player herein is
the tumor suppressor p53 which, under
normal conditions, will lead to apoptosis
of many transformed cells. In cancer cells,
p53 often is mutated and will no longer
function as a tumor suppressor and no
longer pave the road to death. Escape from
apoptosis is one of the six hallmarks of
transformation into a cancerous cell. Apop-
tosis is also important when considering
cancer therapy, because many tumor cells
will override apoptosis induced by che-
motherapeutics (e.g., cisplatin) and be-
come chemoresistant. Thus, it becomes
very important to determine which cellular
pathways and target molecules are in-
volved in chemoresistance to anti-cancer
drugs. The goal is to revert these cancer-
ous cells to a responsive phenotype and to
use a combination therapy to eradicate the
formerly chemoresistant tumor cells by
blocking these survival routes. Apoptosis
can be measured in all cell types and
through several different assays. Com-
monly used apoptosis assays are:
1. DNA fragmentation assay (DNA ladder-
ing): Apoptotic DNA laddering detects
the level of internucleosomal DNA frag-
mentation that occurs during apoptosis.
Often, Southern blots are used to sepa-
rate DNA fragments. Quantification can
be carried out using ethidium bromide,
colorimetry, chemiluminescence, or
radioactive isotopes.
2. Tunel (Terminal deoxynucleotide trans-
ferase dUTP nick end labeling) assay: In
a modified version of the Tunel assay,
apoptotic cells can be measured in a
two-color system using flow cytometry
and microscopy. In this assay, DNA
breaks are labeled by deoxynucleotidyl
transferase using bromodeoxyuridine tri-
phosphate (BrdUTP) and visualized by
anti-BrdU antibody. Propidium iodide is
used to counterstain the total cellular
DNA.
3. Annexin 5 detection assay: Annexin 5
staining allows rapid, specific, and quan-
titative identification of apoptosis in in-
dividual cells. Annexin 5 is a calcium-de-
pendent phospholipid-binding protein.
Early in the apoptotic process, cell sur-
face phospholipid asymmetry is dis-
rupted, and this leads to the exposure of
phosphatidylserine (PS) on the outer
leaflet of the cytoplasmic membrane.
Annexin V preferentially binds to PS,
and can be used as an early indicator of
apoptosis using flow cytometry or in situ
detection. Annexin V conjugated to
either FITC or to biotin can be used for
the detection of cell surface changes dur-
ing apoptosis. In addition, propidium io-
dide can be used on unfixed samples to
determine the population of cells that
have lost membrane integrity, an indica-
tion of late apoptosis or necrosis.
4. Caspase activation assays and other mo-
lecular read-outs for apoptosis: Commer-
cially available caspase antibodies, as
well as caspase activity assays, allow the
rapid measurement of caspase activity in
cells. Caspase activity correlates well
with the apoptotic program. Examples of
other molecular markers used for the in-
dication of apoptosis are Bax, Bcl-2,
BCL-XL and PARP (poly (ADP-ribose)
polymerase).
3.6
Animal Models as the Ultimate Target
Validation
Certainly the most significant method for
target validation is the whole-animal
approach. Several vertebrates have been
used for target validation, and various dis-
ease models have been and will continue
to be developed. Several vertebrate ge-
nomes have been fully or almost complete-
ly characterized which allows the combina-
tion of genomics, proteomics, and systems
biology in these organisms. Of particular
note is the zebrafish (Danio rerio), which
has been mainly used for developmental
studies, though several disease models
have also been established in this verte-
brate. Working with zebrafish has the ad-
vantage of cost efficiency, and larger genet-
ic screens are easy to perform due to the
short reproductive cycle. Currently, the
zebrafish genome is almost entirely se-
quenced, and will most likely be com-
pleted before the predicted deadline in
2005. Its obvious disadvantage is that it is
a fish rather than a mammal, whereas the
mouse, rat and other mammals are cer-
tainly closer to humans. For more in-depth
information on mouse models and knock-
outs, the reader is referred to Part III,
Chapter 4)
3.7
In this chapter we have provided an over-
view of the state-of-the-art techniques used
at early stages of the drug discovery pro-
cess. Our goal was not to supply a compre-
hensive list of all methods used, but more
to focus on novel exciting developments in
the field.
RNA interference (RNAi) has been
termed one of the most exciting discov-
eries in biology in the past decade, and
since its first recognition in 1998 it has
quickly become one of the most powerful
and indispensable tools in molecular biol-
ogy. Using short dsRNA molecules, RNAi
can selectively silence essentially any gene
in the genome. In the laboratory, RNAi is
used routinely to reveal the genetic secrets
of development, intracellular signaling,
cancer, infection, and a full range of other
phenomena. However, can the phenome-
non hailed by the journal Science as the
Breakthrough of the Year in 2002 break
out of the laboratory and lead to novel
therapies as well? Pharmaceutical giants
are hoping so, and several biotech compa-
nies have bet their futures on it, though
not everyone is so optimistic about the fu-
ture of RNAi therapy. At the heart of its
promise as a powerful therapeutic drug
lies the exquisite selectivity of RNAi: like
the fabled magic bullet, an RNAi se-
quence seeks out and destroys its target
without affecting other genes (as was
hoped for antibodies; see also Part V,
Chapter 1). The clinical applications ap-
pear endless: any gene whose expression
contributes to disease is a potential target,
from viral genes to oncogenes, to genes re-
sponsible for heart disease, Alzheimers
disease, diabetes, and many more. How-
ever, for all its promise RNAi therapy is a
long way from entering the clinic. While it
is a proven Wunderkind in the lab, to
date no tests have been performed in hu-
mans, and only modest and circumscribed
successes have been demonstrated in ani-
mals.
To be a successful drug, a molecule
must overcome a long set of hurdles. First,
it should be capable of manufacture at rea-
sonable cost, and administered safely and
conveniently. Then and even more im-
portantly it must be stable enough to
reach its target cells before being degraded
or excreted; it must enter those cells, link
up with its intracellular target, and exert
its effect; and it must exert enough of an
effect to improve the health of the person
taking it. And, finally, it must do all this
without causing significant toxic effects in
either target or non-target tissues. No mat-
ter how good a compound appears in the
laboratory, if it fails to clear any one of
these hurdles it is useless as a biopharma-
ceutical. Stability and delivery are most
likely the major obstacles to successful
RNAi therapy obstacles that are intrinsic
to the biochemical nature of RNA itself, as
well as the bodys defenses against infec-
tion with foreign nucleotides. Delivery will
not be easy for two major reasons. First, a
charged oligonucleotide will not easily
pass through a lipid layer which it must
do in order to enter a cell. The cell has no
desire to take up the RNA, which makes
evolutionary sense, since extracellular
RNA usually signifies a viral infection.
Second, when injected into the blood-
stream, unmodified RNA is rapidly ex-
creted by the kidneys or degraded by en-
zymes. To solve the problem of cell pene-
tration, most researchers have either com-
plexed the RNA with a lipid or modified
the RNAs phosphate backbone to mini-
mize its charge. However, what has been
learned from the antisense field is that
even without other delivery strategies,
when RNA is administered at sufficient
doses, and if it is stable, it will be taken
up by cells. To date, only one antisense
drug has received FDA approval Vitra-
vene, which is used to treat CMV infec-
tions in the eyes of patients with HIV. Vi-
travene is actually a DNA antisense drug,
which binds to viral DNA, and it is un-
clear whether it actually functions by an
antisense mechanism.
The delivery system and cellular uptake
are not only major issues for the develop-
ment of nucleotide-based drugs they ap-
pear even more of a problem for larger
peptide- and protein-based drugs. The use
of transporter peptides has great potential
but, similar to RNAi, this approach has
still to make its way from the bench to the
bedside. However, a successful cellular
transporter system would represent a
breakthrough for the development of novel
biopharmaceuticals, and could even lead to
a renaissance of older and also failed drug
leads.
The examples given here for successful,
approved drugs and for drugs in clinical
trials are not meant to be comprehensive;
rather, they should be seen as examples of
the different types of biopharmaceutical
drugs developed by these new approaches.
One of the most encouraging examples of
a successful novel drug was the tyrosine kin-
ase inhibitor, Gleevec (Imatinib; Novartis).
The success of Gleevec in treating CML as
well as other selected cancers has greatly in-
creased optimism for the broader applica-
tion of kinase inhibitor therapy in cancer,
although to date it remains the only specta-
cularly successful example. Is it simply a
matter of time before kinase inhibitors be-
come more broadly useful, or is CML a
unique disease that does not reflect the true
genetic complexity of other cancers?
Although not comparable to the effective-
ness of Gleevec for CML, inhibitors of other
tyrosine kinases namely Her1 and Her2 of
the EGFR family have led to the develop-
ment of anti-cancer drugs. As described else-
where in this book, the monoclonal antibody
used to block Her2 (herceptin; Genentech/
Roche) has dramatically proven the practi-
cality of the antibody-based drug approach.
These examples show that novel
approaches can lead to a new generation
of effective drugs. For complex diseases
such as cancer, multifaceted treatment and
combination therapy (as has long been
used already) might lead to further future
breakthroughs. Ultimately, the number of
different approaches and their combina-
tion might make the difference. It seems
logical that, the more complex the disease,
the more complex its treatment and cure.
The same logic applies to the drug discov-
ery process but, luckily, not to the biophar-
maceutical itself.
References
1 J. W. Myers, J. T. Jones, T. Meyer, J. E. Ferrel Jr.
Recombinant Dicer efficiently converts large
dsRNAs into siRNAs suitable for gene silenc-
ing. Nat Biotechnology 2003; 21(3), 324328.
2 P. J. Welch, E. G. Marcusson, Q. X. Li, C. Be-
ger, M. Kruger, C. Zhou, M. Leavitt, F. Wong-
Staal, J. R. Barber. Identification and validation
of a gene involved in anchorage-independent
cell growth control using a library of random-
ized hairpin ribozymes. Genomics 2000;
28(13), 26052612.
3 M. Aarts, L. Liu, S. Besshoh, M. Arundine,
J. W. Gurd, Y. T. Wang, M. W. Salter, M. Tymi-
anski. Treatment of ischemic brain damage by
perturbing NMDA receptor-PSD-95 protein in-
teractions. Science 2002; 298(5594), 846850.
4 M. Brenner, R. A. Lerner. Encoded combinator-
ial chemistry. Proc Natl Acad Sci USA 1992;
89(12), 53815383.
5 G. Manning, D. B. Whyte, R. Martinez, T.
Hunter, S. Sudarsanam. The protein kinase
complement of the human genome. Science
2002; 298(5600), 19121934.
6 B. Millauer, L. K. Shawver, K. H. Plate, W. Ri-
sau, A. Ullrich. Glioblastoma growth inhibited
in vivo by a dominant-negative Flk-1 mutant.
Nature 1994; 367(6463), 576579.
7 D. Hanahan, R. A. Weinberg. The hallmarks
of cancer. Cell 2000; 100, 5770.
References 647
Abstract
Genetically altered mice are important
models to study the mechanism of inher-
ited and acquired diseases. Furthermore,
these mice are extremely useful to test the
efficacy of new therapies and to validate
potential drug targets. Although the gen-
eration of mice carrying point mutations
or deletions of specific genes is quite well
established, it is still a relatively long pro-
cedure. Careful planning of the gene tar-
geting is therefore essential. In this chap-
ter we describe the general procedure of
generating genetically altered mice and
discuss potential problems that could be
encountered and their solution.
Abbreviations
BAC bacterial artificial chromosomes
cDNA DNA derived from mRNA by re-
verse transcription
DMEM Dulbeccos modified Eagles medi-
um
DNA deoxyribonucleic acid
ES cell embryonic stem cell
EST expressed sequence tag
(mRNA sequences)
FCS fetal calf serum
kb kilo bases
mRNA messenger RNA
neo neomycin
PAC phage artificial chromosomes
tk thymidine kinase
4.1
Disease-oriented Research
in Genetically Modified Mice
Many diseases are caused or facilitated by
genomic alterations. Studies with human
patients, however, are often difficult to per-
form and to analyze the results due to the
different genetic background, age, disease
history and living conditions, the limited
amount of material for histological and
biochemical analysis, and often also low
numbers of patients. Mouse models for
human diseases have the advantage that
large numbers of genetically identical ani-
mals of the same age and gender can be
handled, that the biology of mice is rela-
tively close to humans in comparison to
fishes, flies or worms, and that they can
be genetically modified. Genetically modi-
fied mice allow the modeling of disease-re-
lated genetic alterations in mice, the study
of exact mechanistic consequences of the
649
4
Genetically Modified Mice in Medical and Pharmaceutical Research
Cord Brakebusch
ISBN: 3-527-31184-X
mutations in vivo, and the design and test-
ing of new therapeutic strategies to fight
these diseases.
First, genetically modified mice can be
used as a disease model. If the genetic al-
teration underlying a disease is known,
one can generate mice carrying exactly the
same mutation. This could be a total loss
of protein, but also a deletion or a point
mutation. If a genetic alteration is sus-
pected to cause or facilitate a disease, this
hypothesis can be tested in mice carrying
this mutation. If no specific gene is sus-
pected of either causing or modifying a
disease, a random mutagenesis screen can
be carried out [13]. Mice with a pheno-
type similar to the human disease will be
analyzed for the mutation they carry. It
will then be tested in human patients,
whether they have a similar genetic altera-
tion. Finally, a genetically modified mouse
might unexpectedly have a phenotype sim-
ilar to a human disease, thereby revealing
the molecular cause of a human illness. In
case the gene mutation is directly causing
the disease, the illness will develop sponta-
neously in mice. In case the mutation is
only modulating disease development, an
increased frequency or severity of disease
development will occur in specific disease
models such as wound healing or tumor
formation. By using these mice, new dis-
ease therapies can be tested and evaluated
in detail, whilst primary cells obtained
from the animals might be used for the
high-throughput screening of new drugs.
Second, mice can be used as models for
therapy. In that case, a potential drug tar-
get is altered in its activity by mutation in
a similar way as the potential drug would
do. Gene targeting allows the generation
of mice that either lack this molecule (cor-
responding to 100% inhibition) or express
instead a constitutively active form of it
(corresponding to 100% activation). Since
inhibition and activation are induced by
genetic alteration, there is no partial effect
and, furthermore, no side effect due to in-
hibition or alteration of other molecules.
These genetically modified mice can be
used in disease models in order to evalu-
ate the in vivo potential of a certain thera-
py, before much time and money is in-
vested in the development of small molec-
ular weight inhibitors or activators.
Although most genes are highly con-
served between man and mouse as has
been revealed by the Genome Project
there are still many differences between
the biology of man and mouse [4]. These
differences can result in a different sus-
ceptibility to certain diseases; for example,
whilst in human skin about five mutations
are required for transformation into tumor
cells, only two to three are necessary in
mice [5,6]. Thus, it should be carefully
tested in each model whether the situation
in the mouse mimics sufficiently that in
humans, and to what degree it might be
necessary to humanize the mouse by ex-
changing specific mouse genes against hu-
mans ones.
Gene targeting allows modification of
the genome in a restricted and defined
way. In contrast to transgenic mice, where
an expression cassette is integrated some-
where in the genome perhaps at multi-
ple locations gene targeting allows the
genome to be changed at a specific site
and in an exactly defined manner. In
knock-out mice, a gene is inactivated so
that no functional protein can be pro-
duced, whereas in knock-in mice a gene
is modified so that either a mutated form
of the protein or an alternative protein is
produced instead of the endogenous mate-
rial. In conditional gene targeting a mo-
lecular switch is introduced into the ge-
nome that allows the restriction of knock-
out or knock-in to specific cells, or for it to
be induced by the exogenous administra-
tion of specific agents. This technique per-
mits the study of mutations for which con-
stitutive mutations would lead to early em-
bryonic lethality or a complex phenotype
due to multiple secondary effects. Further-
more, it reduces the chance that a pheno-
type is partially compensated during devel-
opment. With respect to therapy models, it
allows a genetic alteration to be induced
after disease development, for example to
test whether the inactivation of a certain
molecule might result in tumor remission.
4.2
Generation of Genetically Modified Mice
by Gene Targeting
Directed and specific gene targeting in
mice became possible after two major
breakthroughs. First, the establishment of
embryonic stem (ES) cell lines from early
mouse embryos allowed the cultivation
and manipulation of totipotent murine
cells in vitro [7, 8]. In principle, a complete
mouse can be generated from a single ES
cell. Second, using homologous recombi-
nation it was possible to change the ge-
nome of mammalian cells at specific
places, since long stretches of DNA ho-
mologous to the targeted gene are direct-
ing the targeting construct to the gene of
interest [9]. The combination of these tech-
niques allowed the generation of mice
with restricted, defined mutations in spe-
cific genes [1012].
Today, the generation of gene-targeted
mice has become a standard technique in
many institutions, and detailed manuals
describing the methods are available (Table
4.1) [13, 14]. The procedure remains
highly complex however, and any errors in
the creation of such mice may lead to the
failure of an entire project. Thus, careful
planning and standardized procedures are
essential for the successful creation of
these animals.
Although today the sequence of the
complete human and mouse genome is
largely known, its functional role is far
from understood. The existence of exons,
intron-exon borders and certain promoter
elements is well known, but neither the
functional role nor the importance of
many sequences in introns and outside
genes remains unknown. Thus, the gener-
al rule for all genetic alterations must be
to keep the number of changes as low as
possible in order not to damage any un-
known functional elements. In this way it
is possible to create an artificial phenotype
that is either partially or completely depen-
dent upon these involuntary changes, and
not on alterations introduced into the gene
of interest. For example, different pheno-
types observed in various mouse lines with
knock-outs of the myogenic factor Myf-5
[15] or of the prion protein PrP [16] are
quite likely due to unwanted effects on
other genes in some of the mouse strains.
4.2.1
Analysis of the Gene Structure
Before starting to plan the targeting con-
struct, the gene must first be carefully ana-
lyzed. By comparison of the genomic
mouse sequence (www.ensembl.org/
Mus_musculus/) with cDNA and EST se-
quences (www.ensembl.org/Mus_muscu-
lus/) of the gene of interest, one must de-
termine the exon-intron structure, alterna-
tive promoters, alternative transcription
start sites, alternative splicing, potential al-
ternative translation start sites and codon
phasing, which describes at which position
of the codon triplet an intron is inserted.
Furthermore, it must be checked whether
other genes or promoter, enhancer or si-
lencer elements of neighboring genes are
contained in the gene of interest. A com-
parison to the gene structure of the hu-
man homologue is often helpful.
In order preferentially to obtain a full
knock-out, the exon containing the transla-
tion start site is deleted. If the preceding
exon can splice into a downstream exon
with a potential in-frame translation start
site, it is possible either to increase the se-
quence to be deleted or to delete instead
another exon. If alternative mRNAs are
possible, the exon or exons that are essen-
tial for the gene and as close as possible to
the translation start site must be deleted.
The deletion of exons downstream of
the translation start results in the forma-
tion of a truncated protein. In that case, it
must be checked whether it is likely that
the truncated protein is folded correctly,
and whether it could have a functional
role. If it contains only half of an indepen-
dent folding domain, the truncated protein
will most likely be folded incorrectly, and
Table 4.1 Basic stages in the generation of gene-targeted mice.
The average time for each stage is indicated in parentheses
Analysis of gene structure (few days)
exon-intron structure
alternative promoters
alternative transcription start sites
alternative splicing
alternative translation start sites
codon phasing
neighboring genes
potential function of truncated proteins
Planning of targeting construct (few days)
homologous side arms (total 1012 kb)
identification of homologous recombinants (Southern blot, genomic PCR)
identification of additional heterologous recombinations
Cloning of targeting construct (23 months)
isogenic genomic DNA (e.g., of PAC library)
subcloning of homologous regions
introduction of selection cassette, restriction sites, mutations, loxP sites
reporter genes, etc.
ES cell culture (34 weeks)
electroporation
selection
picking of stable transfectants
freezing
preparation of genomic DNA
identification of homologous recombinants
Generation of homozygous mutant mice (68 months)
blastocyst injection of homologously recombined ES cells
crossing of male chimera for germ line transmission
crossing of heterozygous mice to obtain homozygous mutants
will be degraded quickly. If it contains at
least one full folding unit it might fold
partially, and this might result in a domi-
nant negative mutant form of the protein.
A comparison with known functional do-
mains of the protein helps to evaluate
further any possible functions of the trun-
cated molecule. If antibodies against the
N-terminal region are available, the expres-
sion of truncated proteins could be tested
on protein levels in the mutant mouse.
4.2.2
Planning of the Targeting Construct
A targeting construct is made by two arms
of homologous DNA flanking a selectable
antibiotic resistance marker (Fig. 4.1). In
most cases, neomycin resistance marker is
used as a selectable marker, although
other markers (e.g., hygromycin resis-
tance) may also be utilized. In general, the
longer the arms, the higher the frequency
of homologous recombination [17]. This
relationship is not linear however, so that
above 10 kb the combined size of the ho-
mologous arms becomes less important.
In our laboratory, our aim is to have each
arm at least 2 kb long and for the com-
bined length to be *10 kb, but at least
6 kb. The size of the deleted DNA is less
important and can be rather long, without
affecting the efficiency of recombination.
If one of the arms is short, it is possible to
increase the frequency of homologous re-
combinants among the isolated clones by
adding a thymidine kinase (tk) expression
cassette to the short arm of the construct.
This cassette should be lost during homol-
ogous recombination, whereas it is often
not lost in heterologous recombinants.
With ganciclovir it is possible to select
negatively for those clones that have lost
the tk gene. In order to obtain highest effi-
ciency of homologous recombination, the
homologous DNA in the targeting con-
struct must be isogenous with that of the
ES cells, as even small differences can sig-
nificantly reduce the targeting efficiency.
The second important parameter for the
frequency of homologous recombination,
in addition to the length of the homolo-
gous regions, is accessibility of the tar-
geted locus for the targeting construct and
recombination enzymes. At present, the
factors which influence this are only in-
completely known, and it is not possible to
judge from the DNA sequence about the
accessibility and therefore the probability
of homologous recombination.
When the targeting construct is
planned, consideration should also be
made as to how the homologous recombi-
nants will be identified. As the most sol-
id method to identify homologous recom-
binants, we recommend Southern blot
Fig. 4.1 Gene disruption by targeted mu-
tation. To delete exon 2 (E2), a targeting
construct was prepared with side arms
containing homologous DNA upstream
and downstream of exon 2, respectively.
By recombination events in the upstream
and downstream arms, exon 2 and part
of the surrounding intronic region will be
replaced by a neomycin selection cassette
(neo), resulting in the inactivation of the
gene.
analysis using an external probe outside
the targeting construct (Fig. 4.2). Genomic
DNA will be digested with an enzyme that
cuts three times, downstream of the target-
ing construct, upstream of the mutation,
and inside of the selection cassette. In the
case of heterologous recombination, no se-
lection cassette was introduced to the gene
of interest and therefore also no additional
restriction site. The restriction fragment
detected by the external probe is therefore
the same as for wild-type cells. In the case
of homologous recombination, a new re-
striction site was introduced together with
the selection cassette, and this resulted in
a band of specific length which was recog-
nized by the external probe. The size of
the bands specific for the wild-type and
homologously recombined gene should be
sufficiently different to allow identification
of both. Such a Southern blot analysis
could be performed for both arms of the
targeting construct. The restriction en-
zymes of choice are those that are low-cost
and which cut well under medium- and
high-salt conditions (e.g., EcoRI, EcoRV,
BglII, HindIII, XbaI, BamHI). Since the
wild-type and the homologously recom-
bined fragment are recognized in the
Southern blot by the same probe, the sig-
nal intensities should be similar if the
blotting efficiency is not different. A
stronger wild-type band could be due to a
contamination with wild-type ES cells or
feeder cells. However, differences might
also be caused by strange integration
events, with the duplication of large geno-
mic regions. Therefore, only clones with
Fig. 4.2 Detection of homologous recombinants
by Southern blot. Restriction enzyme A cuts once
in the homologous side arms of the targeting con-
struct and once within the neomycin selection
cassette. For Southern blot, the genomic DNA is
digested with restriction enzyme A and hybridized
with an external probe which stains bands of dif-
ferent size for homologous recombinants or the
wild-type gene. In case of heterologous recombi
nation, when the targeting construct inserts some-
where in the genome, the external probe will de-
tect only the wild-type allele of the gene of inter-
est. An internal probe will detect specific bands
also in case of heterologous recombinations.
Since the size of these bands by chance could be
similar to that of the recombined band, an inter-
nal probe is not used to identify homologous re-
combinants.
similar intensity of wild-type and recom-
bined bands are selected for further inves-
tigation.
If there is a need to insert in some dis-
tance of the selection cassette additional
sequences (e.g., a loxP site or a point mu-
tation), then an attempt should be made
to introduce these sequences together with
restriction sites that can be used for South-
ern blot analysis. Alternatively, genomic
DNA of homologous recombinant clones
can be checked by PCR for the presence of
that sequence.
In order to exclude the unlikely event
that a heterologous recombination oc-
curred together with the homologous re-
combination, a second Southern blot anal-
ysis is carried out with a probe inside the
targeting construct (see Fig. 4.2). This in-
ternal probe will recognize not only ho-
mologous, but also heterologous recombi-
nations.
Conditional gene targeting is possible
using the cre-loxP system; this consists of
the recombinase cre and its corresponding
binding site on the DNA, termed the loxP
sequence (Fig. 4.3) [18, 19]. Originally, the
cre-loxP system is used by the bacterio-
phage P1 to integrate and excise from the
bacteria genome. The DNA region to be
deleted should be flanked by 34 bp-long
loxP sites in the same orientation. Cre re-
combinase molecules bind to each of these
loxP sites, interact with each other, and
catalyze the deletion of the intervening
DNA. This could result in a knock-out, or
in the removal of a stop cassette which
prevented the expression of a gene. The
loxP sites should be introduced in such a
way that they do not interfere with the
promoter, with enhancers/silencers or with
the splicing machinery. It is possible to in-
sert them into the 5' non-coding region,
but only in one orientation, as the other
orientation contains an AUG translation
start site. Homozygous mice carrying a
genomic sequence flanked by loxP sites
(floxed) should have a normal expression
of that gene and a normal phenotype.
The expression of the cre recombinase
determines when, and in which cells, the
deletion occurs. Many transgenic mice are
available where the cre recombinase is ex-
pressed under the control of a tissue-spe-
cific or an inducible promoter. Since cre is
a viral enzyme, there is no natural target
in mammals, and cre-transgenic mice
should have a normal phenotype. Mating
Fig. 4.3 An essential part of the gene
(white box) is flanked by short, directed
loxP sequences (triangles) which are in-
serted in intronic or non-coding regions
in order not to disturb normal expression
of the gene. The Cre recombinase will
bind to the loxP sites, the Cre enzymes
will interact forming a DNA loop, and the
DNA in between the loxP sites will be ex-
cised leaving a single loxP site behind. Ex-
pression of Cre, therefore, determines in
which cells the deletion occurs.
of phenotypically normal floxed mice with
phenotypically normal cre mice results in
offspring which only have an alteration in
the gene of interest in cre-expressing tis-
sues.
Cre function can also be regulated on a
post-transcriptional level. If the cre recom-
binase is fused with a mutated hormone
binding domain of a steroid receptor, the
fusion protein is sequestered into the cyto-
sol and cannot reach the nucleus [20].
Only in the presence of a specific steroid
binding to the hormone binding domain
can the fusion protein translocate to the
nucleus, where it will catalyze the DNA
deletion.
As a selection cassette often disturbs the
normal expression of a gene, it must be
removed in conditional knock-out mice,
and this can be achieved by introducing a
floxed selection cassette. Homologously re-
combined ES cell clones are then transi-
ently transfected in vitro with a cre recom-
binase expression vector. If three loxP sites
were to be introduced, different deletions
would be possible (1?2, 1?3, and
2?3), and normally all are found among
the clones picked. Alternatively, the selec-
tion cassette can be flanked by frt sites
(flirted) that are recognized by the flp re-
combinase [20]. Mice expressing the flp re-
combinase under the control of an ubiqui-
tous promoter can be used to delete the
flirted cassette in vivo.
If the selection cassette also contains a
negatively selectable marker such as tk,
then cells that have lost the cassette can
be selected for. The tk gene must be re-
moved in vitro, as its expression impairs
sperm motility in vivo and this results in
sterile chimeric mice [21]. ES cells without
a selection cassette can easily be retargeted
by the original targeting construct to ob-
tain homozygously recombined ES cells.
4.2.3
Cloning of the Targeting Construct
As targeting vectors are normally in the
range of 15 to 20 kb, cloning may become
rather complicated. As a general strategy,
we first try to subclone the two homolo-
gous arms, and then introduce selection
cassette, loxP sites and other sequences by
directed cloning with two sticky ends.
As a source of isogenic DNA, phage arti-
ficial chromosome (PAC) or bacterial artifi-
cial chromosome (BAC) clones could be
used. Genomic mouse libraries of PAC or
BAC clones are available, and can be used
for screening (www.hgmp.mrc.ac.uk). As a
probe, a genomic sequence of about 0.5
1.5 kb can be used that does not contain
any repetitive sequences, and which could
be tested by programs available on the In-
ternet (www.repeatmasker.org). Positive
clones are ordered and used for subclon-
ing (www.hgmp.mrc.uk).
To subclone a large piece of genomic
DNA, and also to create the knock-out con-
structs quickly, the technique of recombi-
nation cloning can be used [22]. Here, a
vector backbone with 50-bp homologous
regions at both ends is prepared by PCR
and electroporated into bacteria containing
a PAC or BAC clone and an expression
plasmid encoding Red a, b and c which al-
lows efficient homologous recombination
of regions with only 50 bp homologous
DNA.
Based on the sequence of the mouse ge-
nome, it is relatively easy to design a clon-
ing strategy and primers for PCR. How-
ever, it should be taken into account that
the genomic sequence found in the data-
base was derived from different mouse
strains, and it is therefore possible that the
genomic sequence of the ES cells will dif-
fer slightly from the genomic sequence in
the database. For all technical aspects
where the exact sequence is required
(PCR, recombination cloning, restriction
sites), it is therefore useful to confirm it as
early as possible in the planning.
4.2.4
ES Cell Culture and the Generation
of Chimeric Mice
ES cell lines are derived from inner cell
mass of the blastocyst embryo. Different
ES cell lines are available derived from
129Sv and C57Bl6 mice, and have been
successfully used in the creation of geneti-
cally modified mice. They are cultured on
a subconfluent layer of irradiated embryo-
nic fibroblasts as feeder cells, which pro-
vide membrane-bound leukemia inhibitory
factor (LIF) inhibiting ES cell differentia-
tion. In addition, LIF is added also to the
medium. The ES cell medium consists of
DMEM supplemented with 20% fetal calf
serum (FCS), high glucose, sodium pyru-
vate, non-essential amino acids and b-mer-
captoethanol to provide optimal growth
conditions for the ES cells. One crucial
factor at this point is the serum, as differ-
ent serum charges are differentially suited
for ES cells. It is therefore recommended
to test different serum charges on the ES
cells used in the laboratory, and to pur-
chase a large stock of the best serum
charge in order to allow reproducible con-
ditions. With increasing passage number,
the tendency of ES cells to differentiate is
growing, and this in turn decreases the ca-
pacity for germ line transmission. Thus, a
huge stock of low-passage number ES cells
should be frozen. The shorter the time of
in vitro culture, the higher the chance of
germ line transmission. ES cells should be
split in the ratio 1: 5 every 2 days in order
to achieve optimal growth conditions and
to reduce the risk of differentiation. The
medium should be changed every day.
ES cells can be transfected, transduced
with retroviruses or electroporated
(Fig. 4.4). In our laboratory, electroporation
has proved to be successful (410
7
cells +
100 lg linearized DNA in 800 lL PBS;
0.8 kV, 3 lF; Bio-Rad Gene Pulser). After
electroporation, the cells are distributed on
810 cm plates with feeders, and the selec-
tion is commenced about 20 h later. The
preferred selection marker is neomycin,
and selection is performed with
500 lg mL
1
G418. After 23 days, almost
all ES cells are dying, but colonies of surviv-
ing ES cells become visible at about day 5.
On day 6 of the selection, about 360 clones
are picked, expanded, and later frozen and
analyzed. We perform Southern blot analy-
sis routinely, using an external probe to de-
tect homologous recombination, and an in-
ternal probe to detect multiple integrations.
Two to three clones with a homologous re-
combination, a single integration, and equal
intensity of recombined and wild-type band
in the Southern blot, are then injected into
blastocysts to generate chimeric mice. The
blastocysts are 3.5-day-old embryos and con-
tain a large cavity into which the mutated
ES cells are injected. The injected blasto-
cysts are then transferred into pseudopreg-
nant mice to develop further, in time giving
rise to chimeric mice which are derived
partly from the wild-type blastocysts and
partly from the recombined ES cells. Pseu-
dopregnant mice are generated by mating
female mice with sterile males. At 2.5 days
after successful mating (indicated by a vagi-
nal plug), the female mice can receive the
injected blastocysts. ES cells are normally
injected into blastocysts derived from mice
with a different coat color, so that a high
contribution of ES cells to the chimeric
mice is indicated by a high contribution of
ES cell-derived coat color.
The chimeric mice are then tested for
their ability to transmit the recombined
gene to their offspring. Since the ES cells
were derived from male mice and male
contribution to the gonads is dominant,
only male chimeric mice will be tested for
germ line transmission.
If the coat color of the ES cells is domi-
nant over that of the female mice crossed
with the chimeras (e.g., ES cells derived
from agouti-colored 129Sv injected into
blastocysts derived from black C57Bl6),
germ line transmission can be readily seen
by offspring with the coat color of the ES
cell mice.
Chimeric mice can be crossed with mice
from the same strain as the ES cells to ob-
tain corresponding inbred animals. If they
are crossed with mice from a different
strain, outbred mice with variable genetic
background are generated. In some cases
for example, for a disease model that
works only in a specific genetic back-
ground it may be necessary to switch
the mouse strain. This is achieved by
crossing germ line offspring to another
mouse strain. The offspring are genotyped
for the mutation and crossed again with a
wild-type mouse of the desired strain.
Backcrossing mice for 10 generations re-
sults in a rather homogenous genetic
background very similar to that of the de-
sired mouse strain.
Fig. 4.4 The generation of gene-targeted mice.
The targeting construct will be electroporated into
ES cells. After selection for stable transfectants ex-
pressing a specific antibiotic resistance (here:
neomycin resistance, neo), homologous recombi-
nants will be identified and injected into blasto-
cysts. Transfer of injected blastocysts into spe-
cially conditioned (pseudopregnant) foster
mothers results in chimeric mice that consist
partly of cells derived from mutant ES cells (dark),
and partly of cells originating from the wild-type
blastocysts used for injection (white). Male chi-
meric mice are then crossed with wild-type fe-
males. Sperm derived from mutant ES cells (dark)
will result in mice, that in all cells are heterozy-
gous for the mutation. Intercrossing of heterozy-
gous animals will yield homozygous mice that will
be analyzed phenotypically.
blastocyst injection
of mutant ES cells
4.3
Analysis of Genetically Modified Mice
Heterozygous mice have, in most cases,
no or only a very subtle phenotype. If het-
erozygous mice have a defect, it could be
difficult to obtain highly chimeric mice. A
conditional knock-out of the gene would
then be advisable. Heterozygous mice are
crossed and wild-type, heterozygous and
homozygous offspring are compared with
each other. If the genetic background is
mixed, litter mates should always be com-
pared with each other. In the case of
inbred mice, mice of the same age of dif-
ferent litters could also be compared.
For all genetically modified mice, it is
necessary to test first whether homozygous
mutant mice are born at Mendelian ratio
(embryonic lethality?), are viable, grow nor-
mally, are fertile, and have a normal life
span. A histochemical analysis should then
be carried out of all tissues that express the
mutated gene. More specific analyses (e.g.,
immunofluorescence, electron microscopy
or biochemical analysis) of tissues or pri-
mary cells derived from the mutant mice
depends on the phenotype.
If the phenotype of the mice is not too
severe, then disease models may be ap-
plied. Diseases might be induced by the
systemic or topic administration of drugs,
by surgery, by infections, or by matings
with other mice that spontaneously devel-
op diseases.
4.4
Alternative Methods
The sequencing of the mouse genome and
the widespread availability of transgenic fa-
cilities makes it much easier today to gen-
erate gene-targeted mice than was possible
some years ago. Nonetheless, the total
time taken from the targeting construct to
the mutant mouse available for analysis is
still about one year, and even longer for
conditional mice. One must therefore eval-
uate carefully whether alternative methods
might provide similar information, and on
a shorter time scale. Alternatives would be
the use of cellular systems or transgenic
mice expressing dominant negative or con-
stitutively active forms of the protein or
siRNA reducing the expression of the gene
of interest (see Part I, Chapter 10 and Part
III, Chapter 3).
The disadvantages of these methods are
similar to all inhibitors and activators how-
ever: the effect is quite likely only partial,
and there may be unwanted side effects
on other molecules. Consequently, in the
future genetic alteration by targeted muta-
genesis will be an indispensable tool for
both in vivo and in vitro approaches to bio-
pharmaceutical research.
References
1 Inamdar, M. S. (2001) Functional genomics
the old-fashioned way: chemical mutagenesis
in mice. BioEssays 23:116120.
2 Lund, A. H., Turner, G., Trubetskoy, A., Ver-
hoeven, E., Wientjens, E., Hulsman, D., Rus-
sell, R., DePinho, R. A., Lenz, J., van Lohui-
zen, M. (2002) Genome-wide retroviral inser-
tional tagging of genes involved in cancer in
Cdkn2a-deficient mice. Nat. Genet. 32:160
165.
3 Mikkers, H., Allen, J., Knipscheer, P., Ro-
meijn, L., Hart, A., Vink, E., Berns, A., Ro-
meyn, L. (2002) High-throughput retroviral
tagging to identify components of specific sig-
naling pathways in cancer. Nat. Genet. 32:153
159.
4 Pennacchio, L. A. (2003) Insights from hu-
man/mouse genome comparisons. Mamm.
Genome 14:429436.
5 Hahn, W. C., Counter, C. M., Lundberg, A. S.,
Beijersbergen, R. L., Brooks, M. W., Weinberg,
R. A. (2002) Creation of human tumour cells
References 659
with defined genetic elements. Nature 400:
464468.
6 Hahn, W. C., Weinberg, R. A. (2002) Rules for
making human tumor cells. N. Engl. J. Med.
347:15931603.
7 Evans, M. J., Kaufman, M. H. (1981) Culture
of pluripotential cells from mouse embryos.
Nature 292:154156.
8 Martin, G. R. (1981) Isolation of a pluripotent
cell line from early mouse embryos cultured
in medium conditioned by teratocarcinoma
stem cells. Proc. Natl. Acad. Sci. USA
78:76347638.
9 Smithies, O., Gregg, R. G., Boggs, S. S., Kora-
lewski, M. A., Kucherlapati, R. S. (1985) Inser-
tion of DNA sequences into the human chro-
mosomal beta-globin locus by homologous re-
combination. Nature 317: 230234.
10 Doetschman, T., Gregg, R. G., Maeda, N.,
Hooper, M. L., Melton, D. W., Thompson, S.,
Smithies, O. (1987) Targetted correction of a
mutant HPRT gene in mouse embryonic stem
cells. Nature 330:576578.
11 Thomas, K. R., Capecchi, M. R. (1987) Site-
directed mutagenesis by gene targeting in
mouse embryo-derived stem cells. Cell 51:
503512.
12 Schwartzberg, P. L., Goff, S. P., Robertson,
E. J. (1989) Germ-line transmission of a c-abl
mutation produced by targeted gene disrup-
tion in ES cells. Science 246:799803.
13 Talts, J. F., Brakebusch, C., Fssler, R. (1999)
Integrin gene targeting. Methods Mol. Biol.
129:153187.
14 Hogan, B., Beddington, R., Constantini, F.,
Lacy, E. (1994) Manipulating the Mouse Em-
bryo. Cold Spring Harbor Press, New York,
USA.
15 Kaul, A., Koster, M., Neuhaus, H., Braun, T.
(2000) Myf-5 revisited: loss of early myotome
formation does not lead to a rib phenotype in
homozygous Myf-5 mutant mice. Cell 102:17
19.
16 Weissmann, C., Flechsig, E. (2003) PrP knock-
out and PrP transgenic mice in prion re-
search. Br. Med. Bull. 66:4360.
17 Deng, C., Capecchi, M. R. (1992) Reexamina-
tion of gene targeting frequency as a function
of the extent of homology between the target-
ing vector and the target locus. Mol. Cell. Biol.
12: 33653371.
18 Orban, P. C., Chui, D., Marth, J. D. (1992) Tis-
sue- and site-specific DNA recombination in
transgenic mice. Proc. Natl. Acad. Sci. USA
89:68616865.
19 Gu, H., Marth, J. D., Orban, P. C., Mossmann,
H., Rajewsky, K. (1994) Deletion of a DNA
polymerase beta gene segment in T cells
using cell type-specific gene targeting. Science
265:103106.
20 Branda, C. S., Dymecki, S. M. (2004) Talking
about a revolution: The impact of site-specific
recombinases on genetic analyses in mice.
Dev. Cell 6:728.
21 al-Shawi, R., Burke, J., Wallace, H., Jones, C.,
Harrison, S., Buxton, D., Maley, S., Chandley,
A., Bishop, J. O. (1991) The herpes simplex
virus type 1 thymidine kinase is expressed in
the testes of transgenic mice under the con-
trol of a cryptic promoter. Mol. Cell. Biol.
11:42074216.
22 Zhang, Y., Muyrers, J. P., Testa, G., Stewart, A.
F. (2000) DNA cloning by homologous recom-
bination in Escherichia coli. Nat. Biotechnol.
18:13141317.
Abstract
In the late nineteenth century, the German
botanist Brefeld investigated cultures of
the fungus Mucor mucedo, which were pre-
pared on horse dung. He noticed small
sorocarps containing tiny oval spores that
germinated into amoeboid cells and fed on
bacteria, eventually forming new spore-
containing fruiting bodies. Brefeld named
this new species Dictyostelium mucoroides,
and concluded that it belonged to the Myx-
omycetes [1]. Since then, about 70 cellular
slime mold (Dictyostelid) species have
been described, most of them have been
described in detail [2]. The evolutionary
descent of the Dictyostelids is under dis-
cussion, with molecular phylogeny studies
placing them at the root of the Crown
group of eukaryotes, within the clade of lo-
bose amoebae (Amoebozoa) [3]. Thus, the
Dictyostelids are distantly related to the
eukaryotic protozoans, have little in com-
mon with plants, and are among the clos-
est living relatives to animals and fungi.
The Dictyostelids are ubiquitously found
in moderate climates where the solitary
amoebae live in soil on dung and decaying
forest leaves and feed on bacteria or yeasts.
When Dictyostelid cells meet a critical ra-
tio of cell density to available food, they
collect into aggregation centers which
transform into polar, slug-like structures
surrounded by an extracellular slime
sheath (Fig. 5.1). Within the multicellular
structures the individual cells differentiate
into two cell types: prestalk and prespore
cells. Prestalk cells form the future stalk of
the fruiting body, which is the terminal
structure of the differentiation process,
while prespore cells form encapsulated,
dormant spores that locate at the tip of the
fruiting body. Under favorable conditions,
the spores may germinate and begin an-
other round of the asexual life cycle. In
1933, Raper isolated Dictyostelium disco-
ideum from partially composed leaves from
a hardwood forest in North Carolina [4].
Since then, D. discoideum has become the
preferred object of research into Dictyoste-
lid biology. D. discoideum is also well-suit-
ed for studies of fundamental biological
phenomena that play important roles in
human health and disease. For example,
cytokinesis is critical in cell proliferation
and is thus an integral part of the im-
mune response, tissue maintenance, and
cancer. Cell motility is an essential early
event in the metastasis of tumor cells, and
in angiogenesis by endothelial cells. Che-
motaxis and signal transduction by che-
moattractant receptors play a key role in
inflammation, arthritis, asthma, lympho-
cyte trafficking, and in axon guidance. Pha-
661
5
An NIH Model Organism for Biopharmaceutical and Biomedical
Research: The Lower Eukaryote Dictyostelium discoideum
ISBN: 3-527-31184-X
5 Dictyostelium discoideum: Biopharmaceutical and Biomedical Research with a Lower Eukaryote 662
Fig. 5.1 Multicellular stages of Dictyostelium discoi-
deum development, as isolated in 1933 from de-
caying leaves in a forest in North Carolina (photo-
graphs by K. B. Raper) [4]. (A) Aggregation center;
(B) pseudoplasmodium transforming from mound
stage to a migrating slug; (C, D) migrating slugs;
(EG) migrating slugs that just cease migration
and prepare for culmination; (HK) different
stages of culmination; (L) mature fruiting body,
consisting of a stalk that lifts a ball of spores
from the substratum into the air.
gocytosis is a critical process involved in im-
mune surveillance and antigen presenta-
tion. Cell-type determination, cell sorting,
and pattern formation are basic features of
embryogenesis, and alteration of these
events can lead to neoplasms. Many of these
phenomena may be easier to analyze in uni-
cellular D. discoideum than in complex me-
tazoans; moreover, it was also shown that
an embryotoxicity assay could reproduce
the teratogenic activities of valproic acid
analogues previously characterized in ani-
mal models [5]. This assay could in princi-
ple be used to predict the potential embryo-
toxicity of drugs not yet tested in animals
[6], and to suggest a common molecular
mechanism of action for some biopharma-
ceuticals in D. discoideum and in humans
[7, 8]. Consequently, D. discoideum was cho-
sen by the National Institutes of Health
(NIH) as a non-mammalian model organ-
ism for biomedical research (http://
www.nih.gov/science/models/d_discoide-
um/). In this chapter, the state-of-the-art
of biopharmaceutical and biomedical re-
search with D. discoideum is reviewed. First,
the tools available for gene discovery and
analysis are described, after which experi-
ments aimed at optimizing D. discoideum
as an expression host for the production of
therapeutic proteins are summarized. Avail-
able expression systems are described, and
the consequences of peculiar D. discoideum
codon usage on expression efficacy, and re-
cent advances in the fermentation of D. dis-
coideum cells, are discussed. Comment is
also made on post-translational modifica-
tions typical for D. discoideum, as it harbors
the machinery for post-translational protein
modifications such as phosphorylation, acy-
lation, formation of glycosyl phosphatidyli-
nositol anchors, and, in particular, O-linked
and N-linked glycosylation. In addition,
some selected aspects of biomedical re-
search are briefly highlighted: 1) D. discoide-
um as a model for the study of hostpatho-
gen interactions in infectious diseases such
as Legionnaires disease and pseudomonia-
sis; 2) screening future biopharmaceuticals
for potential embryotoxicity in humans
using recombinant D. discoideum strains
carrying reporter genes; and 3) develop-
ment of new concepts for the improvement
of gene transfer vectors in human gene
therapy by studying mobile genetic ele-
ments found in the D. discoideum genome.
Abbreviations
ATIII antithrombin III
bBTC bovine b-cellulin
CbfA C-module-binding factor
CMF conditioned medium factor
CsA contact site A protein
CSP circumsporozoite protein
Ddp2 D. discoideum plasmid 2
GABA c-aminobutyric acid
GFP green fluorescent protein
GPCR G protein-coupled receptor
GPI glycosyl phosphatidylinositol
GST glutathione-S-transferase
GUS b-glucuronidase
hCG human chorionic gonadotropin
IP
3
inositol-1,4,5-trisphosphate
M2 muscarinic acetylcholine receptor
type 2
MSP I merozoite surface protein I
PsA cell surface glycoprotein A
PSF prestarvation factor
REMI restriction enzyme-mediated
integration
REP transactivator of Ddp2 replication
sFasL soluble Fas ligand
tTA tetracycline-regulated trans-
activator
VPA valproic acid
Abbreviations 663
5.1
Introduction
The natural isolate NC-4 of D. discoideum
has a haploid genome of ca. 34 Mb that is
organized into six chromosomes [9]. The ri-
bosomal genes are encoded on about 100
copies per cell of a ca. 88 kb extrachromoso-
mal palindrome [10]. An international ge-
nome sequencing project is currently active
using the axenic D. discoideum strain AX4
(http://dictybase.org/). The sequencing is
carried out at the American Baylor College
of Medicine (Houston, Texas) and in two
European institutions, the Sanger Centre
(Hinxton, UK) and the Institute of Molecu-
lar Biotechnology (Jena, Germany). The
project is supplemented with a cDNA pro-
ject situated in Tsukuba (Japan), and should
be mainly completed by the time that this
book is published. A high-quality sequence
of ca. 30.5 Mb has been assembled. Prelim-
inary annotation predicts 12734 open read-
ing frames (ORFs) (Preliminary Directory
of Dictyostelium Genes, version 3; http://dic-
ty.sdsc.edu/annot-020303.html). Typical fea-
tures of the D. discoideum genome are pre-
sented in Table 5.1.
The recently completed annotation of
chromosome 2 has provided a detailed in-
sight into the structure of the D. discoide-
um genome as a whole [11]. It is highly
enriched in A+T nucleotides (78% on aver-
age), and contains few, small introns (see
Table 5.1). It was estimated that a typical
D. discoideum gene spans ca. 2.6 kb, which
means that 76% of the genome is coding
for cell functions [11].
First comparative analysis of the pre-
dicted genes on chromosome 2 has
strengthened the view that Dictyostelium is
phylogenetically more related to metazo-
ans and fungi than to plants [12]. It also
consolidates our conviction that D. dis-
coideum may provide significant input to
biomedical research. On the one hand, the
D. discoideum genome harbors genes with
similarity to metazoans that are absent in
other microbial models such as yeasts.
One example is a five-membered family of
novel G protein-coupled receptors
(GPCRs), which show high similarity to
mammalian GABA
B
receptors (c-aminobu-
tyric acid receptor type B) that have not
yet been identified in any other non-me-
tazoan species [13]. On the other hand, a
fairly large number of D. discoideum genes
show strong similarity to known human
disease genes. Among those are ortholo-
gues of ATP-binding cassette transporter
Table 5.1 Key features of the D. discoideum genome
a)
Genome size ca. 34 Mb
Ploidy haploid
Number of chromosomes 6
Chromosome sizes
b)
48 Mb
Extrachromosomal DNA
c)
ca. 100 copies of
88 kb palindrome
containing rRNA
genes
G+C content
d)
Exons 28%
Introns 13%
Intergenic 14%
Average whole genome 22%
Average gene length
d)
2.6 kb
Number of gene models
e)
12734
Average intron size
d)
180 bp
Average intergenic regions
d)
790 bp
Genome fraction coding for
exons
e)
60%
Mobile element content
f)
9.6%
a) Use http://dictybase.org to access internet re-
sources of the international Dictyostelium genome
project.
b) Data from Refs. [14, 15].
c) Data from Ref. [10].
d) Calculations based on chromosome 2 data [11].
e) As of March 2003; http://dicty.sdsc.edu/annot-
020303.html.
f) Data from Ref. [16].
proteins, proteins involved in the regula-
tion of cytoskeleton functions, and DNA
repair enzymes [13]. It was estimated,
based on the data on chromosome 2, that
about one-fifth of currently approved hu-
man disease genes may have orthologues
in D. discoideum. Taking this into consid-
eration, despite the evolutionary distance
between cellular slime molds and humans,
the powerful molecular genetics of D. dis-
coideum (discussed below) raise our hopes
that molecular mechanisms of human dis-
eases can be explored in this simple organ-
ism to help accelerate the development of
5.2
The Gene Discovery Tool Box
for Dictyostelium Research
5.2.1
Transformation Methods and Selection
Markers
Dictyostelium discoideum cells can be trans-
formed with plasmid DNA by convenient
methods. The first attempts at this, using
the calcium phosphate method, were re-
ported in the early 1980s [17, 18]. Proto-
cols were subsequently optimized by Firtel
and coworkers [19, 20] and modified in
many laboratories (e.g. Ref. [21]). Up to
2000 transformants can be recovered from
10
7
D. discoideum cells [20], with electro-
poration protocols having been reported
[22, 23]. It has also been noted that the
choice of transformation method can
strongly influence the yield of heterolo-
gously expressed proteins [24]. Hence,
both calcium co-precipitation and electro-
poration should be compared to generate
optimized expression strains for the pro-
duction of biopharmaceutical proteins in
D. discoideum.
Several auxotrophy markers and antibiot-
ic resistance genes are available for the se-
lection of transformed D. discoideum cells.
D. discoideum strains carrying mutations
or deletions in the pyr5-6 gene, which en-
codes UMP synthase, grow in a defined me-
dium (FM) only in the presence of uracil.
Thus, cells transformed with plasmids con-
taining the pyr-5-6 gene can be selected for
uracil prototrophy (ura
+
) in FM medium
lacking uracil [25]. Similarly, the thyA gene
encoding thymidylate synthase can be used
to screen for expression of plasmid-borne
enzyme after transformation into thy
cells
[26]. The pyr5-6 and thyA genes can be
cloned on plasmids under the control of
their native promoters, but can be used only
in strains auxotrophic for thymidine and ur-
acil, respectively [25, 26]. As discussed be-
low, both markers are functional as single-
copy genes and are therefore preferred
markers for gene disruption experiments.
On the other hand, single-copy auxotrophy
markers do not readily support high expres-
sion levels of heterologous genes and are
thus useless for D. discoideum-based expres-
sion systems.
The most successful antibiotic resistance
marker for gene expression efforts in D.
discoideum is the neomycin phosphotrans-
ferase (neo
R
) encoded on the bacterial
transposon Tn5. The neo
R
gene confers re-
sistance to the aminoglycoside antibiotic
G418. The neo
R
gene must be expressed
under the control of an appropriate D. dis-
coideum promoter. High copy numbers of
up to 500 plasmids per genome have been
reported; the plasmids probably amplify in
tandem arrays during the integration pro-
cess [24, 27, 28]. Typical G418 concentra-
tions used for selection of D. discoideum
transformants are in the range of 10 to
40 lg mL
1
.
Hygromycin B [29] is an alternative anti-
biotic resistance marker for the selection
5.2 The Gene Discovery Tool Box for Dictyostelium Research 665
of D. discoideum transformants. The hygro-
mycin phosphotransferase mediates resis-
tance to 2575 lg mL
1
hygromycin. Appli-
cation of this marker is limited by gener-
ally low transformation efficiencies. In pre-
vious reports, hygromycin-resistant clones
could only be established when the hygro-
mycin-resistance gene was placed on an
extrachromosomal vector [29]. Meanwhile,
hygromycin-resistant clones could also be
recovered from integrating expression vec-
tors [24]. Copy numbers per cell up to 300
have been observed with these vectors [24].
Interestingly, notable expression levels of
heterologous genes could only be achieved
if the cells were transformed by electro-
poration instead of calcium phosphate co-
precipitation [24].
Pleomycin [30] and blasticidin S [31, 32]
resistance genes are suitable selection
markers for gene disruption experiments,
as they confer resistance to the antibiotics
as single-copy genes. The blasticidin deami-
nase gene (bsr) is inserted as a single copy
only if low concentrations of 510 lg mL
1
blasticidin are used for selection. Other-
wise, between one and 20 copies per cell
have been observed after selection with
1060 lg mL
1
blasticidin [24, 28].
5.2.2
Generation of Mutants
The study of gene regulatory networks un-
derlying multicellular development is the
focus of interest in D. discoideum research.
Since D. discoideum is haploid, any mutant
defective in a single-copy gene will display
a phenotype. Thus, simply observing mul-
ticellular development of the mutants
identifies genes required for development.
Several mutagenesis methods have been
adapted for the use in D. discoideum, and
some of them will be briefly summarized
in the next sections.
5.2.2.1 Chemical Mutagenesis
Random mutagenesis of the D. discoideum
genome with chemical compounds was
the first method that allowed the isolation
of mutants defective in development (Fig.
5.2). Also, we owe the generation of the
widely used axenic laboratory strain AX3
to this procedure [33]. Incubation of cells
with N'-methyl N'-nitro N-nitrosoguani-
dine increased the natural mutation rate
about 1000-fold and produced mutants
with interesting phenotypes [34]. The iden-
tification of genes inactivated by chemical
mutagenesis requires careful genetic anal-
ysis of the obtained mutants. Although D.
discoideum can, in principle, enter a sexual
reproduction cycle, the generation of dip-
loids in the laboratory is both inefficient
and time-consuming [34]. Nevertheless
this method has allowed the generation of
the first physical maps of the D. discoide-
um chromosomes [34].
5.2.2.2 Gene Disruption and Gene
Replacement (Knock-out)
The transformation of D. discoideum cells
with plasmids carrying DNA fragments
with homology to certain chromosomal re-
gions often results in integration of the
plasmid at these loci via homologous re-
combination. Homologous recombination
operates quite efficiently in D. discoideum,
and is therefore the method of choice for
targeted inactivation of genes. Single-copy
markers such as bsr
R
, pyr5-6, and thyA are
available for selection of mutants. We
should distinguish gene disruption from
gene replacement (Fig. 5.2):
In gene disruption, a single recombina-
tion event leads to integration of the en-
tire plasmid, including the selection
marker, into the targeted gene locus.
In gene replacement the selection marker
is placed between two pieces of DNA ho-
mologous to the targeted chromosomal
locus, and two recombination events are
required to exchange a piece of the tar-
geted genomic locus for the selection
marker (the plasmid is lost; see Fig. 5.2).
Gene disruption was first demonstrated in
1987 for the genes encoding myosin II
heavy chain and a-actinin [35, 36]. Gene
replacement was also first shown for the
gene encoding myosin II heavy chain [37].
Experience from many such experiments
suggests that the selection marker should
be flanked by at least 1 kb of homologous
DNA for efficient gene replacement.
Clearly, the knock-out methods are limited
to single-copy genes and genes dispensa-
ble for growth.
Knock-out approaches function quite ef-
ficiently for some genes, but poorly for
others (frequencies of less than 0.1%) (see
Part III, Chapter 4). Thus, methods were
sought to improve gene replacement effi-
ciency that is, to suppress mutants in
which a single cross-over rather than a
double recombination has occurred. This
was achieved by introducing a positive-
negative selection based on translation
stop codon suppression. In D. discoideum,
approximately 90% of all translation stop
codons are UAA, and the introduction of
an additional tRNA gene that suppresses
UAA (ochre) stop codons is lethal for the
cell [38]. If an ochre suppressor tRNA is
cloned onto a gene replacement vector,
any transformant that received the entire
plasmid by a single recombination event
would be eliminated from the culture due
to expression of the ochre tRNA gene from
the integrated plasmid. According to this
scenario, the presence of an ochre tRNA
gene on a gene replacement vector-en-
riched cells with double recombination
events that exchanged a piece of the tar-
geted genomic DNA for the selection
marker. In practice, this positive-negative
selection reduced the background of non-
targeted transformants about 20-fold [39].
Once a mutant with a defined knock-out
phenotype is isolated, expression of the
cloned gene within the mutant is desirable
in order to restore the wild-type pheno-
type. This is in fact one of the major moti-
vations within the Dictyostelium research
community to develop gene expression
(shuttle) vectors. Such vectors, which were
primarily optimized for expression of ho-
mologous genes for complementation
studies, can also serve as optimized vectors
for the expression of heterologous genes
(see below).
5.2.2.3 Restriction Enzyme-mediated
Integration
Restriction enzyme-mediated integration
(REMI) is a commonly used method for
random mutagenesis that has been
adapted for the use in D. discoideum by
Kuspa and Loomis [40]. A plasmid contain-
ing an appropriate selection marker is lin-
earized with a restriction enzyme, and the
linear plasmid is transformed by electro-
poration along with the restriction enzyme
(Fig. 5.2). It is assumed that the restriction
enzyme will cut certain chromosomal loci
at specific recognition sites. If a trans-
formed plasmid would ligate with its
sticky ends to the double-strand break in-
troduced by the restriction enzyme, mu-
tants can be isolated based on the selec-
tion marker present on the integrated
REMI plasmid.
In most REMI experiments, transforma-
tion frequencies are 20- to 60-fold higher
when functional restriction enzyme is co-
transformed [40]. In order to isolate geno-
mic DNA flanking the integration site of
the REMI plasmid, genomic D. discoideum
DNA from the mutants is digested with a
Fig. 5.2 Molecular genetic techniques used to
generate Dictyostelium discoideum mutants. (a)
Random mutagenesis of the entire genome can
be performed with DNA-damaging agents such as
nitrosoguanidine. Point mutations will eventually
lead to inactivation of genes. After screening for
phenotypes of interest, the gene inactivated in a
particular mutant clone must be identified by sex-
ual or parasexual genetics, or by complementation
with an expression library. (b) Gene disruption is
performed with a plasmid that contains a region
of homology to the chromosomal copy of a gene
that shall be inactivated (gene of interest). A se-
lection marker is inserted on the plasmid down-
stream of the homology region. After transforma-
tion, the homology region recombines with the
chromosomal partner, resulting in integration of
the entire plasmid. After selection for the pres-
ence of the marker mutant, phenotypes can be
analyzed. (c) Gene replacement works with two
homology regions cloned on a vector, with the se-
lection marker inserted between the two. After
transformation, the plasmid-borne homology re-
gions will recombine with their respective part-
ners, such that the chromosomal region between
the two homology regions is exchanged for the se-
lection cassette. (d) Restriction enzyme-mediated
integration (REMI) is a random mutagenesis tech-
nique. A plasmid carrying a selection marker is
linearized with restriction enzyme R1. The vector
is then transformed together with R1. R1 is ex-
pected to cut the chromosomal DNA at its specif-
ic recognition sites, allowing the plasmid to insert
and ligate to the compatible ends. After selection
for marker gene presence, mutants with interest-
ing phenotypes are collected. Genomic DNA is
prepared from such mutants and digested with re-
striction enzyme R2, which does not recognize the
REMI plasmid. The digested genomic DNA is self-
ligated and transformed into E. coli cells. Colonies
will appear due to the presence of an antibiotic
resistance gene and a bacterial origin of replica-
tion present on the REMI plasmid. Genomic DNA
flanking the original site of integration of the
REMI plasmid is recovered on the rescued plas-
mid. This vector is equivalent to a gene replace-
ment vector and can be used directly to verify the
mutant phenotype of the original REMI mutant.
(e) The gene trap (promoter trap) technique is a
modification of both gene disruption and REMI. A
promoter-less fluorescent marker (green fluores-
cent protein, GFP) is cloned on a plasmid to-
gether with a conventional antibiotic resistance
marker. The vector is linearized with restriction
enzyme R1 and transformed into D. discoideum
cells together with R1. Transformants are selected
based on resistance to the antibiotic. Resistant
mutants are screened for GFP fluorescence, which
can only occur after in-frame integration of the
plasmid into a gene that is, downstream of a
chromosomal promoter (P). The plasmid can be
rescued together with flanking genomic DNA by
digesting mutant DNA with enzyme R2, self-liga-
tion, and transformation of bacteria as in REMI.
(f ) Underexpression of essential genes by stop co-
don suppression. The 3 part of a gene of interest
is cloned on a gene disruption vector. A normal
glutamic acid codon somewhere in the coding re-
gion is replaced by an in-frame amber (UAG) stop
codon. A tRNA
Glu
(UAG) suppressor tRNA gene
(supp) is placed on the gene disruption vector to-
gether with a selection marker. After transforma-
tion into D. discoideum cells, the vector will insert
into the gene of interest by homologous recombi-
nation, creating a single intact gene that carries a
premature amber stop codon. After transcription
of that gene, the amber-modified allele can be
translated into a full-length protein at low rates,
due to inefficient amber stop codon suppression.
(g) The steady-state level of a particular protein
can be down-regulated by expressing homologous
double-stranded RNA (dsRNA). In classical anti-
sense RNA-mediated down-regulation, a cDNA is
cloned in inverse orientation on a plasmid down-
stream of a strong promoter, resulting in tran-
scription of antisense RNA that can hybridize to
the homologous mRNA, and either prevent its
translation or trigger its degradation. As a random
mutagenesis method, transformation of an anti-
sense library and screening the transformants for
interesting phenotypes is possible. Phenotypes re-
sult from antisense RNA-mediated mRNA decay
of unknown genes. The targeted genes can be
subsequently identified by rescuing the individual
antisense RNA expression plasmids active in the
respective mutants. RNAi approaches operate by
expressing two pieces of cDNA cloned head-to-tail
on an expression plasmid. After transcription, the
artificial RNA can fold back to form a hairpin
dsRNA that triggers RNAi-mediated degradation
of the homologous mRNA.
3
second restriction enzyme that does not rec-
ognize the REMI plasmid. The plasmid is
circularized by ligation and transformed
into bacteria. Plasmid integration at restric-
tion sites in >70% of recovered transfor-
mants was observed [40]. A major advantage
of the REMI method is that the recovered
REMI plasmid can directly serve as a gene
replacement vector (Fig. 5.2). Thus, the
knock-out phenotype of a successfully
tagged REMI mutant can be reproduced
by reintroduction of the recovered REMI
plasmid into a wild-type strain. The result-
ing knock-out mutant is expected to show
the REMI mutant phenotype.
5.2.2.4 Tagging Gene Disruption (Knock-in)
The gene trap (promoter trap) technology is
a random insertion mutagenesis method.
The idea is to clone a promoter-less reporter
gene (e.g., green fluorescent protein; GFP)
on a plasmid that also contains a selection
marker (see Fig. 5.2). After transformation,
the plasmid inserts randomly into the D.
discoideum genome. If the GFP gene is by
chance fused in-frame with a chromosomal
gene, a fusion product is expressed from
the endogenous promoter and the mutants
can be identified by green fluorescence.
Subsequently, such mutants can be
screened for the phenotype of interest [41,
42]. This method can also be applied as a
REMI mutagenesis; that is, the plasmid is
linearized with a restriction enzyme and
transformed along with the same enzyme
in order to improve mutagenesis efficiency
(as shown in Fig. 5.2).
5.2.2.5 Stop Codon Suppression
As classical knock-out strategies cannot
be applied to essential genes, alternative
methods for controlled underexpression of
such genes are desired. Amber (UAG)
translation stop codons are rarely used in
D. discoideum. Thus, expression of amber
suppressor tRNAs in D. discoideum is well
tolerated, without causing obvious pheno-
types [38, 43]. Suppression efficiencies are
in the range of 741% when the suppres-
sor tRNA gene is placed onto a high-copy
plasmid together with a reporter gene car-
rying an amber mutation [44], but only 2
8% if the suppressor tRNA gene and the
lacZ(amber) gene are placed on plasmids
that cannot be selected for with antibiotics
[38]. This has led to the conclusion that
the efficacy of suppression correlates with
the copy number of the suppressor tRNA
gene, and prompted the consideration that
chromosomal genes engineered to contain
a premature amber stop codon may under-
express the encoded protein at reduced lev-
el due to inefficient suppression by a sin-
gle-copy suppressor tRNA gene. If sup-
pression occurred at 10% efficiency, for ex-
ample, a protein that is essential for
growth would be expressed in a non-func-
tional, truncated form at 90%, and in a
full-length functional form in 10%, of
maximal translation efficiency. This 10%
of remaining functional protein may be
sufficient to support survival of the cells,
whereas its complete absence (after gene
disruption) would kill the affected cell. We
recently tested this hypothesis with the D.
discoideum cbfA gene, encoding a putative
transcription regulator, which could not be
inactivated by gene replacement
approaches [43]. One part of the cbfA gene,
modified to contain an in-frame amber co-
don, was cloned onto a gene disruption
vector together with an amber suppressor
tRNA gene (see Fig. 5.2). The homologous
chromosomal part of the cbfA gene was re-
placed by the cbfA(amber) part by homolo-
gous recombination. This resulted in a sin-
gle, functional cbfA(amber) gene, which al-
lowed expression of a functional, full-
length CbfA protein at less than 5% of
wild-type level (Fig. 5.2).
5.2.2.6 Antisense RNA and RNA
Interference (RNAi)
Dictyostelium discoideum apparently uses
endogenous antisense RNAs to regulate
normal cell functions [45], and the cells
contain a potent RNA interference (RNAi)
machinery [46]. Experimental down-regula-
tion of endogenous D. discoideum genes by
expression of plasmid-borne antisense
RNAs was first demonstrated by Firtels
laboratory for the discoidin I gene family
[47]. These authors inverted a piece of the
discoidin Ia gene such that antisense RNA
was co-expressed with endogenous discoi-
din Ia mRNA under the control of the na-
tive discoidin Ia promoter. Knecht and
Loomis subsequently showed that expres-
sion of mhcA (myosin II heavy chain) anti-
sense RNA produced mutants with the
same phenotype as previously isolated
mhcA knock-out cells [48].
The genome-wide search for new mu-
tants displaying a characteristic phenotype
is impaired by the fact that multicopy genes
and genes essential for growth will likely es-
cape REMI mutagenesis. Since down-regu-
lation of D. discoideum genes by expression
of antisense cDNAs works fairly well, a ge-
nome-wide, random, antisense cDNA ex-
pression mutagenesis approach has been
established by Gomer and co-workers [49,
50]. Shotgun isolation of new mutants
using antisense cDNAs libraries is particu-
larly valuable for the large-scale isolation
of developmental mutants (Fig. 5.2). This
technology may be taken a step further by
expressing antisense cDNA libraries from
promoters active only at specific time peri-
ods of development, thus enabling the isola-
tion of genes important in that particular
stage. An antisense cDNA the expression
of which caused a particular phenotype
can easily be identified by recovering the ex-
pression plasmid from the mutant. A gener-
al drawback in trying to characterize known
or orphan genes by expression of antisense
DNA is that many attempts to down-regu-
late certain genes fail due to negative coun-
ter selection if the targeted gene is required
for growth. In addition, the targeting of
highly expressed genes may work better
than the targeting of weakly expressed
genes. Also, it cannot be excluded that an
antisense cDNA may target several closely
related genes at the same time, thus compli-
cating the interpretation of results.
Expression of artificial cDNAs that are
able to form double-stranded hairpin
RNAs can strongly down-regulate the ex-
pression of endogenous genes by RNAi
[46]. Considering that usual gene replace-
ment techniques are rather time-consum-
ing, the use of RNAi for gene function
analysis in D. discoideum is expected to
boom within the next few years, particular-
ly after completion of the Dictyostelium ge-
nome project and progression from the ge-
nome to the proteome era (see also Part I,
Chapter 10 and Part III, Chapter 3).
5.2.3
Reporter Genes
Any reporter gene useful in other organ-
isms for the in vitro measurement of pro-
moter strength, in situ staining to detect cell
type-specific promoter activity, or visualiza-
tion of protein compartmentation, are also
successfully applied in D. discoideum. Ex-
pression of functional b-galactosidase, fol-
lowed by in situ staining of multicellular
stages, is an outstanding tool for gene func-
tion analysis in D. discoideum [51]. Using
this particular reporter gene assay, it has
been possible to follow the fate of differen-
tiating cell types at all multicellular stages
by fusing the b-galactosidase-encoding lacZ
gene to promoters of genes the disruption
of which had created developmental pheno-
types [52]. Of similar use is the expression
of GFP, a fluorescent marker that made it
possible to observe cell movement in real
time and in living cells [53, 54].
5.2.4
DNA Microarrays
We have by now a near-complete genomic
sequence of D. discoideum in hand. Thus,
it is consequent to perform whole-genome
expression analyses with DNA microarrays
(see Part I, Chapter 3). Currently used mi-
croarrays contain ca. 6000 genes, repre-
senting about one half of the genome [55
58]. Experiments focus on the evaluation
of expression kinetics of known, develop-
mentally regulated genes, as well as the
discovery of new genes showing similar
expression patterns. Since hundreds of D.
discoideum mutants each carrying a sin-
gle known gene knock-out are patiently
waiting in the freezers for analysis, it can
be expected that the work on the determi-
nation of gene regulatory networks con-
trolling multicellular development of D.
discoideum will continue to expand within
the next years.
5.3
Production of Recombinant Proteins
in D. discoideum
5.3.1
Expression Systems
5.3.1.1 Promoters
Actin accounts for ca. 8% of the total pro-
tein of a D. discoideum cell. There are at
least 23 functional actin genes [59], and pro-
moters and transcription terminators iso-
lated from actin genes are popular elements
for the use in expression cassettes. D. dis-
coideum promoters are generally small and
very A+T-rich. For example, a cloned DNA
fragment derived from the actin15 gene
(act15) upstream region, which is one of
the most frequently used promoters for
the expression of D. discoideum and foreign
proteins, is only about 250 bp in length and
composed of 87% A+T [60]. The act6 and
act15 promoters are frequently used to ex-
press proteins in growing D. discoideum
cells and to regulate the expression of anti-
biotic selection markers. Although their ac-
tivities are highest at early development,
both promoters are prematurely activated
when the D. discoideum cells grow in liquid
medium, which is the preferred condition
for transformation [61]. Both promoters
are repressed when the cells grow on bacte-
ria, making them useless for transforma-
tion of wild strains of D. discoideum that
can only feed on bacteria [62]. On the other
hand, repression of the actin promoters can
be used to produce heterologous proteins
that are toxic to D. discoideum cells: protein
accumulation can be repressed by culturing
the cells in association with bacteria, fol-
lowed by a production phase that is started
by removing residual bacteria and suspend-
ing the cells in buffer. However, one may
take into consideration that the advantage
of stronger transcription of actin promoters
at early development (i.e., after suspending
them in buffer) may be outweighed by low
translation rates at this stage due to reduc-
tion of ribosome number in starving cells
[63].
Unfortunately, there is no promoter
available that allows for the induction of
protein expression in response to an extra-
cellular signal. Thus, if regulated expres-
sion of heterologous genes is desired, one
can either use developmentally regulated
promoters, or promoters that respond to
artificial extracellular signals such as DNA-
damaging agents. The most commonly
used developmentally regulated promoter
is derived from the discoidin Ic locus
(disI). Expression of discoidin Ic, a compo-
nent of the extracellular matrix during
multicellular development, is induced by
the quorum sensing factors PSF [64] and
CMF [65]. On the other hand, discoidin I
expression is repressed by folate and food
bacteria during growth [66, 67], and by
cAMP during early development [68].
When cells grow on bacteria in shaken
suspension, no significant discoidin ex-
pression is observed below 3510
6
cells mL
1
, whereas discoidin expression is
detectable at very low cell densities
(10
5
cells mL
1
) when the cells grow in liq-
uid medium [69]. In addition, discoidin ex-
pression is repressed at high cell densities,
probably due to intracellular accumulation
of cAMP in stationary cells [69]. D. dis-
coideum mutant VI88, which is defective
in regulation of discoidin expression [70],
overproduces reporter genes cloned down-
stream of the disIc promoter from 10- to
100-fold, but produces endogenous discoi-
din at even lower cell densities than the
wild-type. Biotechnological production of
proteins using the disIc promoter is a de-
manding task: if fermentation occurs in
liquid culture, protein production is in-
duced at low amounts of biomass, but pro-
duction is repressed at high densities at
the end of batch fermentation. Thus, fed-
batch fermentation may be the method of
choice when working with this promoter.
A commonly used promoter active in
early development is derived from the rasD
gene. This promoter is promising at first
glance, because it is strictly turned off dur-
ing growth and can be rapidly induced by
a single dose of micromolar cAMP at
6 hours after the onset of starvation (i.e.,
suspending the cells in phosphate buffer).
However, the rasD promoter has a disad-
vantage for biotechnological application:
cAMP induction of rasD-mediated gene ex-
pression occurs only in 2040% of the cell
population, which are those cells differen-
tiating into prestalk cells [71]. Thus, induc-
tion of gene expression in this subpopula-
tion is averaged by a majority of non-re-
sponsive cells, leading to an overall weak
expression of the recombinant genes.
The csA gene encodes a cell contact pro-
tein that is essential for multicellular devel-
opment [72]. The csA gene is not expressed
during growth phase, and is strongly in-
duced when the D. discoideum cells aggre-
gate. Since the CsA protein is located on
the cell surface, it contains an authentic se-
cretion signal sequence. Aggregation can be
mimicked by shaking the cells at high den-
sity in a simple phosphate buffer. Thus, the
csA promoter and secretion signal can be
used as an authentic regulatory unit to ex-
press recombinant proteins after accumula-
tion of biomass, and to secrete the recombi-
nant protein into buffer for uncomplicated
purification [73].
Two D. discoideum promoters responsive
to stress conditions may be useful for the
expression of biopharmaceuticals. The pro-
moter from the gene apeA encoding apuri-
nic/apyrimidinic endonuclease responds
with a ca. 7-fold induction of transcription
after exposure to sublethal doses of bleomy-
cin, UV light, or X-rays [74]. The rnrB gene
encodes the catalytic subunit of ribonucleo-
tide reductase. DNA-damaging agents such
as methylmethane sulfonic acid and 4-nitro-
quilone 1-oxide rapidly induce the rnrB pro-
moter. However, the promoter has a rela-
tively high basal activity and the activators
are very toxic at the concentrations required
for full induction. No published experience
is available with either the apeA or rnrB pro-
moter for the large-scale expression of het-
erologous proteins.
5.3.1.2 Secretion Signals
For the large-scale production of biophar-
maceutical proteins by fermentation, it is
of advantage that the transgenic cells are
able to transfer the protein to the extracel-
lular space. This is of particular interest
for batch fermentation of D. discoideum
cells, since protein expression during de-
velopment would allow accumulation of
the recombinant protein into a simple buf-
fer instead of a complex medium. Contin-
uous fermentation of D. discoideum cells, a
promising alternative to batch fermenta-
tion, would also require secretion of the
protein product for purification. There are
two alternative secretion signal sequences
available for use in D. discoideum, and
both are derived from cell surface proteins
produced during multicellular develop-
ment of D. discoideum. The CsA protein is
a contact protein that is attached to the
plasma membrane via a glycosyl phospha-
tidylinositol (GPI) anchor [75]. The secre-
tion signal peptide of the CsA precursor
protein spans 16 amino acids. PsA is a gly-
coprotein which is naturally expressed on
the cell surface of prespore cells [76], and
is also membrane-bound via a GPI anchor.
The secretion signal is 19 amino acids
long; it has been shown experimentally
that the PsA leader is correctly cleaved
from upon transport of an artificial fusion
protein to the plasma membrane [77].
5.3.1.3 Vector Systems
Vectors for protein expression in D. dis-
coideum fall into two groups, as they either
integrate into the genome or remain extra-
chromosomal. The minimal requirements
of an integrating D. discoideum shuttle vec-
tor are: 1) replication signals and antibiotic
resistance genes for cloning in E. coli; 2) a
selection marker cassette (usually neo
R
) in-
cluding a D. discoideum promoter and a
transcription terminator; and 3) a D. dis-
coideum promoter/terminator unit sepa-
rated by a multiple cloning site for inser-
tion of the gene of interest. It has been
found that integrating shuttle vectors con-
taining a neo
R
selection marker amplify to
several hundred copies per cell. Fewer
plasmid copy numbers are required to es-
tablish neomycin resistance when the
act15 is linked to the neo
R
gene, compared
to the act6 promoter [61].
Multitudes of integrating shuttle vectors
have been developed by several laborato-
ries. For example, plasmid pVEII [78] con-
tains a neo
R
selection marker expressed
under the control of the act15 promoter,
and the disIc promoter upstream of a mul-
tiple cloning site and an act8 terminator
for expression of the transgene. This vec-
tor has been successfully used for produc-
tion of heterologous proteins (discussed
later). Vector pDcsA [73], which enables
protein expression during early develop-
ment, contains a neo
R
selection marker
and the csA promoter for expression of a
gene of interest. Protein production is in-
itiated by washing the cells into phosphate
buffer and incubating for 68 hours on a
rotary shaker.
A tetracycline-inducible expression sys-
tem based on an amber suppressor tRNA
gene has recently been described [79]. The
amber suppressor tRNA gene was placed
onto the same plasmid as the tetracycline
repressor on a neomycin-selectable plas-
mid. This vector, cotransformed with a lac-
Z(amber) reporter plasmid, allowed a 330-
fold induction of b-galactosidase activity in
the presence of tetracycline. This system
can be used to express heterologous trans-
genes equipped with a premature amber
codon. As a proof-of-principle, the D. dis-
coideum mhcA gene encoding myosin II
heavy chain was engineered to contain an
amber stop codon, and this gene was intro-
duced into an mhcA null strain. The mhcA
null cells have a very characteristic multi-
nucleated phenotype, and they are unable
to complete multicellular development.
This phenotype could be partially reversed
upon addition of tetracycline to the growth
medium of mhcA
/mhc(amber) transfor-
mants, which resulted in a 5- to 30-fold in-
duction of the MhcA protein [79]. All mu-
tants carrying multicopy suppressor tRNA
genes in combination with the tetracycline
repressor gene showed a low level of basal
activity in the absence of tetracycline. This
may reflect insufficient expression of TetR
and/or poor accessibility of TetR to the nu-
cleus. The latter can be solved by introduc-
ing a nuclear targeting sequence to TetR,
which results in an almost complete accu-
mulation of the modified TetR
NLS
in the
nucleus (T. W., unpublished results).
Blueprints for extrachromosomal vectors
are naturally occurring D. discoideum plas-
mids [80]. Extrachromosomal replication of
plasmid Ddp2 depends on its origin of
replication and a trans-acting factor REP
encoded on the Ddp2 plasmid [81]. The
origin of replication of Ddp2 can be cloned
onto a D. discoideum shuttle vector, and
will mediate extrachromosomal replication
of the plasmid, provided that REP is pres-
ent in the same cell. The rep gene can be
placed on a separate integrating plasmid
that is cotransformed with the extrachro-
mosomal vector [77, 81]. Alternatively, a
production strain can be created by insert-
ing a REP-expressing plasmid into the D.
discoideum genome [82].
The pMUW expression system [77] con-
sists of two plasmids. Plasmid 1 carries
the Ddp2 rep gene and a neo
R
selection
marker. Plasmid 2 does not contain a se-
lection cassette, but has the Ddp2 origin
of replication to keep the plasmid extra-
chromosomal; it also contains a cassette
for the expression of transgenes, consist-
ing of an act15 promoter, a psA leader se-
quence, and a multiple cloning site. Both
plasmids are introduced into D. discoideum
cells by cotransformation. Whereas the
first vector integrates and amplifies to
high copy numbers (ensuring expression
of REP), the second plasmid is extrachro-
mosomal at moderate copy numbers and
supports efficient production and secretion
of the recombinant protein.
Manstein et al. [82] generated a collec-
tion of plasmids that can be used both as
integrating and extrachromosomal vectors.
The latter requires a particular host strain
that expresses the REP protein. The pDXA
and pDXD series offer the act15 or disIc
promoter to regulate production of recom-
binant protein, polyhistidine tags and epi-
tope tags for determination of protein yield
and affinity purification, a Ddp2 origin for
extrachromosomal replication, and a neo
R
selection cassette. In the presence of the
REP protein, these vectors are extrachro-
mosomal at low copy numbers of <10 per
cell [82].
Plasmid pVTL2, which is another exam-
ple of an extrachromosomal plasmid [83],
was used to express the firefly luciferase
gene under the control of various promo-
ters. The luciferase gene can be replaced
for other heterologous transgenes, which
can be expressed from any D. discoideum
promoter introduced into the multiple clon-
ing site. The pVTL2 plasmid was reported
to produce 1050 copies per cell [83].
Blaauw et al. [84] adapted a state-of-the-
art Tet
OFF
system known from gene ex-
pression systems in mammalian cell lines
(see Part IV, Chapter 1). Plasmid 1 (re-
sponse plasmid) remains extrachromoso-
mal, and has a bsr
R
selection cassette and
a tetracycline-responsive minimal promo-
ter consisting of seven copies of the tetra-
cycline operator sequence (tetO) upstream
of an act15 minimal promoter; this is used
to regulate expression of the transgene.
Plasmid 2 (transactivator plasmid) is an in-
tegrating vector that contains a neo
R
selec-
tion cassette and the gene encoding tetra-
cycline-controlled transactivator (tTA
s
) ex-
pressed from the strong act15 promoter.
The tTA
s
protein expressed from plasmid
2 binds to the tetO sequences on plasmid
1 and activates transcription of the trans-
gene in the absence of tetracycline. No sig-
nificant activity of a reporter gene was de-
tected in the presence of low tetracycline
concentrations, and up to 3000-fold induc-
tion (calculated from near-zero background
expression to full expression in the ab-
sence of tetracycline) was observed after
removal of tetracycline from the medium
(i.e., washing the cells). In the fully in-
duced state, the tetO
7
/act15 promoter had
approximately the same activity as an un-
modified act15 promoter [84].
5.3.1.4 Strain Stability
Dictyostelium discoideum shuttle vectors in-
tegrate in tandem arrays of up to 500 co-
pies by a process called co-insertional rep-
lication [85]. Such complex vector loci may
be unstable due to recombination between
adjacent plasmid copies. In many cases, it
is possible to increase the production effi-
cacy of a cell population by transiently
raising the G418 concentration to up to
40 lg mL
1
, which is thought to enrich the
culture with transformants carrying very
high plasmid copy numbers.
It is frequently observed that expression
levels in good production strains cease
after a prolonged culture period. Although
there are no rules for maintaining stable
producer strains, the following precautions
should be taken:
Place the neo
R
selection cassette and
transgene expression cassette onto the
same plasmid.
Clone transformants and test individual
clones for optimal expression.
Prepare stocks of a selected producer
clone immediately.
Avoid accumulation of generations by
holding the cells in culture; rather, grow
cells for batch fermentation from a new
aliquot of stock.
Use high concentrations of antibiotic to
select for cells with high plasmid copy
numbers (e.g., 40 lg mL
1
G418) before
setting up a production culture. After-
wards, batch fermentation can be per-
formed in the absence of the costly anti-
biotic, without decreasing the protein
yield.
5.3.2
Codon Usage
When attempts are made to express het-
erologous proteins in D. discoideum, it is
often observed that high levels of trans-
gene mRNAs accumulate in the transfor-
mants, but very low protein yields are ob-
tained. This points to a special problem
that must be addressed for successful pro-
tein expression in D. discoideum: adapta-
tion of translation start sequences and co-
don frequencies in the transgene. The
average A+T content in the coding regions
of D. discoideum genes is close to 80%.
The third nucleotide position in most D.
discoideum codons is highly biased toward
A+T (85.1% in D. discoideum versus 41.5%
in humans). In D. discoideum, 10 codons
defining a particular amino acid are used
at 2.5% (Table 5.2), whereas the same co-
dons are used at 1038% in humans. In
contrast, for 12 amino acids and the stop
codons, the preferred codons in humans
are the least frequently used codons in D.
discoideum (Table 5.2). It is also clear from
Table 5.2 that, for example, three of six co-
dons for arginine are very rare codons in
D. discoideum, but are used at about 50%
frequency in humans. Particularly rare co-
dons in D. discoideum are CUG, UCG,
AGC, CCC, CCG, ACG, GCG, CGC, CGA,
CGG, GGG (see Table 5.2). By analyzing
these data, it becomes clear that adaptation
of codons in human cDNAs prepared for
expression in D. discoideum is demanding.
It is an open debate whether or not codon
bias serves a gene regulatory function, but
it has been observed in E. coli that most
Table 5.2 Comparison of D. discoideum and human
codon usage
Amino acid Codon Usage (%)
a)
D. discoideum H. sapiens
Phe F TTT 68.7 45.0
TTC 31.3 55.0
Leu L TTA 65.0 7.7
TTG 13.3 12.2
CTT 11.3 12.8
CTC 4.4 19.9
CTA 5.6 9.4
CTG 0.4 38.0
Ile I ATT 61.7 34.7
ATC 14.3 46.8
ATA 24.0 18.5
Met M ATG 100.0 100.0
Val V GTT 55.5 18.1
GTC 8.6 24.0
GTA 30.7 12.4
GTG 5.2 45.5
Ser S TCT 16.1 18.3
TCC 4.2 22.3
TCA 51.5 15.4
TCG 2.5 5.5
AGT 23.2 14.7
AGC 2.5 23.8
Pro P CCT 14.1 28.0
CCC 2.4 33.6
CCA 82.2 27.2
CCG 1.3 11.1
Thr T ACT 35.8 23.8
ACC 13.8 36.6
ACA 48.7 28.7
ACG 1.7 10.9
Tyr Y TAT 84.7 43.6
TAC 15.3 56.4
Ala A GCT 34.2 26.2
GCC 12.1 40.2
GCA 51.8 23.2
GCG 1.9 10.4
His H CAT 86.3 40.7
CAC 13.7 59.3
Gln Q CAA 96.5 27.3
CAG 3.5 72.7
Asn N AAT 89.7 45.4
AAC 10.3 54.6
Lys K AAA 83.7 43.7
AAG 16.3 56.3
Amino acid Codon Usage (%)
a)
D. discoideum H. sapiens
Asp D GAT 91.6 46.0
GAC 8.4 54.0
Glu G GAA 84.0 42.4
GAG 16.0 57.6
Cys C TGT 89.8 45.0
TGC 10.2 55.0
Trp W TGG 100.0 100.0
Arg R CGT 25.3 8.3
CGC 0.3 19.0
CGA 2.2 11.4
CGG 0.2 20.4
AGA 67.7 20.5
AGG 4.3 20.4
Gly G GGT 74.8 16.2
GGC 4.5 34.1
GGA 18.9 25.1
GGG 1.8 24.6
Stop TAA 89.6 21.4
TAG 5.5 17.4
TGA 4.9 61.2
a) Analyzed codons: D. discoideum: 1940951 codons
from 3371 coding regions; Homo sapiens:
28760742 codons from 67943 coding regions as
of November 2003 (for current codon usage ta-
bles, see http://www.kazusa.or.jp/codon/). Bold
numbers compare codons for an amino acid used
at 2.5% in D. discoideum with the corresponding
codon usage in humans. Values shown in italics
denote preferred codons for a particular amino
acid in humans that are the rarest codons for that
amino acid in D. discoideum.
low-usage codons are located within the
first 25 codons of most proteins. Thus, it
has been proposed that adaptation of the
first codons of a heterologous cDNA to the
codon usage of D. discoideum may greatly
improve expression levels [86]. This as-
sumption is supported experimentally by
expression of the biopharmaceutical hu-
man chorionic gonadotropin (hCG) in D.
discoideum cells [87]. Codon optimization
near the translation start and within the
first 1020 codons of the hCG cDNA had
the greatest effect on translation efficiency,
whereas codon optimization in more
downstream regions contributed less. In-
versely, adaptation of codon usage in
downstream regions of a cDNA had lim-
ited or no effect if the first codons were
left unchanged [87]. Thus, it seems that,
in terms of expression efficacy, initiation
of translation is more sensitive to optimal
D. discoideum codons than elongation.
The use of translation stop codons in D.
discoideum is about 90% UAA, 5% UAG,
and 5% UGA (Table 5.2). Reymond et al.
have noted that changing the translation
stop codon in a cDNA from UAG or UGA
to UAA may have a profound positive effect
on protein yield [88]. The polyadenylation
signal AAUAAA is present in many D. dis-
coideum genes in close vicinity to the trans-
lation stop codon, and in many cases it even
overlaps (AAUAAA). Thus, it may be of ad-
vantage to adapt not only the translation
stop codon itself, but also to introduce a
D. discoideum-like polyadenylation signal
overlapping the translation stop codon.
The Kozak model for the initiation of
translation in eukaryotes [89] states that
the ribosome scans along the 5' untrans-
lated region of an mRNA until it selects a
translation start codon to initiate transla-
tion. Selection of an AUG codon to start
translation requires a favorable environ-
ment, which can be described by the gen-
eral consensus sequence GCCGCC(A/G)
CCAUGG in vertebrates (start codon un-
derlined). It has been found that a purine
in position 3 (AUG is designated +1
through +3) is extremely conserved in all
eukaryotes, and that mutation in this posi-
tion will drastically reduce translation effi-
ciency. As long as the purine in the 3 po-
sition is present, however, mutations in
the other positions have only minor effects
on translation efficiencies. In the absence
of a purine in position 3, the G in the +4
position becomes essential [89].
We inspected a list of 4559 putative
ORFs from D. discoideum chromosomes 1
and 2 for nucleotide composition in the 6
through +4 positions relative to the trans-
lation start codons (Fig. 5.3). It transpired
that D. discoideum follows the Kozak rules
in principle, but with considerable differ-
ences. The A+T content of positions 1
through 6 is well above 85%, with nu-
cleotide A being strongly preferred in all
positions from 1 through 6 and in posi-
tion +4 (Fig. 5.3). Thus, a consensus D.
discoideum translation initiation site can be
described as AAAAAAAUG(A/G) (start co-
don underlined). In fact, 13% and 38% of
the 4559 genes analyzed show exactly the
sequence AAAAAAAUG(A/G) and
AAAAUG(A/G), respectively. Hence, it is
imperative to adapt the preferred, G+C-
rich vertebrate translation initiation sites
in order to allow for high-level expression
of heterologous proteins in D. discoideum.
5.3.3
Glycosylation
The cells of D. discoideum harbor the ma-
chinery for post-translational protein modi-
fications such as phosphorylation, acyla-
tion, formation of glycosyl phosphatidyl-
inositol anchors, and, in particular, O-
linked and N-linked glycosylation [90, 91].
In eukaryotes, oligosaccharides attached
to asparagine side chains (N-linked glycans)
fall into three main categories termed high
mannose, hybrid, or complex [92]. High
mannose-type glycans are composed of
mannose (Man) and N-acetylglucosamine
(GlcNAc) that have the general structure
Man
9
GlcNAc
2
with no further sugars at-
tached. Hybrid glycan chains contain addi-
tional GlcNAc and galactose molecules,
and bisecting GlcNAc moieties linked
b1,4 to the b-linked core mannose residue.
Complex N-linked glycans have additional
branches and sugar types, particularly fu-
cose, and often carry terminal sialic acid
moieties. The basic structure of N-linked
glycan chains in D. discoideum is of the high
mannose Man
9
GlcNAc
2
type [93]. Growing
cells and cells in early development have
only high-mannose glycan chains, as they
have low activities of processing a-mannosi-
dases (Fig. 5.4). By contrast, in differentiat-
ing cells a-mannosidases process high-man-
nose glycans down to Man
5
GlcNAc
2
or
Man
4
GlcNAc
2
(Fig. 5.4) [93]. The mannose
glycan chains can be modified with fucose,
sulfate, methyl phosphate [9396], and un-
usual bisecting GlcNAc residues linked
b1,4 to the mannose residue linked a16
to the b-linked core mannose [97] (Fig.
5.4). Yet D. discoideum glycoproteins com-
pletely lack galactose, N-acetylgalactosa-
mine and sialic acid typical for mammalian
complex N-linked glycan chains [93].
O-glycosylation of membrane-bound pro-
teins in D. discoideum involves the attach-
ment of a single GlcNAc to serine or threo-
nine side chains. Rules for determination of
amino acid sequences for O-glycosylation
have been derived experimentally [91] and
Fig. 5.3 Consensus translation start in Dictyoste-
lium discoideum. A total of 4559 predicted genes
from D. discoideum chromosomes 1 and 2 were
inspected for nucleotides in positions 6 to 1
and +4 relative to the ATG translation initiation
codon (designated +1 to +3). Nucleotide frequen-
cies are expressed as a fraction of 100%. At the
bottom of the graph the deduced consensus
translation site for D. discoideum is compared to a
consensus for translation start sites in vertebrates
according to Kozak [89]. [The data are derived
from the ongoing annotation of the D. discoideum
genome and appear courtesy of G. Glckner (IMB
Jena, Germany).]
by in silico methods [98]. It is not clear
whether the acceptor motifs for O-glycosyla-
tion in D. discoideum are similar to those in
mammals. In addition, there is no detect-
able attachment of GalNAc to membrane-
bound proteins as found in mammals [91].
O-linked GlcNAc glycosylation of mamma-
lian proteins expressed in D. discoideum
has been reported [91].
Unusual O-linked glycosylation occurs in
the cytoplasm of D. discoideumcells (for a re-
view, see Ref. [99]). Skp1 is a subunit of the
E3
SCF
-ubiquitin ligase, and carries an un-
usual pentasaccharide the presence of which
is required for transfer of Skp1 to the nu-
cleus of D. discoideum cells. Skp1 is first
modifiedby prolyl hydroxylation, after which
the hydroxyproline is sequentially modified
with a GlcNAc, Galactose, and Fucose to
forman O-linked Fuca1,2Galb1,3GlcNAc tri-
saccharide that is subsequently capped with
a Gala1,6Gal disaccharide. All enzymes in-
volved in these modifications are localized
in the cytoplasm rather than in the secretory
pathway of D. discoideum cells.
5.3.4
Fermentation
Dictyostelids grow by feeding on bacteria or
yeasts. Early methods to isolate and grow
Dictyostelid species were based on boiled
Fig. 5.4 Glycosylation patterns in Dictyostelium dis-
coideum. Depicted are the structures of N-linked
glycan chains during growth/early development
and late development (differentiation) of D. discoi-
deum cells. N-linked oligosaccharides can be
modified with sulfate (SO
4
), methyl phosphate (P-
Me), fucose, and intersecting N-acetylglucosa-
mine. Asn, asparagine; MI and MII, a-mannosi-
dases I and II. (Modified from Ref. [93].)
dung from horse or rabbit as sources for
naturally occurring food bacteria. Later,
more defined media were chosen to prepare
agar plates (e.g., mannite agar, hay-infusion
agar), and isolated Dictyostelid species were
cultured on defined bacteria such as Vibrio
alkaligenes or Escherichia coli [100]. In the
laboratory, Dictyostelids can be readily
grown in association with bacteria, either
on agar plates or in submerged cultures.
However, large quantities of bacteria-free
amoebae for biochemical studies are diffi-
cult to obtain from such cultures. The ma-
jor breakthrough in solving this problem
was made during the late 1970s, when Suss-
man and Sussman announced the isolation
of an axenically growing derivative of D. dis-
coideum isolate NC-4 [101]. Strain AX-1 was
isolated by inoculating spores of D. discoide-
um NC-4 into a complex liquid medium
(CF
3
) containing phosphate buffer, peptone,
yeast extract, and glucose, enriched with liv-
er extract and fetal calf serum [101]. At-
tempts to simplify the CF
3
medium culmi-
nated in the introduction of HL-5 [102],
which is still the medium of choice in Dic-
tyostelium research and biotechnology. HL-
5 contains yeast extract and proteose pep-
tone in a phosphate buffer, supplemented
with glucose. In 1970, Watts and Ashworth
introduced strain AX-2, a derivative of AX-1,
which had adapted optimal growth condi-
tions in HL-5 medium (growth temperature
22238C, generation time 89 hours) [103].
In parallel, Loomis [104] produced another
axenic derivative of D. discoideum NC-4, A3
(also known as A-3 and AX3), by chemical
mutagenesis.
The original HL-5 medium was slightly
modified in several laboratories. For exam-
ple, HL-5C contains casein peptone and
bactotryptone and a slightly different phos-
phate buffer composition [105]. In 1977,
Franke and Kessin [106] published the de-
velopment of a defined minimal medium
that was devoid of yeast extract and pro-
teose peptone. The medium is prepared
from stock solutions of vitamins, amino
acids, and trace metals on the basis of a
phosphate buffer. Glucose is added as car-
bon source. The D. discoideum cells grown
in HL-5 medium require some time to
adapt to the FM medium, but then reach
generation times up to 10 hours. FM is
the medium of choice for cultivation of D.
discoideum cells in fermenters.
In conventional shake-flask cultivation,
D. discoideum cells can reach maximum
cell densities of 1210
7
cells mL
1
in HL-
5 and up to 310
7
cells mL
1
in FM medi-
um. A small-scale industrial facility can
easily increase this number to 510
12
cells
(5 kg) per week. Improvement of the origi-
nal FM medium to compensate limitations
with respect to amino acids (SIH medium)
increased the densities of D. discoideum
cultures to 5610
7
cells mL
1
[105].
Growth of D. discoideum cells in bioreac-
tors in batch and fed-batch mode in a
stirred tank-type bioreactor is a convenient
fermentation method [107]. Under these
conditions, it is possible to accumulate
*36 g cell material (dry weight) from 7 L
of cell culture in about 4 days (E. Flaschel,
personal communication).
Beshay et al. reported conditions for the
short-term continuous cultivation of D. dis-
coideum cells on porous supports (SIRAN
R
beads) in HL-5C medium [108]. D. dis-
coideum cells actively colonized the porous
carrier (Fig. 5.5), after which the colonized
beads can be freely suspended in medium.
Cell densities of free amoebae remained at
about 10
5
per mL for most of the cultiva-
tion time, whereas the cell density on the
SIRAN
R
beads reached up to 10
8
per mL
and remained constant for at least 16 days
of fermentation [108, 109]. By using bro-
ken pumice or CeramTec
R
(a ceramic cata-
lyst support), the immobilized cells reach
very high densities of about 23.510
8
cells mL
1
with respect to the pore vol-
ume. Such cell densities are considerably
higher than the maximal cell densities ob-
tained in suspension culture in HL-5C (1
210
7
cells mL
1
) or in SIH medium (5
6 10
7
cells mL
1
). They are also signifi-
cantly higher than the cell densities ob-
tained so far by colonization of SIRAN
R
and ImmobaSil
R
carriers [110].
5.3.5
Examples of Heterologous Protein
Expression
There are several examples of successful
expression of heterologous proteins in D.
discoideum (for a summary, see Table 5.3).
Some particularly interesting examples are
presented below.
Plasmodium falciparum circumsporozoite
protein (CSP) is an important potential
component of a vaccine against malaria. It
is obvious that P. falciparum should be ex-
pressed in D. discoideum, as the two organ-
isms have quite similar A+T contents of
their genomes (total genome: 78% in D.
discoideum and 81% in P. falciparum, re-
spectively; third codon A+T content 85.1%
in D. discoideum versus 82.8% in P. falci-
parum). The production of CSP in other
systems is limited by low expression rates
and rapid degradation. Fasel et al. [111] ex-
pressed CSP in D. discoideum under the
control of the disIg and rasD promoter. In
all cases the authentic P. falciparum CSP
leader peptide, which was not functional
in D. discoideum, was replaced by the D.
discoideum CsA leader sequence. The CSP
produced with these vectors was mainly
cytoplasmic, and only a small amount of
protein was secreted. It was then found
that the last 23 amino acids of CSP were
responsible for intracellular retention of
the recombinant protein. Removal of this
carboxy-terminal peptide from CSP re-
sulted in efficient secretion of soluble
CSP, whereas its exchange for the D. dis-
coideum carboxy-terminal GPI anchor sig-
nal sequence of D. discoideum CsA re-
sulted in presentation of CSP on the cell
surface of D. discoideum transformants
[88]. Importantly, whole D. discoideum cells
expressing P. falciparum CSP on their cell
surface, when injected into BALB/c mice,
induced an immune response, with the re-
Fig. 5.5 Scanning electron micrographs of Dictyostelium discoi-
deum cells inside SIRAN
beads after continuous bioreactor

operation for 18 days [132]. (Photographs kindly provided by
E. Flaschel, Bielefeld University)
covered antibodies reacting against P. falci-
parum sporozoites. Taking P. falciparum
CSP expression in D. discoideum a step
further, van Bemmelen et al. generated fu-
sions of D. discoideum discoidin I and CSP
[112]. The fusion protein was expressed to
high levels, albeit in the cytoplasmic frac-
tion of D. discoideum cells. However, these
authors took advantage of the lectin-bind-
ing activity of discoidin to efficiently purify
the fusion protein from cell extracts. The
fusion protein was bound to Sepharose-4B
and eluted under mild conditions with ga-
lactose. The affinity-purified CSP was re-
moved from the discoidin tag by thrombin
cleavage and used for immunization ex-
periments [112].
Attempts to produce complex human gly-
coproteins in D. discoideum are exemplified
by antithrombin III (ATIII) production
[113]. This protein has four N-linked poly-
glycan chains, accounting for 9% of its mo-
lecular weight (see Part IV, Chapter 11). In-
terestingly, the authentic (human) signal
peptide of ATIII (32 amino acids) was active
in D. discoideum, allowing the recombinant
protein to accumulate in the cell superna-
tant. The recombinant ATIII was glycosy-
Table 5.3 Heterologous proteins successfully expressed in D. discoideum.
Protein Source Promoter Yield Reference(s)
CSP Plasmodium falciparum disIc, rasD 88, 111
CSP Plasmodium falciparum act6 0.150.3 mg L
1
112
CSP C-terminal fragment Plasmodium falciparum act6 1.23.0 mg L
1
112
CSP C-terminal fragment Plasmodium yoellii act6 0.080.12 mg L
1
112
MSP I fragment Plasmodium vivax act6 0.40.6 mg L
1
112
glycoprotein VP7 Rotavirus SA11 act15 118, 119
hIGFBP6 Homo sapiens act15, disIc 120
bBTC Bos taurus act15 120
GST Schistosoma japonicum act15 17 mg L
1
77, 91
GUS Escherichia coli act15 77
Muscarinic receptor M3 Rattus norvegicus disIc 121
Muscarinic receptor M2 Homo sapiens disIc, rasD 115, 116
Antithrombin III Homo sapiens act6 1 mg L
1
113, 114
sFasL Homo sapiens act15 148205 lg L
1
122
hCG Homo sapiens act15 4 lg L
1
87, 123
FSH Homo sapiens act15 124
Calcium pump HPMCA4b Homo sapiens act15 125
Na
+
, K
+
-ATPase a
1
and b
1
subunits
Gallus gallus act15 126
Aquaporin RD28 Arabidopsis thaliana act15, cotB 127
Glucose transporter GLUT1Rattus norvegicus act15 128
Chimeric myosins Chara corallina act15 129
Profilin Zea mays act15 130
CSP, circumsporozoide protein; bBTC, bovine b-cellulin; FSH,
human follicle-stimulating hormone; GST, glutathione-S-transfer-
ase; GUS, b-glucuronidase; hCG, human chorionic gonadotropin;
hIGFBP6, Insulin-like growth factor-binding protein 6; sFasL,
soluble form of human Fas ligand.
lated but had a lower molecular weight than
authentic ATIII, suggesting pronounced dif-
ferences in glycosylation patterns. The
ATIII protein produced in D. discoideum
was active that is, it showed progressive in-
hibitor activity. However, it displayed very
weak binding affinity to heparin, which
was probably due to false glycosylation
[113]. ATIII was among the first recombi-
nant proteins to be efficiently secreted by
D. discoideum cells; thus, it was straightfor-
ward to set up conditions to produce recom-
binant ATIII in continuous culture. D. dis-
coideum cells were immobilized on a porous
inorganic matrix, and cultured in a packed-
bed fermenter system in HL-5 medium. The
D. discoideum cells stably produced about
500 ng mL
1
of ATIII in the efflux in 1 hour,
with the system remaining stable for at least
10 days [114].
As mentioned previously, GPCRs have
been successfully expressed in D. discoide-
um. The idea was to generate transgenic
D. discoideum strains expressing G protein-
coupled receptors as convenient, easy-to-
handle, and cheap alternatives to expensive
pharmacological testing of receptor ligands
in mammalian cell culture or organ models
(see Part VII, Chapter 3). To do so, the cDNA
encoding the human muscarinic acetylcho-
line receptor M2 (hm2) was cloned down-
stream of the D. discoideum csA leader se-
quence. In addition, codons 3, 5, and 7 of
hm2 were adapted to the D. discoideum co-
don usage. It was found that expression of
functional receptors on the cell surface re-
quired the csA leader [115, 116]. Functional
M2 receptors were inserted into the plasma
membrane and bound ligand with a some-
what lower affinity but similar ligand spec-
ificity as M2 receptors expressed in mamma-
lian cell culture. About 3000 functional M2
receptor molecules were displayed per cell.
An attempt was then made to connect the
functional human M2 receptors to the intra-
cellular signaling cascades of D. discoideum.
Unfortunately, it was not possible to link
these receptors to D. discoideum G proteins
and downstream reporter genes, nor was it
possible to link human G proteins coex-
pressed with M2 receptors to downstream
D. discoideum effector enzymes such as ade-
nylyl cyclase (T. W., unpublished results).
Schistosoma species are trematode para-
sites causing schistosomiasis. The sj26 gene
encoding glutathione-S-transferase (GST)
from S. japonicum was expressed in D. dis-
coideum using the act15 promoter and the
psA leader sequence for secretion of the pro-
tein product [77]. Schistosoma has a moder-
ately biased A+T content (64% of whole ge-
nome; 71.3% A+T in the third codon posi-
tion). Nonetheless, about 9% of the codons
in the sj26 gene are rare codons in D. dis-
coideum. The GST protein was very effi-
ciently secreted; no protein was detected in-
side the cell. Up to 1 mg L
1
of GST was
found in the cell supernatant, representing
about 1% of total secreted protein.
Dittrich et al. [77] performed an interest-
ing comparison of translation efficiencies
as a function of the number of rare co-
dons. The D. discoideum psA gene has
1.4% rare codons, whilst the S. japonicum
sj26 and E. coli gus gene encoding b-glucu-
ronidase contain 8.7 and 20.8% rare co-
dons, respectively. These authors found
that D. discoideum PsA was expressed at
20-fold higher levels than GST, whereas
GUS protein was barely detectable [77].
This result may argue that adaptation of
rare codons throughout an entire cDNA
may have great effects on protein yield.
Taking this a step further, one could pro-
pose the re-synthesis of a gene in vitro
considering the D. discoideum codon
usage, though the results of such an effort
would be unpredictable. As documented
by expression experiments with hCG, opti-
mizing the first 17 codons increased trans-
lation efficiency 4- to 5-fold, and further
codon adaptation in more downstream re-
gions of the hCG cDNA did not provide
additional benefit [87]. In our laboratory,
the gene encoding proteinase K from the
fungus Tritirachium album was entirely
synthesized de novo with the D. discoide-
um-typical A+T content. Expression experi-
ments directly comparing the synthetic
proteinase K gene with its original did not
reveal any difference in produced protein
levels (T. D., unpublished results).
An interesting example for the biotech-
nological application of D. discoideum in
structure biology is presented by Cubeddu
et al. [117]. This study was aimed at the
isotopic labeling of recombinant glycopro-
teins for NMR studies. Food bacteria were
grown in medium containing [
15
N]NH
4
Cl
and [
13
C]glycerol, and a D. discoideum
strain expressing the PsA protein was fed
on the isotopically labeled bacteria. The ex-
pressed protein incorporated 99.9% of iso-
topic label and was good for high-quality
NMR spectra after single-step purification
from cell supernatants.
5.4
Dictyostelium discoideum
in Biomedical Research
The availability of powerful molecular ge-
netics of D. discoideum suggests that this
organism may be useful for addressing
fundamental questions relevant for human
health and disease. We will briefly reflect
three very different examples of the poten-
tial input that D. discoideum can provide in
biomedical research, namely the treatment
of infections caused by opportunistic, in-
tracellularly replicating human pathogens,
the in vivo testing of potential embryotoxic-
ity of drugs under development by the
pharmaceutical industry, and the potential
improvement of current retroviral gene
therapy vectors by studying D. discoideum
retrotransposable elements.
5.4.1
Dictyostelium discoideum
as a Model for Microbial Pathogenesis
In their natural habitats, many intracellu-
larly replicating human pathogens of bac-
terial or fungal origin use protozoans such
as Acanthamoeba castellanii as replication
hosts. Since A. castellanii is similar to D.
discoideum in many aspects, it was tempt-
ing to investigate whether D. discoideum
would be a similarly suitable host to sup-
port replication of such pathogens. If so,
the powerful molecular genetics of D. dis-
coideum might boost the investigation of
hostpathogen interactions.
Legionellae are Gram-negative bacilli in
freshwater that replicate intracellularly in
protozoa such as Hartmanella vermiformis,
Tetrahymena thermophila, and Acanthamoe-
ba castellanii. Infection of humans with Le-
gionella pneumophila can occur after inhala-
tion of aerosol containing the pathogen,
which then infects alveolar macrophages
and causes severe pneumonia (Legion-
naires disease). D. discoideum cells are sus-
ceptible to infection with L. pneumophila,
and intracellular growth of L. pneumophila
in D. discoideum cells is quite similar to
growth of the bacteria in macrophages
(Fig. 5.6), as the pathogen grows in intra-
cellular vacuoles that are associated with
rough endoplasmic reticulum [131, 132].
D. discoideum cells can feed only on non-
virulent strains of L. pneumophila, which
are rapidly degraded and killed by the
amoebae. By contrast, D. discoideum cells
are themselves being killed by virulent
strains of L. pneumophila by a combination
of cytotoxicity and continued intracellular
growth, which supports the hypothesis
that virulence of legionellae depends on
their ability to escape lysosomal degrada-
tion [132]. Growth of L. pneumophila in D.
discoideum depends on Legionella gene
products known to be required for viru-
lence in macrophages [132], suggesting
that genes and gene-regulatory networks
involved in intracellular maintenance and
replication of the pathogen are conserved
in both hosts. Genes from D. discoideum
required for intracellular replication of L.
pneumophila are now being explored. For
example, it has been found that D. dis-
coideum mutants defective in components
of the cytoskeleton, such as coronin, un-
conventional myosins, and profilin, are
more permissive to intracellular replication
of legionellae than wild-type D. discoideum
cells [131, 132].
Pseudomonas aeruginosa is a Gram-nega-
tive opportunistic bacterial pathogen that
causes life-threatening infections in immu-
nocompromised patients. Since P. aerugi-
nosa strains are notoriously resistant to
one or more antibiotics, alternative strate-
gies for fighting P. aeruginosa infections
are desirable. Thus, genetically accessible
models to study P. aeruginosa virulence are
under development. Two laboratories have
shown independently that D. discoideum
cells can be infected with P. aeruginosa
[133, 134]. The ability of P. aeruginosa to
infect and kill D. discoideum cells de-
pended on P. aeruginosa genes known to
be involved in virulence, and inactivation
of such genes changed D. discoideum cells
from potential infection hosts to predators
of the avirulent bacteria [133]. It has been
shown that the D. discoideum model can
robustly predict the virulence of P. aerugi-
nosa strains in a rodent model of acute
pneumonia. This nicely demonstrates that
the D. discoideum system cannot only be
used to dissect virulence mechanisms at
the molecular level, but also to test the
pathogenicity in the run-up to testing in
animals [134].
Cryptococcus neoformans is a yeast-like
fungus that can cause subacute or chronic
meningoencephalitis (cryptococcosis) in in-
dividuals with impaired immunity. C. neo-
formans can infect D. discoideum, and the
mechanisms of interaction between D. dis-
coideum and the pathogen seem to be sim-
ilar to interactions of C. neoformans with
predators of their natural environment
(Acanthamoeba) or human macrophages
[135]. D. discoideum cells are readily in-
fected and killed by virulent strains of C.
neoformans, whereas avirulent C. neofor-
mans strains are ingested and killed by D.
discoideum amoebae [135]. Although inacti-
vation of myosin VII is known to impair
phagocytosis [136], D. discoideum mutants
Fig. 5.6 Transmission electron micrograph of a
Dictyostelium discoideum cell infected with Legionel-
la pneumophila. The bacteria replicate within a sin-
gle vacuole. Scale bar = 1 lm. (Reproduced from
Ref. [133], with permission from Blackwell Science
Ltd.)
defective in myosin VII are more permis-
sive to C. neoformans infection than wild-
type cells. Thus, additional mechanisms
seem to be involved in parasitehost inter-
action that can be nicely explored using
the molecular genetics of D. discoideum.
Importantly, the virulence of C. neoformans
was significantly enhanced after passage
on D. discoideum, supporting the hypothe-
sis that predatorbait interactions may be
important to maintain C. neoformans viru-
lence factors in the wild [135].
The examples described above are prom-
ising for the development of new biophar-
maceuticals to treat important human in-
fections. It can be assumed that the num-
ber of intracellularly replicating pathogens
being tested in D. discoideum will increase.
For example, D. discoideum is also a model
for the study of pathogenicity mechanisms
of Mycobacterium and Salmonella [137,
138]. A major limitation of the D. discoide-
um model is that many pathogens grow
poorly at 20258C, the favored growth
temperature of D. discoideum cells. If a giv-
en pathogen is able to infect and multiply
in D. discoideum, however, the powerful
molecular genetics of D. discoideum will
certainly allow the exploration of host
pathogen interactions at the level of the in-
volved genes. This can be achieved by test-
ing the growth of pathogens on D. dis-
coideum mutants defective in known genes
(the collection of which in laboratories
around the world is huge), or by selecting
mutants out of REMI libraries that do not
support the growth of pathogens.
5.4.2
Use of D. discoideum to Test
for Potential Embryotoxicity in Humans
In previous sections we have emphasized
the power of reporter gene expression under
the control of native D. discoideum promo-
ters to analyze temporal and spatial expres-
sion patterns of genes in the multicellular
development of D. discoideum. In a research
project funded by the European Union
(EU), it was investigated whether certain
pharmaceuticals that are known teratogens
in humans would also show effects on the
development of non-mammalian models,
such as frog eggs, sea urchin embryos, flies,
worms, and zebrafish, or in lower eukar-
yotes such as yeasts and D. discoideum.
The rationale of designing an embryo-
toxicity assay based on D. discoideum devel-
opment was that transgenic D. discoideum
cells expressing b-galactosidase under the
control of developmentally regulated pro-
moters would respond to the exposure to
chemical compounds, with the effects
being measured as alterations in the ex-
pression of the b-galactoside reporters due
to delays in D. discoideum development.
One teratogen tested in all model organ-
isms of the integrated EU-funded project
was valproic acid (VPA). VPA drastically
delayed multicellular development of D.
discoideum cells in a concentration-depen-
dent manner (Fig. 5.7). When D. discoide-
um cells expressing b-galactoside under
the control of prespore or prestalk-specific
promoters were exposed to VPA, expres-
sion of the reporter gene was also sup-
pressed [139, 140]. Thus, a simple colori-
metric test could be used to determine the
state of development after exposure of D.
discoideum cells to chemical compounds
that affected development. It was shown
that the D. discoideum embryotoxicity assay
could reproduce stereoselective activities of
VPA analogues previously characterized in
animal models [140], and the assay could
in principle be used to make predictions
on the potential embryotoxicity of drugs
not yet tested in animals [141].
In addition to being valuable as a cell
system for the pre-screening of future bio-
pharmaceuticals for potential embryotoxici-
ty in humans, the power of D. discoideum
genetics also allows us to explore molecu-
lar targets of embryotoxic compounds. Giv-
en that the complete proteome of D. dis-
coideum is known, the identification of
genes that confer resistance to certain em-
bryotoxic drugs may provide hints of
orthologous targets of these compounds in
human cells, and thus of mechanisms of
action. This is nicely exemplified by a re-
cent study by Williams and colleagues,
who performed a random mutagenesis of
the D. discoideum genome by REMI, and
then screened the transformants for resis-
tance of D. discoideum development to
lithium. Lithium is used in humans for
mood-stabilizing treatment of bipolar af-
fective disorder. It has been hypothesized
that lithium acts by depleting the pool of
inositol-1,4,5-trisphosphate (IP
3
) in neu-
rons, thereby disturbing intracellular sig-
naling cascades. In D. discoideum, lithium
blocks development by depleting the intra-
cellular IP
3
pool [142]. A mutant defective
in the enzyme prolyl oligopeptidase was
resistant to the effects of lithium and had
elevated IP
3
levels. The mutant was cross-
resistant to VPA, suggesting a common
molecular mechanism of action of both
drugs in D. discoideum and in humans
[143].
Fig. 5.7 Effect of valproic acid on Dictyostelium
discoideum development. D. discoideum cells were
incubated on phosphate-buffered agar plates as
detailed recently [139]. Valproic acid was
dispersed in the agar at 0.5 mM (b), 1.5 mM (c)
and 2.0 mM final concentration (d). The control is
without valproic acid (a). Pictures were taken
24 hours after plating. Scale bar = 0.5 mm.
5.4.3
Dictyostelium discoideum Retrotransposons
and Gene Therapy
Somatic gene transfer (gene therapy) is an
experimental strategy of molecular medi-
cine that is aimed at introducing into pa-
tients cells an additional, intact copy of a
gene whose function had been lost by mu-
tation. Lasting even life-long expres-
sion of the therapeutic gene is particularly
desirable if monogenic diseases are being
treated. It has been found in many pre-
clinical gene therapy experiments that
long-lasting expression of foreign genes in
human cells requires insertion of the re-
combinant DNA into the treated cells ge-
nomes. Thus, gene therapy vectors based
on retroviruses are the preferred vectors in
gene therapy (see Part I, Chapters 6 and
7). However, a major disadvantage of retro-
viral vectors is that they integrate ran-
domly, or even have a bias for transcribed
genes [144]. Until recently, it was generally
accepted that the risk of causing insertion
mutagenesis of critical genes by the use of
retrovirus-based gene therapy vector was
negligible. However, a recent report of leu-
kemia-like T-cell proliferation in boys
treated for X-linked severe immunodefi-
ciency [145] has stirred the debate, and the
need for specifically integrating gene ther-
apy vectors was demanded [146].
There are several ideas of how the inte-
gration specificity of gene therapy vectors
could be improved. Thinking of a gene ther-
apy vector as a piece of mobile DNA, it is
not a long way to draw parallels to natural
mobile elements and their often profound
integration preferences for genomic regions
where they can stay without causing harm
to their host. Many mobile elements recog-
nize certain places of their host genome by
specific interaction of the transposase (the
enzyme that catalyzes integration) with
DNA sequences at the integration site. On
the other hand, transposases may select in-
tegration sites by interacting with chroma-
tin proteins or transcription factors asso-
ciated with certain gene loci. In D. discoide-
um, a family of retrotransposons (mobile
elements that amplify via an RNA inter-
mediate) shows strong integration prefer-
ence within a window of ca. 100 bp up-
stream and downstream of tRNA genes
(for a review, see Ref. [147]). Since D. dis-
coideum tRNA genes are scattered on all
chromosomes, the tRNA gene-targeted ret-
rotransposons have permitted an ideal
means of colonizing the entire D. discoide-
um genome. Moreover, they do not cause in-
sertion mutations because the close vicinity
of tRNA genes is devoid of other genes. If
we could uncover the molecular mecha-
nism that allows the retrotransposons to se-
lect tRNA genes for integration, it may be
possible to adapt this to create a new gen-
eration of tRNA gene-targeting vectors that
would circumvent the safety problems asso-
ciated with currently used gene therapy vec-
tors. At present, the first steps towards an
understanding of target selection by tRNA
gene-targeted retrotransposons in D. dis-
coideum cells are being made [148]. Conse-
quently, it will be very interesting to see
whether research with D. discoideum will
contribute further to this exciting field of
biomedical research, and also assist in the
development of modern biopharmaceuti-
cals.
5.5
Conclusions
Dictyostelium discoideum is a microbial or-
ganism that can, within its limits, signifi-
cantly contribute to biopharmaceutical and
biomedical research. As a true eukaryote,
D. discoideum can offer critical properties
5.5 Conclusions 689
required for the expression of biopharma-
ceutical proteins. The major advantages of
D. discoideum as an expression host are
the short generation time, the use of
cheap media and low growth temperature,
fermentation techniques that allow the ac-
cumulation of large amounts of cell mass
in a short period of time, efficient secre-
tion of produced proteins into buffer, the
absence of human pathogens in the cul-
tures, and the ability to store transgenic D.
discoideum spores almost indefinitely with-
out loss in viability. As with every expres-
sion system, D. discoideum also has its pit-
falls. Expression levels are rather low and
unpredictable, the codon usage is exotic,
there are no efficient activatable promo-
ters, glycosylation patterns of human re-
combinant proteins are similar, but not
authentic, and D. discoideum is not yet an
FDA-approved production species. How-
ever, experience with this expression sys-
tem is ever-growing, and D. discoideum
may be well-suited to solve specific expres-
sion problems that cannot be addressed in
other organisms. Important problems such
as the evaluation of pathogenhost interac-
tions and safety problems in gene therapy
can be addressed with the molecular ge-
netics tools to work with, and a completely
sequenced genome in the background. Al-
together, we have shown that D. discoideum
holds great promise as a model for ad-
dressing important questions in biomedi-
cal research and in the development of
Acknowledgments
The studies conducted in the authors lab-
oratory were supported by grants from the
Deutsche Forschungsgemeinschaft (DFG).
References
1 O. Brefeld, Abhandlungen der Senckenbergi-
schen Naturforschenden Gesellschaft Frank-
furt 1869, 7, 85107.
2 K. B. Raper, The Dictyostelids, Princeton Uni-
versity Press, USA, 1984.
3 S. L. Baldauf, A. J. Roger, I. Wenk-Siefert, W. F.
Doolittle, Science 2000, 290, 972977.
4 K. B. Raper, J. Agr. Res. 1935, 50, 135147.
5 J. Tillner, H. Nau, T. Winckler, T. Dinger-
mann, Toxicol. in Vitro 1998, 12, 463469.
6 K. Dannat, J. Tillner, T. Winckler, M. Weiss,
K. Eger, T. Dingermann, Pharmazie 2003, 58,
204210.
7 R. S. B. Williams, M. Eames, W. J. Ryves, J.
Viggars, A. J. Harwood, EMBO J. 1999, 18,
27342745.
8 R. S. B. Williams, L. L. Cheng, A. W. Mudge,
A. J. Harwood, Nature 2002, 417, 292295.
9 W. F. Loomis, D. Welker, J. Hughes, D. Ma-
ghakian, A. Kuspa, Genetics 1995, 141, 147
157.
10 R. Sucgang, G. Chen, W. Liu, R. Lindsay, J.
Lu, D. Muzny, G. Shaulsky, W. Loomis, A.
Kuspa, Nucleic Acids Res. 2003, 31, 23612368.
11 G. Glckner, L. Eichinger, K. Szafranski, J. Pa-
chebat, P. Dear, R. Lehmann, C. Baumgart,
J. F. Abril, G. Parra, R. Guig, et al., Nature
2002, 418, 7985.
12 L. Eichinger, A. A. Noegel, EMBO J. 2003, 22,
19411946.
13 L. Eichinger, Curr. Biol. 2003, 44, 5972.
14 E. C. Cox, C. D. Vocke, S. Walter, K. Y. Gregg,
E. S. Bain, Proc. Natl. Acad. Sci. USA 1990, 87,
82478251.
15 A. Kuspa, W. F. Loomis, Proc. Natl. Acad. Sci.
USA 1996, 93, 55625566.
16 G. Glckner, K. Szafranski, T. Winckler, T.
Dingermann, M. Quail, E. Cox, L. Eichinger,
A. A. Noegel, A. Rosenthal, Genome Res. 2001,
11, 585594.
17 D. A. Ratner, T. E. Ward, A. Jacobson (1981).
In: Developmental biology using purified genes,
pp. 595605, Academic Press, USA, 1981.
18 S. L. Barclay, E. Meller, Mol. Cell. Biol. 1983, 3,
21172130.
19 K. P. Hirth, C. A. Edwards, R. A. Firtel, Proc.
Natl. Acad. Sci. USA 1982, 79, 73567360.
20 W. Nellen, C. Silan, R. A. Firtel, Mol. Cell. Biol.
1984, 4, 28902898.
21 D. A. Knecht, S. M. Cohen, W. F. Loomis, H. F.
Lodish, Mol. Cell. Biol. 1986, 6, 39733983.
22 P. K. Howard, K. G. Ahern, R. A. Firtel, Nucleic
Acids Res. 1988, 16, 26132623.
23 D. Knecht, K. M. Pang, Methods Mol. Biol.
1995, 47, 321330.
24 K. M. Pang, M. A. Lynes, D. A. Knecht, Plasmid
1999, 41, 187197.
25 D. Kalpaxis, H. Werner, E. Boy-Marcotte, M.
Jacquet, T. Dingermann, Dev. Genet. 1990, 11,
396402.
26 J. L. Dynes, R. A. Firtel, Proc. Natl. Acad. Sci.
USA 1989, 86, 79667970.
27 W. Nellen, R. A. Firtel, Gene 1985, 39, 155
163.
28 C. Barth, D. J. Fraser, P. R. Fisher, Plasmid
1998, 39, 141153.
29 T. T. Egelhoff, S. S. Brown, D. J. Manstein, J. A.
Spudich, Mol. Cell. Biol. 1989, 9, 19651968.
30 B. Leiting, A. A. Noegel, Biochem. Biophys. Res.
Commun. 1991, 180, 14031407.
31 K. Sutoh, Plasmid 1993, 30, 150154.
32 H. Adachi, T. Hasebe, K. Yoshinaga, T. Ohta,
K. Sutoh, Biochem. Biophys. Res. Commun.
1994, 205, 18081814.
33 W. F. Loomis, Exp. Cell Res. 1971, 64, 484486.
34 W. F. Loomis. Genetic tools for Dictyostelium
discoideum. In: Methods in Cell Biology, J. A.
Spudich, Ed. pp. 3165. Orlando FL: Aca-
demic Press, 1987.
35 A. De Lozanne, J. A. Spudich, Science 1987,
236, 10861091.
36 W. Witke, W. Nellen, A. Noegel, EMBO J.
1987, 6, 41434148.
37 D. J. Manstein, M. A. Titus, A. De Lozanne,
J. A. Spudich, EMBO J. 1989, 8, 923932.
38 T. Dingermann, N. Reindl, T. Brechner, H.
Werner, K. Nerke, Dev. Genet. 1990, 11, 410
417.
39 A. Morrison, R. Marschalek, T. Dingermann,
A. J. Harwood, Gene 1997, 202, 171176.
40 A. Kuspa, W. F. Loomis, Proc. Natl. Acad. Sci.
USA 1992, 89, 88038807.
41 W. T. Chang, J. D. Gross, P. C. Newell, Plasmid
1995, 34, 175183.
42 P. Fey, E. C. Cox, Curr. Biol. 1997, 7, 909912.
43 T. Winckler, C. Trautwein, C. Tschepke, C.
Neuhuser, I. Zndorf, P. Beck, G. Vogel, T.
Dingermann, J. Mol. Biol. 2001, 305, 703714.
44 T. Dingermann, H. Werner, A. Schutz, I. Zn-
dorf, K. Nerke, D. Knecht, R. Marschalek,
Mol. Cell. Biol. 1992, 12, 40384045.
45 M. Hildebrandt, W. Nellen, Cell 1992, 69, 197
204.
46 H. Martens, J. Novotny, J. Oberstrass, T. L.
Steck, P. Postlethwait, W. Nellen, Mol. Biol.
Cell 2002, 13, 445453.
47 T. E. Crowley, W. Nellen, R. H. Gomer, R. A.
Firtel, Cell 1985, 43, 633641.
48 D. A. Knecht, W. F. Loomis, Science 1987, 236,
10811085.
49 T. Spann, R. Gomer, Mol. Biol. Cell, 1992, 3,
109a.
50 R. H. Gomer, Methods A Companion to Meth-
ods in Enzymology 1999, 18, 311315.
51 T. Dingermann, N. Reindl, H. Werner, M. Hil-
debrandt, W. Nellen, A. Harwood, J. Williams,
K. Nerke, Gene 1989, 85, 353362.
52 R. A. Firtel, R. Meili, Curr. Opin. Genet. Dev.
2000, 10, 421427.
53 G. Gerisch, R. Albrecht, C. Heizer, S. Hodg-
kinson, M. Maniak, Curr. Biol. 1995, 5, 1280
1285.
54 P. Fey, K. Compton, E. C. Cox, Gene 1995,
165, 127130.
55 N. VanDriessche, C. Shaw, M. Katoh, T.
Morio, R. Sucgang, M. Ibarra, H. Kuwayama,
T. Saito, H. Urushihara, M. Maeda, et al., De-
velopment 2002, 129, 15431552.
56 G. Shaulsky, W. F. Loomis, Protist 2002, 153,
9398.
57 N. Iranfar, D. Fuller, W. F. Loomis, Eukaryot.
Cell 2003, 2, 664670.
58 M. Maeda, H. Sakamoto, N. Iranfar, D. Fuller,
T. Maruo, S. Ogihara, T. Morio, H. Urushi-
hara, Y. Tanaka, W. F. Loomis, Eukaryot. Cell
2003, 2, 627637.
59 G. Glckner, L. Eichinger, K. Szafranski, J. Pa-
chebat, P. Dear, R. Lehmann, C. Baumgart,
J. F. Abril, G. Parra, R. Guig, et al., Nature
2002, 418, 7985.
60 R. Hori, R. A. Firtel, Nucleic Acids Res. 1994,
22, 50995111.
61 D. A. Knecht, S. M. Cohen, W. F. Loomis, H. F.
Lodish, Mol. Cell. Biol. 1986, 6, 39733983.
62 B. Wetterauer, P. Morandini, I. Hribar, I. Mur-
gia-Morandini, U. Hamker, C. Singleton, H. K.
MacWilliams, Plasmid 1996, 36, 169181.
63 G. Mangiarotti, F. Altruda, H. F. Lodish, Mol.
Cell. Biol. 1981, 1, 3542.
64 A. Rathi, S. C. Kayman, M. Clarke, Dev. Genet.
1991, 12, 8287.
65 R. H. Gomer, I. S. Yuen, R. A. Firtel, Develop-
ment 1991, 112, 269278.
66 M. Clarke, J. Yang, S. Kayman, Dev. Genet.
1988, 9, 315326.
References 691
67 J. Blusch, P. Morandini, W. Nellen, Nucleic
Acids Res. 1992, 20, 62356238.
68 F. Vauti, P. Morandini, J. Blusch, A. Sachse,
W. Nellen, Mol. Cell. Biol. 1990, 10, 4080
4088.
69 B. W. Wetterauer, K. Salger, C. Carballo-Metz-
ner, H. K. MacWilliams, Differentiation 1995,
59, 289297.
70 B. W. Wetterauer, K. Salger, H. K. MacWil-
liams, Nucleic Acids Res. 1993, 21, 13971401.
71 C. D. Reymond, R. H. Gomer, M. C. Mehdy,
R. A. Firtel, Cell 1984, 39, 141148.
72 A. A. Noegel, G. Gerisch, J. Stadler, M. West-
phal, EMBO J. 1986, 5, 14731476.
73 J. Faix, W. Dittrich, J. Prassler, M. Westphal,
G. Gerisch, Plasmid 1995, 34, 148151.
74 T. M. Freeland, R. B. Guyer, A. Z. Ling, R. A.
Deering, Nucleic Acids Res. 1996, 24, 1950
1953.
75 J. Stadler, T. W. Keenan, G. Bauer, G. Gerisch,
EMBO J. 1989, 8, 371377.
76 A. E. Early, J. G. Williams, H. E. Meyer, S. B.
Por, E. Smith, K. L. Williams, A. A. Gooley,
Mol. Cell. Biol. 1988, 8, 34583466.
77 W. Dittrich, K. L. Williams, M. B. Slade, Bio-
technology 1994, 12, 614618.
78 J. H. Blusch, W. Nellen, Mol. Microbiol. 1994,
11, 331335.
79 K. M. Pang, T. Dingermann, D. A. Knecht,
Gene 2001, 277, 187197.
80 J. E. Hughes, H. Ashktorab, D. L. Welker, Dev.
Genet. 1988, 9, 495504.
81 B. Leiting, I. J. Lindner, A. A. Noegel, Mol. Cell.
Biol. 1990, 10, 37273736.
82 D. J. Manstein, H. P. Schuster, P. Morandini,
D. M. Hunt, Gene 1995, 162, 129134.
83 Y. Z. Yin, B. D. Williamson, C. L. Rutherford,
Gene 1994, 150, 293298.
84 M. Blaauw, M. H. K. Linskens, P. J. M. van
Haastert, Gene 2000, 252, 7182.
85 C. Barth, D. J. Fraser, P. R. Fisher, Plasmid
1998, 39, 141153.
86 P. M. Sharp, K. M. Devine, Nucleic Acids Res.
1989, 17, 50295039.
87 E. B. Vervoort, A. van Ravestein, N. N. M. E.
van Peij, J. C. Heikoop, O. J. M. van Haastert,
G. F. Verheijden, M. H. K. Linskens, Nucleic
Acids Res. 2000, 28, 20692074.
88 C. D. Reymond, C. Beghdadi-Rais, M. Rog-
gero, E. A. Duarte, C. Desponds, M. Bernard,
D. Groux, H. Matile, C. Bron, G. Corradin, et
al., J. Biol. Chem. 1995, 270, 1294112947.
89 M. Kozak, J. Cell Biol. 1989, 108, 229241.
90 E. Jung, A. A. Gooley, N. H. Packer, M. B.
Slade, L. Williams, W. Dittrich, Biochemistry
1997, 36, 40344040.
91 E. Jung, A. A. Gooley, N. H. Packer, P. Karu-
so, K. L. Williams, Eur. J. Biochem. 1998,
253, 517524.
92 R. Kornfeld, S. Kornfeld, Annu. Rev. Bio-
chem. 1985, 54, 631664.
93 D. J. Sharkey, R. Kornfeld, J. Biol. Chem.
1991, 266, 1848518497.
94 R. L. Ivatt, O. P. Das, E. J. Henderson, P. W.
Robbins, Cell 1984, 38, 561567.
95 H. H. Freeze, D. Wolgast, J. Biol. Chem.
1986, 261, 127134.
96 C. A. Gabel, C. E. Costello, V. N. Reinhold, L.
Kurz, S. Kornfeld, J. Biol. Chem. 1984, 259,
1376213769.
97 R. Couso, H. van Halbeek, V. Reinhold, S.
Kornfeld, J. Biol. Chem. 1987, 262, 4521
4527.
98 R. Gupta, E. Jung, A. A. Gooley, K. L. Wil-
liams, S. Brunak, J. Hansen, Glycobiology
1999, 9, 10091022.
99 C. M. West, Cell. Mol. Life Sci. 2003, 60, 229
240.
100 K. B. Raper. The Dictyostelids. Princeton, NJ:
Princeton University Press, 1984.
101 R. Sussman, M. Sussman, Biochem. Biophys.
Res. Commun. 1967, 29, 5355.
102 S. M. Coccuci, M. Sussman, J. Cell Biol.
1970, 45, 399407.
103 D. J. Watts, J. M. Ashworth, Biochem. J. 1970,
119, 171174.
104 W. F. Loomis, Exp. Cell Res. 1971, 64, 484
486.
105 S. I. Han, K. Friehs, E. Flaschel, Proc. Bio-
chem. 2003, 39, 925930.
106 J. Franke, R. Kessin, Proc. Natl. Acad. Sci.
USA 1977, 74, 21572161.
107 S. L. Han, K. Friehs, E. Flaschel, Proc. Bio-
chem. 2003, 39, 585589.
108 U. Beshay, K. Friehs, A. Azzam, E. Flaschel,
Proc. Biochem. 2003, 38, 15211529.
109 U. Beshay, K. Friehs, A. Azzam, E. Flaschel,
Bioproc. Biosyst. Eng. 2003, 26, 117120.
110 Y. Lu, U. Beshay, K. Friehs, E. Flaschel, Proc.
Biochem. 2003, 39, 18591870.
111 N. Fasel, C. Begdadirais, M. Bernard, C.
Bron, G. Corradin, C. D. Reymond, Gene
1992, 111, 157163.
112 M. X. van Bemmelen, C. Beghdadi-Rais, C.
Desponds, E. Vargas, S. Herrera, C. D. Rey-
mond, N. Fasel, Mol. Biochem. Parasitol.
2000, 111, 377390.
113 T. Dingermann, E. M. Troidl, M. Broker, K.
Nerke, Appl. Microbiol. Biotechnol. 1991, 35,
496503.
114 H. Tiltscher, M. Storr, Appl. Microbiol. Bio-
technol. 1993, 40, 246250.
115 G. Voith, T. Dingermann, Biotechnology
1995, 13, 12251229.
116 G. Voith, T. Dingermann, Pharmazie 1995,
50, 758762.
117 L. Cubeddu, C. X. Moss, J. D. Swarbrick,
A. A. Gooley, K. L. Williams, P. M. G. Curmi,
M. B. Slade, B. C. Mabbutt, Prot. Expr. Purif.
2000, 19, 335342.
118 K. R. Emslie, J. M. Miller, M. B. Slade, P. R.
Dormitzer, H. B. Greenberg, K. L. Williams,
J. Virol. 1995, 69, 17471754.
119 K. R. Emslie, M. B. Coukell, D. Birch, K. L.
Williams, J. Biotechnol. 1996, 50, 149159.
120 S. Asgari, S. Arun, M. B. Slade, J. Marshall,
K. L. Williams, J. F. Wheldrake, J. Mol. Micro-
biol. Biotechnol. 2001, 3, 491497.
121 G. Voith, H. Kramm, I. Zundorf, T. Winck-
ler, T. Dingermann, Pharmazie 1998, 53,
707710.
122 Y. Lu, J. C. Knol, M. H. K. Linskens, K.
Friehs, P. J. M. van Haastert, E. Flaschel, J.
Biotechnol. 2004, 108, 243251.
123 J. C. Heikoop, P. D. J. Grootenhuis, M.
Blaauw, J. S. Veldema, P. J. M. van Haastert,
M. H. K. Linskens, Eur. J. Biochem. 1998,
256, 359363.
124 M. H. K. Linskens, P. D. J. Grootenhuis, M.
Blaauw, B. Huisman-de Winkel, A. van Ra-
vestein, P. J. M. van Haastert, J. C. Heikoop,
FASEB J. 1999, 13, 639645.
125 A. B. Cubitt, I. Reddy, S. Lee, J. G. McNally,
R. A. Firtel, Dev. Biol. 1998, 196, 7794.
126 X. Zhou, D. M. Fambrough, J. Membrane
Biol. 1999, 167, 1924.
127 F. Chaumont, W. F. Loomis, M. J. Chrispeels,
6209.
128 N. R. Cohen, D. A. Knecht, H. F. Lodish, Bio-
chem. J. 1996, 315, 971975.
129 T. Kashiyama, K. Ito, K. Yamamoto, J. Mol.
Biol. 2001, 311, 461466.
130 I. Karakesisoglou, M. Schleicher, B. C. Gib-
bon, C. J. Staiger, Cell. Motil. Cytoskel. 1996,
34, 3647.
131 S. Hgele, R. Khler, H. Merkert, M. Schlei-
cher, J. Hacker, M. Steinert, Cell. Microbiol.
2000, 2, 165171.
132 J. M. Solomon, A. Rupper, J. A. Cardelli,
R. R. Isberg, Infect. Immun. 2000, 68, 2939
2947.
133 S. Pukatzki, R. H. Kessin, J. J. Mekalanos,
3164.
134 P. Cosson, L. Zulianello, O. Join-Lambert, F.
Faurisson, L. Gebbie, M. Benghezal, C. van
Delden, L. K. Curty, T. Kohler, J. Bacteriol.
2002, 184, 30273033.
135 J. N. Steenbergen, J. D. Nosanchuk, S. D.
Malliaris, A. Casadevall, Infect. Immun. 2003,
71, 48624872.
136 M. A. Titus, Curr. Biol. 1999, 9, 12971303.
137 C. Skriwan, M. Fajardo, S. Hgele, M. Horn,
M. Wagner, R. Michel, G. Krohne, M. Schlei-
cher, J. Hacker, M. Steinert, Int. J. Med. Mi-
crobiol. 2002, 291, 615624.
138 J. M. Solomon, G. S. Leung, R. R. Isberg, In-
fect. Immun. 2003, 71, 35783586.
139 J. Tillner, T. Winckler, T. Dingermann, Phar-
mazie 1996, 51, 902906.
140 J. Tillner, H. Nau, T. Winckler, T. Dinger-
mann, Toxicol. in Vitro 1998, 12, 463469.
141 K. Dannat, J. Tillner, T. Winckler, M. Weiss,
K. Eger, T. Dingermann, Pharmazie 2003,
58, 204210.
142 R. S. B. Williams, M. Eames, W. J. Ryves, J.
Viggars, A. J. Harwood, EMBO J. 1999, 18,
27342745.
143 R. S. B. Williams, L. L. Cheng, A. W. Mudge,
A. J. Harwood, Nature 2002, 417, 292295.
144 X. Wu, Y. Li, B. Crise, S. M. Burgess, Science
2003, 300, 17491751.
145 S. Hacein-Bey-Abina, C. Von Kalle, M.
Schmidt, M. P. McCormack, N. Wulffraat, P.
Leboulch, A. Lim, C. S. Osborne, R. Pawliuk,
E. Morillon, et al., Science 2003, 302, 415
419.
146 D. B. Kohn, M. Sadelain, C. Dunbar, D. Bo-
dine, H. P. Kiem, F. Candotti, J. Tisdale, I.
Riviere, C. A. Blau, R. E. Richard, et al., Mol.
Ther. 2003, 8, 180187.
147 T. Winckler, T. Dingermann, G. Glckner,
Cell. Mol. Life Sci. 2002, 59, 20972111.
148 P. Beck, T. Dingermann, T. Winckler, J. Mol.
Biol. 2002, 318, 273285.
References 693
Abstract
Coagulation factor IXa (fIXa) is a key
player in the activation cascade of blood
coagulation. Subtle structural characteris-
tics result in an almost latent protease,
and distinguish this enzyme from closely
related coagulation factors like IIa, VIIa
and Xa. Thereby, fIXa can serve its dual
capacity to both boost and throttle the
blood coagulation. Here, we summarize
our insights into the molecular mecha-
nisms of choking and releasing fIXas en-
zymatic activity. Direct evidence for the de-
duced two-step activation theory comes
from a fIXa triple mutant featuring 7000-
fold enhanced catalytic activity. Biopharma-
ceutical applications of genetically modi-
fied fIXa variants will be discussed.
Abbreviations
a activated (e.g., fIXa)
f factor (e.g., fIX)
6.1
Introduction
Quasi-instantaneous, yet delicately balanced
regulation of blood coagulation is an abso-
lutely critical repair mechanismin higher or-
ganisms and its malfunction in either direc-
tion manifests itself in severe disease states.
Blood coagulation is organized by two highly
conserved molecular cascades merging into
one another (see also Part II, Chapters 1 and
3). By utilizing both positive and negative
feedback mechanisms, they modulate and
localize blood coagulation to the site of a
wound. This complex process is impress-
ively shown in a video animation on the
CD-ROM. On a molecular level, blood coag-
ulation involves limited proteolytic events of
high selectivity. Blood coagulation factor IXa
(fIXa) marks the molecular merging point of
both coagulation cascades. As most other co-
agulation enzymes, fIXa is a trypsin-like, vi-
tamin K-dependent serine protease that cir-
culates in the plasma as an inactive single-
chain zymogen (Fig. 6.1) [1, 2].
695
6
Releasing the Spring: Cofactor- and Substrate-assisted Activation
of Factor IXa
ISBN: 3-527-31184-X
Revolution by Evolution
Rational Design for Desire and Scientific Art of Optimization
In marked contrast to trypsin, however,
virtually all coagulation proteases require
(protein) cofactors for enzymatic selectivity
and specificity. fIXa is activated by the cleav-
age of two peptide bonds by either the acti-
vated fVII (fVIIa)tissue factor (TF) complex
or activated fXI (fXIa) [3, 4] to remove a 35-
residue activation peptide. The active en-
zyme, fIXab, consists of two disulfide-linked
chains. The light chain comprises the N-ter-
minal Gla domain with c-carboxylated gluta-
mic acid residues and two epidermal growth
factor-like domains. The heavy chain consti-
tutes the trypsin-like serine protease domain
(see Fig. 6.1). fIXa then activates fX in a re-
action that is dependent on the presence of
calcium ions, a membrane surface and a
nonenzymatic protein cofactor, fVIIIa [2].
The importance of fIXa in hemostasis is re-
flected by the occurrence of the bleeding dis-
order hemophilia B in individuals carrying
mutations in the fIX gene [5]. In the absence
of its activated cofactor fVIIIa, fIXa is an al-
most inactive enzyme against natural and
synthetic substrates. Binding of fVIIIa
ignites a 10
6
-fold selective specificity in-
crease in activation of fX, whereas the activ-
ity with peptidic substrates remains unal-
tered [6, 7]. The latter dramatic substrate-de-
pendent activity modulation distinguishes
fIXa from the related coagulation enzymes
fXa and fVIIa, which, in the presence of their
cofactors, achieve a significant activity en-
hancement with synthetic substrates [8, 9].
In an attempt to understand this unique
behavior of fIXa, we link enzymatic and
structural properties. Here, we review the
substrate preferences of fIXa and homolo-
gous enzymes, and relate them to structural
elements critical for substrate recognition.
Of particular interest, the conformation
of the fIXa 99-loop (chymotrypsinogen
numbering used throughout within the
serine protease domain) deviates consider-
ably from those observed in the related en-
zymes of the fIX gene family [1014].
Kolkman and Mertens further demon-
strated that the 99-loop restricts enzymatic
Fig. 6.1 Exon, domain, and 3d structure of fIX family proteins.
activity towards fX and synthetic substrates
in the absence, but not in the presence, of
the cofactor fVIIIa [15].
The side-chain of Tyr99 adopts different
conformations in the two crystal structures
of fIXa reported to date [11, 16]. In both
structures Tyr99 is in steric conflict with
canonical substrate binding in the S2S4
sites. These observations suggest a critical
role of Tyr99 and spatially neighboring
amino acids in the substrate-dependent ac-
tivity of fIXa. In particular, Lys98 is likely
to electrostatically interfere with the basic
substrate preference of fIXa.
6.2
The Zymogen Form of fIX is Fully Inactive
In vitro, the activation of fIX requires two
distinct enzymatic events. RVV-X initially
cleaves a single peptide bond Arg180
Val181 of single-chain fIX to produce fIX-
aa. Complete activation to fIXab is obtained
autocatalytically by both fIXaa and fIXab by
a second cleavage at Arg145Ala146. This
strict requirement of an already active en-
zyme for the proteolytic activation of fIX
contrasts with the situation for fVII, which
is able to autoactivate without a prior jump
start [1719]. Given the sequence similarity
of both activation peptide cleavage sites, we
expect active fIXa to be able to cleave at both
sites. The need for an exogenous activator
indicates that single-chain fIX is unable to
undergo the conformational transformation
necessary for proteolytic activity [1820].
In the active site mutant rf9a-S195A
RVV-X cleaves only once at the Arg180
Val181 peptide bond, thereby converting
single-chain rf9-S195A into the active con-
formation. Interestingly, however, the two-
chain rf9a-S195A is not further processed
[20]. This contrasts with the situation in
related serine proteases, which display re-
sidual activity when carrying the corre-
sponding mutation of the catalytic residue
Ser195 [2224]. This residual activity re-
sults from the conformational strains in-
duced in the substrates by the enzymes
and suffices to catalyze hydrolysis of the
bound substrate. Consequently, fIXa is un-
able to induce a comparable conforma-
tional strain to bound substrates. These
observations indicate the significance of
the K
M
for the low activity of fIXa.
6.3
Relevance of Tyr99 on the Stability
of the 99-loop
Analysis of the fIXa structure and, in partic-
ular, its comparison with the related struc-
ture of fXa indicated that Tyr99 hinders di-
rect substrate access to the S4 binding site
(Fig. 6.2) [11, 12]. Therefore, we replaced
Tyr99 with several smaller residues. Sur-
prisingly, the activity of these mutants de-
creased [21]. The reactivity was lowest with
the smallest side-chain substitution, Y99A,
refuting the hypothesis that steric hin-
drance of Tyr99 is sufficient to explain the
impeded substrate access to the active site.
Moreover, binding of the S1 inhibitor PABA
was also impaired. This indicates that the
removal of Tyr99 results in complex
changes, affecting not only the S2S4, but
also the S1 site by either blocking access
to the primary specificity site or by even de-
stabilizing its conformation. The side-chain
of Tyr99 is the part of the 99-loop closest to
the S1 pocket. Removal of the bulky side-
chain of Tyr99 probably causes the loop to
slide further into the S2S4 substrate-bind-
ing cleft. As a consequence, the positively
charged side-chain of Lys98 comes closer
to the entrance to S1 and repels basic sub-
strate residues from binding into the pri-
mary specificity pocket.
6.4
Lys98 Hinders Substrate Binding
to fIXa both Sterically and Electrostatically
To elucidate the influence of charge and
size of Lys98 on substrate binding, the ac-
tivity of mutants with uncharged and
shorter side-chains in position 98 was ex-
amined. rf9a-K98M already has a 3-fold in-
creased activity [21], confirming the elec-
trostatic repulsion model where access of
basic substrate residues is stalled by Lys98
(Fig. 6.3, left). The enzymatic activity of
rf9a-K98T is even higher and results in an
approximately 7-fold increase [21], in line
with the assumption that in the wild-type
enzyme both charge and size of Lys98 hin-
der substrate binding. The binding affinity
of the S1 inhibitor PABA is doubled by
the charge removal, but independent of
Fig. 6.2 Comparison of the S2S4 recognition sites (99-loop) of the fIXa active site (orange) with those
of fXa (blue).
Fig. 6.3 A. Electrostatic repulsion model.
While D189 in the bottom of fIXas S1
site defines its substrate preference with
Arg or Lys in P1, access for peptidic sub-
strates is hindered by K98 which acts as
a gatekeeper towards basic residues.
B. In the fIXa-fVIIIa cofactor complex,
however, the electrostatic repulsion con-
flict is resolved towards the correct sub-
strate fX.
the size of the side-chain at position 98
[21]. This indicates that, while the binding
of positively charged P1 residues is electro-
statically hindered by Lys98, the size of the
side-chain at position 98 presumably only
affects the accessibility to the S2S4 site,
but not to the S1 site. As a negative con-
trol we analyzed K98R, which revealed al-
most unaltered enzymatic and binding
properties towards substrates (see Fig. 6.3).
6.5
Tyr177 Locks the 99-loop
in an Inactive Conformation,
which is Released by Cofactor fVIIIa
and Modified by the Physiologic Substrate fX
The unique conformation of the 99-loop in
fIXa relates to several elements, including
a two-residue insertion (Ala95aAla95b)
and a series of more subtle structural char-
acteristics. Its conformation is stabilized
by a hydrogen bond between the carbonyl
of Lys98 and the hydroxyl of Tyr94. Since
fXa has phenylalanine at position 94, this
hydrogen bond is not present in fXa. Also,
Tyr177 and Lys98, both of which are threo-
nines (or serines) in fXa, fVIIa and protein
C [12, 14, 18, 19], would collide if the fIXa
99-loop were to adopt a fXa-like conforma-
tion. Moreover, structure analysis revealed
that Tyr177 stabilizes and locks the 99-loop
in its inactive conformation by interactions
with Asn97 and Asn100.
This rationale was examined by con-
structing a triple mutant rf9a-Y94FK98T
Y177T as well as a chimerical rf9a contain-
ing the 99-loop of fXa and the fXa-like en-
vironment of the loop (Y94FY177T). Both
mutants showed significantly increased
amidolytic activities. Depending on the
substrate, rf9a-99loop(F10)Y94FY177T
achieved a 10- to 15-fold increase in activ-
ity when compared to the wild-type pro-
tein, while the triple mutant showed a dra-
matic 3000- to 7000-fold (!) increase in cat-
alytic activity, with the variation depending
on the used peptidic substrate [21]. Clearly,
while the heterologous fXa loop facilitates
access to the substrate-binding cleft, and
removes the steric clashes caused by Lys98
and Tyr177, it cannot adopt the optimal
conformation as in the triple mutant rf9a-
Y94FK98TY177T.
In the latter, the 99-loop and the 177-
segment probably approach the physiologi-
cally active conformation and open the ac-
tive site cleft for substrate binding. These
data suggest a mechanism of physiological
fIXa activation by cofactor and substrate
binding. In this model, cofactor binding
close to the 177-segment [5, 10] will re-
lease the lock of the inactive 99-loop im-
posed by Tyr177. Only the physiological
substrate fX is able to rearrange the re-
leased 99-loop, paving itself its way into
the active-site cleft of fIXa, thus resem-
bling an example of substrate-assisted cata-
lysis. Both steps of fIXa activation are
combined in the triple mutation, explain-
ing the nonlinear stimulation and high ac-
tivity towards peptidic substrates.
Notably, the conformation of the S1 site
remains mainly unaltered in both multiple
mutants as witnessed by the invariance of
the binding affinity of the S1 site inhibitor
p-benzamidine (K
I
*150 lM) [21]. Thus,
the introduced multiple mutations mainly
affect the S2S4 recognition sites.
6.6
S1 Site Mutations Decrease
the Activity of fIXa
In serine proteases, the S1 site usually
contributes most to substrate recognition
and catalysis. Therefore, we examined the
impact of mutations localized at the S1
site on the fIXa activity, including S190A,
E219G and E219Q. None of them resulted
in improved activity [13, 21].
Glu219 is particularly intriguing because
Gly219 is conserved in almost all trypsin-
like serine proteases, including trypsin it-
self and the other coagulation factors.
Although energetically unfavorable,
Glu219 has an apparently unchanged
backbone conformation compared with
trypsin. We attempted to clarify the impor-
tance of the charge of Glu219 and its salt
bridge with Lys224 [11, 16] for fIXa activity
by replacing it with the isosteric Gln. Both
activity and inhibitor binding were slightly
impaired, indicating a moderate distur-
bance of the S1 site geometry probably
caused by the ionic interaction of Glu219
with Lys224.
In the S190A mutant, the affinity of
benzamidine is reduced approximately 2-
fold, which may be explained by the ab-
sence of a hydroxyl in the S1 site [21].
6.7
Evolutionary Relation of fIXa and fXa
is Reflected in the Dependence
of Activity Changes on Arg/Lys Substrates
The coagulation enzymes fIXa and fXa are
evolutionary closely related. This relation
is also reflected by their parallel change in
activity dependent on Ala/Ser mutations at
position 190 of the enzyme and Arg/Lys at
the substrate P1 site. The Ala190 variant
of either enzyme is twice as active as the
respective Ser190 variant towards the P1-
arginine substrate, while the Ala190 activ-
ity towards P1-lysine substrates is reduced
by a factor of 2 [18, 19]. Moreover, Ser190
in fIXa is encoded by TCN, not AGY. This
observation is inline with the notion that
residue 190 is evolutionary related by sin-
gle-site mutations to the Ala190 in fXa, en-
coded by GCC [25].
The natural substrates of both fIXa and
Xa have arginine in P1 exclusively. Ala190,
the more-active variant, is conserved in
fXa in all species. In contrast, fIXa has a
strictly conserved Ser190, which confers
only half the activity as Ala190. Given the
low absolute activity of fIXa, the activity re-
duction caused by Ser190 is relatively
modest. Nevertheless, Ser190 emphasizes
the evolutionary optimization of fIXa to-
wards an enzyme which is hardly active in
the absence of its cofactor and correct sub-
strate. Ser190 may be considered a choke,
which is released only in the Xase com-
plex. This underlines fIXas critical role in
the blood coagulation cascade.
6.8
By Binding at the 60-loop Ethylene Glycol
Indirectly Reorganizes the 99-loop and
Allosterically Stimulates the Activity of fIXa
We recently showed that fIXa and fVIIa
have a common ethylene glycol-binding site
between residues 60 and 90 that is not ac-
cessible in fXa [18, 19]. Occupation of this
site by ethylene glycol appears to mimic ef-
fects of macromolecular substrate binding
on the 99-loop. Thereby, it stimulates cataly-
sis of synthetic substrates in a way that is
usually only observed for macromolecular
substrates that are able to occupy both sites
(60-loop region and active site cleft) simulta-
neously in the presence of fVIIIa. The de-
creased stimulation of the more active 99-
loop mutants by ethylene glycol [15] sup-
ports the interpretation that the 99-loop is
already reorganized in these mutants, thus
precluding further activity enhancement
upon ethylene binding.
6.9
This chapter illustrates how nature has op-
timized fIXa as a strictly regulated enzyme
with multiple control mechanisms. This
evolutionary optimization reflects the ex-
treme danger of any misfiring of this en-
zyme due to its strategic role at the inter-
section between extrinsic and intrinsic co-
agulation pathway, as well as between initi-
ation and amplification of coagulation. At
the same time, this chapter suggests possi-
ble biopharmaceutical applications of ge-
netically engineered fIXa mutants (and ap-
plying the same approach to other indica-
tions) to better manage situations of he-
matological and other disorders.
References
1 DiScipio, R. G., Hermodson, M. A., Yates,
S. G., Davie, E. W. 1977. A comparison of hu-
man prothrombin, factor IX (Christmas fac-
tor), factor X (Stuart factor), and protein S,
Biochemistry 16, 698706.
2 Davie, E. W., Fujikawa, K., Kisiel, W. 1991.
The coagulation cascade: initiation, mainte-
nance, and regulation, Biochemistry 30, 10363
10370.
3 Fujikawa, K., Legaz, M. E., Kato, H., Davie,
E. W. 1974. The mechanism of activation of
bovine factor IX (Christmas factor) by bovine
factor XIa (activated plasma thromboplastin
antecedent), Biochemistry 13, 45084516.
4 Lindquist, P. A., Fujikawa, K., Davie, E. W.
1978. Activation of bovine factor IX (Christ-
mas factor) by factor XIa (activated plasma
thromboplastin antecedent) and a protease
from Russells viper venom, J Biol Chem 253,
19021909.
5 Giannelli, F., Green, P. M., Sommer, S. S.,
Poon, M., Ludwig, M., Schwaab, R., Reitsma,
P. H., Goossens, M., Yoshioka, A., Figueiredo,
M. S., Brownlee, G. G. 1998. Hemophilia B:
database of point mutations and short addi-
tions and deletions eighth edition, Nucleic
Acid Res 26, 265268.
6 Duffy, E. J., Lollar, P. 1992. Intrinsic pathway
activation of factor X and its activation pep-
tide-deficient derivative, factor Xdes-143191,
J Biol Chem 267, 78217827.
7 McRae, B. J., Kurachi, K., Heimark, R. L., Fuji-
kawa, K., Davie, E. W., Powers, J. C. 1981.
Mapping the active sites of bovine thrombin,
factor IXa, factor Xa, factor XIa, factor XIIa,
plasma kallikrein and trypsin with amino acid
and peptide thioesters: development of new
sensitive substrates, Biochemistry 20, 7196
7206.
8 Rosing, J., Tans, G., Govers-Riemslag, J. W. P.,
Zwaal, R. F. A., Hemker, H. C. 1980. The role of
phospholipids and factor Va in the prothrom-
binase complex, J Biol Chem 255, 274283.
9 Neuenschwander, P. F., Morrissey, J. H. 1994.
Roles of the membrane-interactive regions of
factor VIIa and tissue factor, J Biol Chem 270,
970977.
10 Banner, D. W., DArcy, A., Chne, C., Winkler,
F. K., Guha, A., Konigsberg, W. H., Nemerson,
Y., Kirchhofer, D. 1996. The crystal structure
of the complex of blood coagulation factor
VIIa with soluble tissue factor, Nature 380,
4146.
11 Brandstetter, H., Bauer, M., Huber, R., Lollar,
P., Bode, W. 1995. X-ray structure of clotting
factor IXa: active site and module structure re-
lated to Xase activity and hemophilia B, Proc
12 Brandstetter, H., Khne, A., Bode, W., Huber,
R., von der Saal, W., Wirthensohn, K., Engh,
R. A. 1996. X-ray structure of active site-inhib-
ited clotting factor Xa. Implications for drug
design and substrate recognition, J Biol Chem
271, 2998829992.
13 Hopfner, K. P., Brandstetter, H., Karcher, A.,
Kopetzki, E., Huber, R., Engh, R. A., Bode, W.
1997. Converting blood coagulation factor IXa
into factor Xa: dramatic increase in amidolytic
activity identifies important active site deter-
minants, EMBO J 16, 66266635.
14 Mather, T., Oganessyan, V., Hof, P., Huber, R.,
Foundling, S., Esmon, C., Bode, W. 1996. The
2.8 crystal structure of Gla-domainless acti-
vated protein C, EMBO J 15, 68226831.
15 Kolkman, J. A., Mertens, K. 2000. Insertion
loop 256268 in coagulation factor IX restricts
enzymatic activity in the absence but not in
the presence of factor VIII, Biochemistry 39,
73987405.
References 701
16 Hopfner, K. P., Lang, A., Karcher, A., Sichler,
K., Kopetzki, E., Brandstetter, H., Huber, R.,
Bode, W., Engh, R. A. 1999. Coagulation factor
IXa: the relaxed conformation of Tyr99 block
substrate binding, Structure 7, 989996.
17 Pedersen, A. H., Lund-Hansen, T., Bisgaard-
Frantzen, H., Olsen, F., Petersen, L. C. 1989.
Autoactivation of human recombinant coagu-
lation factor VII, Biochemistry 28, 93319336.
18 Sichler, K., Banner, D. W., DArcy, A., Hopfner,
K.-P., Huber, R., Bode, W., Kresse, G.-B., Ko-
petzki, E., Brandstetter, H. 2002. Crystal struc-
tures of uninhibited factor VIIa map its cofac-
tor- and substrate-assisted activation to specific
interactions, J Mol Biol 322, 591603.
19 Sichler, K., Hopfner, K.-P., Kopetzki, E., Hu-
ber, R., Bode, W., Brandstetter, H. 2002. The
influence of residue 190 in the S1 site of tryp-
sin-like serine proteases on substrate selectiv-
ity is universally conserved, FEBS Lett 530,
220224.
20 Bode, W. 1979. The transition of bovine trypsi-
nogen to a trypsin-like state upon strong li-
gand binding. II. The binding of the pancreat-
ic trypsin inhibitor and of isoleucinevaline
and of sequentially related peptides to trypsi-
nogen and to p-guanidinobenzoatetrypsino-
gen, J Mol Biol 127, 357374.
21 Sichler, K., Kopetzki, E., Huber, R., Bode, W.,
Hopfner, K.-P., Brandstetter, H. 2003. Physio-
logical fIXa activation involves a cooperative
conformational rearrangement of the 99-loop,
J Biol Chem 278, 41214126.
22 Stone, S. R., LeBonniec, B. F. 1997. Inhibitory
mechanism of serpins. Identification of steps
involving the active-site serine residue of the
protease, J Mol Biol 265, 344362.
23 Olson, S. T., Swanson, R., Day, D., Verhamme,
I., Kvassman, J., Shore, J. D. 2001. Resolution
of Michaelis complex, acylation, and confor-
mational change steps in the reactions of the
serpin, plasminogen activator inhibitor-1, with
tissue plasminogen activator and trypsin, Bio-
chemistry 40, 1174211756.
24 Krishnan, R., Sadler, J. E., Tulinsky, A. 2000.
Structure of the Ser195Ala mutant of human
alpha-thrombin complexed with fibrinopeptide
A(716): evidence for residual catalytic activity,
Acta Crystallogr D 56, 406410.
25 Krem, M. M., DiCera, E. 1998. Conserved
water molecules in the specificity pocket of
serine proteases and the molecular mecha-
nism of Na
+
binding, Proteins 30, 3442.
Abstract
Complex multiparameter optimization
problems are quite common to new and
innovative diagnostic applications. For
these applications, the generation of best-
fit biomolecules is particularly challenging,
mainly because of the rapidity with which
desired biomolecules must be created and
diagnostic processes using the newly-cre-
ated molecules developed. For some diag-
nostic applications, rational engineering of
some intensively characterized proteins for
novel functions has been achieved. For en-
gineering most other diagnostic proteins
which have not been well-characterized, a
primary irrational approach to introduce
random mutations into the whole or part
of gene sequence and then screen or select
for expressed variants with desired proper-
ties has become more widely used to en-
hance or to alter specific functions. Direct-
ed evolution is another remarkable
approach for diagnostic product develop-
ment. This approach mimics the natural
evolution process itself as a method for
new function development, and thus the
specified functions are evolved instead of
being designed. Directed evolution also
takes recombination as an additional en-
gine for evolutionary changes and provides
possibly the most effective way to generate
novel biomolecules with desired features
to fulfill requirements of new diagnostic
applications. Undoubtedly, rational design,
molecular irrational design and directed
evolution can work in a synergistic man-
ner to accelerate the development of mod-
ern biopharmaceuticals. This synergism is
highlighted in this chapter by examples of
enhancing PCR (polymerase chain reac-
tion) performance of a B-type DNA poly-
merase from Thermococcus aggregans
through molecular rational design and
evolving the highly active calf intestinal al-
kaline phosphatase (cIAP) through direct-
ed evolution in E. coli.
Abbreviations
AP alkaline phosphatase
bIAP bovine intestinal alkaline phos-
phatase
cIAP calf intestinal alkaline phospha-
tase
ddNTP dideoxy nucleoside triphosphate
dNTP deoxy nucleoside triphosphate
hPAP human placental alkaline phos-
phatase
Pfu Pyrococcus furiosus
703
ISBN: 3-527-31184-X
7
Accelerating Diagnostic Product Development Process
with Molecular Rational Design and Directed Evolution
pNPP p-nitrophenyl phosphate
Tag Thermococcus aggregans
Taq Thermus aquaticus
Tgo Thermococcus gorgonarius
U-DNA uracil-containing DNA
7.1
Introduction
The primary goal of diagnostic testing is
the detection and quantification of disease-
specific analytes ranging from simple spe-
cies (e.g., ions), through complex biomole-
cules such as drugs, hormones and pro-
teins, to complex analytes such as cells
and viruses (see also Part I, Chapter 6;
and Part VII, Chapter 1). As the most ana-
lytes occur at low concentrations in com-
plex biological matrices such as blood,
plasma, sweat, urine, feces, or tissue
biopsy a high analytical sensitivity and
specificity is required in diagnostic testing.
Many of the diagnostic test methods mim-
ic the way in which biological molecules
are recognized in organisms by specific
molecular interactions. For example, in
immunological tests antibodies are used
for the detection of an antigen (e.g., a
virus) (see also Introduction and Part V,
Chapter 6). If the analyte of interest can
be used as a substrate for a specific en-
zyme, an enzymatic assay can be applied
to determine the concentration of this ana-
lyte. Therefore, proteinaceous biomole-
cules are extremely useful for the applica-
tion in diagnostic tests, and further devel-
opment and optimization of these proteins
is continuously ongoing to match new di-
agnostic challenges.
The introduction of molecular biological
methods in protein biochemistry has
achieved steady progress towards the ra-
tional engineering and de novo design of
proteins for novel functions, and several
recent reports have shown these
approaches to be successful [1, 2]. For en-
gineering proteins which have not been
well-characterized, the prospects afforded
by rational design still remain challenging,
because of the need for a precise under-
standing of the rules governing protein
folding, an immense amount of structure
function information for each protein to
be engineered, and particularly for en-
zymes the structural details of enzyme
complex with various ligands and ana-
logues of reaction intermediates. In such
cases, a primary irrational approach to
introduce random mutations into the
whole or part of gene sequence and then
to screen or select for expressed variants
with desired properties has become more
widely used. Through this approach, pro-
tein functions such as specific properties,
bioactivity, substrate specificity, and cofac-
tor requirements can be enhanced or al-
tered (for recent reviews, see [3, 4]). The
use of random mutagenesis offers an in-
teresting alternative approach to engineer
new, tailored functions, which with the ra-
tional design is normally hardly to fulfill.
A remarkable progress in protein engi-
neering for novel functions is the further
development of directed evolution
approaches [510]. Directed evolution
mimics the natural evolution process itself
as a method for new function develop-
ment, and therefore new functions are
evolved instead of being designed. In dif-
fering from rational design and random
mutagenesis approaches in which muta-
tions are considered to be the sole factor
leading to new functions directed evolu-
tion also takes recombination as a crucial
engine for evolutionary changes. By creat-
ing and then screening large libraries of
random variants for specific properties, a
variety of futures can be explored, futures
including new environments (e.g., the evo-
7 Accelerating Diagnostic Product Development Process with Molecular Rational Design and Directed Evolution 704
lution of enzymes for specific in-vitro diag-
nostic buffer systems [11]) or even entirely
altered functions (e.g., enzyme variants se-
lectively reacting with specific diagnostic
molecules [12]). By evolving new functions
and thereby new solutions to in-vitro diag-
nostic design problems, information may
be acquired not only about molecular solu-
tions that might never be revealed or an-
ticipated in classical diagnostics applica-
tion processes, but also about unique in-
sights into the molecular mechanisms of
protein folding, enzyme function, and en-
zyme catalysis.
7.2
Strategies for Optimizing Diagnostic
Proteins
7.2.1
Rational Design
Protein engineering by molecular rational
design is a knowledge-based method to
generate variants of a protein with pre-
dicted features. It requires a detailed
knowledge of the proteins amino acid se-
quence, structure, and mechanism. The
process of rational protein design usually
begins with the choice of a suitable pro-
tein scaffold to be engineered. Sequence
alignment with homologous proteins is
used to identify the critical amino acid re-
sidues in the protein sequence. A com-
parative analysis of biochemical character-
istics of these homologous proteins such
as stability, enzymatic activity, and specific-
ity further supports the identification of
the amino acids to be engineered. A de-
tailed analysis of the protein structure and
its comparison to structures of homolo-
gous proteins results in additional under-
standing of the structurefunction rela-
tionship of the protein of interest. In the
absence of a high-resolution protein struc-
ture, molecular modeling can be used to
generate a working model of the protein
structure. The molecular rational design of
a protein is performed on the level of its
DNA sequence. A polymerase chain reac-
tion (PCR)-based, site-directed mutagene-
sis is the method of choice to replace ami-
no acid residues in specific sequence posi-
tions [13].
Identification of the amino acid se-
quence to be modified requires knowledge
not only of the existing function but, im-
portantly, also of the desired new or im-
proved function. This alteration can be
achieved through single-point mutations,
exchange of the whole domains, or by the
generation of fusion proteins [4]. When
these changes have been made, the mu-
tants are purified and evaluated.
7.2.2
Directed Evolution
Directed evolution, however, mimics the
natural evolution process itself as a meth-
od for new function development, and
therefore new functions are evolved in-
stead of being designed. A significant ad-
vantage of this approach over rational de-
sign methods is that neither structural in-
formation nor a mechanistic roadmap is
required to guide the directed evolution ex-
periment. An overall strategy for directing
the evolution of an enzyme to perform a
new function or an altered function under
new conditions is described in Fig. 7.1.
The generation of new useful enzymes
through this strategy first relies on having
an effective means for accumulating many
small improvements that have resulted
from single mutations and new useful se-
quence motif linkages. Two approaches to
accumulating beneficial mutations se-
quential random mutagenesis (asexual
molecular evolution) and gene recombina-
tion (a sexual approach) can be used
for such purpose. An example of the suc-
cessful implementation of this strategy to
evolve Erwinia sp. creatinase to increase
thermostability has been described [11].
The generation of new useful enzymes
through this directed evolution strategy
also relies on an effective and sensitive
screening method for identifying useful
mutations [6, 14]. Nature explores new
functions through random mutations and
recombination, using natural selection as
the search mechanism. Nature windows
out the unfavorable mutations by linking
the organisms growth rate and reproduc-
tive success to the performance of its com-
ponents. Faster-growing, fitter organisms
eventually dominate, allowing an efficient
search of very large microbial popula-
tions. Similar to natural selection, in-vitro
genetic selections and genetic complemen-
tations offer the advantages that functional
proteins can be identified from very large
libraries simply by growing a population
of transformed cells or variants under spe-
cified selective conditions.
Unfortunately, the features of most en-
zymes used for in-vitro diagnostics often
cannot be linked to the survival or growth
of the host organism. Therefore, mutant
proteins often must be screened rather than
selected. Although relatively time-consum-
ing and often with low throughput, such
screening assays represent the most effec-
tive means for most problems of practical
interest. We have demonstrated, for exam-
ple, that the glucose substrate-specificity of
Acinetobacter calcoaceticus soluble glucose
dehydrogenase can be greatly improved
with this type of liquid screening of individ-
ual enzyme mutants which are arrayed spa-
tially in micro-titer plates [12].
7.2.3
WalkThrough Recombination
During the past few years, an increasing
number of people have recognized the mer-
its of evolutionary search strategies. Only
some practical techniques, however, such
as sequential mutagenic PCR [15], combina-
torial cassette mutagenesis [16], random oli-
gonucleotide mutagenesis [17], DNA shuf-
fling [18, 19], random-priming recombina-
Fig. 7.1 Overall strategy for directing the evolution of an en-
zyme to perform a new function or an altered function under
new conditions.
tion [20], random mutagenesis using muta-
tor strains [21] or mutator plasmids [22],
staggered-extension [23], artificial recombi-
nation [24, 25], and recombined extension
[26] have been successfully applied to these
problems. The development of efficient and
practical experimental techniques to mimic
these key processes is a scientific challenge,
and WalkThrough recombination [27] be-
longs to such newly developed techniques
for directed evolution.
In differing from the above-mentioned
irrational design and directed evolution
techniques, the WalkThrough technique
(Fig. 7.2) presented here involves walking
through a template gene with a mixture of
chain-elongating molecules and chain-ter-
minating molecules (e.g., dNTPs/ddNTPs)
to generate a pool of 3' end-randomly dis-
tributed DNA fragments with a low level
of point mutations. After removing the
original templates and terminating mole-
cules (e.g., ddNMP ends), these short
DNA fragments can prime one another
based on homology under appropriate re-
action conditions, and then re-assemble to
form full-length genes by repeated thermo-
cycling in the presence of thermostable
DNA polymerase. These full-length genes
can be further amplified by conventional
PCR and cloned into a proper vector for
expression of the encoded proteins.
Screening or selection of the expressed
mutants leads to new variants with im-
proved or even novel functions. These im-
proved variants can be used immediately
as partial solutions to a practical problem,
or they can serve as new start points for
further cycles of WalkThrough mutagene-
sis and recombination.
Compared to other techniques used for
protein optimization, the WalkThrough re-
combination technique presented here
shows several advantages for in-vitro pro-
tein optimization:
1. Since the WalkThrough chains are a
population of fragments that each stops
at every position of the template mole-
cules, they are uniform in their posi-
tional preference and lack a sequence
bias. This sequence heterogeneity allows
crossover to occur more randomly.
2. Normal sequential error-prone PCR mu-
tagenesis and DNA shuffling can not ef-
ficiently recombine or dissect two or
more mutations if they are very close to
each other [18]. In contrast, the Walk-
Through approach allows recombination
to occur at every position of templates,
and therefore provides the possibility of
recombining or dissecting two or more
mutations, although they may be very
close to each other.
3. Since DNase I is an endonuclease that
hydrolyzes double-stranded DNA prefer-
entially at sites adjacent to pyrimidine
nucleotides, its use in DNA fragmenta-
tion may result in bias (particularly for
genes with high G+C or high A+T con-
tent) at the step of template gene diges-
tion. The effects of this potential bias on
the overall mutation rate and recombina-
tion frequency have not yet been investi-
gated, but they may be avoided by using
the WalkThrough approach.
4. One of the key steps in this technique is
to control the 3' end of the nascent, sin-
gle-strand DNA synthesized during the
WalkThrough process. Under certain
conditions, this step may also be used
for efficient terminal and/or internal in-
sertion/deletion, resulting in molecules
with different sizes.
By modifying reaction conditions, PCR
can be adjusted for the WalkThrough syn-
thesis using thermostable polymerase for
the short, nascent DNA fragments. To
adapt PCR to the WalkThrough synthesis
provides a more convenient way for more
F
i
g
.
7
.
2
T
h
e
W
a
l
k
T
h
r
o
u
g
h
t
e
c
h
n
i
q
u
e
:
a
n
o
v
e
r
a
l
l
s
c
h
e
m
e
.
nascent DNA fragments, and also makes
the technique more robust. This has been
demonstrated by recombining human pla-
cental alkaline phosphatase (hPAP) and
calf intestinal alkaline phosphatase (cIAP)
for better mutants with improved molecu-
lar properties [27].
7.2.3.1 General Steps of Directed Molecular
Evolution
The gene encoding the protein of interest
is mutated to create a library of mutant
genes, the expression of which leads to a
library of mutant proteins. These protein
variants are screened or selected for a de-
sired property or set of properties. The im-
proved variants are used for further cycles
of mutagenesis/recombination and screen-
ing/selection.
The steps of WalkThrough recombina-
tion involving: 1) the generation of a nu-
cleic acid fragment ladder by nucleic acid
synthesis in the presence of a reaction
mixture comprising template polynucleo-
tides, primers, an enzyme having nucleic
acid-synthesizing activity and a mixture of
nucleic acid chain-terminating and nucleic
acid chain-elongating molecules; 2) remov-
ing the chain-terminating molecules; and
3) re-assembly of the polynucleotide by hy-
bridizing these fragments to one another
or to template polynucleotides in the pres-
ence of a thermostable DNA polymerase,
primers, and a reaction mixture of nucleic
acid chain-elongating molecules.
7.3
Examples
7.3.1
Engineering PCR Performance in B-type
DNA Polymerase from Thermococcus
aggregans
7.3.1.1 Introduction
Application of B-type DNA polymerases in
PCR PCR is an in-vitro method for the
amplification of specific DNA fragments
of defined length and sequence from small
amounts of target nucleic acid. It has be-
come a powerful tool in biological re-
search, medical diagnostic and the devel-
opment of biopharmaceuticals. Due to
their stability at high temperatures, ther-
mostable DNA polymerases are applied in
PCR. Depending on the specific applica-
tions, DNA polymerases with particular
properties are used. The thermostable
DNA polymerase from Thermus aquaticus
(Taq polymerase) has been the first recom-
binant polymerase used in PCR. Due to
their proofreading activity, several thermo-
stable B-type DNA polymerases are applied
in PCR, including Pyrococcus furiosus DNA
polymerase (Pfu polymerase) or Thermococ-
cus gorgonarius DNA polymerase (Tgo poly-
merase).
DNA polymerases can be classified into
at least three families based on their ami-
no acid sequences [28]. Based on the se-
quence homologies to E. coli polymerases
I, II, and III, the families A, B, and C are
defined, respectively. Family B-type DNA
polymerases such as T4 phage DNA poly-
merase, P. furiosus DNA polymerase or
bacteriophage u29 DNA polymerase are
usually monomeric enzymes and possess
a 5' ?3' polymerase activity and a proof-
reading 3' ?5' exonuclease activity. Poly-
merase activity and exonuclease activity
7.3 Examples 709
are synthetic and degradative activities act-
ing on the same substrate, the double-
stranded primer-template DNA.
Thermostable B-type DNA polymerases,
with their associated 3' ?5' exonuclease
activity have found use in PCR applica-
tions requiring high-fidelity amplification.
The 3' ?5' exonuclease activity is responsi-
ble excising mismatched bases in the elon-
gated primer strand. A DNA polymerase
lacking the 3' ?5' proofreading exonu-
clease activity (e.g., Taq DNA polymerase)
can only extend or fail to extend the mis-
matched primer, but cannot correct the er-
ror. Therefore, exonucleolytic proofreading
activity enhances the fidelity of a polymer-
ase, expressed as proportion of misincor-
poration. For Taq DNA polymerase error
rates of 10
4
are reported, whereas Pfu
DNA polymerase a DNA polymerase
with proofreading activity has error rates
of 10
6
. The use of DNA polymerases with
high fidelity in PCR is important for mini-
mizing amplification errors in products
that will be cloned, sequenced, and ex-
pressed. Additionally, a higher yield of
PCR product can be obtained using proof-
reading polymerases. Long PCR fragments
can be amplified using blends of DNA
polymerases with high polymerase proces-
sivity (e.g., Taq DNA polymerase) and
small amounts of B-type DNA polymerase
contributing the proofreading activity [29].
The B-type DNA polymerase of the hy-
perthermophilic T. aggregans (Tag polymer-
ase) has been shown to be suitable for ap-
plications in PCR. However, the wild-type
polymerase is able to yield only low
amounts of PCR product. Additionally, its
application is restricted to the amplifica-
tion of short DNA fragments. In order to
improve the PCR performance of the poly-
merase, mutations were designed based
on the results of mutational analysis ob-
tained with u29 DNA polymerase and Sso
DNA polymerase two polymerases that
are not suitable for PCR. Here, it is shown
that the simultaneous engineering of two
different enzymatic activities within an en-
zyme can improve its application in artifi-
cial biochemical systems such as the PCR.
Structural features of B-type polymerases
Several crystal structures of B-type DNA
polymerases are available. Structures of
the bacteriophage RB69 polymerase were
solved for the apoenzyme and the enzyme
in complex with a DNA primer-template
[30, 31]. Recently, the structure of RB69
polymerase encountering an abasic site
has been determined [32]. Additionally, the
structures of archaeal B-type DNA poly-
merases have been solved from the hy-
perthermophilic archaea T. gorgonarius
[33], Desulforococcus strain Tok [34], Ther-
mococcus sp. 98N-7 [35], and P. kodakaraen-
sis KOD1 [36]. In the structure of the B-
type DNA polymerases distinct domains
were identified: the N-terminal domain,
the exonuclease domain and the polymer-
ase domain including the Palm and Fin-
gers subdomains and the Thumb subdo-
mains. The folding of the domains forms
a disc with three distinct clefts extending
from a central hole. The structure of the
RB69-polymerase containing a primer-tem-
plate demonstrates that one cleft binds
double-stranded primer-template and the
other cleft binds single-stranded template.
A third cleft is perpendicular to the other
two and represents the 3' ?5' exonuclease
cleft. A highly conserved amino acid motif
Y-GG/A was identified in B-type DNA
polymerases. This is located in a solvent-
accessible loop between the polymerase
domain and the exonuclease domain, and
is responsible for the coordination be-
tween the synthesis and degradation in B-
type DNA polymerases. The role of the Y-
GG/A motif was analyzed in the meso-
phile B-type DNA polymerase of bacter-
iophage u29 polymerase. Mutations in this
motif affect the polymerization/exonucleo-
lysis balance due to its importance for the
DNA template binding stability [37]. In an-
other approach, the motif was studied with
the B-type DNA polymerase of the hyper-
thermophilic archaeon Sulfolobus solfatari-
cus (Sso polymerase). A sequence of 70
amino acids involved in enzymeDNA in-
teraction was found to contain the Y-GG/A
motif. A mutational analysis indicated that
this sequence motif is involved in deter-
mining the processivity of the proofread-
ing function [38]. The studies of the Y-GG/
A motif in u29 polymerase and Sso poly-
merase revealed that the coordination of
polymerase activity and exonuclease activ-
ity is strongly affected by specific muta-
tions in the sequence motif. For example,
the mutation Tyr ?Phe results in polymer-
ase mutants with increased polymerase/
exonuclease ratio, whereas the mutation
Tyr ?Ser leads to mutants with low poly-
merase activity and low polymerase/exonu-
clease ratio [37, 38].
7.3.1.2 Results
Engineering of enzymatic activities of Tag
DNA polymerase The B-type DNA poly-
merase of the hyperthermophilic T. aggre-
gans has been cloned, sequenced and ex-
pressed in E. coli [39]. The wild-type Tag
polymerase was shown to be suitable for
applications in PCR. However, only low
yields of PCR products were obtained, and
the application is restricted to the amplifi-
cation of short DNA fragments. In order
to improve the PCR performance, muta-
tions of the Y-GG/A motif were designed
based on the results obtained with mu-
tants of u29 polymerase and Sso polymer-
ase. By sequence alignment the Y-GG/A
motif was identified in the positions 387
389 of the amino acid sequence of Tag
polymerase. Analogous to the mutants de-
scribed for u29 polymerase and Sso poly-
merase [37, 38], the mutations Y387F,
Y387S, and G389A were introduced into
the gene of Tag polymerase by site-directed
mutagenesis. Additionally, the mutants
Y387H, Y387W, and Y387N were generat-
ed. Mutant enzymes were expressed in E.
coli, purified, and their polymerase and
exonuclease activities measured [40]. The
polymerase activity of the wild-type and
mutated enzymes were measured in a
primer extension assay using M13mp9
DNA as substrate. Exonuclease activity was
measured using
3
H-labeled calf thymus
DNA as substrate (Table 7.1).
Based on their polymerase activities, mu-
tants were defined as: 1) mutants with en-
hanced activity (Y387F); 2) mutants with ac-
tivities similar to the wild-type enzyme
(Y387W, Y387H); and 3) mutants with re-
duced activity (Y387N, Y387S, G389A). Ac-
cording to their exonuclease activities, two
groups were defined: 1) mutants having a
wild-type like activity (Y387F, Y387W,
Y387H); and 2) mutants with enhanced ac-
tivity (Y387N, Y387S, G389A). Based on the
ratio of polymerase and exonuclease activ-
ities, the Tag mutants were divided in two
groups: 1) mutants with a ratio similar to
that of the wild-type polymerase (Y387F,
Y387W, Y387H); and 2) mutants with a ratio
<0.1 (Y387N, Y387S, G389A). Mutants of
the first group (including the wild-type poly-
merase) contain an aromatic amino acid in
position 387, while members of the second
group have a non-aromatic amino acid in
this position. Previously, this was shown
for u29 polymerase and Sso polymerase, in
which the mutation Tyr ?Phe resulted in
a high polymerase/exonuclease ratio [37,
38]. A similar result was obtained for the
Y387F mutant of Tag polymerase.
7.3 Examples 711
Performance in PCR Wild-type DNA poly-
merase and the mutant enzymes were
tested for performance in PCR. Fragments
of different lengths were amplified using k
DNA as target. All enzymes except the
mutant G389A were able to perform PCR.
The five mutants with amino acid ex-
changes at position 387 yielded different
amounts of PCR products. In the amplifi-
cation of a 0.5 kb DNA fragment, the mu-
tants Y387H, Y387F, Y387S, and Y387W
showed a significantly higher yield of PCR
product than the wild-type enzyme. The
mutants Y387S and Y387N could amplify
fragments up to 2.5 kb length. The mu-
tants Y387H, Y387F, and Y387W were able
to amplify fragments of up to 7.5 kb
length, but not the wild-type polymerase
and the mutants Y387S and Y387N [41].
As some of the mutations were shown to
alter the exonucleolytic activity of the poly-
merase, the error rates of the wild-type Tag
polymerase and its mutants were deter-
mined under PCR conditions. In a lacI-
based assay, an error rate of 5.010
6
was
found for the wild-type enzyme, while the
mutants with enhanced exonuclease activ-
ity (Y387N and Y387S) revealed *10-fold
improved fidelity rates. Mutated enzyme
with a wild-type-like exonuclease activity
exhibited similar error rates in PCR as the
wild-type polymerase [40].
7.3.1.3 Discussion
In a protein engineering project, point mu-
tations were designed and generated by site-
directed mutagenesis in the Y-GG/A motif
of the thermostable archaeal B-type DNA
polymerase from T. aggregans. Based on
the X-ray structure of Tgo DNA polymerase,
a model of the structure of the Tag DNA
polymerase was built. The sequence motif
is located in a loop between the polymerase
domain and the exonuclease domain. It co-
ordinates the enzymatic elongation and de-
gradation of the single-stranded primer
DNA bound to a template DNA. The muta-
tions were generated based on results pre-
viously obtained with mutants of the ho-
mologous polymerases from S. solfataricus
and bacteriophage u29 two polymerases
that are not suitable for PCR. For the ther-
mostable Tag polymerase it is shown that
the simultaneous engineering of different
enzymatic activities can be used to improve
its application in artificial biochemical sys-
tems such as PCR.
7.3.2
Directed Evolution of Calf Intestinal Alkaline
Phosphatase with WalkThrough
Recombination
7.3.2.1 Introduction
Alkaline phosphatases (AP; EC 3.1.3.1) are
dimeric, zinc-containing, non-specific
phosphomonoesterases which are found in
Table 7.1 DNA polymerase (pol) and exonuclease (exo) activities
of Tag polymerase wild-type enzyme and mutants.
a)
Protein WT Y387F Y387W Y387H Y387N Y387S G389A
Pol (%) 100 160 92 93.6 6.4 17.8 10.7
Exo (%) 100 90 71 98 205 187 236
Pol/Exo ratio 1 1.77 1.29 0.96 0.03 0.09 0.04
a) Data expressed as percentage of the wild-type enzyme activity [40].
all organisms [42]. These enzymes are
characterized by a high pH optimum and
broad substrate specificity. In human and
higher animals, the AP family consists of
two classes: 1) the tissue-specific APs (pla-
cental AP, germ cell AP, and intestinal
AP); and 2) the tissue non-specific APs
(mainly located in the liver, kidney and
bones) [4345].
Through amino-terminal sequencing of
purified AP fractions, Besman and Cole-
man [46] in 1985 proved the existence of
two IAP isoenzymes in bovine intestine,
the cIAP from calf intestine and the bIAP
from intestine of the mature cow. A clear
difference at the amino terminus was de-
scribed between bIAP and cIAP. In 1993,
Weissig et al. [47] successfully cloned bIAP
and intensively characterized this enzyme;
however, cIAP from calf intestine with spe-
cific activities of about 8000 U mg
1
was
not further characterized until the gene
was cloned and expressed in CHO cells
[48] and Pichia pastoris [49].
Native cIAP is a metalloenzyme which
is quite different from E. coli AP. In addi-
tion, the native cIAP processes a different
primary structure, it being a dimeric glyco-
protein of molecular weight 138 kDa [50].
The carbohydrate content accounts *12%
of the total protein weight, whereas E. coli
AP does not contain sugars and has a mo-
lecular weight of only 86 kDa. Moreover,
cIAP has two buried cysteines per mole-
cule of dimer, whereas E. coli AP has no
free SH groups. Even though the structure
of cIAPs active site is probably similar to
that of bacterial AP, rational cloning and
expressing cIAP in E. coli had resulted in
inactive protein, and all attempts function-
ally to express this enzyme in different E.
coli strains were unsuccessful.
Since there was no 3-D structure of
cIAP molecule available, we have chosen
an irrational design approach and recom-
bined cIAP with hPAPs using the Walk-
Through recombination technique in order
to obtain clones which produce the highly
active cIAP in E. coli. Such clones would
be essential for the economic production
of this enzyme, which is due to its very
high specific activity particularly attrac-
tive for biotechnological applications such
as enzyme conjugates for diagnostic re-
agents, protein modifications, or the de-
phosphorylation of DNA.
7.3.2.2 Directed Evolution of cIAP using
WalkThrough Recombination
Recombining (hPAP) and cIAP Walk-
through recombination has been used to
recombine hPAP and cIAP. In addition,
point mutations have also been introduced
during the recombination steps. This mu-
tagenesis and recombination process con-
sists mainly of fragment synthesis, tem-
plate removal, terminator removal, and re-
assembly and amplification steps.
Fragment synthesis is usually a DNA
synthesis reaction where extension is ter-
minated by the incorporation of dideoxy-
nucleotides. The cIAP and hPAP DNA
genes served as templates in this reaction.
The primers were chosen depending on
the length of the two DNA sequences and
the length of the resulting fragments. The
lengths of the synthesized fragments were
controlled by the reaction conditions. A cy-
cle-sequencing reaction was also used for
the fragment synthesis because of its high-
er product yield and easier handlings.
To avoid interference during the reas-
sembly step, the template DNA was com-
pletely removed after this fragment synthe-
sis. In this example, uracil-containing
cIAP and hPAP genes prepared by a PCR
reaction using dUTP instead of dTTP were
used. After the U-DNAs had served as the
7.3 Examples 713
template for the fragment synthesis, the
whole reaction was subjected to uracil
DNA glycosylase treatment to remove the
uracil base at any site where a deoxyuridy-
late had been incorporated. The resulting
abasic sites were subsequently hydrolyzed
by heat treatment.
Since the synthesized fragments termi-
nated with dideoxynucleotides, they do not
carry free 3'-OH ends. These terminators
can be removed by 3' ?5' exonuclease ac-
tivity of several DNA polymerases or nu-
cleases, resulting in 3'-OH ends. In our
case, terminator removal was combined
with the reassembly by employing a mix-
ture of thermostable exonuclease III and
Taq polymerase. Exonuclease III removes
the terminators with its 3' ?5'-exonuclease
activity, and Taq polymerase extends the di-
gested fragment with its 5' ?3'-polymer-
ase activity.
Reassembly is a PCR-like reaction with-
out primers where the fragments anneal
to each other based on their homologies,
and are extended. Through 10 to 40 cycles
of denaturation, annealing and extension,
the WalkThrough fragments grew or reas-
sembled to the full length of the parental
DNA sequences. Amplifying the reas-
sembled DNA took place in a PCR reac-
tion with sequence flanking primers in or-
der to provide enough material for subse-
quent cloning, sequence analysis, and en-
zyme expression. The amplification
product was purified through preparative
gel electrophoresis, cloned into an expres-
sion vector, and expressed in E. coli [27].
Analysis of the random cIAP variants Thir-
teen recombination clones were randomly
chosen and analyzed by sequencing. The
sequence analysis results are summarized
in Fig. 7.3. All these clones contained one
or more recombination events within the
sequenced region. The recombination rate
varied from 2.7 to 7.3 crossovers per kbd
within the analyzed region. The overall
mutation rate was about 0.30% (*3 muta-
tions per kb).
Evolving the expressibility of cIAP in E. coli
In order to screen for expressible cIAP
clones in E. coli, a pool of about 500 hy-
brid hPAP and cIAP transformants was
chosen for further functional evaluation.
All clones were transferred into 96-well
microtiter plate and grown at 37 8C. The
AP activity of each individual cell extract
was prepared and measured by following
the increasing absorbance at 405 nm and
37 8C after adding 30 mM p-nitrophenyl
phosphate (pNPP) as the substrate in a
buffer containing 1.0 M diethanolamine
buffer (pH 9.8), 1 mM MgCl
2
, and 20 mM
ZnCl
2
.
This screening effort led to several cIAP
variants which can produce the highly ac-
tive cIAP in E. coli (Fig. 7.4). Amino acid
substitutions of the cIAP variants are: mu-
tant 1A2B, 1A4B, and 2C9D. All three
cIAP variants contain a short N-terminal
substitution, L20A mutation, and a stretch
of C-terminal replacement. In addition,
clones 1A2B and 1A4B also contain T71A
point mutations.
The cIAP variant 1A4B was purified
with a similar procedure, described by
Beck and Burtscher [51]. The activity of
the purified 1A4B was determined accord-
ing to Mssner et al. [52], but was con-
ducted at 37 8C rather than at 25 8C. The
protein concentration was determined by
measuring the absorbance of protein solu-
tion at 280 nm against buffer. The specific
activity was calculated by forming a quo-
tient of activity relative to the accompany-
ing amount of protein. The specific activ-
ities of the native cIAP and recombinant
hPAP were also determined at the same
time, with the same assay procedure. The
specific phosphatase activity of cIAP vari-
ants expressed in E. coli compared with
those of native cIAP and recombinant
hPAP is summarized in Fig. 7.5. The cIAP
variant 1A4B has a slightly higher specific
activity than that of the native cIAP.
7.3 Examples 715
Fig. 7.3 The sequence analysis results of native ciap and recombinant hpap genes
and the unscreened, recombinant random variants. Mutations and crossovers are
plotted versus the sequence position.
Fig. 7.4 Schematic description of the amino acid residue sub-
stitutions within the screened calf intestinal alkaline phospha-
tase (cIAP) variants which are functionally expressed in E. coli.
| represents synonymous nucleotide substitution.
7.3.2.3 Discussion
Amino acid substitutions in amino acid
positions 20, 71, and 507534 of the native
cIAP shown in Fig. 7.4 are particularly
relevant to functional expression in E. coli.
The amino acid substitution L20A is cru-
cial for the phosphatase expressibility. This
substitution may improve cIAP transloca-
tion and mature to an active phosphatase
in E. coli. Since replacement of amino acid
residues 507534 resulted in a functionally
expressed cIAP in E. coli, we speculate that
the region of 507534 should represent
the so-called glycosylphosphatidylinositol
anchor [53, 54] within the native cIAP.
How the T71A substitution contributes to
the functional expression in E. coli is still
under investigation.
Native cIAP is a non-sialylated glycopro-
tein. Several cIAP variants can be ex-
pressed in E. coli in an active form and
even with higher specific activity, which
showed that glycosylation is not essential
for its enzymatic activity. In differing from
most other asialoglycoproteins, the native
cIAP possesses a glycosylphosphatidylino-
sitol anchor, though this has been replaced
in all active cIAP-E. coli variants, suggest-
ing that the anchor may prevent or impair
cIAP folding in E. coli.
The improved cIAP variant may be essen-
tial for the economic production of this en-
zyme, which is due to its very high specif-
ic activity particularly attractive for bio-
technological applications such as enzyme
conjugates for diagnostic reagents, protein
modification, or dephosphorylation of
DNA. More recently, cIAP has been shown
as being able to detoxify the lipopolysacchar-
ide (LPS)-mediated inflammatory response
[55]. This has provided a new prospect for
cIAP to become a novel therapeutic drug
in the treatment of Gram-negative sepsis
and other LPS-mediated diseases.
Fig. 7.5 Specific activity comparison of the native calf intest-
inal alkaline phosphatase (cIAP), the cIAP variant 1A4B, and
recombinant human placental alkaline phosphatase (hPAP).
7.4
Summary
Today, many routine tests used in the diag-
nosis and management of disease are based
on bioreaction and biorecognition mole-
cules, such as enzymes and antibodies
(see Introduction, and Part V, Chapter 6).
How to improve the crucial properties of
these molecules ultimately determines
whether an enzyme or antibody can be suc-
cessfully used for innovative diagnostic pro-
cesses, lower manufacturing costs, and ro-
bust assay applications (see Part V, Chapter
2). With an increasing number of 3-D pro-
tein structures becoming available in data-
bases, together with the rapid development
of powerful molecular modeling tools, ra-
tional design will become both more effi-
cient and more broadly applicable (see Part
V, Chapter 3). Moreover, with new creative
strategies to increase sequence diversities,
in addition to novel high-throughput
screening techniques, directed evolution of-
fers a great potential to produce new and
much more active and/or specific enzyme
variants (see Part III, Chapter 6). With these
tools, we are also now able to significantly
speed up the development processes for
other biopharmaceuticals, and to meet
broader applications and steadily increasing
performance challenges.
References
1 Ryu, D. D. Y., Nam, D.-H. (2000) Recent Pro-
gress in Biomolecular Engineering. Biotechnol-
ogy Prog. 16, 216.
2 Marshall, S. A., Lazar, G. A., Chirino, A. J.,
Desjarlais, J. R. (2003) Rational design and en-
gineering of therapeutic proteins. Drug Discov-
ery Today 8, 212221.
3 Chen, R. (2001) Enzyme engineering: rational
redesign versus directed evolution. Trends Bio-
technol. 19, 1314.
4 Nixon, A. E., Firestine, S. M. (2000) Rational
and Irrational Design of Proteins and Their
Use in Biotechnology. IUBMB Life (Interna-
tional Union of Biochemistry and Molecular
Biology: Life) 49, 181187.
5 Tobin, M. B., Gustafsson, C., Huisman, G. W.
(2002) Directed evolution: the rational basis
for irrational design. Curr. Opin. Struct. Biol.
10, 421427.
6 Arnold, F. (2001) How enzymes adapt: lessons
from directed evolution. Trends Biochem. Sci.
26, 100106.
7 Chirumamilla, R. R., Reddivari, M., Roger, M.,
Poonam, N. (2001) Improving the quality of
industrially important enzymes by directed
evolution. Mol. Cell. Biochem. 224, 159168.
8 Tao, H., Cornish, V. W. (2002) Milestones in
directed enzyme evolution. Curr. Opin. Chem.
Biol. 6, 858864.
9 Lassner, M. W., McElroy, D. (2002) Directed
Molecular Evolution: Bridging the Gap be-
tween Genomics Leads and Commercial Prod-
ucts. OMICS: J. Integrative Biol. 6, 153162.
10 Turner, N. J. (2003) Directed evolution of en-
zymes for applied biocatalysis. Trends Biotech-
nol. 21, 474478.
11 Shao, Z., Schmuck, R., Kratzsch, P., Kenklies,
J., Weisser, H. (2003) Variants of an Erwinia-
type creatinase. EP 1298213.
12 Kratsch, P., Schmuck, R., Bunk, D., Shao, Z.,
Thym, D., Knappe, W.-R. (2003) Variants of
soluble pyrroloquinoline quinone-dependent
glucose. EP 1332216.
13 Zoller, M. J. (1992) New recombinant DNA
methodology for protein engineering. Curr.
Opin. Biotechnol. 3 (4), 348354.
14 Shao, Z., Arnold, F. H. (1996) Engineering
new functions and altering existing functions.
Curr. Opin. Struct. Biol. 6, 513518.
15 Moore, J. C., Arnold, F. (1996) Directed evolu-
tion of para-nitrobenzyl esterase for aqueous-
organic solvents. Nature Biotechnol. 14, 458
467.
16 Reidhaar-Olson, J. F., Sauer, R. T. (1988) Com-
binatorial cassette mutagenesis as a probe of
the informational content of protein se-
quences. Science 241, 5357.
17 Sneeden J. L. and Loeb, L. A. (2003) Random
oligonucleotide mutagenesis. In: Directed Evo-
lution Library Creation (Eds. F. H. Arnold and
G. Georgiou), pp. 6573. Humana Press.
References 717
18 Stemmer, W. P. C. (1994a) Rapid evolution of
a protein in vitro by DNA shuffling. Nature
370, 389391.
19 Stemmer, W. P. C. (1994b) DNA shuffling by
random fragmentation and reassembly in vi-
tro recombination for molecular evolution.
Proc. Natl. Acad. Sci. USA 91, 1074710751.
20 Shao, Z., Zhao, H., Giver, L., Arnold, F. H.
(1998) Random-priming in vitro recombina-
tion: an effective tool for directed evolution.
Nucleic Acids Res. 26, 681683.
21 Nguyen, A. W., Daugherty, P. S. (2003) Produc-
tion of randomly mutated plasmid libraries
using mutator strains. In: Directed Evolution
Library Creation (Eds. F. H. Arnold and G.
Georgiou), pp. 3944. Humana Press.
22 Selifonova, O., Schellenberger, V. (2003) Evolu-
tion of microorganisms using mutator plas-
mids. In: Directed Evolution Library Creation
(Eds. F. H. Arnold and G. Georgiou), pp. 45
52. Humana Press.
23 Zhao, H., Giver, L., Shao, Z., Affholter, J.,
Arnold, F. H. (1998) Molecular evolution by
staggered extension process (StEP) in vitro re-
combination. Nature Biotechnol. 16, 258261.
24 Shao, Z. (1999) Engineering novel biomole-
cules with artificial recombination. Protein Sci.
8 (Suppl. 2), 55.
25 Volkov, A., Shao, Z., Arnold, F. H. (1999) Re-
combination by in vitro heteroduplex forma-
tion and in vivo repair. Nucleic Acids Res. 27,
e18:ivi.
26 Lee, S. H., Ryu, E. J., Kang, M. J., Wang, E.-S.,
Piao, Z., Choi, Y. J., Jung, K. H., Jeon, J. Y. J.,
Shin, Y. C. (2003) A new approach to directed
gene evolution by recombined extension on
truncated templates (RETT). J. Mol. Catalysis
B: Enzymatic 26, 119129.
27 Shao, Z., Kratzsch, P., Schmuck, R., von der
Eltz, H., Kenklies, J. (2003) A WalkThrough
technique for in vitro/in vivo recombination of
polynucleotide sequences. EP1404825.
28 Braithwaite, D. K., Ito, J. (1993) Compilation,
alignment, and phylogenetic relationships of
DNA polymerases. Nucleic Acids Res. 21, 787
802.
29 Barnes, W. M. (1994) PCR amplification of up
to 35-kb DNA with high fidelity and high
yield from lambda bacteriophage templates.
30 Wang, J., Sattar, A. K., Wang, C. C., Karam,
J. D., Konigsberg, W. H., Steitz, T. A. (1997)
Crystal structure of a pol alpha family replica-
tion DNA polymerase from bacteriophage
RB69. Cell 89, 10871099.
31 Franklin, M. C., Wang, J., Steitz, T. A. (2001)
Structure of the replicating complex of a pol
alpha family DNA polymerase. Cell 105, 657
667.
32 Hogg, M., Wallace, S. S., Doublie, S. (2004)
Crystallographic snapshot of a replicative
DNA polymerase encountering an abasic site.
EMBO J. 23, 14831493.
33 Hopfner, K. P., Eichinger, A., Engh, R. A.,
Laue, F., Ankenbauer, W., Huber, R., Angerer,
B. (1999) Crystal structure of a thermostable
type B DNA polymerase from Thermococcus
gorgonarius. Proc. Natl. Acad. Sci. USA 96,
36003605.
34 Zhao, Y., Jeruzalmi, D., Moarefi, I., Leighton,
L., Lasken, R., Kuriyan, J. (1999) Crystal struc-
ture of an archaebacterial DNA polymerase.
Structure 7, 11891199.
35 Rodriguez, A. C., Park, H.-W., Mao, C., Beese,
L. S. (2000) Crystal structure of a pol alpha
family DNA polymerase from the hyperther-
mophilic archaeon Thermococcus sp. 9 degrees
N-7. J. Mol. Biol. 299, 447462.
36 Hashimoto, H., Nishioka, M., Fujiwara, S., Ta-
kagi, M., Imanaka, T., Inoue, T., Kai, Y. (2001)
Crystal structure of DNA polymerase from hy-
perthermophilic archaeon Pyrococcus kodaka-
raensis. J. Mol. Biol. 306, 469477.
37 Truniger, V., Blanco, L., Salas, M. (1999) Role
of the YxGG/A motif of Phi29 DNA poly-
merase in protein-primed replication. J. Mol.
Biol. 286, 5769.
38 Pisani, F. M., De Felice, M., Rossi, M. (1998)
Amino acid residues involved in determining
the processivity of the 3' ?5' exonuclease ac-
tivity in a family B DNA polymerase from the
thermoacidophilic archaeon Sulfolobus solfatari-
cus. Biochemistry 37, 1500515012.
39 Niehaus, F., Frey, B., Antranikian, G. (1997)
Cloning and characterisation of a thermo-
stable alpha-DNA polymerase from the hyper-
thermophilic archaeon Thermococcus sp. TY.
Gene 204, 153158.
40 Bhlke, K., Pisani, F. M., Vorgias, C. E., Frey,
B., Sobek, H., Antranikian, G. (2000) PCR
performance of the B-type DNA polymerase
from the thermophilic euryarchaeon Thermo-
coccus aggregans improved by mutations in the
Y-GG/A motif. Nucleic Acids Res. 28, 3910
3917.
41 Sobek, H., Frey, B., Antranikian, G., Bhlke,
K., Pisani, F., Mossi, R. (2001) Mutant B-type
DNA polymerases exhibiting improved perfor-
mance in PCR. EP 1132474.
42 McComb, R. B., Bowers, G. N., Posen, S.
(1979) Alkaline Phosphatases. Plenum Press,
New York, NY.
43 Millan, J. L. (1988) Oncodevelopmental expres-
sion and structure of alkaline phosphatase
genes. Anticancer Res. 8, 9951004.
44 Harris, H. (1990) The human alkaline phos-
phatases: what we know and what we dont
know. Clin. Chim. Acta 186, 133150.
45 Schneidinger, B., Meier, T., Schmuck, R.,
Shao, Z. (2004) Conjugate of a tissue non-spe-
cific alkaline phosphatase and dextran, process
for its production and use thereof.
EP1405907.
46 Besman, M., Coleman, J. E. (1985) Isozymes
of bovine intestinal alkaline phosphatase. J.
Biol. Chem. 260, 1119011193.
47 Weissig, H., Schildge, A., Hoylaerts, M. F.,
Iqbal, M., Millan, J. L. (1993) Cloning and ex-
pression of the bovine intestinal alkaline phos-
phatase gene: biochemical characterization of
the recombinant enzyme. Biochem. J. 290,
503508.
48 Manes, T., Hoylaerts, M. F., Mller, R., Lott-
speich, F., Hlke, W., Millan, J. L. (1998) Ge-
netic complexity, structure, and characteriza-
tion of highly active bovine intestinal alkaline
phosphatases. J. Biol. Chem. 273, 23353
23360.
49 Hlke, W., Mller, R., Burtscher, H., Millan,
J. L. (2002) Highly active alkaline phosphatase.
US6406899.
50 Fosset, M., Chappelet-Tordo, D., Lazdunski,
M. (1974) Intestinal alkaline phosphatase.
Physical properties and quaternary structure.
51 Beck, R., Burtscher, H. (1994) Expression of
human placental alkaline phosphatase in
Escherichia coli. Protein Expr. Purif. 5(2), 192
197.
52 Mssner, E., Boll, M., Pfleiderer, G. (1980)
Purification of human and bovine alkaline
phosphatases by affinity chromatography.
Hoppe Seylers Z. Physiol. Chem. 361, 543549.
53 Low, M. G., Zilversmit, D. B. (1980) Role of
phosphatidylinositol in attachment of alkaline
phosphatase to membranes. Biochemistry 19,
39133918.
54 Medof, M. E., Nagarajan, S., Tykocinski, M. L.
(1996) Cell-surface engineering with GPI-an-
chored proteins. FASEB J. 10, 574586.
55 Beumer, C., Wulferink, M., Raaben, W., Fiech-
ter, D., Brands, R., Seinen, W. (2003) Calf in-
testinal alkaline phosphatase, a novel thera-
peutic drug for lipopolysaccharide (LPS)-
mediated diseases, attenuates LPS toxicity in
mice and piglets. J. Pharmacol. Exp. Ther. 307,
737744.
References 719
ISBN: 3-527-31184-X
Part IV
Production of Biopharmaceuticals
Abstract
The first recombinant protein therapeutic
made in cultivated mammalian cells ob-
tained market approval in 1986. This event
made the use of Chinese hamster ovary
(CHO) cells in large-scale bioreactors
known to a wider public. These cells are
now the dominating host system for re-
combinant protein production, as more
than 60% of all new target proteins in the
clinical pipelines of pharmaceutical and
biotechnology companies are being pro-
duced in hamster-derived cells. This chap-
ter will cover aspects of gene transfer, cell
line development and process development
for mammalian protein expression sys-
tems using CHO cells as the main exam-
ple, but also making reference to other
mammalian cells that are used for the
large-scale production of therapeutic pro-
teins. Most importantly, the scientific and
technological insights that resulted in the
rapid and surprising yield improvements
from such processes, bringing the volu-
metric productivity of mammalian cell cul-
ture processes into the gram per liter
range, will be discussed. Not only is this
level of productivity equal to that of micro-
bial systems, but the recombinant proteins
in mammalian cells have all the necessary
secondary modifications that only a higher
eukaryote can execute. The regulatory
framework for the use of mammalian host
systems will also be discussed, as the per-
ceived risks of transmission of adventi-
tious agents to patients have resulted in
stringent rules to which all manufacturers
must adhere.
Abbreviations
ADCC antibody-dependent cellular cy-
totoxicity
BSE bovine spongiform encephalo-
pathy
cGMP current Good Manufacturing
Practice
CIP clean in place
723
1
Manufacture of Recombinant Biopharmaceutical Proteins
by Cultivated Mammalian Cells in Bioreactors
Florian M. Wurm
ISBN: 3-527-31184-X
The Industrys Workhorses Mammalian Expression Systems
FISH fluorescence in situ hybridiza-
tion
GHT glycine, hypoxanthine, and thy-
midine
GS glutamate synthetase
MSCB Master Seed Cell Bank
MSX methionine sulphoximine
MTX methotrexate
MVM minute virus of the mouse
MWCB Manufacturers Working Cell
Bank
PCV packed cell volume
Pip pristinamycin-induced protein
PLA Process Licence Applications
PrP
sc
proteinaceous-infectious parti-
cles, scrapie
S/MARs scaffold or matrix attachment
regions
SEAP secreted alkaline phosphatase
SIP sterilize in place
UCOS ubiquitous chromatin opening
elements
1.1
Introduction
A small number of immortalized mamma-
lian cells have become host systems for
production of kilogram (ton) per year
quantities of complex recombinant pro-
teins for clinical applications. In 2004,
these cells will produce about 6070% of
all recombinant biopharmaceutical pro-
teins. The most popular cells are Chinese
hamster ovary (CHO), Mouse myeloma-de-
rived NS0 (NS0), Baby hamster kidney
(BHK) and human retina-derived immorta-
lized PER.C6 (PER.C6) cells (see also Part
IV, Chapter 3). One compelling reason
why mammalian cells are now so popular
is the exceptional productivity reported in
a few cases. Sophisticated processes with
highly optimized cell lines can provide
grams per liter of recombinant antibodies
or chimeric immunoglobulin-fusion pro-
teins, in extended batch cultures of vol-
umes of up to 12000 L. The know-how
and technology behind large-scale pro-
cesses for mammalian cells have evolved
over 20 years and have resulted in a more
than 100-fold improvement in volumetric
productivity (Table 1.1).
Hundreds of proteins are presently
being produced in mammalian cells for
clinical evaluation. The decision to employ
more complex in vitro cultures of recombi-
nant mammalian cells has been driven by
the need to obtain proteins with complex
biochemical structures and resulting supe-
rior biological activity, reflecting their na-
tive structure and function. Biological ac-
tivity and the pharmacokinetic characteris-
tics of recombinant proteins frequently de-
pend on a number of complex protein
modifications (i.e., proper folding, disul-
fide bridge formation, oligomerization,
proteolytic processing, phosphorylation
and the addition of specific and complex
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 724
Table 1.1 Volumetric productivities from cell culture processes de-
veloped for the production of recombinant proteins for clinical
use: observations over a 20-year period
Volumetric productivity titers of recombinant proteins in cell culture media
Period 19821985 19921995 20022004
Titer range 550 mg L
1
50500 mg L
1
5005000 mg L
1
carbohydrate groups). These and other
protein processing steps can be executed
with high efficiency by mammalian cells.
Thus, Escherichia coli and fungi the pro-
duction hosts of first choice during the
early years of applied DNA technology
are now playing only a minor role for
large-scale expression of proteins for clini-
cal use.
The terms expression and production
used herein refer to secreted proteins only.
To my knowledge, all biopharmaceutical
proteins from mammalian cells were de-
veloped from gene constructs that allow
secretion of the desired protein into the
culture medium. This is one of the most
intriguing advantages of protein produc-
tion in mammalian cell culture. The cell is
truly used as a machine that converts a
given DNA construct into a protein prod-
uct while assuring that the product is se-
creted and thus easily separated from the
majority of the cells contents. Fig. 1.1
shows an SDS-PAGE analysis of a full-
length recombinant antibody in the super-
natant from a CHO culture or in an E. coli
lysate. For the latter, a weak set of IgG-spe-
cific bands and a large number of bands
corresponding to bacterial proteins are ob-
served. In contrast, the antibody product
in the supernatant of CHO cells represents
a much higher percentage of the total pro-
tein.
The use of CHO and NS0 cells for the
large-scale production of recombinant pro-
teins has been facilitated by the fact that
both cell types grow well in single-cell sus-
pension culture with the highest cell den-
sities from an extended batch process re-
ported to be greater than 10
7
cells mL
1
.
This is about five times higher than was
seen 20 years ago. Optimized cell lines
can achieve secretion of 50 pg per cell per
day of recombinant protein. The record on
volumetric titer from recombinant mam-
malian cells (CHO) for a secreted antibody
made in an extended batch process stands
at 4.6 g L
1
[1].
CHO cells, originally derived from the
ovary of a female Chinese hamster,
emerged as a spontaneously immortalized
line, in the hands of Kao and Puck more
than 40 years ago [2, 3]. The cells have
been the object of many studies since
then, and are very well characterized with
respect to a variety of aspects, including
karyotype, chromosome structures, gene
mapping, general culture conditions, cell
physiology and media requirements. For
the purpose of genetic and physiological
studies on the activity of dihydrofolate re-
ductase (DHFR) and its genomic alleles,
during the late 1970s, Urlaub and Chasin
[4] mutagenized the original cell line with
the help of radiation and chemicals. The
goal was to create a mutant cell line lack-
ing functional activity of both alleles of the
DHFR gene locus. The resulting cell line
CHO-DUX (sometimes also referred to as
CHO-DUKX-B11), though not initially in-
Fig. 1.1 SDS-Polyacrylamide gel electrophoresis of
samples of supernatants and lysates, respectively,
of a Chinese hamster ovary process and of an E.
coli fermentation producing full-length human an-
tibodies. (Image provided by Genentech Inc.,
courtesy Drs. Brad Snedecor and Lynne Krum-
men)
tended for this purpose, proved to be an
ideal substrate for a number of pioneering
gene transfer experiments [57].
CHO cells were chosen as the host sub-
strate for the first recombinant protein
from mammalian cells, human tissue plas-
minogen activator (Activase
, tPA) [8],
which achieved market approval in the US
and other global markets in 1986/1987.
Subsequently, other companies chose this
production system because the approval
barriers for a second product from the
same host were considered easier to over-
come, especially with respect to the regula-
tory process in the USA. CHO cells are
unique in many ways and exhibit a full set
of advantages, ranging from ease of intro-
duction of exogenous DNA to capacity for
growth and high-level productivity at large
scales. An excellent selection of articles
concerning the scientific history, general
and cellular biology, cytogenetics and mo-
lecular biology of Chinese hamsters and
CHO cells can be found in a comprehen-
sive compendium edited by Gottesman [9].
For two decades now not only CHO cells
but also NS0 cells have spearheaded the
development of animal cell technology in
an unprecedented way, and are now the
basis of accessory industries that have de-
veloped around the clinical manufacturers.
Numerous companies provide complex
media formulations that boost the growth
and productivity of these cells for large-
scale operations, while eliminating unde-
fined mixtures of growth factors and nutri-
tional components, such as fetal bovine se-
rum or animal tissue-derived peptones.
Any host system different from CHO
will be subject to the same regulatory scru-
tiny. For obvious reasons, due to the evolu-
tionary relatedness of all mammalian spe-
cies, regulatory concerns for the transmis-
sion of unknown disease-causing princi-
ples is higher when utilizing a hamster-
derived cell line than when using a micro-
bial host. Safety concerns were recently
raised to a new level because of the trans-
mission of a bovine prion disease to hu-
mans, causing variant CreutzfeldJakob
disease in hundreds of consumers. This
article attempts to summarize arguments,
issues, advantages, questions and ongoing
research for the industrial production of
high value proteins derived from mamma-
lian cells. Section 1.2 provides an introduc-
tion to the principles of expression of re-
combinant proteins from mammalian
cells. Sections 1.3, 1.4, and 1.5 address, in
a more profound way, the molecular and
cellular biology of gene transfer and gene
amplification of recombinant DNA in
mammalian cells, with the emphasis being
on the genetic stability of recombinant
cells. Section 1.6 discusses process issues
for scale-up and manufacturing, and Sec-
tion 1.7 is an extensive discussion on regu-
latory aspects of these processes. The im-
portance of regulatory issues should not
be underestimated, because most of the
money and time invested in the develop-
ment of a manufacturing scheme based
on mammalian cells will go to addressing
the safety, consistency and quality of the
product.
Issues raised and discussed herein are
not meant to be comprehensive. However,
it is hoped that the most critical points im-
pacting developmental efforts for protein
production with mammalian cells in cul-
ture will be addressed.
1.2
Vectors, Transfections,
and Cell Line Generation
1.2.1
Calcium Phosphate DNA Coprecipitates
for Transfection
Graham and van der Eb [10] showed more
than 30 years ago that exposing cells to
micro-precipitates of DNA and calcium
phosphate allowed the transfer of DNA
into cultivated mammalian cells. This sim-
ple method was termed calcium phosphate
transfection, and a number of modified
versions have become widely used. With
optimized transfection conditions, up to
100% of cells take DNA into their cyto-
plasm. However, in addition to the transit
across the cells plasma membrane, the
transport of DNA from the cytoplasm to
the nucleus seems to be a significant bar-
rier. Axel and collaborators first showed
stable integration of transfected DNA into
chromosomes of mammalian cells [11, 12].
The number of emerging colonies upon
transfection by the classical calcium phos-
phate technique is low: usually between
0.05% and 1% of the transfected cells give
rise to recombinant colonies [13]. Recent
improvements of crucial physico-chemical
parameters of the calcium phosphate
transfection methodology have increased
the frequency of stable recombinant cells
to about 510% of the transfected popula-
tion. In addition, the efficiency of transient
transfection has increased to levels of 50%
or higher [1416]. Calcium phosphate
transfection is still a frequently used meth-
od for the generation of recombinant
CHO cells. Other methods, to be discussed
below, are also appropriate.
With calcium phosphate, and also with
other methods, a large excess of DNA mol-
ecules over the number of cells is nor-
mally used, in the region of 100000 or
more per cell. This is probably necessary,
as much of the DNA will not reach the
nucleus, due to degradation. However, it
has been shown that association of the
DNA with calcium phosphate protects
against nuclease attack during transport to
the nucleus. Eventually, these complexes
become dissociated and nuclear endo- and
exonucleases will have access to the
naked DNA. Moreover, it has been
shown that supercoiled plasmid DNA mol-
ecules will be converted into relaxed (sin-
gle strand cut) and linear molecules (dou-
ble strand cut) within the nucleus after 1
2 hours [17]. This is an essential step for
integration into the linear DNA backbone
of a chromosome.
The mechanism by which the DNA is
transported across the nuclear membrane
(nuclear pores are not large enough for
diffusive transport) and finally to the site
of chromosomal integration is not yet
known. There is evidence however that
disruption of the nuclear membrane dur-
ing mitotic activity of cells is an important
aspect of overcoming this barrier [18]. On
the receiving side of the process of integra-
tion, at least one strand of the chromo-
somal DNA needs to be opened while the
linear plasmid molecule(s) is in suffi-
ciently close proximity to integrate. Nucle-
ar ligases could then mediate covalent
linkages. Little is known about the mecha-
nism affecting the site specificity of inte-
gration. It is assumed that genome DNA
replication and repair could facilitate entry
of exogenous DNA into chromosomal
DNA.
The site of integration is a critical factor
since it influences the transcription rate of
the integrated DNA. The current assump-
tion is that exogenous DNA will integrate
randomly within the genome. This is also
true if the transfection cocktail contains
DNA with homology to sequence seg-
ments of the genome. Gene targeting ex-
periments in mammalian cells is ineffi-
cient; sequence-specific integration is a
rare event. Only 1 in 1000 to 1 in 10000
events result in targeted integration [19,
20]. With respect to actual sites of integra-
tion in CHO cells there are few data avail-
able. If integration in general is random,
only a small proportion of recombinant
cell lines should contain the transferred
DNA within transcriptionally active re-
gions of chromosomes, since very little
DNA of any higher eukaryote is tran-
scribed at any time. However, the fre-
quency in which exogenous DNA inte-
grates into transcriptionally inactive re-
gions of the genome cannot be determined
since cells that do not overcome the selec-
tion step cannot be analyzed.
We performed a small study using fluo-
rescence in situ hybridization (FISH) to de-
termine the integration sites in 12 clonal
cell lines following calcium phosphate
transfection [21]. Interestingly, we found
only a single integration site in each cell
line. There was no preference for a specif-
ic chromosomal region, but there may
have been a slight preference for larger
chromosomes (1 to 3). This bias, however,
is most likely due to the fact that these
three chromosomes represent a large frac-
tion of the genomic DNA.
Nuclear enzymes such as endo- and exo-
nucleases, but also ligases [22] and possi-
bly recombinogenic enzymes [23, 24] act-
ing on the population of plasmid mole-
cules within the nucleus, are responsible
for the modifications of transfected DNA
that eventually becomes integrated into
the genome. These modifications may not
only degrade many plasmid molecules, but
nuclear ligase activity is also responsible
for the creation of larger DNA complexes
containing numerous copies of the plas-
mid DNA. These DNA molecules seem to
be created before integration into the ge-
nome. They provide the physical basis a
genetic link between the selection marker
and the gene(s) of interest in those trans-
fections that utilize separate vectors. In
my laboratory, co-transfections are being
executed routinely, and we have found a
high degree of covalent linkage of all indi-
vidual plasmid molecules when analyzing
the integration sites of stable cell lines
[25]. One should be aware that co-integra-
tion of multiple plasmids is probably a
general phenomenon in eukaryotic cells.
Chen and coworkers reported the genera-
tion of transgenic rice plants receiving and
expressing 13 different plasmids out of 14
that were used in the DNA cocktail. Analy-
sis by Mendelian genetic approaches re-
vealed integration into one locus [26]. It
should be noted here that DNA transfer
methods other than calcium phosphate
might deliver different quantities of DNA
to the nucleus. This may have profound
effects on the structure and copy number
of integrated DNA molecules (see
Sect. 1.4).
Among individual cell lines from a sin-
gle transfection there is usually extensive
heterogeneity in productivity of the recom-
binant protein. The expression levels from
mammalian cell clones generally have a
very wide range, sometimes exceeding two
orders of magnitude [27, 28]. Numerous
laboratories have verified this observation
with CHO, NS0 [29] or PerC6 cells [30]. As
a consequence, the identification of high
producer cell lines is a tedious and labor-
intensive exercise, and requires the screen-
ing of hundreds of individual cell lines. It
is generally necessary to invest between 2
and 4 months of laboratory work into this
task. Only then can the upper range of
sustainable expression for a recombinant
protein be assessed.
1.2.2
Other DNA Transfer Methods
for Mammalian Cells
In addition to calcium phosphate transfec-
tion, methods for DNA transfer into mam-
malian cells by electroporation [31, 32] and
by transfection mediated through cationic
lipids, liposome [3335], biolistics [36] and
polymers [37, 38] have been developed.
Most of these techniques have been re-
ported to mediate higher transfection effi-
ciencies as compared to calcium phos-
phate-mediated DNA transfer. Such claims
must be regarded with some caution, as
all DNA transfer techniques established so
far suffer from high variability due to tech-
nical difficulties. Other factors to cause
major variations in transfection efficiency
are the type of cells used and the condi-
tion of the cells prior to transfection.
Since the mechanism for the transfer of
DNA to cells can vary widely from one
method to another, one suspects that very
different consequences may result within
the cell events over which the experi-
mentation has no or little control. For ex-
ample, the DNA transfer method may af-
fect the plasmid copy number average in
individual clones, depending on the
amount of DNA transferred to the nu-
cleus. Electroporation may result in a low-
er copy number of integrated DNA than
calcium phosphate-based transfection. In
view of the applied selection procedures
for the generation of recombinant cell
lines, very different structures of inte-
grated plasmid DNA may result.
Another factor to be taken into consid-
eration for the integration of plasmid
DNA is the physical form of the trans-
fected DNA. Since linearization of plasmid
DNA is a prerequisite for integration, the
form of the plasmid prior to transfection
will affect the integrated form. When cir-
cular plasmids are used, as was done in
the past, linearization is dependent on cel-
lular enzyme activity. Since eukaryotic
DNA degrading enzymes act on random
sequences, a significant number of mole-
cules will be linearized within the DHFR
sequence, jeopardizing the functionality of
the plasmid. Therefore, opening the circu-
lar plasmid molecule by restriction en-
zyme digestion with appropriate enzymes
prior to transfection is recommended for
improved transfection efficiencies in proto-
cols that aim at integration of plasmid
DNA.
We have seen in our laboratory that line-
arization of plasmids improves the effi-
ciency of stable transfections [15]. Linear
plasmid molecules are also used routinely
to transfer the DNA into a specific (homol-
ogous) DNA sequence of the host genome
(gene targeting). In these experiments, lin-
earization is executed in cutting the ho-
mologous DNA fragment approximately in
half. The molecules for DNA transfer carry
homologous DNA sequences at its ends
[39, 40]. We have used such an approach
in our laboratory by employing (defective)
retroviral DNA sequences derived from the
CHO genome in expression vector cock-
tails. They have improved significantly the
frequency of high producer cell lines upon
transfection [27, 28]. Several conference
presentations have hinted at the use of
gene targeting DNA, though no specific
data have been published in this respect.
This approach requires well-designed vec-
tors in combination with plasmids provid-
ing enzymes favoring targeted integration
such as Bacteriophage P1 Cre recombi-
nase, lambda phage integrase, or yeast Flp
recombinase. These proteins are capable
of exchanging long stretches of DNA, that
are bordered by regions of homology, be-
tween the genome and the vector DNA
[4143].
1.2.3
Vectors for Expression
of Recombinant Proteins
Mammalian expression vectors are gener-
ally made with constitutive promoters. A
strong promoterenhancer cassette, usual-
ly of viral origin, drives the expression of a
cloned recombinant gene [5]. Recently
non-viral [44] or chimeric viral/non-viral
promoters [45] have also been used. These
gene-of-interest constructs can be trans-
fected together with separate vectors that
confer resistance to a selection agent such
as DHFR. Frequently, the selection gene
and the gene-of-interest are inserted into
the same vector. The expectation is that a
better genetic link between the two genes
would be provided this way. This precau-
tion is not necessary due to the abundant
ligase activity in mammalian nuclei (see
above) [46]. In order to increase the chance
of obtaining high-level producer cell lines,
the selective gene can be driven from a
weak promoter. Although this approach is
expected to reduce the efficiency of stable
transfection, cells that survive selection
would be expected to produce more prod-
uct. Another approach is to use a 1: 5 ratio
between selection plasmid and the gene-
of-interest plasmid. This strategy assumes
co-integration of several plasmid mole-
cules, one of which would mediate selec-
tivity. Polycistronic vectors have also been
proposed for obtaining high expressors
[47]. Unfortunately, the different strategies
discussed above have never been formally
compared.
More recently, inducible promoters have
been developed for mammalian expression
vectors because of their potential use in
gene therapy where constitutive expression
would frequently not be desired. One of
the beneficial aspects of induction would
be the separation of culture phases for
large-scale production whereby gene con-
structs beneficial for rapid growth would
be switched on during expansion of cell
populations, and then switched off when
not required. Other gene constructs for
protein expression would be induced dur-
ing the final production phase (see also
the section on host cell engineering) [48,
49]. Inducible promoters can also be used
for some protein products that confer tox-
icity when expressed from a constitutive
promoter in mammalian cells [50, 51].
Another important aspect for high-level
expression is the structure of the mRNA
produced by the integrated vector DNA.
Intron-free cDNA constructs are not ideal
in mammalian cells to obtain efficient cy-
toplasmic transport of the mRNA. Most
expression vectors now include at least
one intron sequence that is usually located
between the promoter/enhancer and the
cDNA coding sequence [45].
Transgene expression in animal cells or
in animals is rapidly silenced in many
cases, probably under the influence of sur-
rounding endogenous condensed chroma-
tin (heterochromatin). This gene silencing
correlates with histone hypoacetylation,
methylation of lysine 9 of histone H3, and
an increase in CpG methylation in the pro-
moter region [52]. Heterochromatin is dif-
ferent in structural organization from eu-
chromatin (transcriptionally active), and
the border between the two chromatin types
has been suggested to be marked by se-
quence elements such as scaffold or matrix
attachment regions (S/MARs) [53]. These
elements and ubiquitous chromatin open-
ing elements (UCOS) [54] have attracted
considerable attention since they are
thought to increase and maintain high-level
production of recombinant proteins. S/
MAR elements act to partition silent re-
gions of the chromosomes from domains
permissive to gene expression [55, 56]. They
also recruit factors such as histone acetyl-
transferases that reconfigure chromatin lo-
cally to adopt a structure that is more per-
missive to gene expression [57, 58].
When inserted into expression vectors or
when co-transfected on separate plasmids,
these elements act to significantly increase
transgene expression [59]. These elements
could therefore decrease the screening
times for the identification of suitable cell
lines.
Recently, the genetic code of MARs has
been broken down to a collection of short
genetic sequences that can be recognized
using bioinformatics. Even more potent
MAR elements have been unraveled from
genomic sequences and are currently en-
tering the recombinant protein field (N.
Mermod, personal communication).
A very recent publication speaks of yet an-
other class of sequence elements that may
mediate and maintain high-level productiv-
ity in mammalian cells upon gene transfer:
Highly conserved anti-repressor elements
of 15002000 base pairs have been identi-
fied and cloned from human genomic li-
braries and inserted into expression vectors.
Some of these elements have shown to al-
low the generation of CHO cell clones pro-
ducing secreted alkaline phosphatase
(SEAP) at 5080 pg per cell per day [60].
Another approach towards inhibition of
silencing is to block deacetylation of his-
tones. Acetylated histones are considered
the primary hallmark of active chromatin
[61]. In recent investigations by Hacker
and colleagues, it was shown that in-
creased expression from silenced trans-
genes in recombinant CHO cell lines
could be mediated by transient expression
of Gam-1, an avian adenovirus protein that
is known to block deacetylase activity [62].
It should be mentioned in this context that
the successful application of butyrate in
large-scale manufacturing processes for
the induction of increased specific produc-
tivity of mammalian cells [63] is thought
to be based, at least in part, on the inhibi-
tion of histone deacetylation [64, 65].
Unfortunately, a reliable and compre-
hensive analysis of the various options for
vector design has never been carried out,
in part because this would be a very diffi-
cult and time-consuming task. Most com-
parisons of promoter/enhancer elements
have been made with model proteins and
in transient transfections since expression
from individual clonal cell lines or from
populations from stable transfections
range dramatically from one experiment to
another, and are also dependent on the
transfection method.
In prokaryotic expression systems, a con-
version of mammalian codon use to bacteri-
al codon use is an accepted and widely used
concept. In general, highly expressed genes
exhibit a codon bias towards more abundant
tRNAs. However, few data are found in the
literature on codon use when human genes
are over-expressed in mammalian host cells.
It is questionable that original exon se-
quences for all the desired proteins of inter-
est will always provide a codon utilization
that is compatible with high-level expres-
sion. In some cases, codon optimization
has been shown to increase transgene ex-
pression dramatically in mammalian cells.
A review of these studies has been provided
by Makrides [66].
1.3
Host Cell Engineering
The use of serum and other undefined
and complex media additives for the
growth of mammalian cells may result in
problems of reproducibility in the process.
To avoid batches with negative impact on
cells, testing procedures must be incorpo-
rated before use of these additives. Serum
and animal-derived growth factors cannot
be heat-sterilized prior to use a typical
process stage for the elimination or reduc-
tion of infectious agents. In general there-
fore, chemically well-defined materials are
preferred as substrates for mammalian cell
culture processes. Variable and undefined
additives to culture media are expected to
be removed and to be replaced eventually
by chemically defined components that
have little or no variations in quality and
are of higher purity.
Serum serves as a source of growth fac-
tors, and one way to avoid this is to have
them produced by the host cell line. Such
engineered hosts can subsequently be em-
ployed as a platform for generic pro-
duction processes. The hope is to generate
hosts that would achieve superior growth
rates, that survive death-inducing insults
in extended batch processes, for example
when media components are exhausted,
and that have higher productivity. Onco-
genes, cell cycle genes (cyclines), hormone
genes (insulin-like growth factor) and anti-
apoptotic genes [6773] have been indivi-
dually transferred into the cellular genome
resulting in novel and possibly superior
production hosts [74]. More recently, due
to observed limitations in protein folding
and secondary modification in cell lines,
chaperone genes and genes encoding en-
zymes for glycosylation have been trans-
fected into mammalian cells, and this has
resulted in superior protein quality and
quantity in bioreactors [75]. Insights into
the genomic organization and function of
mammals will dramatically increase in the
years to come. It can be expected that this
information will provide leads towards a
more efficient use of mammalian cells for
protein production. Most likely, targeted
knock-out mutations as well as designed
enhancements of metabolic pathways for
efficient nutrient use, will make mamma-
lian cells even more useful for production
purposes.
Targeted transgene expression control in
mammalian cells is another exciting new
opportunity for host cell engineering. Tet-
racycline has been used in the Tet-on, Tet-
off system, developed by Bujard and collea-
gues [76]. In order selectively to use inde-
pendent gene control of two different gene
activities in the same cells, Fusseneggers
group developed a repressible as well as
an inducible system based on the repres-
sor Pip (pristinamycin-induced protein)
[77]. Such systems allow control over
growth and productivity. Rapid cell mass
expansion would be a first goal for the
generation of biomass, followed by the
growth arrest and boost of high-level pro-
ductivity [78, 79]. Host cell engineering for
metabolic benefits and improved produc-
tivity has already been shown in a hybrido-
ma cell line by introduction of the gluta-
mine synthetase gene, resulting in inde-
pendence of cells from glutamine addition
to the medium and in a reduction of the
waste product ammonium [80].
Another highly promising aspect of host
cell engineering concerns the improve-
ment in post-translational protein modifi-
cation and processing. A number of thera-
peutic antibodies produced in CHO cells
have been successful products. Yet, the ef-
ficacy of these antibodies can probably be
improved by enhancing the potency of
their natural immune effector functions.
In particular, the affinity of the interaction
between the antibody Fc region and Fc-
gamma receptor appears to be crucial for
in vivo biological activity [81]. These molec-
ular interactions are affected by the pres-
ence of carbohydrates at conserved sites in
the antibody Fc region [82]. Engineering
the Fc oligosaccharides can be explored as
a means to enhance Fc-gamma receptor
binding and the associated immune effec-
tor functions. Umaa and co-workers were
the first to demonstrate that recombinant
DNA-based technology could be used to
manipulate the apparatus of cells for sec-
ondary modification, thus generating anti-
bodies with a modified glycosylation pat-
tern and an associated increased immune
effector function. These authors have de-
veloped stable over-expression of a-1,4-N-
acetylglucosaminyltransferase-III in recom-
binant antibody-producing CHO cells in
order to generate IgGs with high levels of
bisected, non-fucosylated oligosaccharides
in the Fc region, and to obtain large in-
creases (over two orders of magnitude) in
antibody-dependent cellular cytotoxicity
(ADCC) [75].
1.4
Gene Transfer and Gene Amplification
in Mammalian Cells
The underlying principles for use of mam-
malian cells as recipients of protein encod-
ing DNA vectors and for the most popular
way to improve productivity, by experimen-
tally induced gene amplification, are de-
scribed in this section. It is well accepted
that these principles are applicable to any
mammalian systems used in the industry,
albeit with the appropriate modifications
in methods.
A mutant CHO cell line, lacking DHFR
activity can be cultivated in the presence
of glycine, hypoxanthine, and thymidine
(GHT). When these cells are transfected
with a functional DHFR gene, cells that
have acquired the gene can be selected
and expanded in media lacking GHT. A
second expression cassette for a product of
interest (for example a protein with thera-
peutic value) can be included on the
DHFR plasmid or on a separate vector(s)
(Fig. 1.2). It should be noted that co-trans-
fection of several plasmids is possible due
to an apparently unlimited capacity for the
uptake of foreign DNA by mammalian
cells. Surprisingly, integration of individu-
ally co-transfected DNAs at the same site
in the genome seems to be the rule in
mammalian cells [25]. In spite of this ex-
perience, which has been verified in nu-
merous cell lines created by the co-trans-
fection of individual vectors, an interesting
approach was proposed recently to tightly
link the expression of DHFR with expres-
Fig. 1.2 Generation of stable
cell lines using DHFR-minus
CHO cells. The example here
uses co-transfection of restric-
tion enzyme linearized plas-
mids. Clones appear 23
weeks after exposure to selec-
tive environmental conditions
here, culture of cells in me-
dia lacking hypoxanthine and
thymidine.
sion of an antibody. DNA segments con-
taining the coding sequences of each a
half of a DHFR-protein were integrated
into the two vectors containing cassettes
for a heavy and light chain, respectively, of
an antibody gene. The functional assembly
of the two halves of the DHFR-protein was
assured by the addition of leucine-zipper
sequences to the respective DHFR-protein
segments [83]. While the average expres-
sion level derived from these clones is re-
ported to be good and in the range of 10
25 pg per cell and day, the screening of a
large number of cell lines will be always
required in order assure satisfying produc-
tivities. Typical screening efforts will evalu-
ate 100 to 500 individually established cell
lines, preferably from several independent
transfections.
While DHFR is still the most frequent
selection approach with CHO cells, other
selection systems can be used (e.g., anti-
biotics such as neomycin, hygromycin or
puromycin), as well as fluorescence pro-
teins [84]. An important consideration for
the choice of a selection agent is the de-
gree of selectivity (stringency). The more
stringent the selective agent used, the
smaller will be the number of obtainable
clones. However, a more stringent agent
will select for colonies of cells that express
the resistance marker gene at higher levels
and, frequently, also the desired gene of
interest. The DHFR system is even
when the gene is driven for example by a
relatively strong SV-40 promotor a rather
stringent selection system and will pro-
duce, after transfection of cells, fewer
clones than would selection with the anti-
biotics neomycin [85].
With both DHFR selection and the glu-
tamate synthetase (GS) system, expression
of both the selection gene and the gene of
interest can be augmented by exposing re-
combinant cells to drugs that block the ac-
tivity of the product of the selection gene
(see also Part IV, Chapter 4). For DHFR,
the drug methotrexate (MTX) has been
used successfully in a large number of
cases [8688]. MTX is a folate derivative
that blocks DHFR activity completely and
irreversibly. Usually, after 23 weeks of ex-
posure to MTX, a majority of cells die
while a few survive that are resistant to
MTX toxicity due to elevated expression of
DHFR. Essentially any given level of MTX
is overcome by a small number of cells
that produce more DHFR than would be
inhibited by the given intracellular quan-
tity of MTX. It was found that these cells
frequently contain chromosomally inte-
grated plasmid sequences in a higher copy
number than observed in cells before ex-
posure to MTX. Stepwise treatment with
elevated concentrations of MTX can be re-
peated several times and may result in the
isolation of cells that contain dramatically
increased copy numbers of the transferred
genes. The phenomenon of MTX-mediated
gene amplification had been observed be-
fore the use of recombinant DNA technol-
ogy, most notably in cancer patients [89].
CHO cell lines containing several hundred
to a few thousand copies of transfected
plasmid DNA have been established [50,
90]. In most cases, the amplified segments
contain the gene of interest, but large seg-
ments of 100 to 10000 kilobases of the
surrounding region have also been ampli-
fied in the process [91, 92]. Most ampli-
fied cells produce more product than the
unamplified host cells did previously.
However, the improvement of specific pro-
ductivity (up to 10- to 20-fold) is highly
variable when studying individual clones
[93], and also varies from product to prod-
uct [94].
The principle for MTX-driven amplifica-
tion also applies to other immortalized cell
lines [95]. However, DHFR gene transfer
followed by amplification in MTX works
best with cells that lack a functional en-
dogenous DHFR gene. A popular
approach with NS0 cells utilizes GS as a
selective gene. This system relies on the
fact that NS0 cells express very low levels
of GS. These cells require either an exoge-
nous source of glutamine or an exogenous
GS gene in order to survive in the absence
of glutamine. A specific and irreversible
inhibition of GS can be mediated by the
addition of methionine sulphoximine
(MSX) to the culture medium. At a con-
centration of 10 to 100 lM MSX, resistant
clones can be identified in selected NS0
cell populations that have amplified the
transgene complex containing the GS gene
and the desired gene(s) of interest [96, 97].
The GS system can also be applied to cells
such as CHO that have a normal level of
glutamate synthetase. In this case, the
starting concentration of MSX needs to be
higher than that used for the selection of
recombinant NS0 clones in order to block
the endogenous GS and to select for
clones that over-express the exogenous GS
gene [98]. Unfortunately, the literature
does not provide any information on the
cytogenetics of gene amplification in the
GS system.
With recent publication of the genome
sequences of man, mouse and rat, we have
learned that mammalian genomes are ex-
ceptionally dynamic due to the presence of
repetitive sequences, remnants of retroviral
genomes and transposable elements. This
phenomenon can be termed sequence
mobility. Mammalian cells particularly
immortalized cells have an even more
intrinsic genomic fluidity. This becomes
evident when studying chromosome num-
bers in metaphase spreads [99]. It appears
that immortalized cell populations will di-
verge, even when established as clonal cell
lines, in the number and structure of their
chromosomes within very short time
frames (weeks to months). Due to se-
quence mobility and chromosomal insta-
bility immortalized mammalian cells are
ideal substrates for experimentally induced
gene amplification.
1.4.1
Cytogenetics of CHO Cell Lines, Genetic
and Production Stability
Subpopulations and clones of MTX-treated
CHO cells may contain the transfected
DNA sequences at very high copy num-
bers [90, 100]. Studies concerning genetic
features of these amplified DNA se-
quences within the CHO genome have
been ambiguous. It is still controversial to-
day, whether continued presence of a se-
lective agent (i.e., MTX) in long-term cul-
tures of recombinant CHO cells is re-
quired for production stability. Whereas
some studies suggest that MTX is required
for the stable production of recombinant
proteins [101], others indicate that continu-
ous cultivation of clonal cell lines in MTX
might not be necessary [102]. Cytogenetic
studies, using FISH [103] were performed
in my own laboratory in the late 1980s
with clonal and non-clonal recombinant
CHO cell lines, and showed that continu-
ous exposure to MTX at the same concen-
tration, under which the cell populations
were initially established, promotes genetic
instability at the chromosomal level [104,
105]. Cell lines selected at micromolar con-
centrations of MTX showed elongated
chromosomal structures that hybridize to
probes representing transfected DNA.
Some of the chromosomes that contained
a large number of tightly arranged bands
of fluorescence differed dramatically, most
notably in length, from the normal CHO
chromosome. Frequently, we found chro-
mosomes with transgenic DNA up to the
very end of one arm, indicating the ab-
sence of a telomere. If in fact these chro-
mosomes do not have functional telo-
meres, then this raises questions about the
stability of the amplified DNA during con-
tinuous subcultivation.
In summary, when cultivating these cell
lines in the absence of MTX, unique and
characteristic integrations were found in
9599% of metaphase spreads. We contin-
ued to cultivate these cell lines for ex-
tended periods in the absence of MTX,
and occasionally performed FISH analyses.
We found that the chromosomal structures
described above were stable within the ob-
servation period (a minimum of 60 days,
and in one case of 160 days). Cytogeneti-
cally, these observations indicate a high de-
gree of genetic stability of chromosomally
amplified sequences in the absence of
MTX. Equivalent observations as those
presented above have been made recently
by Kim and Lee [106].
1.4.2
Transgene Structure
and Locus Determination
Southern hybridization of genomic DNA
is a useful tool to determine the molecular
structure of integrated plasmid DNA.
Using suitable restriction enzymes, this
technique will provide information on the
integrity of the transgene. Estimates of the
copy number of the chromosomally inte-
grated DNA can be established when
using in the same experiment known
quantities of plasmid DNA restricted with
the same enzyme as that used for the
genomic DNA. Southern hybridization
may also provide information on the ques-
tion of whether one or more than one in-
tegration locus exists for the plasmid se-
quences. However, conclusions on multi-
plicity of integration must be made with a
degree of caution. Aberrations from the ex-
pected signals can be due to post-integra-
tion rearrangement in a fraction of the cell
population or in the initial co-integration
of a few copies of the plasmid sequences
that had been subjected to nuclease attack,
resulting in the deletion of the restriction
enzyme site used for the analysis.
All Southern hybridizations are based
on DNA extracted from thousands of indi-
vidual cells. Even if the cell lines are based
on a cloning step, one must be aware
that none of the cell lines is clonal in the
most narrow sense: Genetic variations oc-
cur very rapidly in immortalized cells, due
to their inherent chromosomal instability.
FISH can be used to gather knowledge
about the degree of chromosomal amplifi-
cation (not to be confused with copy num-
ber estimates) and the chromosomal loca-
tion of the recombinant DNA. In order to
provide some useful information, FISH
studies must be supplemented by a statis-
tical analysis of identified integration sites
(and structures observed). They should
also take into consideration the time point
of analysis with respect to the total time of
cultivation of the cells. A reasonable value
may be gained from a FISH study per-
formed shortly after cells have been
thawed from a bank of cells stored in liq-
uid nitrogen. Depending on the culture
conditions (with or without MTX or se-
rum) and the length of cultivation time,
the results of FISH analyses may vary con-
siderably. Despite this problem, in the
studies discussed above we were able to
determine the identity of one recombinant
cell line from another by identifying a
chromosomal marker containing hybridiz-
ing DNA which was present in a large
fraction of the individual cells of the popu-
lations studied. The identifying chromo-
somal markers containing recombinant se-
quences were termed master integrations
which we found to be the genetically
stable entities in the cell lines.
1.5
Production Principles for Mammalian Cells:
Anchorage-dependent Cultures
and Suspension Cultures
Process scientists and their managers must
decide which type of production system to
choose for the product in question. Chiefly,
the anticipated scale of operation for the
manufacturing process drives the choice.
A number of very successful recombinant
products from mammalian cells such as er-
ythropoeitin (Epogen
) are given to tens of

thousands of patients. Epogen
is a protein
hormone developed by Amgen for the treat-
ment of dialysis patients with chronic renal
failure. The single dose necessary for treat-
ment is relatively small (about 100 lg/pa-
tient). In contrast, treatment with another
protein therapeutic, the recombinant anti-
body Herceptin
, developed by Genentech,
requires multiple doses over weeks and
months with a maintenance dose of about
150 mg per patient. Herceptin
is a huma-
nized IgG directed against the Her2 recep-
tor that is over-expressed in a percentage
of breast cancer patients (see also Part I,
Chapter 5). The number of patients treated
per annum is approximately the same for
these two products. It is clear that a 1000-
to 10000-fold difference in the annual quan-
tities of Epogen
and Herceptin
needed
will require entirely different decisions on
the scale and mode of operation when de-
veloping the manufacturing processes for
these two products.
Processes for recombinant proteins from
mammalian cells can be established on
the basis of two cellular growth modes: ad-
herent and suspension cultures. CHO and
other hosts such as BHK and HEK293
cells can be grown in either mode. NS0
cells that were derived from a mineral oil-
induced plasmacytoma in mice will only
grow in suspension, and will not firmly at-
tach to a surface that is exposed to mixing
induced shear force.
1.5.1
The Rollerbottle Process
In the case of erythropoeitin and a few
other hormone-type protein products, a
process based on rollerbottles appears to
be sufficient to supply the market with
product. In this case, adherent cells are
cultivated on the inner surface of a cylin-
drical bottle having a volume of 1, 2 or
3 L. A typical 2-L bottle provides an inner
surface of 850 cm
2
, but there are varia-
tions of these bottles that provide extended
areas for attachment of cells. A simple and
reproducible process can be established
with minimal initial investment in equip-
ment using such rollerbottles. Provided
that there are sufficient human resources
available, this process can be easily scaled
up since the number of rollerbottles
handled in parallel determines scale.
Cells thawed from a cell bank can be ex-
panded by subcultivation into a fixed num-
ber of rollerbottles. The standard 2-L roller-
bottle is usually filled with 300500 mL of
medium. The remaining volume provides
the necessary oxygen, while the closed bot-
tles are slowly rolled at about 1 r.p.m. in
an incubator at 378C. A sufficiently large
number of rollerbottles containing an ade-
quate cell population represents the start-
ing point for several production cycles. For
scale-up, the cells from a single confluent
rollerbottle can seed up to 20 rollerbottles.
From freshly seeded rollerbottles to conflu-
ent rollerbottles requires 36 days, depend-
ing on seeding density, growth rate, and
composition of the medium. Since adher-
1.5 Production Principles for Mammalian Cells: Anchorage-dependent Cultures and Suspension Cultures 737
ence of cells to the inner surface of the
rollerbottle is required, serum is frequently
used in such a process at a concentration
of 110%, providing necessary attachment
factors to the cells. Adherence can also be
assured in media lacking serum if fibro-
nectin and other cell attachment factors
obtained from animal sources are added to
the culture medium. Within two to three
subcultivations, starting from a seed vial
obtained from the working cell bank, a
sufficiently large number of inoculated
rollerbottles can be generated that can con-
stitute a production phase. A part of the
cell mass generated in the last subcultiva-
tion cycle can be used for the generation
of seed culture for the subsequent produc-
tion cycle.
Media for the production phase are
usually richer in nutrient content than the
seed train medium in order to maintain vi-
ability and productivity of the cells for a
minimum of 12 weeks. For facilitating re-
covery and purification of the product and
for cost reduction, serum is not used for
the production phase. The attachment of
cells to the surface of the rollerbottle will
not be compromised by such a modifica-
tion in the medium. However, gentle han-
dling is required or the sheets of cells will
detach. Upon incubation of the confluent
cells with the enriched, serum-free medi-
um, the secreted product will be harvested,
leaving the adherent cells inside the bottle.
Sometimes, a refeeding with fresh medi-
um for a second production cycle is possi-
ble, on the condition that the product of
the first and the second harvest will be
similar in composition and quality. Since a
standard 2-L rollerbottle will contain about
300 mL of medium for harvest, 1000 roller-
bottles will provide from 300 L to 600 L of
supernatant. Over a one-year period, a
manufacturing process based on this
schedule will deliver 1500030000 L of
cell-free culture medium containing the
product of interest. Product concentrations
in the 50 to 200 mg L
1
ranges are possi-
ble, thus providing the protein in the kilo-
gram range annually. Such a process is la-
bor-intensive, requires the repeated use of
trypsin for detachment of cells, and is at
considerable risk of contamination with
adventitious agents through handling.
Epogen
(erythropoietin) has been de-

veloped on the basis of a rollerbottle pro-
cess. Todays Epogen
process is essen-
tially a robot-based manufacturing proce-
dure whereby all the critical handling
steps including the seeding of cells, fill-
ing of bottles with media and harvesting
of cell culture fluids are executed within
air-filtered environments and without hu-
man interaction.
A variation of the above process involves
stirred tanks or hollow-fiber bioreactors for
growth of the seed culture. The growth of
CHO cells in both the suspension and ad-
herent modes allows streamlining the ro-
llerbottle production process. The seed cul-
ture for the rollerbottle production phase
can be generated in spinner flasks or in
bioreactors. The advantage of such a pro-
cess is that fewer subcultivations are
needed to generate sufficient cell mass. It
also reduces the risk of contamination by
adventitious agents. Hollow-fiber bioreac-
tors can also be used in which very high
cell densities can be achieved through the
continuous perfusion of the reactor with
fresh medium. Several reactor volumes of
fresh medium can be perfused through
such a system, resulting eventually in cell
densities approaching tissue-like character.
These rather compact systems provide suf-
ficient cells to seed a very large number of
production rollerbottles. Again, the goal of
such an approach is to reduce human in-
teraction and the risk of contamination.
1.5.2
Adherent Cell Culture in Bioreactors
Van Wezel [107] proposed the use of water-
suspended polymer spheres termed mi-
crocarriers for the culture of adherent
cells in stirred-tank bioreactors. The pur-
pose of growing cells in stirred bioreactors
instead of on fixed surfaces is to allow for
easier scale-up and increased homogeneity
in supply of nutrients in media, but also
in supply of oxygen and carbon dioxide ex-
change. Several processes have been devel-
oped in the human and animal vaccine in-
dustry using the microcarrier concept, al-
ways with cells that have a high anchorage
dependency [108]. These cells serve as sub-
strates for the multiplication of viruses
such a measles, polio or mumps [109].
CHO cells are being used for the produc-
tion of several human recombinant pro-
teins, most notably at Serono, on microcar-
riers in stirred bioreactors [110]. These
processes date from the early phase of re-
combinant mammalian cell culture tech-
nology and require the use of serum, at
least for parts of the process. For scale-up,
cells are seeded at a density of about one
to five cells per bead, and these will subse-
quently grow to confluency on the beads.
Widely used microcarriers are Cytodex
TM
1
and Cytodex 3, both marketed by Amer-
sham Biosciences. Cytodex 1 is based on a
cross-linked dextran matrix which is sub-
stituted with positively charged N,N-di-
ethylaminoethyl groups. The charged
groups are distributed through the micro-
carrier matrix. Cytodex 3 consists of a thin
layer of denatured collagen chemically
coupled to a matrix of cross-linked dex-
tran. The denatured collagen layer is sus-
ceptible to digestion by a variety of pro-
teases including trypsin and collagenase,
allowing for removal of cells from the mi-
crocarriers while maintaining maximum
cell viability. Once cells have been de-
tached, additional carriers can be added,
while both cells and carriers are gravity
settled in the bioreactor. The microcarrier
approach has certain advantages with re-
spect to harvesting of product from cell
culture fluids, but also for perfusion pro-
cesses where fresh medium is added to a
culture while spent medium is withdrawn.
Spier and Kadouri have reviewed the evo-
lution of commercial production processes
based on anchorage-dependent cultivation
[111].
Processes without the use of microcar-
riers are however less cumbersome, since
the transfer of cells from one scale to the
next and thus reseeding of fresh carriers is
tedious and complicates processes beyond
need, especially when the preferred host
cells for recombinant protein production
can now easily be cultivated without any
matrix.
Microcarriers will remain important in
the field of vaccine production, since sev-
eral viral products are dependent on
strictly adherent cell lines and in tissue en-
gineering. For the latter, different cell
types are needed to reconstruct multi-
layered organs and tissues. Macroporous
carriers and matrices can be used to gener-
ate structured cellular complexes that are
molded into functional organ/tissue sys-
tems (see also Part I, Chapter 15).
1.5.3
Stirred-tank Bioreactor Processes
Increased worldwide needs for recombi-
nant biopharmaceutical proteins drive ma-
jor investments into the construction of
new bioreactor facilities. In addition to
those companies that produce their own
protein pharmaceuticals in large-scale
manufacturing plants, a few contract man-
ufacturers offered in 2004 a bioreactor ca-
pacity of about 130000 L. These contract
manufacturers serve an increasingly com-
petitive market. Projections state a capacity
shortfall of about 400000 L for the year
2006 [112]. The clinical and commercial
success that recombinant proteins have
had during the past 10 years has clearly
stimulated many newcomers in the field
to try to develop similar clinical targets,
thereby creating a demand for bioreactor
capacity which exceeds current availability.
Most of the successful antibody or anti-
body-like proteins are given to patients in
rather large doses (hundreds of milligrams
to grams per patient) and thus require
very large facilities for manufacturing.
Likewise, new markets are generated after
the approval of a given product that widen
the application of a biopharmaceutical.
Off-label use increases the product de-
mand even further.
Suspension culture of mammalian cells
is the most popular approach for large-
scale manufacturing [113]. This approach
using CHO cells and a few other cell lines
now dominates the domain of mass pro-
duction of recombinant protein products.
With the exception of blood-derived cells,
most of the other cells used in the indus-
try were of fibroblast or epitheloid charac-
ter and were therefore initially anchorage-
dependent. In the early 1980s, the adapta-
tion of CHO cells to suspension culture
was a tedious process, mostly because of
the lack of media formulations that facili-
tate suspension growth. Today, multiple
factors seem to have made the transition
from adherent to suspension culture much
easier. For example, cell culture media
have been developed which support the
growth of cells in suspension better than
earlier formulations based on DMEM and
Hams F12. Also, the selection of cell pop-
ulations in media with reduced serum and
calcium concentrations has resulted in cell
lines that support the transition from ad-
herent to suspension growth more readily.
Some scientific reports have claimed facili-
tated serum-free suspension growth due to
genetic modification [114]. However, these
advances must be regarded with caution
as non-modified cells do readily grow now
in optimized suspension media. When
handled correctly and when using appro-
priate media formulation, seeding densi-
ties and stirring conditions, the transition
of CHO cells from adherent to suspension
cultures, even without genetic modifica-
tion, can be executed in a few weeks.
In a simple bioreactor-based process,
the scale-up to very large volumes can oc-
cur rather rapidly. This is usually executed
by diluting the entire volume of one bio-
reactor into 520 volumes of fresh medi-
um held prewarmed in a larger reactor
(Fig. 1.3). Within 1015 days, a suspension
culture at the 50-L scale can be used to in-
oculate a 10000 L reactor. It is a major
goal of process development work to opti-
mize media for the production phase.
Such a medium needs to support good
growth initially in order to achieve the
highest cell density possible, and then it
needs to provide the nutritional basis and
physiological balance to maintain viability
and productivity for extended periods. The
periods for production (614 days) usually
exceed in time the typical subcultivation
periods of 35 days. While the termination
(i.e., harvest) of such a culture is driven
mainly by plant capacity and volumetric
productivity, the other important issue to
consider is the quality of the derived prod-
uct. The continuously changing composi-
tion of the culture medium during the
production phase can affect the quality of
earlier synthesized product through degra-
dative activities mediated by cell-released
enzymes. Also, a diminishing supply of
nutrients as energy providers or as build-
ing blocks for the synthesized product will
most likely change the molecular composi-
tion of recombinant proteins. The most
probable alteration of protein being made
early or late in the production process
would involve the structure and extent of
glycosylation. This topic is discussed in
more detail elsewhere in this book (see
1.5.4
Batch and Extended-batch Perfusion
Batch and extended-batch processes have
achieved unprecedented productivity.
These are the results of many months if
not years of work that went into the de-
velopment of the manufacturing process.
This development work is summarized in
a very simplified way in the Fig. 1.4, start-
ing from gene transfer to cells and ending
with the establishment of a well character-
ized masterbank. Scientists from Genen-
Fig. 1.3 Diagram of a simple batch (or extended-
batch) process with suspension cells. Cells are ob-
tained from a Master- or Working Cell bank
(MCB/WCB) and inoculated into spinners for a
defined subcultivation period (every 34 days,
usually for up to 100 days or more). For mainte-
nance purposes of the culture, the cells in the
spinner are referred to as the seed train. Cells
from spinners (15 L volume, filling volume up to
40%) are used to inoculate bioreactors at increas-
ing scales of operation, until the final volume for
production is obtained. Cells in vessels with in-
creasing volume are referred to as the inoculum
train. The final and largest vessel is used for pro-
duction purposes.
tech reported in March, 2004 on volu-
metric titers of more than 4 g L
1
of a se-
creted antibody-product in the supernatant
of CHO cells in large-scale bioreactors. A
single production run, executed at a vol-
ume of 10000 L can therefore produce
more than 30 kg of purified product (as-
suming a recovery yield of about 70%). Re-
peating such a production run successfully
20 times each year would provide 600 kg
of product. A handful of companies have
invested heavily in large-scale production
facilities, with several having up to six par-
allel trains for scale-up. The largest mam-
malian bioreactor system is presently
being constructed by Roche in Basle, with
a bioreactor volume of 25000 L. Just 20
years ago, 5 mg L
1
from CHO cells was
considered sufficient to justify investments
into the development of a CHO platform
for recombinant protein production.
Clearly, the success of several antibody and
antibody-fusion products in clinic, for the
treatment of cancer and diseases such as
rheumatoid arthritis, has driven huge in-
vestments in order to assure market sup-
ply. Therapeutic antibodies are projected to
obtain six to eight market approvals per
year and to reach sales in the USA of $20
billion by 2010.
The batch process is considered the
most simple and thus most robust pro-
duction process for stirred bioreactors. The
term batch is connected to the very last
phase of the process, the phase during
which accumulated product is maintained
in a final production vessel. Since all man-
ufacturing cell lines used so far drive the
expression of the product gene from con-
stitutive promoters, product will be synthe-
sized during earlier phases of the process,
but not harvested. One popular approach
is to define the entire process from the
thawing of cells from a bank to the pro-
duction vessel as three separate phases.
These are the seed train, the inoculum
train, and the final production phase (see
Fig. 1.3). In the step preceding produc-
tion, the cells in a smaller bioreactor are
cultivated to maximal cell density and then
transferred along with the exhausted
growth medium into the production reac-
tor. The timing of cell culture subcultiva-
tion and the target density of inoculation
of the subsequent culture step are the sub-
jects of process development questions
and must be determined on a case-by-case
basis. The production process begins when
cells and fresh medium are mixed in the
reactor, and it ends at a predetermined
time-point when the synthesis of recombi-
nant protein diminishes due to exhaustion
of nutritional components in the medium
and/or accumulation of toxic end products
of cellular metabolism. Usually, with CHO
and NS0 cells the production phase lasts
for between 7 and 14 days after inocula-
tion of the reactor, depending on the sus-
ceptibility of the proteins to degradative
enzymes, as well as a number of other
process-related factors. The advantage of
such a process is obvious. Provided that
the inoculating cell mass can be generated
reproducibly, the resulting production pro-
cess will show a high degree of similarity
with respect to cell growth, viability, and
quantity and quality of the product being
synthesized.
The issue of reproducibility of process
parameters and of achievable product
quantity and quality is of highest signifi-
cance as this will ultimately be evaluated
by the regulatory agencies. For Investiga-
tional New Drug applications (IND) and
Process Licence Applications (PLA), rather
specific requirements must be met with
respect to the minimal number of product
batches analyzed. For INDs, no less than
three product runs are recommended.
Shorter batch processes (57 days) have
the advantage of generating more data
within a given time frame. This can be a
very important cost factor, since the time
period necessary to acquire and evaluate
necessary data from the new process will
eventually affect the overall time necessary
for entering the market.
On the other hand, there are a number of
arguments that would sway process devel-
opment decisions to another direction.
The option to prolong the synthetic activ-
ities of cells in the production vessel would
capitalize on the process investment
which allowed generating the necessary cell
mass for the production vessel. Increasingly
therefore, extended-batch or perfused-batch
cultures are used. With a longer-lasting pro-
duction phase in cell culture, feeding addi-
tional medium components becomes neces-
sary. Clearly, extending the process for a
considerable period of time (e.g., from 8 to
14 days or longer) only makes sense when
the return for this investment in labor
and in occupation of the production facility
result in a sufficiently high increase in
product concentration within the vessel.
There are various ways that medium and
medium components can be added to a cul-
ture that had been initiated a few days ear-
lier in the same tank. This might be done
by feeding (batch wise) highly concentrated
mixtures of essential amino acids and other
medium components, thereby not signifi-
cantly affecting the volume in the tank. Al-
ternatively, a culture can be started in the
production vessel at half or so of the work-
ing volume, after which standard concentra-
tion medium can be pumped slowly and
continuously or batch-wise into the tank un-
til the final working volume has been
reached. The choice of either mode or
combinations thereof is in the hand of
the process development scientist, who
must evaluate carefully any advantages
and disadvantages. No matter what princi-
ple will be used for extending the produc-
tion phase, the ultimate overall result will
always be a tank that contains in one
batch the entire protein population for
subsequent recovery and purification.
Continuously perfused production pro-
cesses represent an entirely different philo-
sophy for manufacturing. Here, the goal is
to achieve the highest cell concentrations
possible within smaller tanks that hold
the cells. Sometimes the reasoning used is
that up-front investment for manufactur-
ing equipment is reduced and product
quantities can be quite high from such
processes. The much-improved knowledge
base in technology and in the physiology
of mammalian cells in culture have made
this more complicated approach to manu-
facturing attractive. Perfused cultures can
be maintained for many weeks and
months, with product harvests occurring
repeatedly throughout that period. A pro-
tein of high interest to the pharmaceutical
industry for several decades the antihe-
mophilic Factor VIII (see also Part II, Chap-
ter 3) is reliably being manufactured
using perfusion technology with BHK cells.
The glycoprotein, which is probably the
largest secreted single peptide chain protein
ever produced in bioreactors, is harvested
continuously through Bayers cell retention
technology that allows cells to be returned
to the bioreactor. This process, when run
for up to 6 months, improves the yields of
fragile proteins that would be degraded if
left in the fermentor for the typical time
used in fed-batch processes. While the pro-
duction of Cognate
Factor VIII [115] has

pioneered the use of perfusion technology
for recombinant proteins with mammalian
cells in a non-CHO cell, other products
from CHO cells have also been approved
using this technology [116].
Finally, many monoclonal antibodies
have been produced for some years at the
laboratory scale in perfusion systems
using hollow-fiber technology. Hence, the
transfer of this technology to production
scale for pharmaceutical manufacture has,
for some time, been obvious [117, 118]. A
diagnostic monoclonal antibody for imag-
ing in patients with prostate cancer (Pros-
taScint
TM
) was the first FDA-approved
product to have been produced by a hol-
low-fiber perfusion process [119]. The ap-
plication of this and other perfusion tech-
nologies using immobilization of mamma-
lian cells on macroporous carriers produc-
ing recombinant antibodies or other
proteins is likely to be yet another option
for future manufacturing processes [120].
1.6
Large-scale Transient Expression
Mammalian cells have been used as produc-
tion hosts in some of the first pioneering ex-
periments for the design of novel proteins.
One of these proteins for example, the
first chimeric therapeutic candidate, the fu-
sion of the CD4-receptor and a human im-
munoglobulin was first designed as mam-
malian gene construct and then generated
by transient expression in HEK-293 cells
[121]. Another designed human therapeutic
the thrombolytic biopharmaceutical pro-
tein drug TNKase
, a mutagenized tissue-
plasminogen activator was developed
based on hundreds of TPA-variants that
have been expressed initially by transient
expression [122]. While these investigations
in the early 1990s were carried out on a
small scale (i.e., 110 mL cell culture and
microgram quantities of protein), large-
scale transient expression from mammalian
cells is a new technology addressing an ur-
gent need in biotechnology for the rapid
production of recombinant proteins in the
milligram to gram range. With better tech-
nologies for the reliable growth of mamma-
lian cells, and with better nucleic acid trans-
fer systems, the opportunity arose to ex-
plore transient expression in mammalian
cells beyond the laboratory scale. In addi-
tion, many companies in the field of so-
matic gene therapy, using artificial or mod-
ified virus vectors, depend on transient
DNA transfer to mammalian cells as one
of the key manufacturing steps for their
products [123]. Other than with stable ex-
pression, vector DNA is not required to inte-
grate into the chromosome DNA of the host
cell, but remains shortly (transiently) in the
nuclear environment where at least some
transgene DNA is utilized as templates for
transcription into mRNA [124]. The highly
improved DNA transfer systems developed
over the past 10 years (see also Part VI,
Chapter 6) allow to supply frequently 50%
or more of cells in a population with suffi-
cient DNA. The most popular large-scale
transient expression systems are based on
non-viral DNA delivery and utilize calcium
phosphate [125] and PEI [126, 127] as vehi-
cles, and the preparation of these vehicles
with DNA has been modified for use with
stirred single cell suspensions in bioreac-
tors. Calcium phosphate and PEI are both
cheap components an important consid-
eration for scale-up. Several groups have re-
ported the scale-up of transient expression
to bioreactors of 10 to 100 L [128, 129],
mainly for the production of research mate-
rials used in pre-clinical research. The
yields from these exploratory experiments
are in the range of 1 to 50 mg L
1
for anti-
bodies [130], and referred in one report to
the expression of recombinant protein at
100 mg L
1
from 100-L scale operations
with transiently transfected CHO cells
[131]. These yields are clearly far below
those observed with highly optimized pro-
duction processes that have proven their ro-
bustness and reproducibility in large-scale
operations at the 1000 or 10000 L scale.
Why then the need to engage into the devel-
opment of an alternative technology?
The reason is speed. At only days after the
availability of an expression vector, milli-
grams to hundreds of milligrams of a re-
combinant protein can be delivered into
the hands of the researcher. Vectors for tran-
sient expression do not require a selection
marker the goal is to deliver DNA to a
maximal number of cells in the population.
Several vectors can be transfected simulta-
neously into cells and will be expressed si-
multaneously. With calcium phosphate as
a vehicle, it was shown that approximately
20000 plasmid molecules per cell can be de-
livered [132]. After a few days, the copy
number of plasmid molecules will decline
in the nucleus and the production of mRNA
ceases. Depending on the protein at the
time point of the highest accumulated yield,
the product is harvested and cells are dis-
carded. A new production can be re-started
at any time when sufficient fresh cells can
be provided and a new DNA-vehicle prepa-
ration is ready for transfection.
Large-scale transfection requires signifi-
cant quantities of DNA. With both calcium
phosphate and PEI, approximately 12 mg
of plasmid DNA are usually needed per li-
ter of suspension culture. Media and cul-
ture conditions for large-scale transient
transfection are under further develop-
ment, as are the vehicle preparation tech-
niques. With calcium phosphate as a vehi-
cle, a small concentration (12%) of fetal
bovine serum may be required for high
transfection efficiency. Here as well, it is a
goal to generate processes that are low or
free of undefined components.
It remains to be seen whether transient
expression technologies will eventually be
used under conditions for clinical produc-
tion and thus provide eventually products
for human medical use.
1.7
Regulatory Issues
All mammalian cells used for the large-
scale production of recombinant proteins
are considered immortalized, as they can
be grown continuously for an indefinite
period if correct culture conditions are pro-
vided. This is an exceptional characteristic
for animal-derived cells, since the tissues
and organs of animals are constructed of
cells with a defined lifespan. The limited
lifespan of cells in animals was detected
first by Hayflick [133], and is linked
among other reasons to a declining telo-
merase activity on chromosomes of so-
matic cells, but not in germ cells.
The climate for permission by regulatory
agencies, particularly by the FDA in the
United States to use immortalized CHO
cells for the production of recombinant
proteins was not favorable in the early
1980s. Discussions about risks associated
with the use of mammalian cells were
controversial and had been initiated more
than two decades earlier [134] when a first
generation of classical biological products
(i.e., vaccines and the natural interferons)
were developed on the basis of primary
monkey kidney cells, human diploid cells
and, later, transformed mammalian cells.
The manufacturers of recombinant pro-
teins for clinical applications and regula-
tory agencies were in agreement that it
was extremely important to minimize
eventual risks associated with the use of
recombinant mammalian cell hosts. Risks
were seen in tumour principles, carried
by the DNA of the host and in adventi-
tious agents (viruses, mycoplasma, etc.)
that could infect the host cell lines and
thus eventually be transmitted to patients
receiving products from those hosts. Also,
the consistency and quality of the recombi-
nant proteins were discussed in the con-
text of risk assessment and risk control.
The result of a long series of scientific dis-
cussions in journals and at conferences
held over a decade was that stringent con-
trols, regulations and monitoring proce-
dures were enforced as a prerequisite for
manufacture of proteins from such cells
[135]. A balance had to be found between
the almost assured clinical benefit of some
of those first recombinant products and
the perceived risks associated with the un-
avoidable necessity to produce them in
tumour cells.
The first product recombinant tissue
plasminogen activator (Activase
-rtPA)
proved to be a good candidate to achieve
such a balance, since the benefit the sav-
ing of lives of heart-attack patients out-
weighed by far the anticipated risks. How-
ever, approval was achieved only after a
large amount of data were provided to the
regulatory agencies which showed: 1) that
consistently only a minute quantity of
CHO DNA (<10 pg per dose, later relaxed
to <100 pg per dose) was present in the fi-
nal product; 2) that the product itself could
be produced with a high degree of repro-
ducibility; and 3) that it was produced with
a purity not achieved before in any biologi-
cal derived from mammalian cell culture.
1.7.1
Bacterial and Fungal Contamination
The prevention of bacterial or fungal infec-
tions in cell culture and recovery systems
can be assured, to a high degree of confi-
dence, by the use of a piping and vessel
system which maintains absolute contain-
ment of the sterile medium fluids. The
equipment used must be of a nature to al-
low cleaning and sterilizing by Clean in
place (CIP) and Sterilize in place (SIP)
procedures (usually high-quality stainless
steel). Most cell culture processes require
complex media containing amino acids,
vitamins, protein hormones and fetal bo-
vine serum. Some of the components of
mammalian cell culture media cannot be
autoclaved, and thus sterility (freedom
from viruses and microbial organisms)
cannot be assured to a 100% confidence
level. To exclude the introduction of bacte-
rial and fungal contamination through raw
materials, prior testing and, in addition,
filtration through membranes of 0.2 lm or
even 0.1 lm pore size into pre-sterilized
containers is employed. It is, of course,
well understood that most (small) viruses
and prions cannot be excluded through fil-
tration procedures.
Rigorous testing of the Master Seed Cell
Bank (MSCB) and the Manufacturers
Working Cell Bank (MWCB) (for a review,
see Ref. [136]), which is accomplished by
analyzing cells of a number of representa-
tive cryovials, assures that the production
cell line itself is not contaminated with
viruses, bacteria, mycoplasma, and fungi.
Sterility testing of the cell line must be
carried out in appropriate media lacking
antibiotic or anti-fungal compounds, for
obvious reasons. Virus testing is per-
formed in suitable cell systems that are va-
lidated for each of the individual virus spe-
cies. Mycoplasmas, which represent the
smallest living cells, are frequent contami-
nants of cells derived from patients and
from animal sources. They can remain un-
detected in cell culture for extended peri-
ods of time, and are therefore more threat-
ening to cultures for large-scale processes
than typical bacteria that multiply rapidly.
1.7.2
Prions
The use of sera or other products derived
from bovine sources in culture media repre-
sents a potential risk of transfer of the caus-
ative agent for bovine spongiform encepha-
lopathy (BSE) to patients. Therefore, regula-
tory agencies request detailed information
on the origin and processing of products de-
rived from bovine sources. For example,
sera obtained from countries in which
BSE was diagnosed, even in a small number
of animals, are considered unacceptable by
regulatory agencies. Companies have, in
most cases anticipating these regulations,
assured their supply of bovine-derived pro-
cess materials from countries such as New
Zealand or Australia, where BSE has not
been reported so far. The US had been con-
sidered a BSE-free country until recently
(2003) when a single cow was diagnosed
with BSE. In view of new insights into the
molecular biology of prion diseases, one
must consider now that these agents are
more widely present in nature than pre-
viously thought. Disease risk perception
rose in Europe, in the US and elsewhere,
and has initiated public safety discussions
and even the implementation of stringent
process regulations. The use of components
of animal origin in media including the use
of amino acids purified from animal
sources is considered increasingly unac-
ceptable. The high degree of concern re-
garding BSE is based on findings that: 1)
a small compound PrP
sc
(proteinaceous-in-
fectious particles, scrapie) is likely to be re-
sponsible for the disease; 2) transfer of the
bovine disease to human populations as a
variant Creutzfeld-Jakob disease has oc-
curred in hundreds of cases; 3) detection
of the causative agent is only possible with
rather sophisticated techniques, and then
only in tissues that are typically highly af-
fected; and 4) the inactivation of infectivity
of PrP
sc
is difficult. Even autoclaving proce-
dures (1218C, 20 min) do not completely
eliminate infectivity. Reviews on prion biol-
ogy were published by Prusiner (1997) [137]
and Aguzzi et al. (2004) [138].
1.7.3
Viral Contaminants
Why were hamster cells chosen as a host
for making human recombinant proteins
in the early 1980s? A strong argument for
favoring non-human cell lines over human
cell lines for the production of proteins is
the fact that certain life-threatening hu-
man viruses cannot be propagated at all,
or multiply only poorly, in non-human cell
lines. CHO cells do not support the repli-
cation of pathogenic viruses such as polio,
herpes, hepatitis B, HIV, measles, adeno-
viruses, rubella, and influenza. Thus, the
risk of a viral adventitious agent of being
involuntarily carried along with the prod-
uct of interest can be considered extremely
low. Wiebe et al. tested a total of 44 hu-
man pathogenic viruses for replication in
CHO cells and found only seven (reo
1,2,3, mumps, and parainfluenza 1,2,3)
that were able to infect these cells [139].
Exclusion of these virus species and others
that can be propagated on CHO cells, such
as the parvovirus MVM (Minute Virus of
the Mouse), can be assured to a high de-
gree of certainty through testing. Tests can
be performed with all materials which en-
ter the manufacturing process and which
would support the viability of the virus
in question. Tests are also obligatory with
fluids which contain the product of inter-
est.
Sterility filtration of fluids containing a
variety of raw materials, some of which
may have been exposed to viruses, does
not prevent the introduction of viral con-
taminants into the process. Only recently
have membranes with pore sizes small en-
ough to exclude passage of virus particles
become available for industrial-scale opera-
tions. However, these membranes cannot
be introduced into existing processes,
without complex consequences on regula-
tory issues. Therefore, testing still appears
to be the most efficient method to exclude
viruses which may reside within the host
cell line itself, or which could be intro-
duced via the biologically derived raw ma-
terials required for cell culture process.
Specialized service companies have devel-
oped, in close collaboration with the phar-
maceutical client companies, batteries of
validated test procedures. They utilize cell
culture systems in which supernatants or
lysates of the production cells are co-culti-
vated with the corresponding virus-sensi-
tive substrates. Since cells in MSCBs and
in MWCBs are of the highest importance
for the cell culture production process,
samples from these are the first to be con-
sidered for the rather expensive and time-
consuming testing exercise.
A complementary approach to virus
safety is the design of virus kill and re-
moval steps of the protein recovery pro-
cess. These include the physical and chem-
ical principles of separating (theoretical)
viral contaminants from the product, or in-
activating them. Again, appropriate testing
procedures and the demonstration of inac-
tivation and removal of model viruses, as
discussed by Wiebe et al. [139] is consid-
ered a major provision for the safety of re-
combinant products from hamster cells.
While the argument remains a strong
one non-human host cells for human
protein drugs for biosafety reasons it
should be noted that recently (2002) a hu-
man cell line was approved for the produc-
tion of a recombinant protein. A human
embryo kidney cell line, transformed by a
shared adeno-virus DNA (Human Embryo
Kidney 293-cells), was used to produce Ac-
tivated Protein C (E. Lilly).
1.7.4
Product Consistency, Quality, and Purity
Within the past two decades, a rich collec-
tion of methods has become available for
the analysis of purified proteins. In addi-
tion, most of these methods have been op-
timized and fine-tuned to very high sensi-
tivities and resolution. When employed as
routine analytical procedures during the
manufacturing process, they are able to as-
sure high quality and consistency of pro-
tein products [140145]. Nonetheless, the
manufacturer of a biopharmaceutical pro-
tein has one major concern: Will the es-
sential characteristics of a product that has
demonstrated its efficacy in clinical trials
remain the same when produced over
many years in a defined manufacturing
process? It seems surprising, but due to
the large size and complexity of proteins
under study for clinical use today, their
structure and function within the human
body may not be fully understood by the
manufacturer after completion of clinical
trials. This is especially true for the newer
generation of pharmaceutical proteins that
are larger in size, and often contain multi-
ple polypeptides and/or specialized do-
mains with secondary modifications.
Subtle changes which sometimes are dif-
ficult to detect due to inherent heterogene-
ity in protein populations may result in
a loss or modification of activity and could
pose risks to the patient. In order to re-
duce this possibility, batteries of in-process
controls and tests are an inherent part of
the production of clinical biopharmaceuti-
cal proteins. In the following section, sen-
sitive analytical techniques are outlined
and discussed. The objective is to: 1) pre-
vent the occurrence of even small changes
in the procedures for production of the
product; and 2) enable the detection and
exclusion from the final product variants
differing in a major way from that tested
in clinical trials.
The first level of protection against inad-
vertent changes of the product rests in ef-
fective management of the production pro-
cess over time. The challenge is that of
any mature industry: to produce large
amounts of material at competitive cost,
while ensuring that product consistency
and quality are maintained. Clearly, manu-
facturing teams and their supervisors un-
dertake serious efforts to reproduce the
manufacturing process to the utmost de-
tail in every production run. Defining and
describing each of the various steps in the
form of detailed protocols achieve this. Al-
most every aspect of the procedure is
documented, these documents establish-
ing, in the form of cGMP (current Good
Manufacturing Practice) protocols, the ba-
sis for the overall procedure. A sign-off
procedure by supervisors represents an in-
tegral part of this procedure, assuring that
the operating personnel for the manufac-
turing process are in fact controlling, as-
sessing and executing it according to the
established protocols. At critical points of
the overall process, the signatures are pre-
requisites to allow the progression of the
process to proceed to the consecutive
steps.
The time period of cell line and process
development, leading to the establishment
of the manufacturing process, is important
for the definition of critical check points.
During this period, knowledge about pa-
rameters and steps is acquired that can re-
sult in product changes. Once certain lim-
its of variations of process conditions have
been identified (within which no change
was observed), the cGMP protocol is
drafted and finalized. Specific events de-
fined in precise terms as part of the man-
ufacturing protocols can trigger a more
elaborate investigation. Supervisors and
managers can even order an interruption
of the manufacturing process. In extreme
cases, crude product batches are withheld
from further processing and are discarded.
Quality control (QC) is an integral part
of the manufacturing process for recombi-
nant products (see also Part VII, Chapter
1). A comprehensive approach, utilizing
independent validated techniques, is ap-
plied to assess the quality and identity of
the product from various angles (for a re-
view, see also Ref. [146]). It is the goal of
QC efforts to assure that products made
over years of manufacturing will meet the
stated specifications in terms of identity,
quantity, activity, and purity.
In principle, it is no longer difficult to
produce large quantities of highly purified
recombinant proteins, especially when pro-
teins are secreted into the medium. How-
ever, methods to produce recombinant pro-
teins are still part of a young technology,
since its basis is the manipulation of ge-
netic material in the laboratory. Those ma-
nipulations involve the creation of plasmid
vectors and their transfer into mammalian
cells cultivated in vitro. A major concern
has been the fidelity (amino acid sequence
identity) of the final product, particularly
in view of the high degree of ignorance
with respect to gene transfer mechanisms
in higher eukaryotes. It also appears that
transfected DNA may have a somewhat
elevated propensity for mutation during or
following transfer into mammalian cells
[147].
Based on a history of experience with
this technology for more than 20 years, it
can be stated that this young technology
is very reliable. It has been suggested that
rigorous and extensive nucleic acid-based
tests most notably a complete sequence
assessment of the integrated DNA (or
transcribed RNA from recombinant cells)
should be performed [148, 149]. How-
F
i
g
.
1
.
4
O
u
t
l
i
n
e
o
f
p
r
o
c
e
s
s
d
e
v
e
l
o
p
m
e
n
t
e
f
f
o
r
t
s
f
o
r
a
l
a
r
g
e
-
s
c
a
l
e
m
a
n
u
f
a
c
t
u
r
-
i
n
g
p
r
o
c
e
s
s
o
n
t
h
e
b
a
s
i
s
o
f
m
a
m
m
a
l
i
a
n
c
e
l
l
s
.
U
s
u
a
l
l
y
,
m
a
n
y
c
a
n
d
i
d
a
t
e
c
e
l
l
l
i
n
e
s
m
u
s
t
b
e
e
v
a
l
u
a
t
e
d
f
o
r
p
r
o
d
u
c
t
i
v
i
t
y
a
n
d
l
o
n
g
-
t
e
r
m
s
t
a
b
i
l
i
t
y
f
r
o
m
d
i
f
f
e
r
-
e
n
t
t
r
a
n
s
f
e
c
t
i
o
n
s
.
T
h
e
w
a
v
y
l
i
n
e
s
i
n
d
i
c
a
t
e
s
u
b
c
u
l
t
i
v
a
t
i
o
n
s
.
O
v
e
r
t
i
m
e
,
t
h
e
m
a
j
o
r
i
t
y
o
f
t
h
e
s
e
c
a
n
d
i
d
a
t
e
c
e
l
l
l
i
n
e
s
a
r
e
e
l
i
m
i
n
a
t
e
d
t
o
o
b
t
a
i
n
a
f
i
n
a
l
c
a
n
d
i
-
d
a
t
e
l
i
n
e
f
o
r
e
s
t
a
b
l
i
s
h
i
n
g
a
M
a
s
t
e
r
S
e
e
d
C
e
l
l
B
a
n
k
t
h
a
t
l
e
a
d
s
i
n
t
o
c
G
M
P
m
a
n
u
f
a
c
t
u
r
i
n
g
.
I
n
o
r
d
e
r
t
o
s
e
c
u
r
e
t
h
e
s
u
p
p
l
y
o
f
c
e
l
l
s
f
o
r
e
x
p
e
r
i
m
e
n
t
a
l
w
o
r
k
,
s
m
a
l
l
c
e
l
l
b
a
n
k
s
a
r
e
g
e
n
e
r
a
t
e
d
a
l
o
n
g
t
h
e
w
a
y
.
ever, it appears that these complex, expen-
sive and time-consuming tests are not nec-
essary. As pointed out above, powerful and
reliable new methods have emerged in
analytical protein chemistry, and these
have increased the capacity to characterize
purified protein preparations to a high de-
gree of sensitivity and resolution. These
methods represent efficient tools to assess
identity, including the amino acid se-
quence, quantity, potency, and purity of
the product immediately prior to adminis-
tration to the patient [150].
This is not to say that mutants may not
occasionally emerge. A 1000-L reactor con-
tains usually more than 10
12
cells. Muta-
tions will occur at a frequency similar to
that in mammalian genomes (1 bp change
in 10
9
bp for each generation) [151]. Of
course, such mutation will be scattered
over the entire genome, and it is highly
unlikely that individual and specific vari-
ants of a given protein product will
emerge in the population of protein mole-
cules. This is of course different if a mu-
tant plasmid DNA molecule was inte-
grated into the genome of the host cell at
the time of gene transfer.
A telling example for the power of in-
process controls and the associated bio-
chemical assays to identify variants in a
population of molecules is given by Sliw-
kowski et al. [152]. This describes the de-
tection of an amino-acid exchange mutant
of a recombinant monoclonal antibody
(anti-Her2). This variant was detected early
during process development efforts when
using a MTX-amplified clone of CHO
cells. The variant represented about 10%
of the total population of antibody mole-
cules. The origin of this mutant remains
mysterious, but it seems it was the result
of an early event during the development
of the cell line.
1.8
Concluding Remarks
The technology to use mammalian cells
for recombinant biopharmaceutical protein
production is, surprisingly, still in its in-
fancy. Much must be done to establish
production processes in a more straightfor-
ward way, and also to make them more
productive. CHO, NS0, BHK, and PER.C6
cells have been developed that express, in
highly optimized manufacturing processes,
several grams per liter of secreted proteins,
usually antibodies or antibody-fusion pro-
teins. Thus, recent claims [97] that cells of
lymphoid origin (e.g., NS0 cells) are espe-
cially equipped for the secretion of pro-
teins and therefore are preferential high
producers must be questioned. It appears
that immortalized mammalian cells of
whatever origin have tremendous plastic-
ity, both for the uptake of foreign DNA, al-
lowing high level protein synthesis under
bioreactor conditions. Even with the newly
obtained yields in highly optimized pro-
cesses of several grams per liter, one
should not feel that the end of the oppor-
tunity for further improvement has been
reached. Indeed, yields of 1020 g L
1
and
higher product concentrations should be
possible in the near future, particularly if
one considers the fact that batch and ex-
tended-batch processes obtain, at best, cell
densities of about 10
6
mL
1
. These cell
numbers correspond to about 2% biomass
with respect to the total volume in the
bioreactor (2% packed cell volume (PCV)).
Highly developed microbial processes
achieve 2030% PCV.
A data explosion is occurring presently
in biology and in biomedical research.
Knowledge gained from genomics will also
lead us to a much better understanding of
the biochemistry and physiology of mam-
malian cells. As a result, there is every rea-
son to take a highly very optimistic view
that mammalian cells will continue to be
preferred as hosts for recombinant protein
production.
References
1 Andersen, D. (2004) Scientific report at a cell
culture conference in Cancun.
2 Kao, F.-T., Puck, T. T. (1968) Genetics of so-
matic mammalian cells VII. Induction and
isolation of nutritional mutants in Chinese
hamster cells. Proc. Natl. Acad. Sci. USA
60:12751281.
3 Puck, T. T. (1985) Development of the Chinese
hamster ovary (CHO) cell for use in somatic
cell genetics. In: M.M. Gottesman (Ed) Molec-
ular Cell Genetics, John Wiley & Sons, New
York, pp. 3764.
4 Urlaub, G. and Chasin, L. A. (1980) Isolation
of Chinese hamster cell mutants deficient in
dihydrofolate reductase activity. Proc. Natl.
Acad. Sci. USA 77:42164220.
5 Ringold, G., Dieckmann, B., and Lee, F.
(1981) Co-expression and amplification of di-
hydrofolate reductase cDNA and the Escheri-
chia coli XGPRT gene in Chinese hamster
ovary cells. J. Mol. Appl. Genet. 1:165175.
6 Kaufman, R. J. and Sharp, P. (1982) Amplifi-
cation and Expression of Sequences Cotrans-
fected with a Modular Dihydrofolate Reduc-
tase Complementary DNA Gene. J. Mol. Biol.
159:601621.
7 Kaufman, R. J., Wasley, L. C., Spiliotes, A. J.,
Gossels, S. D., Latt, S. A., Larsen, G. R., and
Kay, R. M. (1985) Coamplification and coex-
pression of human tissue-type plasminogen
activator and murine dihydrofolate reductase
in Chinese hamster ovary cells. Mol. Cell.
Biol. 5:17501759.
8 Pennica, D., Holmes, W. E., Kohr, W. J., Har-
kins, R. N., Vehar, G. A., Ward, C. A., Bennett,
W. F., Yelverton, E., Seeburg, P. H., Heyneker,
H. L., Goeddel, D. V., and Collen, D. (1983)
Cloning and expression of human tissue-type
plasminogen activator cDNA in E. coli. Nature
301:214221.
9 Gottesman, M. M. (1985) Molecular Genetics.
John Wiley & Sons, New York.
10 Graham, F. L. and van der Eb, A. J. (1973) A
new technique for the assay of infectivity of
human adenovirus 5 DNA. Virology 52:456
467.
11 Wigler, M., Silverstein, S., Lee, L.-S., Pellicer,
A., Cheng, Y.-C., and Axel, R. (1977) Transfer
of purified Herpes Virus thymidine kinase
gene to cultured cells. Cell 11:223232.
12 Pellicer, A., Wigler, M., and Axel, R. (1978)
The transfer and stable integration of the
HSV thymidine kinase gene into mouse cells.
Cell 14:133141.
13 Robins, D. M., Ripley, S., Henderson, A., and
Axel, R. (1981) Transforming DNA integrates
into the host chromosome. Cell 23:2939.
14 Chen, C. and Okayama, H. (1987) High effi-
ciency transformation of mammalian cells by
plasmid DNA. Mol. Cell. Biol. 7, 8:27452752.
15 Jordan, M., Schallhorn, A. and Wurm, F. M.
(1996) Transfecting mammalian cells: optimi-
zation of critical parameters affecting calcium-
phosphate precipitate formation. Nucleic
Acids Res. 24, 4:596601.
16 Jordan M., Khne, C. and Wurm, F. M. (1997)
Principles of a scaleable transient transfection
and expression technology for mammalian
cells. In: M. J. Carrondo, B. Griffiths and J. P.
Moreira (Eds) Animal Cell Technology from
Vaccines to Genetic Medicine, pp. 4750.
Kluwer Academic Publishers.
17 Finn, G. K., Kurz, B. W., Cheng, R. Z. and
Shmookler, R. J. (1989) Homologous plasmid
recombination is elevated in immortally trans-
formed cells. Mol. Cell. Biol. 9:40094017.
18 Grosjean, F., Batard, P., Jordan, M. and
Wurm, F.M. (2002) S-phase synchronized
CHO cells show elevated transfection effi-
ciency and expression using CaPi. Cytotech-
nology 38, 1/2:5762.
19 Mansour, S. L., Thomas, K. R., and Capecchi,
M. R. (1988) Disruption of the proto-oncogene
int-2 in mouse embryo-derived stem cells: a
general strategy for targeting mutations to
non-selectable genes. Nature 336:348352.
20 Zheng, H. and Wilson, J. H. (1990) Gene tar-
geting in normal and amplified cell lines. Na-
ture 344:170173.
21 Wurm, F. M., Johnson, A., Ryll, T., Khne,
C., Scherthan, H., Glaab, F., Lie, Y. S., Petro-
poulos, C. J. and Arathoon, W.R. (1996) Gene
transfer and amplification in CHO cells effi-
cient methods for maximising specific produc-
tivity and assessment of genetic consequences.
Ann. N.Y. Acad. Sci. 782:7078.
22 Perucho, M., Hanahan, D., and Wigler, M.
(1980) Genetic and physical linkage of exoge-
nous sequences in transformed cells. Cell
22:309317.
23 Finn, G. K., Kurz, B. W., Cheng, R. Z., and
Shmookler Reis, R. J. (1989) Homologous
Plasmid Recombination is Elevated in Im-
mortally Transformed Cells. Mol. Cell. Biol. 9,
9:40094017.
24 Folger, K. R., Wong, E. A., Wahl, G., and Ca-
pecchi, M. R. (1982) Patterns of integration of
DNA microinjected into cultured mammalian
cells: evidence for homologous recombination
between injected plasmid molecules. Mol.
Cell. Biol. 2, 11:13721387.
25 Jordan, M. and Wurm, F. M. (2003) Co-trans-
fer of multiple plasmids/viruses as an attrac-
tive method to introduce several genes in
mammalian cells. In: S. C. Makrides (Ed.)
Gene Transfer and Expression in Mammalian
Cells. New Comprehensive Biochemistry, Vol-
ume 38, pp. 337357.
26 Chen, L., Marmey, P., Taylor, N. J., Brizard, J.-
P., Espinoza, C., DCruz, P., Huet, H., Zhang,
S., Kochko, A., de Beachy, R. N., and Fauquet,
C.M. (1998) Expression and inheritance of
multiple transgenes in rice plants. Nature Bio-
technol. 16:10601064.
27 Wurm, F. M. (2004) Production of recombi-
nant protein therapeutics in cultivated mam-
malian cells. Nature Biotechnology 22,
11:13931398.
28 Wurm, F. M., Johnson, A., Lie, Y. S., Etchever-
ry M. T., and Anderson, K. P. (1992) Host Cell
Derived Retroviral Sequences Enhance Trans-
fection and Expression Efficiency in CHO
Cells. In: Spier, R. E., Griffiths, J. B., Mac-
Donald, C. (Eds) Animal Cell Technology: De-
velopments, Processes and Products. Butter-
worth-Heinemann, Oxford, UK, pp. 3541.
29 Barnes, L. M., Bentley, C. M., and Dickson,
A. J. (2004) Molecular definition of predictive
indicators of stable protein expression in re-
combinant NS0 myeloma cells. Biotechnol.
Bioeng. 85:115121.
30 Jones, D., Kroos, N., Anema, R., von Mont-
fort, B., Vooys, A., van der Kraats, S., van der
Helm, E., Smits, S., Schouten, J., Brouwer, K.,
Lagerwerf, F., van Berkel, P., Opstelten, D.-J.,
Logtenberg, T., and Bout, A. (2003) High level
expression of recombinant IgG in the human
cell line PER.C6. Biotechnol. Prog. 19:163
168.
31 Chu, G., Hayakawa, H., and Berg, P. (1987)
Electroporation for the efficient transfection of
mammalian cells with DNA. Nucleic Acids
Res. 15:13111326.
32 Barsoum, J. (1990) Introduction of stable
high-copy-number DNA into Chinese hamster
ovary cells by electroporation. DNA Cell Biol.
9, 4:293300.
33 Felgner, P. L. and Ringold, G. M. (1989) Cat-
ionic liposome-mediated transfection. Nature
337:387388.
34 Behr, J.-P., Demeneix, B., Loeffler, J.-P., and
Perez-Mutul, J. (1989) Efficient gene transfer
into mammalian primary endocrine cells with
lipopolyamine-coated DNA. Proc. Natl. Acad.
Sci. USA 86:69836986.
35 Muller, S. R., Sullivan, P. D., Clegg, D. O., and
Feinstein, S.C. (1990) Efficient transfection
and expression of heterologous genes in PC12
cells. DNA Cell Biol. 9, 3:221229.
36 OBrien, J. and Lummis, S. C. R. (2004) Biolis-
tic and diolistic transfection: using the gene
gun to deliver DNA and lipophilic dyes into
mammalian cells. Methods 33:121125.
37 Brokx, R. and Gariepy, J. (2004) Peptide- and
polymer-based gene delivery vehicles. Methods
Mol. Med. 90:139160.
38 Putnam, D., Gentry, C. A., Pack, D. W., and
Langer, R. (2001) Polymer-based gene delivery
with low cytotoxicity by a unique balance of
side-chain termini. Proc. Natl. Acad. Sci. USA
98(3):12001205.
39 Jasin, M. and Berg, P. (1988) Homologous in-
tegration in mammalian cells without target
gene selection. Genes Dev. 2:13531363.
40 Smithies, O., Gregg, R. G., Boggs, S. S., Kora-
lewski, M. A., and Kucherlapati, R. S. (1985)
Insertion of DNA sequences into the human
chromosomal beta-globin locus by homolo-
gous recombination. Nature 317:230234.
41 Bode, J., Benham, C., Ernst, E., Knopp, A.,
Marschalek, R., Strick, R., and Strissel, P.
(2000) Fatal connections: when DNA ends
meet on the nuclear matrix. J. Cell Biochem.
35:322.
42 Bode, J., Gtze, S., Ernst, E., Hsemann, Y.,
Baer, A., Seibler, J., and Mielke, C. (2002) Ar-
chitecture and utilization of highly expressed
genomic sites. In: S. C. Makrides (Ed) Gene
transfer and expression in mammalian cells.
Elsevier Science B.V., pp. 551571.
References 753
43 Wilson, T.J. and Kola, I. (2001) The LoxP/CRE
system and genome modification. Methods
Mol. Biol. 158:8394.
44 Gopalkrishnan, R. V., Christiansen, K. A.,
Goldstein, N. I., DePinho, R. A., and Fisher,
P. B. (1999) Use of the human EF-1alpha pro-
moter for expression can significantly increase
success in establishing stable cell lines with
consistent expression: a study using the tetra-
cycline-inducible system in human cancer
cells. Nucleic Acids Res. 27(24):47754782.
45 Kim, S. Y., Lee, J. H., Shin, H. S., Kang, H. J.,
and Kim, Y. S. (2002) The human elongation
factor 1 alpha (EF-1 alpha) first intron highly
enhances expression of foreign genes from
the murine cytomegalovirus promoter. J. Bio-
technol. 93:183187.
46 Wilson, J. H., Berget, P. B., and Pipas, J. M.
(1982) Somatic cells efficiently join unrelated
DNA segments end to end. Mol. Cell. Biol.
2:12581269.
47 Balland, A., Faure, T., Carvallo, D., Cordier, P.,
Ulrich, P., Fournet, B., de la Salle, H., and Le-
cocq, J. P. (1988) Characterisation of two dif-
ferently processed forms of human recombi-
nant factor IX synthesised in CHO cells trans-
formed with a polycistronic vector. Eur. J. Bio-
chem. 172(3):565572.
48 Meents, H., Enenkel, B., Werner, R. G. and
Fussenegger, M. (2002) p 27Kip1-mediated
controlled proliferation technology increases
constitutive sICAM production in CHO-
DUKX adapted for growth in suspension and
serum-free media. Biotechnol. Bioeng. 79,
6:619627.
49 Kaufmann, H. and Fussenegger, M. (2003)
Metabolic engineering of mammalian cells for
higher protein yield. In: S.C. Makrides (Ed)
Gene transfer and expression in mammalian
cells. Elsevier Science B.V., pp. 547569.
50 Wurm, F. M., Gwinn, K. A., and Kingston,
R. E. (1986) Inducible overexpression of the
mouse c-myc protein in mammalian cells.
Proc. Natl. Acad. Sci. USA 83:54145418.
51 Weber, W., Fux, C., Daoud-El Baba, M., Keller,
B., Weber, C. C., Kramer, B. P., Heinzen, C.,
Aubel, D., Bailey, J. E., and Fussenegger, M.
(2002) Macrolide-based transgene control in
mammalian cells and mice. Nature Biotech-
nol. 20:901907.
52 Richards, E. J. and Elgin, S. C. (2002) Epige-
netic codes for heterochromatin formation
and silencing: rounding up of the usual sus-
pects. Cell 108:489500.
53 Girod, P.-A. and Mermod, N. (2003) Use of
scaffold/matrix-attachment regions for protein
production. In: S. C. Makrides (Ed) Gene
transfer and expression in mammalian cells.
Elsevier Science B.V., pp. 359379.
54 Antoniou, M., Harland, L., Mustoe, T., Wil-
liams, S., Holdstock, J., Yague, E., Mulcahy, T.,
Griffiths, M., Edwards, S., Ioannou, P. A.,
Mountain, A., and Crombie, R. (2003) Trans-
genes encompassing dual-promoter CpG is-
lands from the human TBP and HNRPA2B1
loci are resistant to heterochromatin-mediated
silencing. Genomics 82(3):269279.
55 Stief, A., Winter, D. M., Stratling, W. H., and
Sippel, A. E. (1989) A nuclear attachment ele-
ment mediates elevated and position-indepen-
dent gene activity. Nature 341:343345.
56 Phi-Van, L., von Kries, J. P., Ostertag, W., and
Stratling, W. H. (1990) The chicken lysozyme
5' matrix attachment region increases tran-
scription from a heterologous promoter in
heterologous cells and dampens position ef-
fects on the expression of transfected genes.
Mol. Cell. Biol. 10:23022307.
57 Martens, J. H., Verlaan, M., Kalkhoven, E.,
Dorsman, J. C., and Zantema, A. (2002) Scaf-
fold/matrix attachment region elements inter-
act with a p300-scaffold attachment factor A
complex and are bound by acetylated nucleo-
somes. Mol. Cell. Biol. 22:25982606.
58 Fernandez, L. A., Winkler, M., and Grosschell,
R. (2001) Matrix attachment region-dependent
function of the immunoglobulin mu enhancer
involves histone acetylation at a distance with-
out changes in enhancer occupancy. Mol. Cell.
Biol. 21:196208.
59 Zahn-Zabal, M., Kobr, M., Imhof, M. O., Cha-
tellard, P., de Jesus, M., Wurm, F., and Mer-
mod, N. (2001) Development of stable cell
lines for production or regulated expression
using matrix attachment regions. J. Biotech-
nol. 87:2942.
60 Kwaks, T. H. J., Barnett, P., Hemrika, W.,
Siersma, T., Sewalt, R. G. A. B., Satijn, D. P. E.,
Brons, J. F., van Blokland, R., Kwakman, P.,
Kruckeberg, A. L., Kelder, A. and Otte, A. P.
(2003) Identification of anti-repressor ele-
ments that confer high and stable protein pro-
duction in mammalian cells. Nature Biotech-
nol. 21:553558.
61 Mutskov, V. and Felsenfeld, G. (2004) Silenc-
ing of transgene transcription precedes
methylation of promoter DNA and histone
H3 lysine. EMBO J. 23:138149.
62 Hacker, D. L., Derow E. and Wurm, F. M.
(2005) The CELO Adenovirus Gam1 Protein
Enhances Transient and Stable Recombinant
Protein Expression in Chinese Hamster Ovary
Cells. J. Biotechnology 117:2129.
63 Gorman, C. M., Howard, B. H., and Reeves, R.
(1983) Expression of recombinant plasmids in
mammalian cells is enhanced by sodium bu-
tyrate. Nucleic Acids Res. 11:76317648.
64 Riggs, M. G., Whittaker, R. G., Nuemann, J. R.,
and Ingram, V. M. (1977) n-Butyrate causes
histone modification in HeLa and Friend ery-
throleukaemia cells. Nature 268:462464.
65 Cuisset, L., Tichonicky, L., and Delpech, M.
(1998) A protein phosphatase is involved in
the inhibition of histone diacetylation by so-
dium butyrate. Biochem. Biophys. Res. Com-
mun. 246:760764.
66 Makrides, S. C. (1999) Components of vectors
for gene transfer and expression in mamma-
lian cells. Protein Express. Purif. 17:183202.
67 Shirahata, S., Teruya, K., Seki, K., Mori, T.,
Ohashi, H., Tachibana, H., and Murakami, H.
(1992) Oncogene-activated production of re-
combinant proteins for animal cells. In: Spier,
R.E., Griffiths, J.B., MacDonald, C. (Eds) Ani-
mal Cell Technology: Developments, Processes
and Products. Butterworth-Heinemann, Ox-
ford, pp. 5459.
68 Lee, K. H., Guerini-Sburlati, A., Renner, W. A.,
and Bailey, J. E. (1997) Deregulated expression
of cloned transcription factor E2F-1 in Chi-
nese Hamster Ovary cells shifts protein pat-
terns and activates growth in protein-free me-
dium. In: Carrondo, M.J.T. et al. (Eds), Ani-
mal Cell Technology from vaccines to genet-
ic medicine. Kluwer Academic Publishers,
pp. 621623.
69 Ishaqe, A. and Al-Rubeai, M. (2002) Role of
vitamins in determining apoptosis and extent
of suppression by bcl-2 during hybridoma cell
culture. Apoptosis 7:231239.
70 Pak, S. C. O., Hunt, S. M. N., Bridges, M. W.,
Sleigh, M. J., and Gray, P. P. (1996) Super
CHO a cell line capable of autocrine growth
under fully defined protein-free conditions.
Cytotechnology 22:139146.
71 Simpson, N. H., Singh, R. P., Emery, A. and
Al-Rubeai, M. (1999) Bcl-2 overexpression re-
duces growth rate and prolongs G1 phase in
continuous chemostat cultures of hybridoma
cells. Biotechnol. Bioeng. 64:174186.
72 Sunstrom, N.-A., S., Gay, R. D., Wong, D. C.,
Kitchen, N. A., Deboer, L., Gray, P. P. (2000)
Insulin-like growth factor-1 and transferrin
mediate growth and survival of Chinese ham-
ster ovary cells. Biotechnol. Prog. 16:698702.
73 Renner, W. A., Lee, K. H., Hstzimanikatis, V.,
Bailey, J. E., and Eppenberger (1995) Recombi-
nant cyclin E expression activates proliferation
and obviates surface attachment of Chinese
hamster ovary (CHO) cells in protein-free me-
dium. Biotechnol. Bioeng. 47:476482.
74 Arden, N., Nivitchanyong, T., and Beten-
baugh, M. J. (2004) Cell engineering blocks
cell stress and improves biotherapeutic pro-
duction. Bioprocess. J. 2:2328.
75 Umaa, P., Jean-Mairet, J., Moudry, R., Am-
stutz, H., and Bailey, J. E. (1999) Engineered
glycoforms of an antineuroblastoma IgG1
with optimized antibody-dependent cellular
cytotoxic activity. Nature Biotechnol. 17:176
180.
76 Gossen, M. and Bujard, H. (1992) Tight con-
trol of gene expression in mammalian cells by
tetracycline-responsive promoters. Proc. Natl.
Acad. Sci. USA 89:55475551.
77 Fussenegger, M., Morris, R. P., Fux, C., Ri-
mann, M., von Stockar, B., Thompson, C.J.,
and Bailey, J. E. (2000) Streptogramin-based
gene regulation systems for mammalian cells.
Nature Biotechnol. 18:12031208.
78 Simpson, N. H., Singh, R. P., Perani, A., Gol-
denzon, C., and Al-Rubeai, M. (1999) In hy-
bridoma cultures, deprivation of any single
amino acid leads to apoptotic death, which is
suppressed by expression of the bcl-2 gene.
Biotechnol. Bioeng. 59:9098.
79 Mazur, X., Fussenegger, M., Renner, W. A.,
and Bailey, J. E. (1998) Higher productivity of
growth-arrested Chinese hamster ovary cells
expressing the cyclin-dependent kinase inhibi-
tor p27. Biotechnol. Prog. 14:705713.
80 Birch, J. R., Boraston, R. C., Metcalfe, H.,
Brown, M. E., Bebbington, C. R. and Field,
R. P. (1994) Selecting and designing cell lines
for improved physiological characteristics. Cy-
totechnology 15:1116.
81 Clynes, R. A., Towers, T. L., Presta, L. G., and
Ravetch, J. V. (2000) Inhibitory Fc receptors
modulate in vivo cytotoxicity against tumor tar-
gets. Nature Med. 6:443446.
References 755
82 Wright, A. and Morrison, S. L. (1997) Effect of
glycosylation on antibody function: implica-
tions for genetic engineering. Trends Biotech-
nol. 15:2632.
83 Bianchi, A. A. and McGrew, J. T. (2003) High-
level expression of full-length antibodies using
trans-complementing expression vectors. Bio-
technol. Bioeng. 84:439444.
84 Hunt, L., M. De Jesus, M. Jordan, and F.M.
Wurm (1999) GFP expressing mammalian
cells for fast, accurate, non-invasive cell
growth assessment in the kinetic mode. Bio-
technol. Bioeng. 65:201205.
85 Wurm, F. M., unpublished observation.
86 Griffiths, B. and Wurm, F. (2002) Mammalian
cell culture. Encycl. Phys. Sci. Technology
9:3147.
87 Wurm, F. M., Johnson, A., Ryll, T., Khne, C.,
Scherthan, H., Glaab, F., Lie, Y. S., Petropou-
los, C. J., and Arathoon, W. R. (1996) Gene
transfer and amplification in CHO cells. Ann.
N.Y. Acad. Sci. 782:7078.
88 Gandor, C., Leist, C., Fiechter, A., and Assel-
bergs, F. A. (1995) Amplification and expres-
sion of recombinant genes in serum-indepen-
dent Chinese hamster ovary cells. FEBS Lett.
377:290294.
89 Biedler, J. L. and Spengler, B. A. (1976) A novel
chromosome abnormality in human neuro-
blastoma and antifolate-resistance Chinese
hamster cell lines in culture. J. Natl. Cancer
Inst. 57, 3:683689.
90 Wurm, F. M., Gwinn, K. A., and Kingston,
R. E. (1986) Inducible overexpression of the
mouse c-myc protein in mammalian cells.
91 Coquelle, A., Pipiras, E., Toledo, F., Buttin,
G., and Debatisse, M. (1997) Expression of
fragile sites triggers intrachromosomal mam-
malian gene amplification and sets bound-
aries to amplicons. Cell 89:215225.
92 Stark, G. R., Debatisse, M., Giulotto, E., and
Wahl, G. M. (1989) Recent progress in under-
standing mechanisms of mammalian DNA
amplification. Cell 57:901908.
93 Zettlmeissl, G., Wirth, M., Hauser, H. J., and
Kpper, H. A. (1988) Isolation of overproducing
recombinant mammalian cells by a fast and
simple selection procedure. Gene 73:419426.
94 Wurm, F. M., unpublished observation.
95 Hendricks, M. B., Luchette, C. A., and Banker,
M. J. (1989) Enhanced Expression of an im-
munoglobulin based vector in myeloma cells
mediated by coamplification with a mutant
dihydrofolate reductase gene. Biotechnology
7:12711274.
96 Bebbington, C. R., Renner, G., Thomson, S.,
King, D., Abrams, D., and Yaranton, G. T.
(1992) High level expression of a recombi-
nant antibody from myeloma cells using a
glutamine synthetase gene as an amplifiable
selectable marker. Bio/Technology 10:169
175.
97 Barnes, L. M., Bentley, C. M. and Dickson,
A. J. (2000) Advances in animal cell recombi-
nant protein production: GS-NS0 expression
system. Cytotechnology 32:109123.
98 Bebbington, C. R. (1991) Expression of anti-
body genes in non lymphoid mammalian
cells. Methods: A companion to Methods in
Enzymology 2:136145.
99 Hsu, T. C. (1961) Chromosomal evolution in
cell populations. Int. Rev. Cytol. 12:69121.
100 Nunberg, J. H., Kaufman, R. J., Schimke,
R. T., Urlaub, G., and Chasin, L. A. (1978)
Amplified dihydrofolate reductase genes are
localized to a homogeneously staining region
of a single chromosome in a methotrexate-
resistant Chinese hamster ovary cell line.
101 Weidle, U. H., Buckel, P. and Wienberg, J.
(1988) Amplified expression constructs for
human tissue-type plasminogen activator in
Chinese hamster ovary cells: instability in
the absence of selective pressure. Gene
66:193203.
102 Vitek, J. A. (1987) Similarity in dynamics of
single and double minute chromosomes in-
cidence and number of chromosomal aber-
rations during long-term treatment of a hu-
man cell line with methotrexate. Neoplasma
34, 6:665670.
103 Pinkel, D., Straume, T., and Gray, J. W.
(1986) Cytogenetic analysis using quantita-
tive, high sensitivity fluorescence hybridiza-
tion. Proc. Natl. Acad. Sci. USA 83:2934
2938.
104 Pallavicini, M. G., DeTeresa, P. S., Rosette,
C., Gray, J. W., and Wurm, F. M. (1990) Ef-
fects of Methotrexate (MTX) on Transfected
DNA Stability in Mammalian Cells. Mol.
Cell. Biol. 10:401404.
105 Wurm, F. M., Pallavicini, M. G., and Ara-
thoon, R. (1992) Integration and stability of
CHO amplicons containing plasmid se-
quences. Dev. Biol. Stand. 76:6982.
106 Kim, S. J. and Lee G. M. (1999) Cytogenetic
analysis of chimeric antibody-producing
CHO cells in the course of dihydrofolate
reductase-mediated gene amplification and
their stability in the absence of selective
pressure. Biotechnol. Bioeng. 64, 6:741749.
107 Van Wezel, A. L. and van der Velden-de
Groot, C. A. M. (1978) Large scale cultivation
of animal cells in microcarrier culture. Pro-
cess Biochem. 13:6.
108 Van Wezel, A. L., van der Velden-de Groot,
C. A. M., de Haan, H. H., van den Heuvel,
N., and Schasfoort, R. (1984) Large scale ani-
mal cell cultivation for production of cellular
biologicals. Dev. Biol. Stand. 60:229236.
109 Berry, J. M., Barnabe, N., Coombs, K. M.,
and Butler, M. (1999) Production of reovirus
type-1 and type-3 from Vero cells grown on
solid and macroporous microcarriers. Bio-
technol. Bioeng. 62(1):1219.
110 Loumaye, E., Dreano, M., Galazka, A.,
Holes, C., Ham, L., Munafo, A., Eshkol, A.,
Giudice, E., De Luca, E., Sirna, A., Antonetti,
F., Giartosio, C-E., Scaglia, L., Kelton, C.,
Campbell, R., Chappel, S., Duthu, B., Cym-
balista, S., and Lepage, P. (1998) Recombi-
nant follicle stimulating hormone: develop-
ment of the first biotechnology product for
the treatment of infertility. Hum. Reprod.
Update 4, 6:862881.
111 Spier, R. and Kadouri, A. (1997) The evolu-
tion of processes for the commercial exploi-
tation of anchorage-dependent animal cells.
Enzyme Microb. Technol. 21:28.
112 Wood Mackenzie, Horizons, Pharmaceuti-
cals Issue 6 (January 2003).
113 Birch, J. R. and Arathoon, W. R. (1990) Sus-
pension culture of mammalian cells. In:
A. S. Lubiniecki (Ed) Large Scale Mamma-
lian Cell Culture Technology. Dekker Inc.
N.Y., pp. 251270.
114 Afandi, V. and Al-Rubeai M. (2003) Stable
transfection of CHO cells with the c-myc
gene results in increased proliferation rates,
reduces serum dependency and induces an-
chorage independence. Cytotechnology 41:1
10.
115 Bdecker, B. G. D., Newcomb, R., Yuan, P.,
Braufman, A., and Kelsey, W. (1994) Produc-
tion of recombinant Factor VIII from perfu-
sion cultures: I. Large Scale Fermentation.
In: Spier, R.E., Griffiths, J.B., Berthold, W.
(Eds), Animal Cell Technology, Products of
Today, Prospects for Tomorrow, pp. 580590.
116 Annual Report 1995, ARES-SERONO: Folli-
cle Stimulating Hormone.
117 van Wedel, R. J. (1987) Mass culture of
mouse and human hybridoma cells in hol-
low-fiber culture. In: S. Seaver (Ed) Com-
mercial Production of Monoclonal Anti-
bodies. Dekker, New York, pp. 159173.
118 Davies, J. M., Lavender, C. M., Bowes, K. J.,
Hanak, J. A. J., Combridge B. S., and Kings-
land, S. L. (1995) Human therapeutic mono-
clonal Anti-D antibody produced in long-
term hollow-fibre culture. In: Beuvery, E.C.
et al. (Eds) Animal Cell Technology: Develop-
ments towards the 21st century. Kluwer Aca-
demic Publishers, pp. 149153.
119 Heimbuch, S. (1996) Press Release of Cellex
Biosciences: Cellex Biosciences perfusion
equipment used for 1st injectable product
produced in hollow fiber licensed by FDA.
October 30, 1996.
120 Preissmann, A., Bux, R., Schorn, P., and
Noe, W. (1995) Comparative study for the
propagation of anchorage-dependent cells
using different forms of macroporous car-
riers. In: Beuvery, E.C. et al. (Eds) Animal
Cell Technology: Developments towards the
21st century. Kluwer Academic Publishers,
pp. 841845.
121 Byrn, R. A., Mordenti, J., Lucas, C., Smith,
D., Marsters, S. A., Johnson, J. S., Chamow,
S. M., Wurm, F. M., Gregory, T., Groopman,
J. E. and Capon D. J. (1990) Biological Prop-
erties of a CD4 Immunoadhesin. Nature
344, 6267:667670.
122 Bennett, W. F., Paoni, N., Botstein, D., Jones,
A. J. S., Keyt, B., Presta, L., Wurm, F. M. and
Zoller M. (1991) Functional Properties of a
Collection of Charged-to-Alanine Substitu-
tion Variants of Tissue-Type Plasminogen
Activator. J. Biol. Chem. 266, 8:51915201.
123 Ostrove, J. M., Iyer, P., Marshall, J., and Va-
cante, D. (1998) Comparison of manufactur-
ing techniques for Adenovirus production.
In: Merten, O.W., et al. (Eds) New Develop-
ments and New Applications in Animal Cell
Technology. Kluwer Academic Publishers,
pp. 515521.
124 Gorman, C. M., Gies, D. R., McCray, G.
(1990) Transient Production of Proteins
Using an Adenovirus Transformed Cell Line.
DNA Protein Eng. Technol. 2:128.
References 757
125 Meissner, P., Kulangara, A., Pick, H., Chatel-
lard, P. and Wurm, F. M. (2001) Transient
Gene Expression: Recombinant Protein Pro-
duction with Suspension-Adapted HEK-
293EBNA Cells. Biotechnol. Bioeng. 75,
2:197203.
126 Boussif, O., Lezoualch, F., Zanta, M.,
Mergny, M., Scheman, D., Demeneix, B.,
and Behr, J.-P. (1995) A versatile vector for
gene and oligonucleotide transfer into cells
in culture and in vivo: Polyethyleneimine.
127 Derouazi, M., Girard, P., Van Tilborgh, F.,
Muller, N., Bertschinger, M. and Wurm, F.
M. (2004) Serum-free large scale transient
transfection of CHO cells. Biotechnol.
Bioeng. 84, 7:537545.
128 Schlaeger, E. J., Legendre, J. Y., Trzeciak, A.,
Kitas, E. A., Christensen, K., Deuschle, U.,
and Supersaxo, A. (1998) Transient transfec-
tion in mammalian cells. In: Merten, O. W.
et al. (Eds) New Developments and New Ap-
plications in Animal Cell Technology.
Kluwer Academic Publishers. pp. 105112.
129 Girard, P., Derouazi, M., Baumgartner, G.,
Bourgeois, M., Jordan, M., Jacko, B. and
Wurm, F. M. (2002) 100 Liter-transient trans-
fection. Cytotechnology 38:1521.
130 Wurm, F. M. and Bernard, A. (2001) Transi-
ent gene expression from mammalian cells
a new chapter in animal cell technology?
Cytotechnology 35:155156.
131 Reilly, D. (Genentech), conference report
2004.
132 Batard, P., Jordan, M., Chatellard, P. and
Wurm, F. M. (2001) Transfer of high copy
number plasmid into mammalian cells by
calcium phosphate transfection. Gene
270:6168.
133 Hayflick, L. and Moorhead, P. S. (1961) The
serial cultivation of human diploid cell
strains. Exp. Cell. Res. 25:585621.
134 Hayflick, L. (1997) SV 40 and Human Can-
cer, letter. Science 276:336337.
135 Food and Drug Administration, Center for
Biologics Evaluation and Research (1987)
Points to Consider in the Characterization of
Cell Lines Used to Produce Biologicals, re-
vised 1993.
136 Wiebe, M. E. and May, L. H. (1990) Cell
banking. In: Lubiniecki, A. S. (Ed) Large
Scale Mammalian Cell Culture Technology.
Dekker, New York, pp. 147160.
137 Prusiner, S. B. (1997) Prion diseases and the
BSE crisis. Science 278:245251.
138 Aguzzi, A. and Polymenidou, M. (2004)
Mammalian Prion Biology: One century of
evolving concepts. Cell 116:313327.
139 Wiebe, M. E., Becker, F., Lazar, R., May, L.,
Casto, B., Semense, M., Fautz, C., Garnick,
R., Miller, C., Masover, G., Bergman, D.,
and Lubiniecki, A. S. (1989) A multifaceted
approach to assure that recombinant tPA is
free of adventitious virus. In: Spier, R. E.,
Griffiths, J. B., Berthold, W. (Eds) Advances
in Animal Cell Biology and Technology for
Bioprocesses. Croy, pp. 6871.
140 OConnor, J. V., Keck, R. G., Harris, R. J., and
Field, M. J. (1994) In: Brown, F., Lubiniecki
A. S. (Eds) Genetic Stability and Recombi-
nant Product Consistency. Dev. Biol. Stand.
Karger, Basel, pp. 165173.
141 Garnick, R. L. (1992) Peptide mapping for
detecting variants in protein products. Dev.
Biol. Stand. 76:117130.
142 Lu, H. S., Tsai, L. B., Kenney, W. C. and Lai,
P.-H. (1988) Identification of unusual re-
placement of methionine by norleucine in
recombinant interleukin-2 produced by E.
coli. Biochem. Biophys. Res. Commun.
156:807813.
143 Chloupek, R. C., Harris, R. J., Leonard, C. K.,
Keck, R. G., Keyt, B. A., Spellman, M. W.,
Jones, A. J. S., and Hancock, W. S. (1989)
Study of the primary structure of recombi-
nant tissue plasminogen activator by re-
versed-phase high-performance liquid chro-
matographic tryptic mapping. J. Chromatogr.
463:375396.
144 OConnor, J. V. (1993) The use of peptide
mapping for the detection of heterogeneity
in recombinant DNA-derived proteins. Bio-
logicals 21:111117.
145 Stephenson, R. C. and Clarke, S. (1989) Suc-
cinide formation from aspartyl and asparagi-
nyl peptides as a model for the spontaneous
degradation of proteins. J. Biol. Chem.
264:61646170.
146 Anicetti, V. and Hancock, W. S. (1994) Analy-
tical Considerations in the Development of
Protein Purification Processes. In: R. Harri-
son (Ed) Protein Purification Process Engi-
neering. Marcel Dekker, Inc., New York.
147 Calos, M., Lebkowski, J. S., and Botchan,
M. R. (1983) High mutation frequency in
DNA transfected into mammalian cells.
148 Galibert, F. (1990) Stability of a gene recom-
binant: what does it mean and how to check
for it. Biologicals 18:221224.
149 Galibert, F. (1994) Assessing genetic stability
at the nucleic acid level. In: Brown, F., Lubi-
niecki, A. S. (Eds) Genetic Stability and Re-
combinant Product Consistency. Dev. Biol.
Stand. Karger, Basel, pp. 2730.
150 Wurm, F. M., Petropoulos, C. J. and OCon-
nor, J. V. (1996) Manufacture of proteins
based on recombinant CHO cells: Assess-
ment of genetic issues and assurance of con-
sistency and quality. In: E. R. Schmidt, Th.
Hankeln (Eds) Transgenic Organisms and
Biosafety Horizontal Gene Transfer, Stabil-
ity of DNA, and Expression of Transgenes.
Springer, New York, pp. 283304.
151 Drake, J. W. (1991) Spontaneous mutation.
Annu. Rev. Genet. 25:125146.
152 Harris, R. J., Murnane, A. A., Utter, S. L.,
Wagner, K. L., Cox, E. T., Polastri, G. D.,
Helder, J. C., and Sliwskowski, M. B. (1993)
Assessing genetic heterogeneity in produc-
tion cell lines: Detection by peptide mapping
of a low level Tyr to Gln sequence variant in
a recombinant antibody. Bio/Technology,
pp. 12931297.
References 759
Abstract
Complex glycosylated biopharmaceutical
proteins are typically produced in mamma-
lian cells, and the majority originate from
Chinese hamster ovary (CHO) cells and
mouse NS0 cells. The development of
mammalian super-producer cells from
these starter cell lines is an unpredictable
and time-consuming effort, requiring the
identification of rare clones which com-
bine integration of the expression unit into
a highly active genomic locus with superi-
or folding, processing and secretion cap-
abilities. Fine tuning the selection and vec-
tor, which includes new cellular promo-
ters, allows us to reproducibly generate
productive clone pools of CHO cells suit-
able for immediate production of test ma-
terial and improves identification of supe-
rior clones. Alternatively, the fast and reli-
able generation of clones is achieved by
site-specific cassette exchange based on
heterospecific flp sites. We have expanded
the strategy to use the strong IgH locus of
the G-line, a human/mouse heterohybrido-
ma: replacement of the endogenous hu-
man IgM heavy chain gene provides the
environment for efficient transcription, se-
cretion and a mostly human glycosylation
pattern for Ig fusion proteins. As a new
platform alternative to CHO and NS0,
which supports the production of fully hu-
man proteins, we evaluate human de-
signer cell lines of various tissues created
directly from primary cells.
Abbreviations
CMV cytomegalovirus
GS glutamine synthetase
IRES internal ribosome entry site
LCR locus control region
MSX methionine sulfoximine
MTX methotrexate
2.1
Mammalian Cells as a Workhorse to
Produce Protein-based Biopharmaceuticals
The majority of biopharmaceutical pro-
teins are complex glycoproteins. Among
them, monoclonal antibodies have experi-
enced tremendous growth over recent
years with some products reaching
blockbuster status (see also Part V, Chap-
ters 1 and 2). They are followed by cyto-
kines and fusion proteins truncated re-
761
2
Alternative Strategies and New Cell Lines for High-level Production
ISBN: 3-527-31184-X
ceptors or ligands equipped with addi-
tional effector domains (see also Part V,
Chapters 6 and 7). Replacement therapies
using recombinant versions of human gly-
coproteins represent the major treatment
option for many monogenic genetic dis-
eases (see also the Introduction to this
book). All these proteins contain multiple
domains, and have substantial require-
ments for folding and post-translational
processing. Their function is often depen-
dent on, or at least modulated by, carbohy-
drate structures. The glycosylation pattern
is a crucial factor for correct protein fold-
ing, intracellular trafficking and secretion,
as well as for in vivo clearance rate, immu-
nogenicity, proteolytic stability and full bio-
logical activity of the recombinant glyco-
protein (see also Part IV, Chapter 7) [14].
Moreover, therapeutic glycoproteins may
be rendered antigenic upon exposure of
epitopes that are normally masked by oli-
gosaccharides (see also Part VI, Chapter
3). Whereas lower eukaryotic systems such
as yeast can cope with some aspects of
folding, proteolytic processing and phos-
phorylation (see also Part IV, Chapter 13),
only mammalian cells perform carboxyla-
tion, isoprenylation, and add the expected
N- and O-linked sugars (see also Part IV,
Chapter 12).
This capability comes at a high price:
mammalian cell lines are substantially
more demanding with respect to media
and fermentor design (see also Part IV,
Chapter 1). Lower cell densities and prod-
uct yields per cell result in comparatively
low volumetric productivity. Under these
conditions manufacturing costs become a
substantial parameter affecting the success
or failure of a biopharmaceutical. In addi-
tion, slow replication of mammalian cells
(duplication time 3648 h) compared to
prokaryotes and lower eukaryotes increases
the time required for establishment of pro-
ducer lines and generation of clinical ma-
terial. However, modern expression vectors
and cell lines as well as improved culture
media and process designs have raised
yields from below 100 mg L
l
to 5 g L
l
for
individual antibodies (see also Part IV,
Chapter 16). Although already at a very
high degree of complexity, careful analysis
of the existing technology, rational design
of cell lines and modulation of biochem-
ical pathways is expected to boost this
number even further. New approaches cap-
able of improving yields or shortening
time lines are of great importance. This
chapter will summarize general strategies
in mammalian cell line development,
highlight the most essential factors, and
provide a more detailed description of al-
ternative approaches exploring new uncon-
ventional cell substrates and locus-specific
gene targeting.
2.2
The Cell Line of Choice
Any mammalian cell line has the basic
machinery to express and secrete recombi-
nant protein, and huge numbers of cell
lines with suitable growth properties are
available from various tissues and species.
The small number of cell lines industrially
used for manufacturing is, therefore, sur-
prising. Two hamster cell lines, the Chi-
nese hamster ovary cell line (CHO) and
the baby hamster kidney cell line (BHK),
and two genetically related mouse cell
lines, the myeloma NS0 derived from
BALB/c mice, and the hybridoma SP2-0, a
fusion of the myeloma with B cells from
the same mouse strain, supply most of the
mammalian cell-based biopharmaceuticals,
whether marketed or still under develop-
ment. Once commonly accepted as pro-
ducers, a large body of information about
2 Alternative Strategies and New Cell Lines for High-level Production of Biopharmaceuticals 762
these cell lines has accumulated and al-
lowed us to build improvements on top of
sophisticated existing technology, further
increasing the acceptance of the respective
cell lines. Moreover, clinical studies and
marketed products have provided substan-
tial safety information about CHO and
NS0 cell lines, resulting in a higher level
of acceptance by regulatory agencies such
as the FDA (see also Part VII, Chapters 4
and 5).
The production cell lines were selected
mainly for their growth properties: they
are propagated in synthetic or chemically
defined media with a doubling time of 24
36 h. However, originating from natural
tumors (plasmacytoma, NS0) or embryonic
tissue (CHO), these cells have lost most
differentiated features. This also includes
loss of the highly specialized expression
and secretion apparatus of differentiated
cells.
In contrast, in living organisms, most of
the secretory proteins are provided by ter-
minally differentiated resting cell types
equipped with a unique set of transcrip-
tion factors to activate specific promoters
and induce complex adaptations in the en-
doplasmic reticulum and Golgi apparatus.
Examples are plasma cells, secretory cells
of the pituitary, pancreatic island cells and
hepatocytes. Special pluripotent precursor
cells or stem cells (see also Part I, Chap-
ters 11 and 12) are required to maintain
homeostasis. Proliferation and efficient
production of secretory proteins seem to
be mutually exclusive. This conflict may
be specifically addressed in new or engi-
neered producer cell lines. One example is
the separation of growth and production
in a biphasic process: the cell is engi-
neered to express a protein inducing dif-
ferentiation or blocking cell cycle in a
drug-regulated fashion. The cyclin kinase
inhibitor p27 which prevents phosphoryla-
tion of Rb causing arrest in the G
1
phase
of the cell cycle may serve as an example.
Taking the current selection of producer
cells into account, it may seem that the
mammalian species of origin does not
have any impact on the quality of the
product. However, much care is taken that
human biopharmaceuticals contain human
coding sequences. Whereas the first anti-
bodies applied in clinical trials were de-
rived from mouse genes and created a se-
vere human anti-mouse antibody response
[5], todays antibody therapeutics are
mainly constituted of human sequences
(see also Part V, Chapter 2). They originate
from phage-displayed antibody libraries ex-
pressing variable domains of human ori-
gin or from transgenic mice in which IgG
genes are replaced with their human coun-
terparts [68]. While this improves the
pharmacological features substantially,
even these advanced biopharmaceuticals
may induce an immunologic response or
suffer more rapid clearance. Post-transla-
tional modifications of human or huma-
nized immunoglobulins produced on cells
of nonhuman origin may contribute to
this phenomenon. Although mammalian
cells in general provide complex N- and O-
linked glycosylation (sugars attached to as-
paragine and threonine residues of the
polypeptide chain), the specific pattern de-
pends on the tissue type and species of
origin as well as on cell culture conditions
[911] (see also Part IV, Chapter 1).
For instance, proteins produced in
mouse cells carry glycans containing Gal
a13Gal residues, which are missing in
human cells [12]. A high titer of anti-Gal
a13Gal antibodies in humans [13] causes
a rapid clearance of proteins carrying this
residue in their glycans. Antibodies pro-
duced in CHO cells which lack Gal a1
3Gal residues still require high dosages.
Therefore, it is likely that other post-trans-
lational modifications are involved in a
specific human immune response against
antibodies with human primary sequence,
but produced in CHO cells.
There are even indications that not just
species, but individual tissues, provide a
specific glycosylation pattern with func-
tional implications. For instance, brain-de-
rived glycoproteins are reported to contain
a higher degree of fucosylation and high
amounts of bisecting N-acetylglucosamine
[14], whereas in blood-derived glycopro-
teins a high rate of terminal sialic acid is
evident which is likely to be required to
protect the protein from clearance via the
hepatic asialoglycoprotein receptor.
Human cells or cells with human glyco-
sylation machinery should minimize these
problems. However, for many years the
regulatory hurdles for human cells have
been even stronger than those for rodent
cells. The lack of a species barrier allowing
easier transfer of adventitious agents was
considered as a major limitation. On the
other hand, it can be argued that infection
with human pathogenic agents is likely to
result in a full-blown pathogenic effect in
human cells that is easy to detect, whereas
the agent may be dormant in rodent cells.
For all new cell lines, whether of animal
or human origin, the risk of transmission
of prion-based diseases is addressed with
strict documentation requirements and the
lack of contact with any potentially in-
fected bovine material (see also Part I,
Chapter 6). So far, only one such cell line,
PER.C6, a transformed human retinoblast,
has entered the market (see also Part IV,
Chapter 3).
2.3
Pushing Expression Levels Impact of
Vector Design and Cell Clone Selection
During the 1980s, multiple strong promo-
ters and enhancers were described, and
functional models for the relationship be-
tween the core promoters and upstream
elements were proposed [15, 16]. Most of
these promoters are of viral origin (from
human or mouse cytomegalovirus, SV40
or Rous sarcoma retrovirus). Their core
promoter activity is dominated by a TATA
box 2030 bp upstream of the start site,
which directs accurate transcription initia-
tion via binding of a protein called TBP
(TATA-binding protein), recruitment of as-
sociated factors and formation of the poly-
merase II pre-initiation complex. The core
promoter was found to be functionally se-
parated from the enhancer, a collection of
transcription factor-binding sites acting in-
dependent of position and orientation, and
mediating promoter strength via removal
of nucleosomal repression. Despite the 10-
to 50-fold different promoter activity in
transient assays (expression measured 23
days post-introduction of recombinant
DNA), stable producer clones containing
the strongest promoter [human cytomega-
lovirus (hCMV) IE] have no clear advan-
tage over clones derived with other viral
promoters. Moreover, expression levels
vary greatly between individual clones con-
taining the same vector and in many
clones expression declines with prolonged
propagation. One explanation for this ob-
servation is that viral promoters integrated
into the host genome preferentially be-
come inactivated by DNA methylation [17]
or progressive deacetylation of histones
H3 and H4 [1820]. Both processes are
linked: DNA methylation induces deacety-
lation of histones making the region inac-
cessible to transcription factors and exten-
sive acetylation is able to prevent methyla-
tion at promoter sites [21]. The hCMV IE
promoter, one of the most active and fre-
quently used promoters in cell line estab-
lishment, is affected so strongly that only
very few stable CHO clones maintain ex-
pression at a medium or higher level. Spe-
cific sequences such as the chicken HS4
insulator adjacent to the promoter/enhan-
cer can protect from both methylation and
histone deacetylation [22]. The search for
stable and highly expressing clones after
random integration of the vector, which
makes cell line generation so tedious and
time consuming, simply identifies rare
genomic sites with functions similar to
those mentioned above. Once found, time-
efficient approaches can be established by
using these same advantageous locations
for other transgenes. This makes homolo-
gous recombination and integration by
site-specific recombinases so attractive in
cell line design. Typically, a reporter gene
(such as b-galactosidase) linked to a site
for recombinases such as flp or cre is used
to identify a preferable locus for integra-
tion (see also Part III, Chapter 2). During
a secondary transfection in the presence of
recombinase, the gene of interest is in-
serted at the predetermined position and
the test gene is inactivated. In this chapter
we describe a flp recombinase-based ex-
change system applied to a selected locus
in CHO and to the highly active immuno-
globulin locus of a human mouse hetero-
hybridoma.
Alternatively, sequences proposed to sta-
bilize or increase expression may be in-
serted into the vector. Multiple such ele-
ments have been described such as ubiqui-
tous chromatin opening (UCOE) element
or the EASE element [23]; US Patent
6,312,951). Comparable to matrix attach-
ment regions, insulators or locus control
regions (LCR), these elements act in cis
(upon the same DNA molecule) in stably
transformed cell lines by rendering the
DNA accessible to transcription indepen-
dent of the site of integration and/or by
protecting CpG islands in the proximity of
promoters from methylation. In contrast
to LCR regions, however, the elements act
in a tissue-independent manner. It is no
surprise that the effect of these elements
was demonstrated and is most pronounced
with inactivation-sensitive promoters such
as hCMV promoter.
We and others have isolated regions
from cellular genes that are strong promo-
ters and enhancers and in addition trans-
fer the property of locus-independent ex-
pression and prevent transgene deactiva-
tion. As an example for this strategy, 12 kb
of upstream and 3 kb of downstream areas
of the hamster EF1 a gene have provided
stable expression levels exceeding those of
the hCMV promoter by at least an order of
magnitude [24]. In contrast, a 1.3-kb re-
gion of the human EF1 a promoter en-
ables only moderate expression levels.
In addition to the transgene cassette, ex-
pression vectors typically contain selection
marker genes. They primarily serve to
eliminate untransfected and transiently
transfected cells after transfection, and
help to generate a clone pool from which
high producers can be selected. However,
they may also be used to substantially en-
rich the fraction of high producers. It is
believed that a transcriptional link between
the marker and the gene of interest is re-
quired to achieve this goal. For this strat-
egy, both genes are placed on a bicistronic
message and driven by a single strong pro-
moter. While the gene of interest posi-
tioned close to the cap site at the 5'-end of
the message is expressed in a cap-depen-
dent manner, expression of the marker in
the second position is ensured by an inter-
nal ribosome entry site (IRES) often taken
2.3 Pushing Expression Levels Impact of Vector Design and Cell Clone Selection 765
from picorna viruses (encephalomyocardi-
tis virus or poliovirus). Despite the pres-
ence of the IRES element, expression of
the marker gene is impaired. We have
found that this reduced marker expression
is most critical to rich selectivity for high
producers with increasing drug concentra-
tions. We have achieved the same effect by
expressing both genes as separate tran-
scripts located in close proximity. Marker
gene expression in IRES-based constructs
strongly depends on the nature of the
gene of interest in the first position [25].
This complicates selection as appropriate
drug concentrations have to be determined
for each new protein. In contrast, generic
selection strategies can be applied when
separate transcription units are used.
The nature of the marker itself is crucial
to the efficacy of the selection process.
One class represented by neomycin phos-
photransferase (npt), hygromycin B-phos-
photransferase (hpt) or blasticidin deami-
nase (bda) and puromycin N-acetyl-trans-
ferase (pac) encodes enzymes to inactivate
drugs blocking protein biosynthesis; the
other auxotrophic markers such as gluta-
mine synthetase (GS) (see also Part IV,
Chapter 4) and dihydrofolate reductase
(DHFR) (see also Part IV, Chapter 1) en-
codes metabolic enzymes which eliminate
specific nutritional requirements. Auxo-
trophic markers require target cell lines
deficient in the respective genes like CHO
dhfr
cell line clones (DUXB11 and DG44)

[26, 27] from which the gene has been
mutated or deleted, or myeloma cells pos-
sessing very little GS activity per se [28].
Drug inhibition of the enzyme [methio-
nine sulfoximine (MSX) and methotrexate
(MTX) for GS and DHFR, respectively] in
multiple steps induces amplification of the
marker gene and the colocalising trans-
gene [28, 29]. This time-consuming pro-
cess has provided most of the earlier pro-
duction cell lines. Used at a single drug
concentration, selection with MSX or MTX
eliminates low producing clones. The com-
bination of markers from both classes and
the use of stable cellular promoters has al-
lowed us to generate CHO clone pools
reaching up to 14 pgcell
1
day
1
of a re-
combinant glycoprotein. This strategy even
competes with the specific targeting ap-
proaches described below.
2.4
A Single CHO High-producer Clone
for Multiple Products
The investment into a producer cell is sub-
stantial and increasing when a cell line en-
ters later stage phases of clinical develop-
ment. Defining and fine tuning media and
processes often takes more than a year,
and requires a larger team of process engi-
neers. The time is well spent because volu-
metric productivity can often be increased
by an order of magnitude (see also Part
IV, Chapter 1). This effort focuses on a
particular producer clone rather than a
starting cell line. With every new product
candidate introduced into a given starting
cell line the investment has to be repeated
from scratch. It would be intriguing to ex-
change one protein for another and keep
most of the features (e.g., high expression
level) of the particular clone. As discussed
above, the genomic locus harboring the
foreign gene substantially contributes to
productivity. In addition, the preferred
clone has adapted to efficient protein fold-
ing, glycosylation and secretion and has
escaped the unfolded protein response [30,
31], a protective biochemical pathway in-
duced by stressful overexpression of pro-
teins. This selection therefore cannot be
carried out with an empty producer cell
line.
We have explored flp recombinase-
mediated exchange of one transgene for
another to approach this issue. As a test
gene we used the adipostatic hormone lep-
tin linked to an IgG4 Fc-domain (hobFc)
which represents the large group of Fc fu-
sion proteins. The gene was cloned down-
stream of a human CMV promoter linked
to the first intron of EF1a gene. An SV40-
driven blasticidin gene was used for selec-
tion. The two expression units were
flanked by target recognition sites (frt) for
flp recombinase (see also Part III, Chapter
2). To favor gene exchange over excision
we have used mutated frt sites differing in
there core sequence F3 and F5 [32]: these
sites efficiently recombine with identical
frt sites, but fail to interact with each
other. A third heterospecific site (wild-type
frt) was inserted between the leptin gene
and its promoter to allow exchange of the
gene only. We positioned a promoterless
ATG-deficient neo gene outside of the re-
placement cassette. The exchange vector
was equipped with a minimal promoter as
well as an in-frame ATG to activate the neo
gene allowing selection for correct ex-
change (Fig. 2.1). As proof for this strategy
we introduced a promoterless gfp gene and
the signals for activation of the neo gene
into a clone pool harboring the target vec-
tor (Fig. 2.2) and found that gfp expression
was activated in all neo resistant cells. To
select a superior starter clone the target
vector was introduced into a pre-selected
CHO DUXB11 clone by electroporation,
the transfection method providing the high-
est degree of single-copy integration events
[33, 34]. Screening of 1500 clones yielded
three clones with productivities between 6
and 10 pg cell
1
day
1
. While leptinFc
could be reproducibly exchanged by the
gene of interest in several of the clones,
we unexpectedly observed variations in the
expression level among clones originating
from a single individual recipient cell.
Some, but not all, variations were reflected
by different RNA levels (Fig. 2.3). This het-
erogeneous expression can be attributed to
the perturbance of the architecture of a pre-
Fig. 2.1 Gene replacement at a predefined locus.
The test gene leptinFc (hobFc) and the primary
selection marker blasticidin deaminase (bda) re-
siding between heterospecific frt sites (frtF3, frtwt
and frtF5) are exchanged for the target gene. The
minimal promoter and an in-frame ATG present in
the targeting vector activate the neomycin resis-
tance gene (npt) used to select for correct recom-
binants.
viously stable locus caused by epigenomic
phenomena. Moreover, the nature of the
transgene itself influenced the level of ex-
pression as well as the degree of variation.
While the system was not able to completely
replace clone screening, it still provides a
substantial advantage for the evaluation of
multiple product candidates: medium or
high producers are obtained with high re-
producibility within 5 weeks after screening
of 10 or 20 clones. Morphologic features
and growth parameters of secondary clones
usually are inherited from the primary
clones. Therefore, the search for a process-
friendly starter clone may render this flp
system even more valuable. For some pro-
teins superior productivity has been already
achieved: CHO clones secreting up to
40 pg cell
1
day
1
were generated with the
flp strategy for a human proteoglycan.
Fig. 2.2 Highly efficient gene targeting in CHO cells. A pro-
moterless gfp gene is activated in all cell clones surviving
G418 selection after flp-mediated recombination in a pool of
clones carrying the targeting vector.
Fig. 2.3 Heterogeneity of expression from a single
genomic locus. A leptinFc-expressing starter
clone was targeted with a promoterless a
1
-anti-
trypsin gene. Individual daughter clones were ana-
lyzed for daily cellular productivity at a defined
cell density (10
6
mL
1
). Levels of a
1
-antitirypsin
RNA relative to endogenous hamster b-actin RNA
were determined in a SYBR Green-based real-time
PCR assay. Variable productivity dependent and in-
dependent of RNA levels is observed.
2.5
The G-line: Use of the Immunoglobulin
Locus of a Human/Mouse Heterohybridoma
for Heterologous Gene Expression
A well-characterized natural cell line with
long-term, high-level protein secretion and
a known locus responsible for this expres-
sion may provide a template for more reli-
able heterologous gene expression as an al-
ternative to identification of superior geno-
mic loci in common cell lines after ran-
dom integration of a target vector and
large-scale screening. We have explored a
human mouse heterohybridoma, which ex-
pressed a human IgM antibody in a stable
configuration for over 2 years, as a poten-
tial protein producer cell line.
The heterohybridoma CB03 was created
by fusing human B lymphocytes from a pa-
tient with chronic thrombocytopenia, ob-
tained by therapeutic splenectomy, with
the mouse myeloma line P3X63Ag8. CB03
secretes human autoantibodies of the IgMk
type which react with human platelets, and
double- and single-stranded DNA [35]. The
hybridoma was shown to secrete the anti-
body in a stable manner at a rate of
45 pg cell
1
day
1
over a period of 2 years.
This is in striking contrast to the majority
of heterohybridomas, which tend to quickly
lose human chromosomes, resulting in un-
stable immunoglobulin expression. More-
over, expression was preserved when the
cell line was cultivated in high-density fer-
mentation systems of the CellPharm family
(Unisyn Technologies) and remained stable
in five independent fermentor runs which
had a mean duration of 66 days. Expression
did not decrease below 30 pg cell
1
day
1
when the medium was exchanged for a pro-
tein-free medium in a continuous fermenta-
tion run.
This suggests that the immunoglobulin
loci of CB03 are highly accessible and stable,
and are, therefore, well suited to drive a het-
erologous transgene. To assess the possibil-
ity to target these loci, thereby not only intro-
ducing a transgene cassette, but also abolish-
ing IgM expression, the cell line was sub-
mitted to spectral karyotype analysis to
find out whether the respective loci are pres-
ent as single copies. The typical CB03 cell
contains 6994 chromosomes with a domi-
nance of mouse chromosomes. Via hybridi-
zation with specifically labeled human chro-
mosome libraries, eight complete human
chromosomes [4, 5, 7, 10, 14, 17, 18 and
22] and fragments of others [4, 811, 14
and 16], each linked to a mouse chromo-
some, were identified (Fig. 2.4). Since a com-
plete and a partial copy of chromosome 14
(the chromosome harboring the IgH genes)
were found, in-situ hybridization with an
IgH probe was performed and a single copy
of the human IgH region was identified.
With a single chromosome 22 present, a sin-
gle copy of the Igk locus was expected as
well.
Using the known cDNA sequence and
the IMGT database, we have identified V1
2, D1, J6 and l as the elements participat-
ing in constitution of the heavy chain
gene. A sequence map of the rearranged
IgH locus of CB03 was built including se-
quences located upstream of the V
H
pro-
moter region of V12. Using polymerase
chain reaction (PCR) primers located
2000 bp upstream of the transcription start
point for V12 and within J
H
6, the pre-
dicted structure was confirmed. A typical
targeting vector for the IgH locus was con-
structed using a proofreading PCR system.
This vector consists of a short flank
(1930 bp), which represents sequences up-
stream of the V
H
promoter, the promoter it-
self, the transcriptioninitiationpoint and the
RNAleader sequence without the start codon
and a long flank (7400 bp) ranging from J
H
6
to C
H
1 spanning the entire C
l
intron.
2.5 The G-line: Use of the Immunoglobulin Locus of a Human/Mouse Heterohybridoma 769
As for the CHO approach, the targeting
vector was equipped with the blasticidin
gene for selection and hobFc as the repor-
ter gene. Either the CMV/EF1 fusion pro-
moter or the endogenous V
H
promoter
were used to drive the reporter. The ex-
pression units were again flanked by het-
erospecific flp sites to allow for secondary
exchange (Fig. 2.5).
Fig. 2.4 Chromosome analysis of the heterohybri-
doma CB03. GTG banding (upper left), spectral
karyotype analysis (upper right) and identification
of human chromosomes by hybridization with
specifically labeled human chromosome libraries
(lower panel). Eight complete human chromo-
somes [4, 5, 7, 10, 14, 17, 18 and 22] were identi-
fied. In addition, fragments of human chromo-
somes 4, 811, 14, and 16 were found, each
linked to a mouse chromosome
The composition of the k gene locus
was investigated and a targeting construct
was designed using a similar approach.
The V
k
319 gene (upstream) and J
C
2
(downstream) are the components forming
the active k gene of CB03. The flanks were
limited to 4000 (V
k
) and 4500 bp (J
C
2) in
order to exclude highly repetitive se-
quences located further upstream and
downstream. Hygromycin and a
1
-antitryp-
sin were used as selection marker and re-
porter, respectively. An independent gene
replacement system similar to that of the
heavy chain, but based on alternative frt
sites and an inactive histidinol resistance
marker, was included.
In order to screen clone numbers large
enough to detect homologous recombi-
nants, we developed direct cell staining
techniques for secreted IgM and IgG (Fc)
using Texas Red- and AMCA-labeled anti-
bodies. By optimizing the concentration
and incubation time, we were able to form
antibody precipitates in situ in the absence
of methylcellulose or agarose, which are
usually employed to limit product diffusion
(Fig. 2.6). This procedure allowed not only
the identification of high producers, but
also the quick isolation of these clones by
micro-capillary picking. From approxi-
mately 800 clones, 32 with intense IgG
staining were identified. Of these clones,
14 showed no staining for IgM, which was
confirmed by Western blot for a total of 11
clones. PCR tests using primer pairs located
outside the targeting vector and in the
transgene region confirmed a homologous
targeting event in the clones analyzed. In-
terestingly, and in confirmation for our
strategy, expression remained stable over 3
months in clones originating from homolo-
gous recombination, whereas several clones
resulting from random integration lost ex-
pression. We observed the highest expres-
sion levels of 25 pg cell
1
day
1
for clones
containing the hCMV/EF1a hybrid promo-
Fig. 2.5 Structure of the human germ line (a),
and rearranged (b) and targeted (c) IgH l locus of
the G-line. The targeting vector used in homolo-
gous recombination contains the leptinFc repor-
ter gene (hobFc) and the blasticidin deaminase
gene flanked by heterospecific frt sites and the in-
active neomycin phosphotransferase gene. The
cassette is flanked by regions homologous to the
IgH locus at either side.
ter. The resulting cell line was named G-
line.
Based on the presence of the set of hu-
man chromosomes in the heterohybrido-
ma, a glycosylation pattern (Fig. 2.7) differ-
ent from that of mouse myelomas such as
NS0 was expected. Thus, we analyzed the
oligosaccharide structure of hobFc gener-
ated in a roller bottle process. The single
N-linked oligosaccharide chain located in
the Fc region was sialylated at 37%, a rate
close to average sialylation on antibodies
in human blood. Sialic acids were mainly
N-acetylneuraminic acid, typical for hu-
man cells. Only 2% were represented by
N-glycolylneuraminic acid, the immuno-
genic form dominating in mouse myelo-
ma cells. a13Gal structures, not made in
human cells and recognized by pre-exist-
ing antibodies, were only found in 1.3% of
the glycans. This suggests that the G-line
indeed executes glycosylation in a way that
better resembles the pattern of human
compared to mouse cell lines. Moreover,
we observed an unexpectedly low degree
of core fucose (Tab. 2.1). The inhibition of
fucosylation, a secondary effect from artifi-
cial introduction of b1,4-N-acetylglucosa-
minyltransferase III into CHO cells, is
known to enhance Fc effector functions
such as antibody-dependent cellular cyto-
toxicity [36]. For some applications this
may enhance potency of Fc-fusion proteins
or antibodies generated from the G-line.
Based on the presence of frt sites and se-
lection systems in the targeting vectors, the
G-line allows the simple introduction of sec-
ondary target genes via recombinase
mediated cassette exchange. Multiple glyco-
proteins have been introduced into the IgH
locus. For a
1
-antitrypsin (aat) introduced
into the IgH locus expression levels reached
9 pg cell
1
day
1
. In general, expression lev-
els of secondary transgenes were compar-
able to those achieved in CHO cells. Abso-
lute levels as well as the degree of homoge-
neity are transgene dependent. The G-line
seems particularly well suited for Fc-fusion
proteins which are more difficult to produce
in other systems.
Fig. 2.6 Identification of recombinant clones by
direct immunostaining. Colonies were incubated
with Texas Red-labeled antibodies against IgG at
concentrations allowing the formation of immune
precipitates at colonies expressing leptinFc
(left, upper large colony) in contrast to colonies
without expression (bottom, small colony). Phase
contrast (4) of the same area for comparison
(right).
Fig. 2.7 Typical structure of the complete bian-
tennary N-linked glycan of an antibody. GlcNAc,
N-acetyl-D-glucosamine; Fuc, L-fucose; Man, D-
mannose; Neu Ac/Gc, N-acetyl-D-neuraminic acid
or N-glycolyl-D-neuraminic acid. In CHO cells, bi-
secting GlcNAc (mannose position 4) is absent
and neuraminic acids are missing in a high per-
centage of the glycans.
Table 2.1 Differences in glycan structures between production cell lines.
Feature Impact G-line CHO NS0 Human
Sialylation Proteolytic sensitivity
clearance rate
37% Variable Variable 3540%
N-acetylneuraminic acid Human glycostructure 98% High Low 100%
N-glycolylneuraminic
acid
Immunogenic 2% Low >50% No
26 linkage Unknown No No No Variable
a13Gal Pre-existing antibodies
(increased clearance rate)
1.3% Variable High No
Bisecting N-acetyl-D-
glucosamine
Inhibits core fucosylation ND No No 10%
Lack of core
fucosylation
Increased antibody-
dependent cellular
cytotoxicity and
Fcc-binding
60% 5% 1050% 5%
G
0
structures Increased dimerization;
G
2
increases complement-
dependent cytotoxicity
4.3% Variable Variable Low
2.6
Human Designer Cell Lines
The quest for genuine human protein pro-
cessing is best addressed with human cell
lines.
Natural tumor cells growing well in vitro
appear to be the first choice. The strongest
regulatory arguments against these cell
lines have been a poorly documented his-
tory, propagation in mice during establish-
ment and risk of transferring oncogenes
to recipients. These hurdles can be over-
come when cell lines are made directly
from primary cells from a well-documen-
ted source using a set of weak and well-
known oncogenes. Allowing a better risk
assessment, the use of such designer cell
lines is permitted for the manufacture of
live-attenuated viral vaccines or viral vec-
tors. Some of the criteria also apply to cell
lines intended for glycoprotein manufac-
turing. Whereas tumorigenicity and the
transfer of oncogenes are much better con-
trolled through minimal levels of contami-
nating DNA in protein preparations, a
well-documented cell history is just as crit-
ical. Although the designer cell approach
provides a theoretical solution, there are
substantial practical obstacles. Human pri-
mary cells have been quite refractory to
transformation: at least four different bar-
riers protect against tumor formation.
Blocking of cell cycle progression by Rb.
Apoptosis induction via p53 caused by
deregulation of the Rb pathway.
Growth factor dependence for cell cycle
progression and apoptosis prevention.
Telomere shortening and crisis in the
absence of telomerase [37] (see also Part
1, Chapter 1).
While it became feasible to transfect pri-
mary cells at reasonable and in some cases
high efficiencies, stable integration of for-
eign DNA is extremely inefficient com-
pared to established cell lines. Often sev-
eral of the immortalizing/transforming
genes have to be introduced simulta-
neously to prevent the immediate induc-
tion of apoptosis.
To mirror the process of oncogenic
transformation in vitro, typically three to
four genes are applied: SV40 large T anti-
gen to inactivate Rb and p53 pathways, the
catalytic subunit of the telomerase enzyme
hTERT to maintain telomere length, and
v-ras to abolish the strong dependency on
external growth factors. This process has
been reviewed as a major milestone in
cancer research [37]. However, these fac-
tors are not suitable for designer cell lines.
Transforming genes of SV40 and virally ac-
tivated ras are considered strong and dan-
gerous oncogenes for recipients of thera-
peutic preparations. The same applies to
other genes frequently used in research
projects, such as E6 and E7 of human pa-
pilloma virus type 16 or 18, viruses highly
associated with malignant cervical tumors.
One exemplary set of proteins suitable
for immortalization originates from adeno-
viruses (see also Part I, Chapter 6 and Part
IV, Chapter 3). The adenoviral E1A (12S
and 13S) proteins bind Rb and family
members, and act as a transcriptional
modulator, while the E1B protein 55k con-
verts p53 from a transcriptional activator
to a repressor. The second E1B protein
(19 k) is a homolog to Bcl-2 and blocks
apoptosis interacting with bax and bad
genes. These proteins have an extremely
high safety profile: the E1 proteins of
group C adenoviruses have never been as-
sociated with human tumors despite an al-
most complete exposure of the human
population to the virus causing the com-
mon cold (over 90% of the human popula-
tion in Europe has neutralizing antibod-
ies). Moreover, the E1A genes possess tu-
mor-inhibiting, apoptosis-inducing features
when introduced in the absence of E1B.
This desirable feature makes immortali-
zation using this strategy more challenging
compared to introduction of SV40 large T
antigen. Only few cell lines have been gen-
erated. The first cell line, HEK293 [38], was
made by transfecting sheared adenovirus
DNA (see also Part IV, Chapter 12). It origi-
nates from embryonic kidney, but is be-
lieved to be of neuronal origin [39]. The
911 [40] and PER.C6 [41] carry a defined
adenovirus E1fragment linked to heterolo-
gous promoter and poly(A) signals (see also
Both are based on embryonic retino-
blasts, a cell type highly susceptible to im-
mortalization. A third published cell type
is an E1-immortalized amniocyte [42].
We have developed human cell lines
from several tissues using multiple vectors
and approaches: new cell lines have been
generated by transfection with immortaliz-
ing cellular or viral genes followed by con-
tinuous passage, subcloning and adapta-
tion to serum-free conditions. These proce-
dures have been carried out in designated
laboratories, separated from other cell cul-
ture activities.
Although we were able to identify and
expand clones from other tissues as well,
neuronal precursor cells were most sus-
ceptible to transformation, yielding up to
six clones from a single transfection of
210
6
cells. By continuous cultivation for
8 months, cells derived from some of the
cell clones stabilized and became more
homogenous in size and cell morphology.
Doubling time in fetal calf serum-contain-
ing culture was reduced from above 72
down to 40 h when cells were kept at high-
er densities. Adaptation to serum-free
growth in suspension, an ultimate require-
ment for any new cell line applicable to
production of biopharmaceuticals, was suc-
cessful only for a fraction of clones and
served as a criterion for further develop-
ment. In contrast to CHO cells for which
a stepwise reduction of serum concentra-
tion and transfer to serum-free medium
(so called weaning) is recommended [43],
we succeeded when cells were transferred
directly to appropriate serum-free media
(Fig. 2.8) (see also Part II, Chapter 3).
In order to assess their capacity to pro-
duce and process recombinant proteins we
introduced the a
1
-antitrypsin gene driven
by a mouse CMV/EF1-a hybrid promoter
together with selection markers into one
of the cell clones derived from neuronal
precursors. Twelve individual clones were
isolated and specific cellular productivity
was determined. Several of the clones se-
creted more than 75 pg cell
1
day
1
, at least
6-fold more than the best CHO or G-line
based producers isolated for this protein
so far. Titers of 0.5 g L
1
accumulated over
17 days of stationary culture in T-flasks.
Fig. 2.8 Neuronal cell line NCA1 growing in se-
rum-free medium in suspension. Hoffmann mod-
ulation contrast, 20. Dead cells are stained blue
due to Trypan blue uptake.
2.7
Mammalian cells allow the production of
complex biopharmaceutical proteins. Ad-
vances in the generation of stable recombi-
nant clones, media formulation, and pro-
cess and fermentor design have signifi-
cantly increased yields over the past two
decades. Exciting networks across the pro-
teome and transcriptome are elucidated
that show the enormous potential still hid-
den in mammalian cells. The near future
will definitely show further improvement
with cell lines specifically designed and
metabolically engineered for industrial-
scale production of biopharmaceuticals.
References
1 Berg DT, Burck PJ, Berg DH, Grinnell BW.
1993. Kringle glycosylation in a modified hu-
man tissue plasminogen activator improves
functional properties. Blood 81, 13121322.
2 Helenius A. 1994. How N-linked oligosacchar-
ides affect glycoprotein folding in the endo-
plasmic reticulum. Mol Biol Cell 5, 253265.
3 Fiedler K, Simons K. 1995. The role of N-gly-
cans in the secretory pathway. Cell 81, 309312.
4 Sareneva T, Pirhonen J, Cantell K, Julkunen I.
1995. N-glycosylation of human interferon-
gamma: glycans at Asn-25 are critical for pro-
tease resistance. Biochem J 308, 914.
5 Shawler DL, Bartholomew RM, Smith LM,
Dillman RO. 1985. Human immune response
to multiple injections of murine monoclonal
IgG. J Immunol 135, 15301535.
6 McCafferty J, Griffiths AD, Winter G, Chis-
well DJ. 1990. Phage antibodies: filamentous
phage displaying antibody variable domains.
Nature 348, 552554.
7 Bruggemann M, Spicer C, Buluwela L, Rose-
well I, Barton S, Surani MA, Rabbitts TH. 1991.
Human antibody production in transgenic
mice: expression from 100 kb of the human
IgH locus. Eur J Immunol 21, 13231326.
8 Lonberg N, Taylor LD, Harding FA, Trouns-
tine M, Higgins KM, Schramm SR, Kuo CC,
Mashayekh R, Wymore K, McCabe JG. 1994.
Antigen-specific human antibodies from mice
comprising four distinct genetic modifica-
tions. Nature 368, 856859.
9 Goochee CF, Monica T. 1990. Environmental
effects on protein glycosylation. Biotechnology
8, 421427.
10 Hahn TJ, Goochee CF. 1992. Growth-asso-
ciated glycosylation of transferrin secreted by
HepG2 cells. J Biol Chem 267, 2398223987.
11 Wright A, Morrison SL. 1997. Effect of glyco-
sylation on antibody function: implications for
genetic engineering. Trends Biotechnol 15, 26
32.
12 Borrebaeck CA. 1999. Human monoclonal an-
tibodies: the emperors new clothes? Nat Bio-
technol 17, 621.
13 Galili U, Anaraki F, Thall A, Hill-Black C, Radic
M. 1993. One percent of human circulating B
lymphocytes are capable of producing the nat-
ural anti-Gal antibody. Blood 82, 24852493.
14 Hoffmann A, Nimtz M, Getzlaff R, Conradt
HS. 1995. Brain-type N-glycosylation of asia-
lo-transferrin from human cerebrospinal fluid.
FEBS Lett 359, 164168.
15 Mueller-Storm HP, Sogo JM, Schaffner W.
1989. An enhancer stimulates transcription in
trans when attached to the promoter via a pro-
tein bridge. Cell 58, 767777.
16 Kermekchiev M, Pettersson M, Matthias P,
Schaffner W. 1991. Every enhancer works with
every promoter for all the combinations
tested: could new regulatory pathways evolve
by enhancer shuffling? Gene Expr 1, 7181.
17 Chen C, Yang MC, Yang TP. 2001. Evidence
that silencing of the HPRT promoter by DNA
methylation is mediated by critical CpG sites.
J Biol Chem 276, 320328.
18 Wolffe AP, Pruss D. 1996. Targeting chroma-
tin disruption: transcription regulators that
acetylate histones. Cell 84, 817819.
19 Grunstein M. 1997. Histone acetylation in
chromatin structure and transcription. Nature
389, 349352.
20 Wade PA, Pruss D, Wolffe AP. 1997. Histone
acetylation: chromatin in action. Trends Bio-
chem Sci 22, 128132.
21 Cervoni N, Szyf M. 2001. Demethylase activity
is directed by histone acetylation. J Biol Chem
276, 4077840787.
22 Mutskov VJ, Farrell CM, Wade PA, Wolffe AP,
Felsenfeld G. 2002. The barrier function of an
insulator couples high histone acetylation lev-
els with specific protection of promoter DNA
from methylation. Genes Dev 16, 15401554.
23 Antoniou M, Harland L, Mustoe T, Williams
S, Holdstock J, Yague E, Mulcahy T, Griffiths
M, Edwards S, Ioannou PA, Mountain A,
Crombie R. 2003. Transgenes encompassing
dual-promoter CpG islands from the human
TBP and HNRPA2B1 loci are resistant to het-
erochromatin-mediated silencing. Genomics
82, 269279.
24 Running Deer J, Allison DS. 2004. High-level
expression of proteins in mammalian cells
using transcription regulatory sequences from
the Chinese hamster EF-1 gene. Biotechnol
Prog 20, 880889.
25 Mizuguchi H, Xu Z, Ishii-Watabe A, Uchida
E, Hayakawa T. 2000. IRES-dependent second
gene expression is significantly lower than
cap-dependent first gene expression in a bicis-
tronic vector. Mol Ther 1, 376382.
26 Urlaub G, Chasin LA. 1980. Isolation of Chi-
nese hamster cell mutants deficient in dihy-
drofolate reductase activity. Proc Natl Acad Sci
USA 77, 42164220.
27 Urlaub G, Mitchell PJ, Kas E, Chasin LA, Fu-
nanage VL, Myoda TT, Hamlin J. 1986. Effect
of gamma rays at the dihydrofolate reductase
locus: deletions and inversions. Somat Cell
Mol Genet 12, 555566.
28 Bebbington CR, Renner G, Thomson S, King
D, Abrams D, Yarranton GT. 1992. High-level
expression of a recombinant antibody from
myeloma cells using a glutamine synthetase
gene as an amplifiable selectable marker. Bio-
technology 10, 169175.
29 Kaufman RJ, Sharp PA. 1982. Amplification
and expression of sequences cotransfected with
a modular dihydrofolate reductase compli-
mentary DNA gene. J Mol Biol 159, 601621.
30 Harding HP, Calfon M, Urano F, Novoa I,
Ron D. 2002. Transcriptional and translational
control in the Mammalian unfolded protein
response. Annu Rev Cell Dev Biol 18, 575599.
31 Cudna RE, Dickson AJ. 2003. Endoplasmic re-
ticulum signaling as a determinant of recom-
binant protein expression. Biotechnol Bioeng
81, 5665.
32 Schlake T, Bode J. 1994. Use of mutated FLP
recognition target (FRT) sites for the exchange
of expression cassettes at defined chromosom-
al loci. Biochemistry 33, 1274612751.
33 Mielke C, Maass K, Tummler M, Bode J.
1996. Anatomy of highly expressing chromo-
somal sites targeted by retroviral vectors. Bio-
chemistry 35, 22392252.
34 Baer A, Schubeler D, Bode J. 2000. Transcrip-
tional properties of genomic transgene inte-
gration sites marked by electroporation or ret-
roviral infection. Biochemistry 39, 70417049.
35 Jahn S, Niemann B, Winkler T, Kalden JR,
von Baehr R. 1994. Expansion of a B-lympho-
cyte clone producing IgM auto-antibodies en-
coded by a somatically mutated V
H
I gene in
the spleen of an autoimmune patient. Rheu-
matol Int 13, 187196.
36 Sburlati AR, Umana P, Prati EG, Bailey JE.
1998. Synthesis of bisected glycoforms of re-
combinant IFN-beta by overexpression of beta-
1,4-N-acetylglucosaminyltransferase III in Chi-
nese hamster ovary cells. Biotechnol Prog 14,
189192.
37 Weitzman JB, Yaniv M. 1999. Rebuilding the
road to cancer. Nature 400, 401402.
38 Graham FL, Smiley J, Russell WC, Nairn R.
1977. Characteristics of a human cell line
transformed by DNA from human adenovirus
type 5. J Gen Virol 36, 5974.
39 Shaw G, Morse S, Ararat M, Graham FL. 2002.
Preferential transformation of human neuronal
cells by human adenoviruses and the origin of
HEK 293 cells. FASEB J 16, 869871.
40 Fallaux FJ, Kranenburg O, Cramer SJ, Hou-
weling A, Van Ormondt H, Hoeben RC, Van
Der Eb AJ. 1996. Characterization of 911, a
new helper cell line for the titration and pro-
pagation of early region 1-deleted adenoviral
vectors. Hum Gene Ther 7, 215222.
41 Fallaux FJ, Bout A, van der Velde I, van den
Wollenberg DJ, Hehir KM, Keegan J, Auger
C, Cramer SJ, van Ormondt H, van der Eb AJ,
Valerio D, Hoeben RC. 1998. New helper cells
and matched early region 1-deleted adenovirus
vectors prevent generation of replication-com-
petent adenoviruses. Hum Gene Ther 9, 1909
1917.
42 Schiedner G, Hertel S, Kochanek S. 2000. Ef-
ficient transformation of primary human am-
niocytes by E1functions of Ad5, generation of
new cell lines for adenoviral vector produc-
tion. Hum Gene Ther 11, 21052116.
43 Kim EJ, Kim NS, Lee GM. 1999. Development
of a serum-free medium for dihydrofolate re-
ductase-deficient Chinese hamster ovary cells
(DG44) using a statistical design: beneficial ef-
fect of weaning of cells. In Vitro Cell Dev Biol
Anim 35, 178182.
References 777
Abstract
The PER.C6
human cell line was gener-

ated by immortalizing retina cells with the
E1 genes of human adenovirus type 5.
Master and Working cell banks were laid
down and characterized in detail. Initially,
the cell-line was used for the efficient and
safe manufacture of recombinant adenovir-
al vectors for use in gene therapy and as
vaccines. In total, six adenoviral vectors
manufactured on PER.C6 are currently in
clinical trials in the US and in Europe, of
which one is used as a vaccine. In addi-
tion, PER.C6 is used for the manufacture
of classic vaccines such as the influenza
virus and West-Nile virus vaccines. The lat-
est application of PER.C6 is in the field of
protein production. A monoclonal antibody
manufacture process has been developed to
determine the growth and metabolic prop-
erties of PER.C6 and to investigate the yield
and quality of the produced proteins. This
chapter details the history of the PER.C6
cell line, the generation of antibody-produc-
ing PER.C6 cells, and the performance of
these cells in production processes. In gen-
eral, PER.C6 can be easily adapted to se-
rum-free medium and can grow to very
high cell concentrations in fed-batch (>10
7
cells mL
1
) and, in particular, continuous
perfusion (>10
8
cells mL
1
). Specific pro-
ductivity can be maintained at these high
cell concentrations, resulting in high prod-
uct yields. In addition, the high cell densi-
ties have no impact on product quality. Such
cell densities are novel in the industry and
will have a significant impact on the cost
of manufacturing biopharmaceutical pro-
teins, in particular those that are difficult
to manufacture.
Abbreviations
APAC analytical protein A chromatog-
raphy
CBER Center for Biologics Evaluation
and Research
cIEF capillary isoelectric focusing
CMV cytomegalovirus
CSPR cell specific perfusion rate
DCW dry cell weights
DHFR dehydrofolate reductase
DMEM Dulbeccos modified eagle
medium
E1 transcription unit
ELISA enzyme linked immuno sorbent
assay
FBS fetal bovine serum
G418 geneticin
HER human embryonic retina
779
3
PER.C6
Cells for the Manufacture of Biopharmaceutical Proteins

Rodney Gagne, Jose Coco Martin, Nico Oosterhuis, Dirk-Jan Opstelten, and Abraham Bout
ISBN: 3-527-31184-X
HPAEC-PAD high performance anion ex-
change chromatography
with pulsed amperometric
detection
HPLC high performance liquid
chromatography
HP-SEC high-performance size ex-
clusion chromatography
IEF Isoelectric focusing
LC liquid chromatography
LC-MS liquid chromatography
coupled mass spectroscopy
MALDI matrix assisted laser de-
sorption ionization
MALDI-TOF matrix assisted laser de-
sorption ionization-time of
flight
MCB Master Cell Bank
MDM2 morse double mint 2
MS mass spectrometry
MW molecular weight
N/D not detected
NS0 hybrid cells
PGK phosphoglycerate kinase
PNG peptide:N-glycosidase
PrPsc prior specific protein
(scrapy)
SDS-PAGE sodium dodecyl sulfate
polyacrylamide gel electro-
phoresis
SP2 hybrid cells
UVA 280 ultaviolet absorption at a
wavelength of 280 nm
VCD viable cell density
VPR volumetric production rate
3.1
Introduction
The PER.C6
human cell line was gener-

ated by the immortalization of primary re-
tina cells with E1 sequences of human ade-
novirus serotype 5. The cell line was initially
developed for the safe production of phar-
maceutical grade recombinant human ade-
noviral vectors. Such vectors are currently
used for vaccine and gene therapy purposes.
In addition, the cell line is exploited for the
manufacture of classical vaccines including
influenza and West Nile Virus vaccines.
More recently, PER.C6 cells were evalu-
ated for the production of therapeutic pro-
teins, the global market of which has
grown rapidly over the past five years, with
an average annual growth rate of approxi-
mately 21% and sales reaching approxi-
mately $41 bn in 2003 (AS Insights 2003;
Reuters Business Insight 2003). Further-
more, with approximately one-third of all
pipeline candidates currently in clinical de-
velopment, this growth looks set to con-
tinue. One group of therapeutic proteins
the monoclonal antibodies has shown
particularly rapid growth in recent years,
increasing from approximately 1% of ther-
apeutic protein sales in 1995 to 14% in
2001. There are currently 15 approved
antibodies on the market and many more
in the late stages of clinical development.
The production of therapeutic proteins
is commonly performed using mammalian
cell lines, most commonly Chinese ham-
ster ovary (CHO), but including also NS0
and SP2/0 cell lines. Mammalian cell lines
are currently responsible for more than
60% of all licensed products. Their impor-
tance is due to an ability to perform the
correct complex post-translational modifi-
cations required by many therapeutic pro-
teins for their physiological activity. How-
ever, a drawback of mammalian cells is
that yields are typically low compared to
bacterial and yeast systems, while develop-
ment and manufacturing costs are high. A
major goal of process development groups
over recent years has therefore been the
increase of product yields (from both up-
stream and downstream process improve-
ments) and the reduction of development
3 PER.C6

and manufacturing costs and timelines
(see also Part IV, Chapter 1).
Due to the high doses required for many
antibody therapies, high product yields are
particularly desirable. Yields of 12 g L
1
are current industry standard targets. Yields
above 2 g L
1
have also been achieved, but at
present these do not generally result in sig-
nificant cost savings due to current limita-
tions in downstream processing. However,
as progress is made in this area, demands
for higher yields can be expected. At the
same time, there is also a drive to reduce de-
velopment costs and timelines. One goal is
to reduce the time to clinic by reducing
the time taken to generate production cell
lines and to generate the material for pre-
and early clinical phase studies. Current
timelines may vary slightly depending on
the individual situation, cell line, antibody,
etc., but typically range from 14 to 16
months for cDNA to production of clinical
trial material. However, these may be ex-
pected to decrease in the coming years.
The more efficient use of process develop-
ment resources is another major driver, par-
ticularly for projects up to early clinical
phase studies where the risk of failure is
highest and where it may be necessary to
run a number of projects in parallel.
An approach adopted by many has been
the development of platform technologies.
The aims of such a platform include for ex-
ample, the provision of technology required
to generate cell lines with high cell-specific
productivity, to ensure the selection of pro-
duction cell lines that perform well in the
desired production process, to develop
cost-effective production media, and to pro-
vide high-yielding, efficient and cost-effec-
tive production and purification processes
suitable for large-scale manufacture.
Of particular importance is the develop-
ment of generic processes. By developing
processes that are generic, timelines can
be shortened and development costs re-
duced for each new cell line that is gener-
ated. For example, the development of
generic production and purification pro-
cesses removes the need to perform
lengthy and costly process development
for each new cell line, thus reducing costs,
timelines and allowing multiple projects to
proceed simultaneously. They may also act
as a basis for development of the final
manufacturing process, thus minimizing
investment in process development for
Phase III and beyond. Moreover, the inclu-
sion of a generic production process in the
cell line generation program allows the se-
lection of cell lines that perform optimally
in the desired final production process.
It is the aim of Crucell and DSM Biolog-
ics to establish the human PER.C6 cell
line as a platform for the production of
therapeutic proteins, with particular em-
phasis on monoclonal antibodies. The
approach taken has been to develop an in-
tegrated production platform that com-
bines the rapid generation of high-yielding
production cell lines with high-yielding
generic production (batch, fed-batch and
perfusion) and purification processes and
a metabolically characterized host cell line.
Data generated from the metabolic charac-
terization of PER.C6 cell lines was used to
design generic, high-yielding batch, fed-
batch and perfusion production processes,
matched to the metabolic requirements of
the cells. Cell lines are evaluated as early
as possible in the desired production pro-
cess so that lead clones are selected that
match and will perform optimally in the
desired production process.
The investigations described in this
chapter provide an overview of clone gen-
eration, fed-batch and perfusion process
development, as well as detailing the his-
tory of the PER.C6 cell line, and how it
has been characterized. These studies have
been conducted in an alliance between
Crucell Holland N.V. (Leiden, The Nether-
lands) and DSM Biologics (Groningen,
The Netherlands and Montreal, Canada).
3.2
Generation of PER.C6 Cells
3.2.1
Immortalization of Cells by E1 Proteins
of Human Adenovirus
Human adenoviruses are associated only
with mild disease in healthy humans [1].
Adenoviruses have a DNA genome of ap-
proximately 36 kb that encodes proteins of
the virus capsid, and proteins that dedicate
the cell to replicate the viral genome and
synthesize viral proteins. Among the latter
are the so-called E1 gene products of ade-
novirus. It has long been known that the
isolated E1 genes can immortalize primary
human cells [2]. This property of E1 genes
of adenovirus was used to generate the
PER.C6 cell line. The E1 region of adeno-
virus 5 consists of two transcriptional
units, E1A and E1B. The E1A transcription
unit encodes two proteins, which are gen-
erated by alternative splicing. The proteins
3 PER.C6

Fig. 3.1. (A) E1A-induced cell proliferation. E1A
proteins release E2F from Rb, which subsequently
induces CyclinE/cdk2 gene expression and pushes
cells into S-phase. (B) E1A-induced cell cycle ar-
rest and apoptosis counteracted by E1B. E1A pro-
teins activate p53, which leads to cell cycle arrest
and apoptosis. These effects are counteracted by
E1B 55K, which binds directly to p53, and E1B
21K, which inactivates cytochrome-c.
are acidic in nature, are 243 and 289 ami-
no acids long, respectively, and are located
in the nucleus of the cell. A detailed de-
scription of the E1 regions of adenovirus
and the function of E1A proteins of adeno-
virus type 5 is available on (http://
www.geocities.com/jmymryk.geo/).
The E1B region generates one RNA,
which is translated into two proteins, with
molecular weights of 21 and 55 kDa.
For efficient immortalization and trans-
formation of primary cells, both the E1A
and E1B regions are required, although it
has been described that the E1A region by
itself can immortalize rodent cells [3] and
occasionally human cells [4], with very low
efficiency. Expression of E1A alone usually
results in the induction of programmed
cell death (apoptosis), which can be pre-
vented by co-expression of E1B [5].
E1A proteins affect major cellular pro-
cesses such as cell cycle control, differen-
tiation, apoptosis, and transformation. The
immortalization of primary cells occurs by
binding of E1A proteins to the tumor sup-
pressor protein pRB, p107 and p130, as
well as to the co-activator p300 [6]. These
proteins have in common that they can
form complexes with E2F transcription
factor proteins, leading to inactivation of
the E2F factors. The binding of E1A leads
to the release of E2F transcription factors
from the complexes, which results in acti-
vation of cellular genes that have E2F
binding sites in their promoters (Fig.
3.1A). Amongst these genes is cyclinE/
cdk2 that stimulates the cell to enter the
cell cycle. However, the strong prolifera-
tion signal of E1A causes the cell to acti-
vate p53 (Fig. 3.1B). p53 is complexed to
MDM2, which renders it inactive. E1A
mediates the induction of p14ARF protein
expression, which inhibits the activity of
MDM2, thereby causing release of p53 [7].
P53 is an activator or transcription of
genes that cause cell cycle arrest and apop-
tosis of the cells. So E1A stimulates cells
to proliferate but also induces a stress re-
sponse in the cell, leading to growth arrest
or apoptosis. The stress response is coun-
teracted by the E1B proteins (Fig. 3.1B).
E1B 55K, which is located in the nucleus,
forms a complex with p53 and thereby in-
activates it. E1B 21K interferes with the
apoptotic effects of p53-induced Bax pro-
tein in the cell [8]. Bax is a pro-apoptotic
protein, that causes release of cytochrome-
c from mitochondria, which in turn causes
caspase-mediated apoptosis. E1B 21K is a
homologue of the cellular anti-apoptotic
protein Bcl2 [8]. It inhibits Bax-induced re-
lease of cytochrome-c from mitochondria,
thereby preventing apoptosis.
3.2.2
Generation of PER.C6 Cells
The DNA construct pIG.E1A.E1B (Fig. 3.2)
was used for making PER.C6 cells. The E1
genes are driven by the human phosphogly-
cerate kinase (PGK) promoter, which is a
known house-keeping promoter [9] and
the poly(A) sequences are derived from
the Hepatitis B surface antigen gene [10, 11].
The primary cells selected for transfec-
tion with the E1 construct were human
embryonic retina (HER) cells, which can
be immortalized relatively easily by E1 of
human Ad5 [4, 12, 13] and Ad12 [14].
Fig. 3.2 DNA construct used to generate PER.C6
cells. In this construct, the E1A gene is driven by
the human phosphoglycerate kinase (PGK) pro-
moter. Transcription is terminated by the hepatitis
B virus surface antigen poly(A) sequences.
Primary HER cells have a limited life
span, and can be cultured only a few pas-
sages, after which the cells senesce. Trans-
fection of HER cells with E1 constructs re-
sults in transformation and immortaliza-
tion of the cells, reflected by focus forma-
tion in the cultures. This is easily
recognized by both macroscopic and mi-
croscopic examination of the cultures.
Such foci can be isolated and cultured
further. In this way, PER.C6 cells were iso-
lated after transfection of primary HER
cells with pIG.E1A.E1B [15, 16]. The cells
were apparently immortalized also, with-
out passing through a crisis phase.
Transformation and immortalization of
primary cells with E1 sequences of adeno-
virus guarantees: 1) a stable expression of
E1 proteins, as the cells need E1 expres-
sion for growth; and 2) that no external se-
lection marker is needed to distinguish E1
expressing from non-expressing cells.
PER.C6 cells stably express the E1 pro-
teins. In particular, the 21K and 55K E1 pro-
teins that counteract apoptosis and p53-
mediated cell cycle arrest, respectively are
expressed to high levels as compared to,
for example, HEK293 cells [15]. We assume
that this makes the cells relatively insensi-
tive to apoptosis, and may be one of the fac-
tors that make the PER.C6 cells grow to
high cell densities and support production
of a wide variety of proteins, without further
manipulation of the cells.
At passage number 29, a research Mas-
ter Cell Bank was laid down, which was
extensively characterized and tested for
safety (including sterility testings). Re-
search cell banks were made at passage
numbers 33 and 36.
The characterization and safety testing
of the cell banks has been described exten-
sively elsewhere [16]. In brief, the identity,
sterility, viral safety, absence of PrPsc pro-
tein, tumorigenicity and genetic character-
ization, including chromosome analysis,
has been performed.
A description of the history of the cell
line as well as study protocols and re-
ports of all safety studies carried with the
cell line has been filed as a Biologics
Master File at CBER.
3.3
PER.C6 Cells for the Manufacture
The first step in the manufacturing train is
the generation and selection of a high-pro-
ducing cell line. This has been performed
3 PER.C6

Fig. 3.3 Overview of the generation of PER.C6 clones produc-
ing high levels of monoclonal antibodies.
in total more than 15 times, mainly for
IgG1 (j and k light chains, and for both al-
lelic variants of IgG1 heavy chains [17]). In
addition, IgM and IgA, as well as F(ab)
2
fragments have been expressed. A brief de-
scription of the selection of high-producing
cell lines (summarized in Fig. 3.3), as well
as adaptation to serum-free suspension con-
ditions is presented, followed by a summary
of the generic fed-batch and perfusion pro-
cesses that have been developed for PER.C6
cells producing recombinant protein.
3.3.1
Vector Construction and Transfection
The first step in the production of monoclo-
nal antibodies is generation of the expres-
sion construct. For antibody generation,
the antibody construct depicted in Fig. 3.4
has mostly been used. Here, expression of
both the light chain and the heavy chain
genes is driven by a cytomegalovirus
(CMV) promoter that has been modified to
obtain high levels of gene expression in
PER.C6 cells. Adherent PER.C6 cells, cul-
tured in medium containing fetal bovine se-
rum (FBS) are transfected with this con-
struct. Cells that contain a stably integrated
construct are selected using G418 (Geneti-
cin). G418-resistant colonies are transferred
to 96-well plates.
3.3.2
Primary and Secondary Screens
A total of 300400 clones is isolated and
transferred to 96-well plates and cultured
in DMEM supplemented with 10% FBS.
After 510 days, culture supernatants are
sampled and screened for the presence of
IgG either by Protein A HPLC or by ELI-
SA. Production titers from two indepen-
dent screening rounds are used to rank
the transfectants, and the top 2030 are se-
lected and expanded for cryopreservation
and further evaluation. Selection pressure
is maintained until cryopreservation, after
Fig. 3.4 Expression plasmid
encoding IgG heavy and light
chains used for transfection
into PER.C6 cells.
which G418 is removed from the culture
medium.
The highest ranked cell lines selected
from the primary screens are then screened
in 6-well plates using DMEM plus 10%
FBS. Cells are seeded at 0.5 10
6
cells per
well in duplicate and incubated for 4 days
at 378C and 10% CO
2
. Culture supernatants
are then harvested and the IgG concentra-
tion determined by Protein A HPLC or ELI-
SA. Final cell concentrations are measured
and the data used to calculate cell-specific
production rates. The cell lines with the
highest cell specific production rate are se-
lected for adaptation to serum-free condi-
tions. Fig. 3.5 illustrates the results of a typ-
ical secondary screen performed for an in-
ternal clone generation program at Crucell.
Fig. 3.5A shows the final antibody concen-
tration, and Fig. 3.5B the cell-specific pro-
ductivity. There is usually a good correlation
between volumetric and cell-specific pro-
ductivity; that is, cells with a high volu-
metric productivity show a correspondingly
high specific productivity, and vice versa.
Occasionally, a cell line with a high specific
3 PER.C6

Fig. 3.5 Results of a secondary screen of 57
clones from a cell-line generation program con-
ducted at Crucell. The screen was performed in 6-
well plates using DMEM + 10% FBS. Cells were
seeded at 0.5 10
6
mL
1
and the supernatant was
harvested at day 4. (A) volumetric productivity;
(B) specific productivity.
Fig. 3.6 Southern blot indicating the copy number of DNA en-
coding the light chain in different IgG expressing PER.C6
clones. Plasmid copies are measured to a standard compris-
ing a known amount of plasmid DNA in a background of hu-
man chromosomal DNA.
productivity will show a low volumetric pro-
duction due to poor growth, or vice versa. In
this example, 20 clones were selected for
adaptation to serum-free conditions.
Selected PER.C6 cell lines possess low
copy numbers, typically below 10 copies
per cell as measured by Southern analysis
(Fig. 3.6).
3.3.3
Selection Serum-free
Adherent cells are trypsinized and re-sus-
pended directly in shake flasks containing
serum-free medium. Cells are cultured
every 24 days and incubated at 378C, 5%
CO
2
and 100 r.p.m. The adaptation period
for PER.C6 cell lines using this strategy is
typically quite short, up to a maximum of
1520 days. Fig. 3.7 shows two typically
observed adaptation profiles for antibody-
producing cell lines. Once adaptation is
complete and a stable doubling time is ob-
served, cell lines are evaluated in the de-
sired production process, typically batch or
fed-batch. Growth, production and metabo-
lite profiles are characterized. The product
is purified and analyzed by SDS-PAGE (no
major contaminating and/or unexpected
bands), IEF (conform previous produced
material), HPLC-SEC (>90% monomer)
and glycan analysis (correct galactosyla-
tion). A selection of one to three lead cell
lines is then made based on process per-
formance (growth rate, productivity, meta-
bolic profile) and product quality. Fig. 3.8
shows the final antibody concentrations
from batch culture for the 20 selected
clones. In this example, seven clones
showed yields above 0.5 g L
1
and were se-
lected for further evaluation in fed-batch.
Fig. 3.7 Typical adaptation profiles for two different antibody-producing PER.C6 cell lines (A
and B). Cells cultured in the presence of 10% FBS are transferred directly to serum-free media
in Erlenmeyer shake flasks. Cells are passaged and population doubling time is calculated.
Fig. 3.8 Antibody yields from batch cultures for 20
antibody-producing cell lines adapted to serum-
free media.
Fig. 3.9A shows the final antibody concen-
tration, Fig. 3.9B specific productivity, and
Fig. 3.9C the doubling time for each se-
lected cell line.
Once final selection has been made,
cryopreserved cell stocks are prepared in
serum-free medium.
3.3.4
Sub-cloning
Cell lines that are carried forward as po-
tential production lines are sub-cloned.
Cells are plated at an average of 0.3 cells
per well in 96-well plates, and out-growing
colonies screened, expanded, frozen, and
tested as described for the initial clones.
3.3.5
Cell-line Generation Timelines
The aim of cell-line generation is rapidly
to select high-yielding cell lines that per-
form optimally in the desired production
process. The process from transfection to
final selection of the lead clone (including
evaluation in batch or fed-batch) takes 67
months (see Fig. 3.10). The aim is to move
as quickly as possible to serum-free condi-
tions and to make the final selection based
on performance in one of the generic pro-
duction processes, whether batch, fed-
batch, or perfusion. The inclusion of such
generic production processes in the selec-
tion program not only ensures that the
lead cell line that will perform optimally
in a production environment, but also re-
duces the amount of process development
work required for each new cell line. An
3 PER.C6

Fig. 3.9 Overview of the seven antibody-producing cell lines
yielding more than 500 mg L
1
in batch culture (A), specific
productivity ranging from 12 to 22 pg per cell per day (B), and
population doubling time ranging from 24 to 34 hours (C).
additional feature of the PER.C6 cell line
that reduces the timeline for cell-line gen-
eration is the easy and rapid adaptation to
serum-free conditions (see also Fig. 3.7),
which typically requires less than 3 weeks
by a direct adaptation in shake flask. Final-
ly, expression in PER.C6 cells does not in-
volve amplification of gene copy number,
as for example in CHO DHFR
cell lines.
As a result, recombinant PER.C6 cell lines
can be relatively quickly selected and eval-
uated in the required production system,
without the time normally needed for am-
plification.
3.4
Fed-batch Process Development
A generic fed-batch process has been de-
veloped for the production of monoclonal
antibodies in PER.C6 cells. The process
typically results in a 3- to 4-fold increase
in antibody yields compared to the batch
process, with yields of 13.5 g L
1
after 16
18 days. The feed strategy is based on the
metabolic requirements of the PER.C6 cell
line. Metabolic characterization of several
antibody-producing cell lines identified nu-
trients and medium components which
are important for the maintenance of
growth and productivity. These were as-
sembled in a nutrient concentrate consist-
ing of glucose, phosphate and amino acids
and a component concentrate consisting
of vitamins, lipids, trace elements, salts
and growth factors.
The feed strategy involves the addition
of these nutrients based on cell-specific re-
quirements in order to supply the nutri-
ents only as required by the culture and to
limit overflow metabolism or the build-up
of nutrients or metabolites that may result
in reduced process performance (antibody
yields) and product quality [1825].
In addition to a controlled feed strategy,
physico-chemical process parameters have
been optimized for process efficiency. For
example, the growth rate of PER.C6 cells
is optimal at pH 7.3 (Table 3.1). The cell-
specific rates of nutrient utilization are
highest at that pH (Table 3.1) however,
with values for glucose, glutamine and
phosphate for example up to two- or three-
fold higher than at pH 6.9. This increased
rate of nutrient utilization at pH 7.3 does
not result in higher maximum cell yields
or cell-specific productivities, and can thus
be regarded as metabolically less efficient.
It also has a significant influence on the
Fig. 3.10 Timelines for the generation of stable antibody-producing PER.C6 cell lines.
design and efficiency of a fed-batch pro-
cess, as a feed strategy at pH 7.3 would in-
volve the addition of two to three times
the nutrient concentrations as for a pro-
cess at pH 6.9. This would give increased
osmolality and result in reduced process
performance. The problem with operating
a process at pH 6.9 is the sub-optimal
growth rate compared with pH 7.3, which
results in a longer process. This was over-
come by controlling the starting pH of cul-
tures to 7.3, but then operating without a
low limit pH control. In PER.C6 cell cul-
tures this resulted in a pH drift down to
approximately 6.9 during growth, which
led to a culture that showed optimal
growth rates and nutrient utilization pro-
files. Operating the process with such a
pH drift also reduces lactate accumulation.
PER.C6 cells possess a lactate transport
system that is a proton symport system
and thus is dependent on a low extracellu-
3 PER.C6

Fig. 3.11 Glucose utilization and lactate production profiles
for: (A) a batch culture operated with no low limit pH control
(initial culture pH 7.3); and (B) a batch culture operated with
pH control at 7.3.
Table 3.1 Summary of metabolic and growth data for antibody-
producing PER.C6 cell line (antibody A) in batch culture at differ-
ent culture pH values.
pH 6.9 pH 7.3 No low limit pH control
qGlc 0.6 1.6 0.7
qGln 0.17 0.32 0.18
qPhos 0.05 0.12 0.07
Nv (max.) 9.8 10.2 10.5
Avge T(d) day 14 38 28 31
qAb 1416 1416 1416
lar pH. When cultures are operated with
no low limit pH control therefore, there is
a period of lactate release and the pH de-
creases. As this occurs, lactate transport
starts and extracellular lactate concentra-
tions plateau and begin to decrease. How-
ever, if pH is maintained at 7.3, no lactate
transport is observed and lactate accumu-
lates in the culture (Fig. 3.11).
A typical feed strategy involves the addi-
tion of four to six bolus feed additions at
regular intervals during a 16- to 18-day
process. Similar growth and production
profiles are observed for all antibody-pro-
Fig. 3.12 (A) Cell and (B) antibody profiles for batch (open
symbols) and fed-batch cultures for a PER.C6 cell-line express-
ing antibody A. The data represent an average of eight 2-L
bioreactor runs.
ing antibody B. The data represent an average of three 2-L
bioreactor runs.
3 PER.C6

ing antibody C. The data represent an average of 18 shake
flask runs.
Fig. 3.15 Nutrient profile during a fed-batch culture of a
PER.C6 cell-line expressing antibody A. The data show the
concentration of (A) glutamine (closed triangles), glucose
(closed squares); (B) leucine (open squares), cystine (closed
circles); and (C) serine (open triangles).
ducing cell lines that have been evaluated
in the fed-batch process. Figs. 3.12, 3.13
and 3.14 show the growth and production
profiles of three different antibody-produc-
ing cell lines (cell lines A, B, and C, re-
spectively) in batch and fed-batch. Table
3.2 shows a summary of the same three
cell lines (A, B, and C) and a fourth (cell
line D) evaluated in batch and fed-batch,
and includes the average specific produc-
tivity. Fig. 3.15 illustrates selected nutrient
profiles during the fed-batch for one of the
cell lines (cell line A), showing that con-
centrations of the added nutrients remain
stable during the fed-batch and that the
feed is accurately matched to the metabolic
requirements.
3.4.1
Supplemented Batch Process
Metabolic data from the fed-batch develop-
ment was used to develop a supplemented
batch process involving the addition of up
to three of the feeds added in the fed-batch
process, to the culture medium prior to in-
oculation of the cells. Final antibody yields
are not as high as for the fed-batch pro-
cess, typically an increase of 2-fold over
batch yields compared to 3- to 4-fold in-
creases for the fed-batch. However, the
process offers a relatively simple way of
obtaining increased antibody yields.
Fig. 3.16 shows a supplemented batch cul-
ture for clone C where the final antibody
concentration reached 1.3 g L
1
.
Table 3.2 Summary of results obtained with four PER.C6 cell lines expressing different IgG1
Batch [g L
1
] Fed-batch [g L
1
] Process length [days] qAb [pg/cell/day]
Antibody 1 0.4 1.3 18 1215
Antibody 2 0.5 1.4 18 1012
Antibody 3 0.6 2.1 18 1518
Antibody 4 0.6 1.8 16 1619
Fig. 3.16 (A) Cell and (B) antibody profiles for a PER.C6 cell-
line expressing antibody C in batch and supplemented batch
cultures. Supplementation of the batch was made prior to in-
oculation, with 50% of the feed added to a typical fed-batch
culture. The data represent the average of four shake flasks.
3.5
Operation of PER.C6 Cells in Continuous
Perfusion
Perfusion presents a number of advan-
tages over other modes of cultivation, such
as increased volumetric productivity and
rapid removal of easily inactivated prod-
ucts from the culture environment [26].
These homogeneous perfusion systems
can be operated under conditions of total
cell retention, or with the removal of part
of the biomass through culture bleeding
[27]. Total cell retention facilitates kinetics
studies and, as demonstrated in other
studies [28], prevents unnecessary cell divi-
sion, allowing for cells to produce product
at higher rates [29].
A perfusion process has been developed
for the production of monoclonal antibod-
ies in PER.C6 cells. The process typically
results in a more than 30-fold increase in
antibody yield compared to the batch pro-
cess, and a 10-fold increase compared to
the fed-batch process. As for the develop-
ment of the fed-batch process, the perfu-
sion strategy is based on the metabolic
and physico-chemical requirements of the
PER.C6 cell line.
3.5.1
Initial Assessment
The first feature investigated was the im-
pact of cell perfusion rate (CSPR) on cell
culture performance using a spinfilter as
the cell retention device. Fig. 3.17 shows
the viable cell concentration of a PER.C6
cell line expressing antibody A operated
under the same conditions with two differ-
ent CSPR implemented at the 7-L scale.
It was observed that a doubling of the
CSPR resulted in a 65% increase in the
upper cell density. The fact that such an
increase was achieved suggested that sig-
nificant improvements in cell density (and
related productivity) could be obtained by
modification of the CSPR. By day 18, both
of these cultures were prematurely ended
due to clogging of the spinfilter. Upon in-
spection of the spinfilter material, a cake
of cells was observed to have gathered,
which suggested that the high cell densi-
ties achieved with this culture are not
compatible with a standard spinfilter op-
eration as a cellular retention device.
A complete metabolic characterization of
these cultures was performed (data not
3 PER.C6

Fig. 3.17 Cell concentration (10
6
cells mL
1
) versus culture
time (days) for PER.C6 cell line expressing antibody A oper-
ated at two different cell perfusion rates (CSPR).
shown); post analysis of a modified perfu-
sion strategy was assessed.
3.5.2
Stage 1 Development
Fig. 3.18 details the cell profile achieved dur-
ing round 1 development of a continuous
perfusion process using a PER.C6 cell line
expressing antibody A. The viable cell
concentration reached approximately
30 10
6
mL
1
by day 12 and remained con-
stant at this level for a further 20 days. At
this point, a series of actions were taken
which resulted in the cell concentration in-
creasing to approximately 80 10
6
mL
1
by
day 38. Conditions were then kept constant
for a 14-day period during which a constant
cell concentration was maintained. On day
52, the same process change as operated
at day 32 was implemented which led to a
further increase in the viable cell concentra-
tion to 155 10
6
mL
1
by day 53. During the
next 24 hours, the cell retention device be-
came unusable due to the high cell concen-
trations, and this led to termination of the
run. The cell viability throughout the pro-
cess was greater than 95%.
The specific production profile (q
Ab
) of
the cells was stable for the majority of the
process, with an average of 15 pg per cell
per day. Higher values at the end of the
process were due to cell retention device
failure. High volumetric production rates
(VPR) were achieved during this process,
with an average VPR of 0.76 g L
1
per day
calculated over the entire process, 0.9 g L
1
per day at approximately 80 10
6
mL
1
and an upper VPR of 3 g L
1
per day
achieved at 150 10
6
mL
1
.
3.5.3
Stage 2 Development
Round 2 development consisted of identify-
ing the critical features necessary to achieve
the extreme cell densities. Fig. 3.19 shows
the viable cell concentration (10
6
cells
Fig. 3.18 Viable cell density (VCD; 10
6
cells mL
1
), cell-specif-
ic production rate (qAb; pg per cell per day) and volumetric
production rate (VPR; g L
1
per day) versus culture time
(days) for PER.C6 cell line expressing antibody A in a continu-
ous perfusion.
mL
1
) versus culture time (days) achieved
using the round 2 modified process.
The progression gained from each of the
stages can be visualized in Fig. 3.19. It
should be noted that biphasic growth profile
observed in process B is reduced to a mono-
phasic growth profile by process C. In addi-
tion, there is significant reduction in the
process time required to achieve cell con-
centrations of 100 10
6
cell mL
1
. Process
C requires approximately 16 days to achieve
cell concentrations of 100 10
6
cells mL
1
,
while process B required approximately 40
days to reach 80 10
6
cells mL
1
.
In summary, and to the present authors
best knowledge, the cell concentrations
achieved with the PER.C6 cell line (see
Fig. 3.20) are the highest reported value to
date for a mammalian cell line. These re-
sults propel the overall productivity for a
PER.C6 cell line to the highest reported
values for an antibody-producing process.
3 PER.C6

Fig. 3.19 Viable cell concentration (10
6
cells mL
1
) versus
culture time (days) of the optimized modified process com-
pared to the initial experiments.
Fig. 3.20 PER.C6 cell concen-
tration with corresponding dry
cell weights (DCW) (g L
1
)
obtained during a perfusion
mode of operation.
3.6
Characterization of Antibodies Produced
by PER.C6 Cells
One of the cornerstones in the development
of biopharmaceuticals is the availability of
reliable, sensitive and accurate analytical
methods to characterize recombinant prod-
ucts [30]. Typically, these methods are di-
vided into several categories: product quan-
tity, identity, purity, and potency testing [31].
Product characterization methods can be
general for any recombinant protein or spe-
cific to a particular drug substance. In addi-
tion, new analytical technologies and modi-
fications to existing technologies should be
used whenever the methods can add valu-
able data to process development and lead
to a better understanding of the conse-
quences of process changes to ensure the
safety and efficacy of the product in pa-
tients. These methods should be set up dur-
ing the process development phase of a bio-
pharmaceutical product. This section sum-
marizes the analytical results obtained in
parallel with the development of processes
for the production of monoclonal antibodies
in the PER.C6 cell line. General analytical
methods that are used to release materials
from production such as sterility testing or
in-process testing such as process-related
impurities and indicators are beyond the
scope of the present review, or are shown
as part of process development.
Quantity is an important product charac-
teristic, and is usually the first method to
be implemented during process develop-
ment. There are a number of physico-
chemical tests available to measure the
antibody content of cell culture samples,
including ELISA, UVA
280
and affinity
chromatography [32]. Typically, ELISA
methods to assay IgG content are very sen-
sitive and used to quantify low levels of
crude biological products such as those
seen in clone selection or in microscale
processes; a large number of these meth-
ods are commercially available [33]. Several
ELISA methods have already been success-
fully implemented to quantify antibody ti-
ters produced in the PER.C6 cell line.
Our method of choice to quantify re-
combinant antibodies produced in PER.C6
cells is that of analytical protein A chroma-
tography (APAC). This method is auto-
mated, rapid (total run time of 3.8 min),
precise (standard deviation <5%), and sen-
sitive (quantitation limit 25 lg mL
1
), and
generally outperforms ELISA methods, ex-
cept for sensitivity.
There are a number of features of
monoclonal antibodies that are identified
as critical quality attributes. Much of the
discussion between biotechnology compa-
nies and regulatory agencies centers on
the choice of the appropriate methods to
demonstrate product consistency for lot re-
lease. In the present project, since the no-
velty of the technology is the use of hu-
man cells in the production of recombi-
nant antibodies, the focus of the character-
ization studies is general characteristics
such as glycosylation and protein profiles
[34]. In particular, the exhibition of the
predictive nature and consistency of glycan
and protein heterogeneity is important in
order to validate the PER.C6 cell line as an
attractive expression system for the manu-
facture of antibodies.
It was necessary to purify the monoclonal
antibodies that were produced in PER.C6
cells in order to examine the structural fea-
tures of these products. To accomplish this,
a two-step purification method was exe-
cuted whereby cell culture material was
loaded onto a semi-preparative protein A
column and the antibodies were eluted at
low pH [35]. The eluted antibodies were in-
jected directly on a desalting column and
then concentrated in order to prepare the
samples for characterization analyses. The
purification of the antibodies produced in
PER.C6 cells was straightforward, and re-
coveries exceeded 75%. The samples were
very pure after the affinity step, as shown
by SDSPAGE analysis and high-perfor-
mance size exclusion chromatography
(HP-SEC). The level of higher molecular
weight impurities in the samples was very
low (typically <1.5%). Once purified, sam-
ples were subjected to a number of identity
assays including gel and capillary electro-
phoresis and glycosylation analysis.
3.6.1
SDS-PAGE
The conventional slab gel technique SDS-
PAGE is an important technique for the
routine identity and purity analysis of bio-
technology products [36]. A typical SDS-
PAGE analysis, run under reducing condi-
tions with Coomassie blue staining, is
shown in Fig. 3.21A. The heavy and light
chains of the purified antibody samples
are clearly seen in all of the loaded sam-
ples. Molecular weight (MW) markers that
encompass the bands of interest are also
loaded onto the gel. No differences have
been detected between cultures run in
batch, fed-batch or perfusion modes.
3.6.2
Isoelectric Focusing
Isoelectric focusing (IEF) is another routine
slab gel technique for the identification of
biotechnology products [37]. The isoelectric
points (pI) of the produced protein and its
variants can be monitored. The presence
of several charge variants is a common fea-
ture of recombinant antibodies and can be
the result of, for example, deamidation or
differences in processing of C-terminal ly-
sine residues [38, 39]. The consistency of
the charge heterogeneity can be monitored
by using IEF. A typical IEFanalysis is shown
in Fig. 3.21B. The pI values of the protein
isoforms can be identified for all of the sam-
ples. The IEF patterns for antibodies pro-
duced in batch, fed-batch and perfusion
modes with PER.C6 cells are very similar.
Unlike its conventional slab gel counter-
part, capillary isoelectric focusing (cIEF) is
automated, precise and quantitative [40]. A
typical cIEF profile of a purified human
IgG1 that was produced in PER.C6
cells
is shown in Fig. 3.22. In this case, five iso-
forms can be identified and quantified.
The major isoform has a pI value of 8,
which is typical for human IgG1.
During the process development of mono-
clonal antibodies production in PER.C6
3 PER.C6

Fig. 3.21 (A) SDS-PAGE analysis under reducing
conditions of human IgG1 produced in PER.C6
cells. Lane 1: MW markers; lanes 2 and 4: IgG1
produced in batch process; lanes 3, 5, and 6:
IgG1 produced in fed-batch process. (B) IEF gel
of human IgG1 produced in PER.C6 cells. Lane 1:
pI markers; lanes 2 and 4: IgG1 produced in fed-
batch process; lanes 3 and 5: IgG1 produced in
batch process.
cells, several in-process samples were taken
from the production vessel, purified, and
subjected to cIEF analysis. Typical results
of these analyses are shown in Table 3.3. Ex-
cellent quality control is seen in both fed-
batch and perfusion modes, since the cIEF
profiles of samples taken at the mid and
end-points of the runs were very similar. In-
ter-assay variation is 3%for each isoformof
the same sample. Minor differences in cIEF
profiles can be seen for samples taken from
the three modes of manufacturing. Presum-
Fig. 3.22 A typical capillary isoelectric focusing profile of puri-
fied human IgG1 from PER.C6 cells. The peaks at pI values of
10.1, 8.6, and 7.5 are pI markers.
Table 3.3 Typical pI values of the five isoforms for purified human
IgG1 samples produced in batch, fed-batch, and perfusion
modes.
Sample Peak 1 [%] Peak 2 [%] Peak 3 [%] Peak 4 [%] Peak 5 [%]
Batch end 3 6 59 24 8
Fed-batch mid N/D 4 66 21 9
Fed-batch end 4 7 67 17 5
Perfusion early N/D 4 87 6 3
Perfusion mid N/D N/D 87 10 3
Perfusion late N/D N/D 87 10 3
Relative peak areas were obtained by the integration of the major
isoforms in the cIEF electropherograms of the samples.
N/D=not detected.
ably, the charge heterogeneity is due to dif-
ferences in deamidation levels caused by
the different production conditions.
3.6.3
Glycan Analyses
All eukaryotic recombinant expression sys-
tems produce therapeutic proteins that are
glycosylated in their native state for in vivo
activity [41]. The expressed glycoprotein con-
tains glycosylation variants, called glyco-
forms, which are the subject of a number
of recent reviews [4245]. They are also dis-
cussed in detail in this book (Part IV, Chap-
ter 1, 2, 7; and Part VI, Chapter 2). Since gly-
cosylation of biological molecules is
achieved through a complex, post-transla-
tional pathway involving several enzymes,
carbohydrate structures are very sensitive
to even subtle differences in the environ-
ment in which they are formed. The main
factors that influence the oligosaccharide
profile of glycoproteins are cell line, cell cul-
ture medium, bioreactor parameters, har-
vest time, and manufacturing site changes.
In the present case, all IgG1 molecules
contain a conserved N-glycosylation site at
asparagine 297 in the constant region of
the molecule. In human serum, the major
glycoforms of IgG1 contain a biantennary
structure with a core fucose [46]. The three
most common glycans found in IgG1 mol-
ecules, G0, G1, and G2, are shown in
Fig. 3.23.
Purified antibody samples from develop-
ment runs were subjected to PNGase F
treatment to release the glycans, and these
were analyzed by either MALDI-MS or
HPAEC-PAD to determine the glycan
structures.
IgG1 produced by PER.C6
cells in batch
culture show a similar galactosylation pro-
file to human serum IgG [34], with approxi-
mately 30% G0, 50% G1, and 20% G2
(Fig. 3.24). This can be compared to CHO-
produced antibodies, which are typically
3 PER.C6

Fig. 3.23 Structures of the three most common
glycans found in human IgG1 molecules, G0, G1,
and G2.
Fig. 3.24 MALDI-TOF traces of glycans of antibody B produced in batch and fed-batch.
produced predominantly in the G0 form.
For example, Hills et al. [47] reported a ga-
lactosylation profile for an antibody pro-
duced in NS0 and CHO cells of approxi-
mately 60% G0, 35% G1, and 5% G2 for
the NS0 cell line, and 63% G0, 33% G1,
and 4% G2 for the CHO cell line. A small
decrease in galactosylation is typically ob-
served in the fed-batch-produced antibody,
with the percentage of G0 glycoforms in-
creased from 30 to 45%, and the percentage
of G1 and G2 glycoforms decreased from
50% and 20% to 40% and 15%, respectively.
Fig. 3.25 shows the galactosylation profile of
three different antibodies produced in batch
and fed-batch culture from three different
PER.C6 cell lines. This reduction in galacto-
sylation is likely due to the different culture
conditions between the two processes, such
as an increase in process length and in final
Fig. 3.25 Galactosylation profile of three different monoclonal
antibodies produced in batch and fed-batch processes. The
profiles were made by MALDI-TOF analyses.
Table 3.4 Glycan profile by HPAEC-PAD chromatography of hu-
man IgG1 produced in PER.C6 cells in the three manufacturing
modes. The normal ranges of the three major isoforms are pre-
sented
Sample G0 [%] G1 [%] G2 [%]
Batch 1827 5559 1525
Fed-batch 1827 5559 1525
Perfusion 1120 5357 2033
ammonia concentrations (ammonia con-
centrations typically reach 1012 mM by
the end of the fed-batch compared to 3
4 mM for the batch process). The effects
of ammonia on glycosylation have been well
reported [4850], and appear to be due to
the activity of ammonia as a weak base, in-
creasing the pH of the lumen of the Golgi
body. In general, the distribution of the
G0, G1, and G2 isoforms varies only slightly
between cell culture runs. Inter-assay varia-
tion is 0.5% for each isoform of the same
sample. Only slight differences have been
observed between the cell lines analyzed to
date in either batch and fed-batch modes.
Little or none bisecting N-acetylglucosa-
mine or sialic acid is present, and no evi-
dence of structures that may be immuno-
genic in humans, such as high-mannose
or hybrid structures has been detected.
These structures have been reported for gly-
coproteins produced in non-human cell
lines, such as NS0 (a mouse myeloma cell
line) [51]. An interesting observation is that
the recombinant antibody is generally more
galactosylated in perfusion mode (Table
3.4), where the cell counts are the highest
(*10
8
cells mL
1
).
3.6.4
Peptide Mapping
Peptide mapping using liquid chromatogr-
aphy (LC) coupled with either UV detection
or mass spectrometry (MS) is a powerful
technique to study the primary structure
of the antibody, and to further investigate
post-translational or chemical modifica-
tions. The peptide map is a chromato-
graphic finger-print which is obtained after
reduction/alkylation and subsequent pro-
teolytic digestion of the antibody (Fig.
3.26). Thus obtained UV-patterns are used
routinely to screen for structural integrity
after process changes or for quality control
of production lots [52]. LC-MS is applied
to confirm the amino acid sequence and
to detect and identify modifications.
The primary sequences of PER.C6 cell-de-
rived IgG1 and IgG4 antibodies were con-
firmed in LC-MS peptide maps, and no
changes compared to the sequence expected
from DNA-transcription were detected. The
presence of the typical glycan structures
(see Fig. 3.23) could be confirmed, and it
was demonstrated that non-glycosylated
heavy chains were not present. Modifica-
tions of the N- and C-terminus of the heavy
chains were observed in both IgG1 and
IgG4 antibodies. In all cases, the N-terminal
glutamine residue was converted by cycliza-
3 PER.C6

Fig. 3.26 Peptide map of a
PER.C6-derived IgG1 using
coupled with UV detection.
tion into a pyroglutamate residue, which is
a common chemical modification of anti-
bodies. The C-terminal modification is
caused by removal of the C-terminal lysine
residue from the heavy chain. This is due
to the activity of carboxypeptidases, and is
a frequently observed characteristic of pro-
teins produced in mammalian cell culture
[53]. Both the pyroglutamate conversion
and removal of the lysine residue can con-
tribute to charge heterogeneity of the prod-
uct. However, in the described antibodies
the conversions were 100%, resulting in
homogeneous N- and C-termini. The
charge heterogeneity observed in the IEF-
analyses is most likely caused by deamida-
tion of asparagine residues. The exact de-
amidation sites and the percentage of de-
amidated forms could be established with
LC-MS.
3.6.5
Summary Biochemical Analyses
A number of analytical methods have been
used to characterize monoclonal antibodies
produced in PER.C6 cells. It has been
shown that the antibodies can be quanti-
fied and purified easily, and several quality
attributes can be maintained in batch, fed-
batch, and perfusion production modes.
The protein and carbohydrate structures of
the antibodies are completely human in
nature, and excellent control of these attri-
butes during process development has
been observed.
In line with this, the bio-activity of
PER.C6 cell-produced monoclonal antibod-
ies was shown to be equal or better than
antibodies produced in CHO or murine
cell lines (results not shown).
3.7
Conclusion
The PER.C6 cell line was generated from
retina-derived primary human cells, which
were immortalized by insertion of the ade-
novirus E1 gene. In comparison to CHO
and NS0 cells, for example, the PER.C6
cell line has been used for only a few years
in protein production. Nonetheless, it is al-
ready an attractive expression platform
that exhibits many favorable characteristics
for the production of IgG and other pro-
teins. These include the very high cell
numbers and hence high yields that are
obtained, the rapid generation of high-ex-
pressing clones that match a generic and
robust fed-batch process, and the high
quality of the antibodies produced. These
issues are discussed below.
3.7.1
Rapid Generation of High-producing Clones
The transfection of PER.C6 cells with ex-
pression plasmids is very efficient, as is
the subsequent generation of stable cell
lines (an overview of the process is shown
in Fig. 3.3). Importantly, high expression
levels of recombinant antibodies are ob-
served in the absence of gene amplification,
giving a considerable time advantage over
cell lines which require amplification for ef-
ficient protein expression. High-expressing
PER.C6 cell lines contain between two and
10 copies of antibody genes per genome,
compared to hundreds of copies in ampli-
fied cell lines. A high gene copy number
is associated with instability of expression
over time. In contrast, expression from
PER.C6 cell lines producing antibodies is
very stable over several months.
Because the amplification of gene copy
number is not required in PER.C6 cells,
and thanks to the rapid and easy adaptation
3.7 Conclusion 803
to serum-free and animal component-free
media, the timelines for cell line generation
are minimal 67 months from transfec-
tion to final lead clone selection for anti-
body-expressing PER.C6 cell lines. These
cell lines are selected based on their perfor-
mance in generic batch or fed-batch pro-
cesses, thereby ensuring the selection of cell
lines that perform optimally in the desired
production process and reducing the invest-
ment in process development required for
the rapid generation of material for pre-clin-
ical and Phase I clinical studies.
3.7.2
High Numbers of Viable Cells
The generic production processes are based
on the metabolic and physico-chemical re-
quirements of PER.C6 cells. The specific
productivity of antibody-producing cell lines
analyzed to date ranges typically from 10 to
20 pg per cell per day. This does not com-
pare with highly amplified CHO cell lines,
for example, but it does result in high batch
(0.40.8 g L
1
), fed-batch (1.32.2 g L
1
) and
perfusion (up to 1 g L
1
per day) due to the
high cell numbers obtained in these pro-
cesses, which is on a par with levels ob-
tained in industry with CHO and other cell
lines. The reason for the high viable cell
numbers may be related to the high-level ex-
pression of E1B proteins, which are known
to be anti-apoptotic.
3.7.3
Consistent Product Quality of the Antibodies
Antibodies produced on PER.C6 cells show
consistent product quality, as measured by
IEF, SDSPAGE and glycan analysis in
batch, fed-batch, and perfusion processes.
Typical glycan profiles for antibodies pro-
duced in PER.C6 cells are similar to human
serum IgG, with a ratio of G0: G1: G2 of ap-
proximately 30%: 55%: 15% (see Fig. 3.25).
Galactosylation is slightly reduced in the
fed-batch, with an increase in G0 to 40
45% and a decrease in G1 and G2 to 45
50% and 10%, respectively. This decrease
is probably due to the influence of ammo-
nia, which reaches a higher concentration
in the fed-batch.
To date, the profiles obtained are similar
for all antibodies produced in PER.C6
cells, independent of the production levels
obtained. Thus, the outcome of a produc-
tion process becomes highly predictable.
By contrast, the majority of CHO-de-
rived IgGs contain low levels of galactose,
which may diminish the antibodys ability
to initiate effector functions. In addition,
in CHO cells sialic acid is added only via
an a(2-3) linkage, whereas the sialic acid
linkage in human serum may be a(2-6) or
a(2-3). NS0 cells exhibit similar character-
istics, but may also add an extra galactose
to an existing terminal galactose via an
a(1-3) linkage. Humans lack the enzyme
that adds this structure, and such a Gal
a(1-3) structure is highly immunogenic in
humans: indeed, it is estimated that 1% of
circulating Ig is directed against this moi-
ety. Glycans with high-mannose structures
and hybrid structures have also been ob-
served on IgGs produced in CHO and
NS0 cells. However, no such structures
have been identified in antibodies pro-
duced in PER.C6 cells.
3.7.4
Future Prospects
As yet, the PER.C6 cell line has been used
for protein production for only a relatively
short period of time, but the high cell den-
sities obtained, the generic fed-batch pro-
cess and consistent product indicate that
the cell line has vast potential. Unequalled
high cell densities (>15010
6
cells mL
1
) ob-
3 PER.C6

tained in continuous perfusion can be used
to manufacture very large amounts of pro-
teins in a small-scale reactor. In addition,
this procedure may be used to produce un-
stable proteins in an efficient manner.
The process also demonstrates that the
maximum cell densities obtained in fed-
batch can be improved significantly, and
hence further increase the yields. The ulti-
mate aim is a production platform on
which large quantities of high-quality pro-
tein are produced at low cost, thereby al-
lowing more people to benefit from effec-
tive, but expensive, biopharmaceuticals.
References
1 Horwitz, M. S., Adenoviridae and their replica-
tion, in: B. N. Fields and D. M. Knipe (Eds)
Virology, 1990, Raven Press, Ltd, New York,
pp. 16791740.
2 Graham, F. L., J. Smiley, W. C. Russell, and R.
Nairn, Characteristics of a human cell line
transformed by DNA from adenovirus type 5.
J. Gen. Virol., 1977, 36, 5972.
3 Houweling, A., P. v. d. Elsen, and A. v. d. Eb,
Partial transformation of primary rat cells by
the left-most 4.5% of adenovirus 5 DNA.
Virology, 1980, 105, 537550.
4 Gallimore, P. H., R. J. A. Grand, and P. J. Byrd,
Transformation of human embryo retinoblasts
with simian virus 40, adenovirus and ras on-
cogenes. Anticancer Res., 1986, 6, 499567.
5 Teodoro, J. and P. Branton, Regulation of
apoptosis by viral gene products. J. Virol.,
1997, 71, 17391746.
6 Sang, N., J. Caro, and A. Giordano, Adenoviral
E1A: everlasting tool, versatile applications,
continuous contributions and new hypotheses.
Front. Biosci., 2002, 4, 407413.
7 Zhang, Y. and Y. Xiong, Mutations in human
ARF exon 2 disrupt its nucleolar localization
and impair its ability to block nuclear export
of MDM2 and p53. Mol. Cell, 1999, 3(5), 579
591.
8 Cuconati, A. and E. White, Viral homologues
of BCL-2: role of apoptosis in the regulation
of virus infection. Genes Dev., 2002, 16, 2465
2478.
9 Singer-Sam, J., D. H. Keith, K. Tani, R. L. Sim-
mer, L. Shively, S. Lindsay, A. Yoshida, and
A.D. Riggs, Sequence of the promoter region
of the gene for X-linked 3-phosphoglycerate
kinase. Gene, 1984, 32, 409417.
10 Valerio, D., M. G. C. Duyvesteyn, B. M. M. Dek-
ker, G. Weeda, T. M. Bervens, L. v. d. Voorn,
H. v. Ormondt, and A. J. v. d. Eb, Adenosine
deaminase: characterization and expression of
a gene with a remarkable promoter. EMBO J.,
1985, 4, 437443.
11 Simonsen, C. and A. Levinson, Analysis of
processing and polyadenylation signals of the
hepatitis B Virus surface antigen gene by
using simian virus 40-Hepatitis B Virus chi-
meric plasmids. Mol. Cell Biol., 1983, 3,
22502258.
12 Vaessen, R. T. M. J., A. Houweling, A. Israel,
P. Kourilsky, and A. J. v. d. Eb, Adenovirus
E1A-mediated regulation of class I MHC ex-
pression. EMBO J., 1986, 5, 335341.
13 Fallaux, F. J., O. Kranenburg, S. J. Cramer, A.
Houweling, H. v. Ormondt, R. C. Hoeben, and
A. J. v. d. Eb, Characterization of 911: a new
helper cell line for the titration and propaga-
tion of early-region-1-deleted adenoviral vec-
tors. Hum. Gene Ther., 1996, 7, 215222.
14 Byrd, P., K. Brown, and P. Gallimore, Malig-
nant transformation of human embryo retino-
blasts by cloned adenovirus 12 DNA. Nature,
1982, 298, 6971.
15 Fallaux, F. J., A. Bout, I. v. d. Velde, D. J. M. v. d.
Wollenberg, K. Hehir, J. Keegan, C. Auger,
S. J. Cramer, H. v. Ormondt, A. J. v. d. Eb, D.
Valerio, and R. C. Hoeben, New helper cells
and matched early region 1-deleted adenovirus
vectors prevent generation of replication-com-
petent adenoviruses. Hum. Gene Ther., 1998,
9, 19091917.
16 Nichols, W. W., R. Lardenoie, B. J. Ledwith, K.
Brouwer, S. Manam, R. Vogels, D. Kaslow, D.
Zuidgeest, A. J. Bett, J. Chen, M. Van der Kaa-
den, S. M. Galloway, R. B. Hill, S. V. Machotka,
C. A. Anderson, J. B. Lewis, D. Martinez, J.
Lebron, C. Russo, D. Valerio, and A. Bout,
Propagation of adenoviral vectors: use of
PER.C6 cells, in: D. T. Curiel and J. T. Douglas
(Eds) Adenoviral vectors for gene therapy,
2002, Elsevier: San Diego, pp. 129167.
17 Paterson, T., J. Innes, L. McMillan, I. Down-
ing, and M. C. Carter, Variation in IgG1 heavy
chain allotype does not contribute to differ-
ences in biological activity of two human anti-
References 805
Rhesus (D) monoclonal antibodies. Immuno-
technology, 1998, 4, 3747.
18 Ljunggren, J. and L. Haggstrom, Glutamine
limited fed-batch culture reduces the overflow
metabolism of amino acids in myeloma cells.
Cytotechnology, 1992, 8, 4556.
19 Ljunggren, J. and L. Haggstrom, Catabolic
control of hybridoma cells by glucose and glu-
tamine limited fed-batch cultures. Biotechnol.
Bioeng., 1994, 44, 808818.
20 Xie, L. Z. and D. I. C. Wang, Fed-batch cultiva-
tion of animal cells using different medium
design concepts and feeding strategies. Bio-
technol. Bioeng., 1994, 43, 11751189.
21 Xie, L. Z. and D. I. C. Wang, Stoichiometric
analysis of animal cell growth and its applica-
tion in medium design. Biotechnol. Bioeng.,
1994, 43, 11641174.
22 Sanfeliu, A., C. Paredes, J. J. Cairo, and F.
Godia, Identification of key patterns in the
metabolism of hybridoma cells in culture. En-
zyme Microb. Technol., 1997, 21, 421428.
23 Zhou, W. C., C. C. Chen, B. Buckland, and
J. Aunins, Fed-batch culture of recombinant
NS0 myeloma cells with high monoclonal
antibody production. Biotechnol. Bioeng.,
1997, 55, 783792.
24 Zhou, W. C., J. Rehm, A. Europa, and W. S.
Hu, Alteration of mammalian metabolism by
dynamic nutrient feeding. Cytotechnology,
1997, 24, 99108.
25 Europa, A. F., A. Gambhir, P. C. Fu, and W. S.
Hu, Multiple steady states with distinct cellu-
lar metabolism in continuous culture of mam-
malian cells. Biotechnol. Bioeng., 2000, 67,
2534.
26 Mercille, S., M. Johnson, S. Lanthier, A. A. Ka-
men, and B. Massie, Understanding factors
that limit the productivity of suspension based
perfusion cultures operated at high medium
renewal rates. Biotechnol. Bioeng., 2000,
67(4), 435447.
27 Banik, G. G. and C. A. Heath, An investigation
of cell density effects on hybridoma metabo-
lism in a homogeneous reactor. BioProcess
Eng., 1994, 11, 229237.
28 Brosie, D. d. l., M. Nosieux, R. Lemieux, and
B. Massie, Long-term perfusion culture of hy-
bridoma: a grow or die cell cycle system. Bio-
technol. Bioeng., 1991, 38, 781787.
29 Richieri, R. A., L. S. Williams, and P. C. Chou,
Cell cycle dependency of monoclonal antibody
production in asynchronous serum-free hybri-
doma cultures. Cytotechnology, 1991, 5, 972
976.
30 Schenerman, M. A., B. R. Sunday, S. Kozlows-
ki, K. Webber, H. Gazzano-Santoro, and
A. Mire-Sluis, CMC Strategy Forum Report:
Analysis and structure characterization of
monoclonal antibodies. BioProcess Int., 2004,
February, 4252.
31 International Conference on Harmonisation
Guidance on Specifications: Test procedures
and acceptance criteria for biotechnological/
biological products. Fed. Reg., 1996, 64, 2733.
32 Krips, D. M., R. O. Sitrin, and C. N. Oliver, A
very rapid 2 min protein A HPLC assay for
monoclonal antibodies. FASEB J., 1991, 5,
A465.
33 Brown, M. A., L. M. Stenberg, U. Persson, and
J. Stenflo, Identification and purification of
vitamin K-dependent proteins and peptides
with monoclonal antibodies specific for gam-
ma-carboxyglutamyl (Gla) residues. J. Biol.
Chem., 2000, 275, 1979519802.
34 Jones, D., N. Kroos, R. Anema, B. van Mont-
fort, A. Vooys, S. van der Kraats, E. van der
Helm, S. Smits, J. Schouten, K. Bouwer, F.
Lagerwerf, P. van Berkel, D.-J. Opstelten, T.
Logtenberg, and B. Bout, High-level expres-
sion of recombinant IgG in the human cell
line PER.C6. Biotechnol. Prog., 2003, 19, 163
168.
35 Biedermann, K., Sabater, M., Sorensen, J.,
Fiedler, H., and Emborg, C, Quantitative bind-
ing studies of a monoclonal antibody to im-
mobilized protein-A. Bioseparation, 1991, 2,
309314.
36 Shapiro, A., E. Vinuela, and J. Maizel, Molecu-
lar weight estimation of polypeptide chains by
electrophoresis in SDS polyacrylamide gels.
Biochem. Biophys. Res. Commun., 1967, 28,
815.
37 Righetti, P. G., Isoelectric Focussing: theory,
methodology and applications. 1983, Elsevier
Biomedical Press, Amsterdam.
38 Harris, R. J., B. Kabakoff, F. D. Macchi, F. J.
Shen, M. Kwong, J. D. Andya, S. J. Shire, N.
Bjork, K. Totpal, and A.B. Chen, Identification
of multiple sources of charge heterogeneity in
a recombinant antibody. J. Chromatogr. B,
2001, 752, 233245.
39 Perkins, M., R. Theiler, S. Lunte, and M.
Jeschke, Determination of the origin of charge
heterogeneity in a murine monoclonal anti-
body. Pharm. Res., 2000, 17, 11101117.
3 PER.C6

40 Wehr, T., R. Rodriguez-Diaz, and M. Zhu, Re-
cent advances in capillary isoelectric focus-
sing. Chromatographia Suppl., 2001, 53, S47
S58.
41 Jenkins, N., R. B. Parekh, and D. C. James,
Getting the glycosylation right: Implications
for the biotechnology industry. Nature Bio-
technol., 1996, 14, 975981.
42 Cumming, D. A., Glycosylation of recombi-
nant protein therapeutics: Control and func-
tional implications. Glycobiology, 1991, 1,
115130.
43 Wright, A. and S. Morrison, Effect of glycosy-
lation on antibody function: implications for
genetic engineering. Trends Biotechnol., 1997,
15, 2632.
44 Jefferis, R., Glycosylation of Human IgG anti-
bodies: Relevance to therapeutic applications.
BioPharmacology, 2001, September, 1927.
45 Raju, T. S., Glycosylation variations with ex-
pression systems and their impact on biologi-
cal activity of therapeutic immunoglobulins.
BioProcess Int., 2003, April, 4453.
46 Raju, T. S., J. B. Briggs, S. M. Borge, and
A. J. S. Jones, Species-specific variation in gly-
cosylation of IgG: Evidence for the species-
specific sialylation and branch-specific galacto-
sylation and importance for engineering re-
combinant glycoprotein therapeutics. Glyco-
biology, 2000, 10, 477486.
47 Hills, A. E., A. K. Patel, P. N. Boyd, and D. C.
James, Control of therapeutic antibody glyco-
sylation, in: A. Bernard, et al. (Ed) Animal
Cell Technology: Products from cells, cells as
products, 1999, Kluwer Academic Press, Dor-
drecht, The Netherlands, pp. 255257.
48 Borys, M. C., D. H. Linzer, and E. T. Papoutsa-
kis, Ammonia affects the glycosylation pat-
terns of recombinant mouse placental lacto-
gen-I by Chinese hamster ovary cells in pH
dependent manner. Biotechnol. Bioeng., 1994,
43, 505514.
49 Andersen, D. C. and C. Goochee, The effect of
ammonia on the O-linked glycosylation of
granulocyte colony stimulating factor by Chi-
nese hamster ovary cells. Biotechnol. Bioeng.,
1995, 47, 96105.
50 Gawlitzek, M., U. Valley, and R. Wagner, Am-
monium ion and glucosamine dependent in-
creases of oligosaccharide complexity in re-
combinant glycoproteins secreted from culti-
vated BHK-21 cells. Biotechnol. Bioeng., 1998,
57, 518528.
51 Hills, A. E., A. Patel, P. Boyd, and D. C. James,
Metabolic control of recombinant monoclonal
antibody N-glycosylation in GS-NS0 cells. Bio-
technol. Bioeng., 2001, 75, 239251.
52 Bongers, J., J. J. Cummings, M. B. Ebert,
M. M. Federici, L. Gledhill, D. Gulati, G. M.
Hilliard, B. H. Jones, K. R. Lee, J. Modzanows-
ki, M. Naimoli, and S. Burman, Validation of
a peptide mapping method for a therapeutic
antibody: what could we possibly learn about
a method we have run 100 times? J. Pharm.
Biomed. Anal., 2000, 21, 10991128.
53 Harris, R. J., Processing of C-terminal lysine
and arginine residues of proteins isolated
from mammalian cell culture. J. Chromatogr.
A., 1995, 705, 129134.
References 807
Abstract
Mammalian cell culture is becoming in-
creasingly important for the production of
high-volume biopharmaceutical proteins.
This is driving improvements in process ef-
ficiency. This chapter provides examples of
improvements in both the creation of cell
lines and in cell culture optimization, focus-
ing particularly on experience with the glu-
tamine synthetase (GS) expression system.
Abbreviations
ACF&PF animal component-free and
protein-free
ADCC antibody dependent cellular
cytoxicity
cdk cyclin-dependent kinase
FACS fluorescence-activated
cell sorter
GS glutamine synthetase
hCMV human cytomegalovirus
IVC time integral of the viable cell
concentration
MIE major intermediate early
MSX methionine sulfoximine
NS0 non-secreting murine myeloma
OUR oxygen uptake rate
SV40 simian virus 40
TIMP tissue inhibitor of metallo-
proteinases
4.1
Introduction
Mammalian cell culture is an established
technology for the manufacture of biophar-
maceuticals. In recent years, there has been
a significant increase in the number of pro-
teins particularly monoclonal antibodies
which are used in relatively large volumes:
this has been a significant driver for im-
provements in manufacturing processes.
Substantial progress has been made, both
in the optimization of upstream processes
and in the design of expression systems,
for the creation of highly productive cell
lines. A high-yielding biopharmaceutical
protein manufacturing process is the result
of using a number of approaches that affect
the cell line per se, the cell culture process,
product recovery, and purification activities.
Improved cell lines are the result of increas-
ing the efficiency of gene expression and
protein secretion, together with the use of
stringent selection protocols to isolate the
rare high producers. The optimization of
809
4
Use of the Glutamine Synthetase (GS) Expression System for the
Rapid Development of Highly Productive Mammalian Cell Processes
ISBN: 3-527-31184-X
cell culture processes by, for example, im-
proving media and by developing advanced
feeding strategies that support high space-
time yields of viable biomass has increased
substantially the product concentrations
achieved in the bioreactor [13] (see also
Part IV, Chapter 1). Antibody concentra-
tions of 5.1 and 4.3 g L
1
have been reported
in fed-batch culture of GS-NS0 and GS-
CHO respectively [4, 5] (see also Part IV,
Chapter 16). This chapter reviews an ex-
pression technology based on the use of
the glutamine synthetase gene and the inte-
gration of this technology into the develop-
ment of high-yielding, large-scale manufac-
turing processes for biopharmaceuticals.
4.2
Cell Line Construction and Selection
4.2.1
Choice of Cell Line
The enzyme glutamine synthetase (GS) cat-
alyzes the formation of glutamine from glu-
tamic acid and ammonia, driven by the hy-
drolysis of ATP. Glutamine has multiple
roles in cell metabolism, particularly as an
energy source, protein constituent and as
a nitrogen donor in purine and pyrimidine
synthesis. Cell lines that do not produce
GS have an absolute requirement for gluta-
mine and do not grow in glutamine-free
culture media. This provides the basis for
using the enzyme as a selectable marker
in gene expression vectors. The cloning of
the GS gene from Chinese hamster ovary
(CHO) cells was described by Sanders and
Wilson [6]. The utility of GS as a selectable
marker is increased by the availability of an
efficient inhibitor of GS, methionine sulfox-
imine (MSX), which can be used to improve
the stringency of selection, to select for gene
amplification and to inhibit enzyme activity
in those cell lines which produce endoge-
nous GS. Most myeloma and hybridoma
cells have an absolute requirement for glu-
tamine. In contrast, many other cell types
such as BHK-21, L-cells and the widely used
CHO do not require glutamine, provided
that glutamic acid is present in the culture
medium. In these cases GS can still be used
as a selectable marker, but it is necessary to
use a specific inhibitor of GS (e.g., MSX) to
inhibit the endogenous enzyme. The toxici-
ty of MSX at very low concentrations (3 lM)
for wild-type CHO cells has been demon-
strated [6]. Since the enzyme is used as a
dominant selectable marker, it is not neces-
sary to create relevant mutant host cells.
GS expression vectors designed for use in
mouse NS0 cells [1] and CHO cells [7, 8]
have been described. The NS0 cell line
was chosen because it has an absolute re-
quirement for glutamine in contrast to
other lymphoid cells, which gave a relatively
high frequency of glutamine-independent
variants. Other factors contributing to the
choice of this cell line included the ease
with which it could be grown in serum-free
suspension culture and, given its B-cell line-
age, an expectation that it has the machin-
ery for efficient antibody secretion.
In principle, selection based on GS can
be used with a wide variety of cell lines,
and in practice the most commonly used
are NS0 and CHO. Whilst GS-NS0 has
been used most often for antibody produc-
tion, GS-CHO has been used to express a
large range of proteins in CHO cells. In
addition to antibodies (e.g., [9]), enzymes
(e.g., [10]), interleukins [11] and mem-
brane-bound proteins (e.g., [12]) are exam-
ples of the range of proteins produced
using the GS expression system. For many
proteins and particularly many non-anti-
body products the glycosylation proper-
ties of CHO are preferable to those found
in NS0. It has been found, for example,
4 Use of the Glutamine Synthetase (GS) Expression System 810
that NS0 has a limited ability to add sialic
acid to glycoproteins, and this can have a
significant effect on the clearance rate of
the protein in vivo. Flesher et al. [13] com-
pared the in vivo clearance profile of a sol-
uble form of the membrane receptor
CTLA4 produced in CHO and NS0 cells.
A correlation was observed between the
quantity of N-acetylneuraminic acid in the
product and in vivo clearance rates. Prod-
uct made in NS0 cells had no detectable
N-acetylneuraminic acid and exhibited an
accelerated clearance rate. Baker et al. [14]
compared the glycosylation of recombinant
tissue inhibitor of metalloproteinases
(TIMP) made in GS-CHO and GS-NS0;
these authors found significant differ-
ences, and in particular a high proportion
(30%) of the NS0 glycans terminated in a-
1,3-linked galactose. In addition a high
proportion of the sialic acid in NS0 materi-
al was in the form of N-glycolylneuraminic
acid as opposed to N-acetylneuraminic
acid. Lifely et al. [15] compared the glyco-
sylation of the monoclonal antibody Cam-
path
-1H made in CHO, GS-NS0 and rat

Y0 myeloma cells. The glycan profiles of
CHO and NS0 were similar, although NS0
antibody was underglycosylated to a signif-
icant extent. The Y0-derived antibody had
fucosylated and non-fucosylated glycans
containing a bisecting GlcNAc, and was
observed to have enhanced ADCC activity.
4.2.2
Expression Vector Design
The ability to express the product at a high
level is the critical issue for any manufactur-
ing process using recombinant cell lines.
Consequently, expression vectors have been
developed that, through a combination of
suitable promoters and favorable RNA pro-
cessing signals, can achieve high levels of
transcription from the genes of interest.
Strong promoters used to drive expression
of the genes of interest are generally of viral
origin or from highly expressed genes in a
mammalian cell. A number of different vir-
al promoters have been evaluated for use
with the GS expression vectors [8]. These
authors screened different promoters using
a test system based on the transient trans-
fection of CHO-K1 cells with the gene for
TIMP. The efficiency of the transcription
units was in the order hCMV>SV40 early>-
hybrid Moloney murine leukemia virus-
SV40 promoter >SV40 late. Expression of
the TIMP gene driven by the hCMV promo-
ter produced five to ten times more TIMP
than the SV40 early promoter. The hCMV
promoter fragment chosen consists of the
complete enhancer-promoter and 5'-UTR
(untranslated region) of the major inter-
mediate early (MIE) gene. This promoter,
unlike the SV40 early promoter, is highly ef-
ficient in most cell types including lympho-
cytes (see Ref. [1]). Various vector construc-
tions were evaluated in NS0 cells, and the
system chosen is shown in Fig. 4.1. Product
gene expression is driven by the hCMV-MIE
promoter. A polylinker site is incorporated
downstream of the hCMVsequence to allow
incorporation of the product gene, and an
SV40 polyadenylation site is situated down-
stream of the polylinker. For the expression
of monoclonal antibodies, both the heavy
and light chain genes are contained in the
same vector, with the heavy chain gene
downstream of the light chain gene. The
GS gene, which is upstream of the product
genes, is driven by the SV40 early promoter.
4.2.3
Increasing Transcription
Several options exist to increase transcrip-
tion. In early expression systems this was
generally by gene amplification. Gene am-
plification is usually achieved by construct-
ing the expression vector so that the genes of
interest are linked to an amplifiable gene.
Although high-yielding GS-NS0 and GS-
CHO cell lines can be isolated without am-
plification particularly if attention is paid
to the stringency of selection several
groups have used amplification, especially
for CHO cells.
For the GS expression system, amplifica-
tion is achieved by applying increasing
concentrations of the GS inhibitor, MSX.
Laubach et al. [10], for instance, comment
that amplification (by growth in the pres-
ence of 400 lM MSX) of two GS-CHO cell
lines making inducible nitric oxide
synthase resulted in a 3- to 4-fold increase
in productivity. The amplification of a GS-
CHO cell line making a monoclonal anti-
body increased the antibody concentration
from 110 to 250 mg L
1
[16].
Bebbington et al. [1] found that, for a
GS-NS0 cell line, amplification using MSX
was accompanied by an increase in copy
number of the vector from one to four.
Similar results have been described by
other groups [17, 18]. This is much lower
than the levels observed with the dihydro-
folate reductase (DHFR)-CHO system [17].
In CHO cell lines the copy numbers of
DHFR-linked genes can be as high as
1000, whereas in NS0 clones the copy
number of GS-linked genes rarely exceeds
20. For a GS-CHO cell line, amplification
with 200 lM MSX increased the copy
number from 5 to 200 [16].
Peakman et al. [17] studied the impact
of amplification upon the productivity of
both GS-CHO and GS-NS0 cell lines. For
GS-NS0 cell lines expressing a recombi-
nant antibody, the copy number of either
the GS or immunoglobulin heavy chain
genes may be the same in different clones,
but it does not follow that mRNA levels
will be the same. Although amplification
Fig. 4.1 Glutamine synthetase
(GS) expression vector.
of DHFR-CHO and GS-NS0 cell lines may
result in markedly different copy numbers,
the two cell lines may still express approxi-
mately the same amount of antibody due
to mRNA levels being virtually identical.
Therefore, for cell lines generated using
the GS vector system, amplification of
copy number is not as critical for generat-
ing high-producing clones as with the
DHFR expression system. With the GS ex-
pression system, the position of integra-
tion of the transfected DNA is the impor-
tant factor in determining whether the cell
line will ultimately be a high producer.
The high copy numbers of the expres-
sion vector seen upon amplification espe-
cially with the DHFR expression system
may increase the cell-specific productivity,
but it can also have a detrimental effect
upon other properties of the cell. Amplifica-
tion of the desired gene will frequently re-
sult in poor growth performance of the re-
sulting cell population and may alter cellu-
lar metabolism. These effects have been
seen in both GS-NS0 and DHFR-CHO cell
lines [17, 19]. Gu et al. [20] suggested that
the poor growth seen upon amplification
of the DHFR gene is not due to increased
expression of a recombinant protein; rather,
it is a consequence of the higher metabolic
burden imposed upon the cell by the in-
crease in specific DHFR activity. The prob-
lem of poor growth (low values for the max-
imum viable cell concentration and time in-
tegral of the viable cell concentration) can
be counteracted to some extent by using
the combination of good growth character-
istics and a high specific production rate
as selection criteria. Amplification and the
resulting variation in copy number can
also alter the inherent stability of expres-
sion, and often requires the continued
presence of the amplification agent [16,
21]. If the selective agent is required in
the production bioreactor, it will be neces-
sary to demonstrate that the purification
process removes this compound from the
bulk drug product.
Expression from the hCMV-MIE promo-
ter in CHO-K1 cells can also be enhanced
by expression of the adenovirus 5 E1A
transactivator or a mutant of E1A that has
lost the oncogenic transformation function
[22]. After optimization of E1A expression,
since over-expression led to inhibition of
cell growth, it was possible to raise expres-
sion levels of TIMP in non-amplified GS
cell lines to levels which were previously
only achievable after vector amplification.
CHO cell lines constitutively expressing
the transactivator were created. The trans-
activator increased the specific production
rate of recombinant protein product in
both a transient expression system, and in
stable cell lines either by incorporating it
into a producer cell line or by transfecting
the product gene into a host cell line that
already contained E1A. Vector amplifica-
tion did not produce higher-producing cell
lines in E1A-containing cells. Likewise, the
introduction of E1A into an amplified GS
cell line making TIMP failed to enhance
productivity.
4.2.4
Selection of High-producers
without Amplification
Amplification of the number of copies of
expression vector can lead to an increase
in productivity of the cell line, but this is
not the only method for generating high-
yielding cell lines. The GS system [1] and
some variants of the DHFR system [23] do
not rely upon amplification to achieve
high productivities. Instead, these systems
rely upon insertion of the antibody con-
struct into a transcriptionally active region
to achieve high productivities. Specific pro-
duction rates for non-amplified antibody-
producing GS cell lines of about 65 pg per
cellday, and an antibody concentration of
more than 1.8 g L
1
, can be obtained in
fed-batch bioreactors [24].
Currently, the generation of high-yield-
ing cell lines is typically achieved by
screening large numbers of transfectants,
or a combination of amplification plus
screening. There are problems with these
approaches due to the time needed to ob-
tain a cell line with an acceptable produc-
tivity, the potential instability of amplified
cell lines, and the deleterious effect of am-
plification upon other characteristics of
cell line. The reason it is difficult to obtain
high-yielding cell lines is that large regions
of the genome are organized into hetero-
chromatin, which is believed to be tran-
scriptionally inactive. The level of mRNA
expression from a vector that integrates
into the heterochromatin will be low. Since
there are only a few loci within the ge-
nome capable of expressing the selectable
marker gene and the linked gene(s) of in-
terest gene at high levels, it follows that
the probability of integration into such a
transcriptionally active locus is low. Thus,
large numbers of transfectants normally
have to be screened to isolate those few
clones where the vector has integrated into
transcriptionally active loci, with concomi-
tant high product expression levels.
Several approaches have been developed
to reduce the time needed to obtain high-
yielding cell lines. These approaches ex-
ploit the importance of the chromosomal
locus in determining the level of gene ex-
pression to increase the proportion of
transfectants with the expression vector in-
tegrated into a transcriptionally active lo-
cus by up to 10-fold. Since these
approaches generate a clone with only one
or at most a few copies of the expression
vector [1, 25], the problems associated with
amplification are eliminated.
One approach is to use site-specific re-
combination of the gene(s) of interest into
a known transcriptionally active locus. Ex-
pression vectors can be constructed that
contain a specific targeting sequence that
will direct the vector to integrate by ho-
mologous recombination into a particular
active site. Such a sequence has been iden-
tified in the immunoglobulin locus of NS0
cells [26]. Vectors containing this sequence
are targeted to the immunoglobulin locus
in more than 50% of high-producing GS-
NS0 clones.
A corollary of this approach is to take
the sequences flanking the transcription-
ally active locus and incorporate them into
the expression vector. Thus, the vector
should create a favorable environment for
expression independent of its integration
site in the genome. Vectors based on ubiq-
uitous chromatin opening elements [27] or
the flanking sequences of the Chinese
hamster elongation factor-1a gene [28]
have been described.
An alternative approach is to transfect the
cells with a conventional expression vector
(i.e., randomly to integrate the expression
vector into the genome), but then bias the
selection method so that only transfectants
where the vector integrated into a transcrip-
tionally active site are progressed. This can
be done by using a selection system that
only allows transfectants producing suffi-
cient levels of the selectable marker gene
product to proliferate. Expression systems
using a selectable marker gene with either
the weak SV40 promoter [1] or an impaired
Kozak sequence upstream of the marker
gene [23] are included in this class of selec-
tion system. Linkage of the antibody con-
struct to the selectable marker gene results
in the over-production of antibody, as both
genes are integrated into a transcriptionally
active locus. The choice of selection condi-
tions is extremely important for the success
of this approach. The data in Table 4.1 show
that increasing the selection stringency for
GS cell lines (through increased MSX con-
centration) reduces the number of stable
transfectants but, by optimizing the trans-
fection conditions, it is possible to maintain
the number of transfectants generated. A
higher selection stringency resulted in shift
in the position of the median antibody con-
centration (from about 45 mg L
1
to about
90 mg L
1
) and interquartile range (from
25 to 70 mg L
1
, to about 60 to 130 mg L
1
). This approach shows that it is possible
to increase the average productivity without
restricting the number of transfectants.
The function of the expression vectors
described in the previous sections is to
generate cell lines with high specific pro-
duction rates of the protein of interest.
However, a transfectant with a high specif-
ic production rate does not necessarily re-
sult in a cell line that performs well in the
production process. Hence, a sufficient
number of cell lines need to be generated
to allow for the attrition in numbers when
screening for other desired characteristics.
By definition, transfectants with the
highest productivities are rare: this is
shown in Fig. 4.2. The figure shows the
probability of finding a transfectant that
produced antibody at a defined concentra-
tion. The probability of finding a primary
transfectant producing 150 mg L
1
is about
0.0005. The majority of transfectants
(90%) produced less than 90 mg L
1
anti-
body, and only 1.5% produced more than
150 mg L
1
. The issue is therefore, how
can the hit rate for finding highly produc-
tive cell lines or the number of hits be in-
creased?
Finding these rare events requires the
combination of a number of approaches.
The simplest approach is to screen more
Table 4.1 Influence of transfection and selection conditions upon
the yield of stable antibody-producing GS-CHO transfectants
Electroporation condition Selection condition
MSX [lM]
Number of stable transfectants
per 510
6
cells electroporated
250 V, 400 lF 25 68
50 32
275 V, 650 lF 25 124
50 57
300 V, 900 lF 25 197
50 70
Fig. 4.2 Productivity distribution of antibody con-
centrations for primary GS-CHO transfectants.
Ninety-two primary GS-CHO transfectant colonies
were transferred from 96-well to 24-well plates
and grown for 14 days: the mean concentration at
harvest was 48 mg L
1
. A log-normal probability
density function was fitted to the antibody con-
centration data.
transfectants, but how many? Simulations
of screening experiments using the
150 mg L
1
cut-off suggest that, over the
long term, at least 500 transfectants would
have to be screened to avoid any individual
screening experiment having no transfec-
tants over 150 mg L
1
(A.J.R., unpublished
results). To obtain tens of transfectants
above this cut-off, then several thousand
transfectants should be screened.
Conventional methods for the screening
of cell lines after cloning are labor-inten-
sive, and this limits the number of cell
lines that can be screened. Increasingly,
robotics are being used to automate the
liquid handling and cell transfer stages,
but this does not address the need to
screen large numbers of transfectants to
identify sufficient high producers to screen
against the additional growth criteria that
contribute to high productivity in a manu-
facturing process. Flow cytometry can be
used to identify cells making high levels of
the target product, while fluorescence-acti-
vated cell sorting (FACS) can be used to
collect cells aseptically with the desired
characteristics from large heterogeneous
populations. Cells can be sorted into large
populations (bulk sorting), from which
cell lines can be isolated by conventional
cloning methods, or by single cell sorting.
A number of FACS-based approaches
have been reported for the isolation of cell
lines secreting high levels of antibody.
These include encapsulating the secreting
cells in a biotinylated agarose droplet, which
captures the secreted antibody [29], trapping
the secreted protein in the membrane [30],
or using a matrix constructed on the cell
surface to trap the secreted antibody [31].
Holmes and Al-Rubeai [32] used a surface
capture methodology to isolate clones with
higher specific production rates from the
GS-NS0 cell line 6A1(100)3. On average,
the sorted clones had a specific production
rate which was 25% higher than the origi-
nal GS-NS0 population. Racher [33] de-
scribed a modification of this surface cap-
ture method that uses Protein A immobi-
lized on the cell surface as a capture method
for monoclonal antibodies.
The identification of high producers is the
first step in isolating high-producing cell
lines. The next step is to screen the pool
of high producers against criteria that fit
the cell line to the manufacturing process.
Fig. 4.3 shows a schematic for a cell line se-
lection program for GS cell lines. High-pro-
ducing transfectants are identified and ex-
panded through static into suspension cul-
ture. Once acceptable and reproducible
growth is achieved, the cell lines are adapted
to animal component and protein-free
(ACF&PF) medium. Initially, transfectants
are screened against productivity criteria:
once in suspension culture, the cell lines
are screened against additional criteria.
Typically, several criteria are used to select
the production cell line. The criteria in-
clude: a high specific production rate;
growth characteristics such as the magni-
tude of the time integral of the viable cell
concentration (IVC) and maximum cell con-
centration; product concentration at har-
vest; cell line stability; and product quality.
The importance of screening prior to cell
line selection in a system that has relevance
to the manufacturing process was demon-
strated by Brand et al. [34]. These workers
found there to be poor correlation between
productivity of recombinant myelomas in
static culture (cloning plates and flasks)
and agitated suspension culture.
A key feature of any selection scheme is
that it is important either to undertake the
screening in an acceptable model of the
manufacturing process, or to know the pre-
dictive power of the screen. In Fig. 4.4, the
cell lines are evaluated in suspension cul-
ture using the same media, feeds and sub-
culture regimes as used in the manufactur-
ing process. Using this approach, it is possi-
ble routinely to obtain cell lines producing
more than 1 g L
1
. Fig. 4.5 shows the pre-
dictive power of the screening process out-
lined in Fig. 4.3. The data in Fig. 4.5 are ob-
tained from the cell lines eventually chosen
for the manufacture of seven randomly cho-
Fig. 4.3 Schematic of a cell line selection pro-
gramme for glutamine synthetase (GS) cell lines.
High-producing transfectants are identified and
expanded through static into suspension culture,
and then adapted to chemically defined medium.
For the suspension phase, the cell lines are grown
in Erlenmeyer flasks. Initially, transfectants are
screened against productivity criteria: once in sus-
pension culture, the cell lines are screened against
productivity, growth and product quality criteria.
Fig. 4.4 Data are from six cell line construction programs. The data are the antibody concentra-
tions achieved by panels of 10 GS-NS0 cell lines during the fed-batch assessment phase of the
programme outlined in Fig. 4.3.
sen antibodies. In general, the values ob-
tained in bioreactors are 0.8- to 1.2-fold
the values obtained in the Erlenmeyer flask
assessment, although there are exceptions.
The Erlenmeyer flask fed-batch model used
in the assessment to select GS cell lines for
the GMP manufacture of antibodies ap-
pears to have high predictive power.
In summary, although the transfectants
with the highest specific production rates
are (by definition) rare, the current
approaches are successful in making pro-
ductive cell lines. By using a combination
of expression vectors with strong promo-
ters and a stringent selection system it is
possible to construct and then select high-
producing, non-amplified, cell lines. Selec-
tion against productivity criteria should
not be the only consideration. Multiple se-
lection criteria should be used in an ac-
ceptable scale-down model of the manufac-
turing process to select cell lines that fit
the manufacturing cell culture process.
4.3
Cell Line Stability
An important issue in creating cell lines is
to maintain stability over the number of
generations required for manufacturing
in practice, several tens of generations
from a cell bank for a fed-batch process
at 10- to 20 000-L scale. Stability will be in-
fluenced by factors such as copy number
and site of integration of the foreign
gene(s). One would expect non-amplified
lines with low copy number to be more
stable than amplified lines with high copy
number. In general it is possible to isolate
stable, non-amplified lines which do not
require MSX to maintain productivity. In
the case of amplified cell lines MSX may
be required. Hassell et al. [16] monitored
the stability of GS-NS0 and GS-CHO cells
making antibodies. Two amplified GS-NS0
cell lines making different antibodies were
stable in shake-flask culture in both the
presence and absence of MSX. In contrast,
an amplified CHO cell line was stable in
the presence of MSX but not in its ab-
sence. Other groups have also reported on
the need to maintain MSX selection in
amplified GS-CHO cells. Cosgrove et al.
[35] used MSX to maintain stability of an
CHO cell line producing insulin receptor
ectodomain; MSX was added to culture
media until the final production step. Sim-
ilarly, Guerini et al. [12] found that expres-
sion of an ATPase in an amplified CHO
cell line was stable for 6 months in the
presence of MSX, but decreased substan-
tially over 30 to 40 generations in its ab-
sence.
The stability of cell lines will be depen-
dent not only on the characteristics of the
cell line and gene inserts, but also on cul-
ture conditions. Bird et al. [36] presented
evidence that stability of a GS-NS0 cell
line was reduced under conditions where
Fig. 4.5 Relationship between productivity charac-
teristics of the lead cell lines making seven differ-
ent antibodies when evaluated during the fed-
batch assessment phase of Fig. 4.3 and in the
production bioreactor. The data are either the anti-
body concentration or the specific production rate
of the bioreactor culture (Qp) normalized to the
values obtained in the fed-batch assessment
phase that uses Erlenmeyer flasks.
glutamine accumulated for example,
when cultures were grown in hollow-fiber
devices. The authors found that the pres-
ence of 60 lM glutamine was sufficient to
cause instability, presumably overcoming
the selection pressure of the glutamine-
free medium.
The mechanisms underlying instability
in recombinant cell lines are poorly under-
stood. Barnes et al. [37, 38] determined
that there may be molecular features of
transfectants that predicate instability.
These authors studied a series of GS-NS0
cell lines making an anti-CD38 monoclo-
nal antibody. Although copy number re-
mained constant in these cell lines, there
was a loss in expression of mRNA during
prolonged culture. This did not result in
loss of productivity in all of the cell lines.
It seems that productivity was not influ-
enced provided that levels of antibody
mRNA remained above a critical threshold
value.
4.4
Cell Engineering to Increase Productivity
4.4.1
Delaying Apoptosis
An important tenet for achieving highly
productive processes is the achievement
within the bioreactor of a high viable cell
concentration and its subsequent mainte-
nance for an extended period. The latter
requires the death rate to be minimized.
This section describes cell engineering
approaches evaluated with GS cell lines to
minimize the death rate.
Al-Rubeai et al. [39] showed that the ma-
jor cause of cell death in animal cell cul-
ture is through the induction of apoptosis
(programmed cell death) pathways by
chronic, rather than acute, insults. As
apoptosis can be induced by a variety of
insults and is mediated by several path-
ways, diverse environmental and genetic
strategies to limit cell death have been pro-
posed (for a review, see Ref. [40]). There
are a number of perceived advantages
from increasing cell robustness by engi-
neering apoptosis resistance. These in-
clude increased space-time yields for viable
biomass with a concomitant increase in
product concentrations, enhanced survival
in nutrient limited conditions, and more
efficient clarification of the feedstock prior
to downstream processing.
Since apoptosis can be induced by nutri-
ent deprivation, one approach to limit its
extent is to prevent nutrient limitation.
The use of fed-batch operations can delay
the onset of apoptosis in GS-NS0 cell lines
and substantially reduce its extent [3, 41],
thereby increasing the IVC. However, the
use of a fed-batch process does not com-
pletely eliminate apoptosis. Another
approach to increasing process productiv-
ity is to engineer resistance to apoptosis
into the cell lines.
As activation of the apoptotic pathways
results in destruction of the cell, the path-
ways must be tightly regulated. The best
understood regulatory mechanism involves
the Bcl-2 family of proteins. Some mem-
bers of the Bcl-2 family stimulate apopto-
sis (e.g., Bax, Bak and Bid), whilst others
have an anti-apoptotic function (e.g., Bcl-2
and Bcl-x
L
). Bcl-2 family members have
been postulated to inhibit apoptosis by a
number of mechanisms [42, 43].
The anti-apoptotic properties of Bcl-2
family members have been used to protect
industrially important cell lines, including
GS-NS0 and GS-CHO cells [44, 45], from in-
sults typically experienced during cell cul-
ture operations. Interestingly, although Tey
et al. [45] report that over-expression of
Bcl-2 protects a GS-NS0 cell line against
apoptosis, Murray et al. [46] reported no
benefit in another GS-NS0 cell line. These
workers found that this NS0 cell line ex-
pressed Bax and Bcl-x
L
. Given that Bcl-x
L
is a sequence and functional homologue
of Bcl-2, they postulated that Bcl-2 is redun-
dant in the NS0 cell background [46]. These
authors postulate further that cell lines such
as NS0 express only a subset of genes im-
portant in apoptosis. Modulation of death
characteristics in such cells will have to take
account of the expression profile of such
genes and their regulatory interactions.
A number of authors also report an in-
crease in product concentration achieved
in cultures of the Bcl-2 over-expressing cell
lines compared to the control cell line [45,
47, 48]. A fed-batch culture of the Bcl-2
over-expressing GS-NS0 cell line 6A1-bcl2
made more antibody than the parental cell
line 6A1(100)3: the antibody concentration
increased from about 26 mg L
1
at harvest
to about 38 mg L
1
[45]. Again, there are
also reports of no benefit [44]: the antibody
concentration achieved by a GS-CHO over-
expressing Bcl-2 was similar to that of the
control cell line at about 40 mg L
1
,
although differences in growth kinetics
were seen.
At least one alternative to over-expressing
Bcl-2 family proteins has also been evalu-
ated in NS0 cell lines. Studies in an anti-
body-producing GS-NS0 cell line using the
specific inhibitor Z-VAD-fmk, which targets
a range of caspases, showed that although
the extent of apoptosis was reduced there
was no benefit to productivity [49].
The data from studies of over-expressing
Bcl-2 in GS-NS0 cell lines are contradic-
tory, whilst no improvement in antibody
concentration was seen with the use of
caspase inhibitors. These observations,
coupled with the complexity of the circuits
controlling apoptosis, suggest that apopto-
sis will have to be modulated at several
sites simultaneously if a substantial in-
crease in product concentration is to be
achieved. However, this will increase the
metabolic load upon the cell.
Most of the studies of Bcl-2 over-express-
ing cell lines are limited in that, although
they used industrially important cell lines,
the systems used were only simple models
of modern biopharmaceutical manufactur-
ing processes. Characteristics of modern
commercial cell culture processes include
the use of serum- or protein-free media
and feeding strategies that support high vi-
able cell concentrations (ca. 10
7
mL
1
) for
extended periods (more than 240 h). In
contrast, the media used in the reports de-
scribed above often contained serum and
were not highly developed. The develop-
ment of media, feeds and processes may
have eliminated or postponed the appear-
ance of the insults that trigger apoptosis.
For example, we [50] have evaluated the
Bcl-2 over-expressing GS-NS0 cell line
6A1-bcl2 [45] in a scale-down model (10 L)
of a state-of-the-art fed-batch process used
to manufacture therapeutic antibodies at
the 5000-L scale. Again, as reported by Tey
et al. [45], over-expression of Bcl-2 resulted
in a substantial increase in the space-time
yield of viable biomass and protects
against apoptosis. However, unlike the re-
sults of Tey et al. [45] with the same cell
lines, no improvement in antibody concen-
tration was seen, with both cell lines pro-
ducing 500 to 700 mg L
1
antibody.
4.4.2
Manipulating the Cell Cycle
to Increase Productivity
Studies with hybridomas [51, 52] showed
an inverse correlation between growth rate
and specific antibody production rate.
Methods to achieve growth arrest whilst
maintaining high viability therefore have
potential for improving specific production
rates. Thus, an ideal production process
would involve a period of rapid cell growth
to a high viable cell concentration, with
the cells in a physiological state capable of
maintaining a high specific production
rate but with a low death rate. This phase
is then preserved by induction of a sus-
tained growth arrest. It is hypothesized
(e.g. [53]) that the cell diverts metabolism
from growth-associated processes to main-
tenance processes, which include the syn-
thesis of constitutively expressed recombi-
nant proteins. This section describes
approaches evaluated with GS cell lines to
arrest growth.
The cell cycle and cell proliferation are
controlled by the activity of cyclin-depen-
dent kinases (cdks) (for a review, see Ref.
[54]). The cdks are activated by association
with cyclin regulatory subunits and phos-
phorylation, and inhibited by binding of
inhibitors such as p21
CIP1
and p27
KIP1
.
The inhibitor p21
CIP1
inhibits cdk2, which
is known to have a role in the G
1
/S transi-
tion: over-expression of p21
CIP1
in a variety
of cell lines results in G
1
-phase cell cycle
arrest.
Al-Rubeai and co-workers have investi-
gated the effect of expressing p21
CIP1
in
both GS-CHO and GS-NS0 cell lines [55,
56]. In one study, an antibody-producing
GS-CHO cell line was engineered to ex-
press inducibly the p21
CIP1
cdk inhibitor
[56]. Upon induction, cell growth was ar-
rested and the specific production rate in-
creased, the largest increase being from
about 60 pg to about 250 pg per cellday.
However, the induced cells actually pro-
duced less antibody than the non-induced
cells, most likely because the loss in viable
biomass outweighed the increase in specif-
ic production rate. If p21
CIP1
was induced
at higher cell concentrations (above ca.
5 10
5
mL
1
), cell death was observed. In-
duction of apoptosis in growth-arrested
cells is a possibility, and it has previously
been shown for the parental GS-CHO cell
line that growth arrest induced apoptosis
that could be protected against by over-ex-
pression of Bcl-2 [44].
Over-expression of both a cdk inhibitor
and an anti-apoptosis protein has been
evaluated with a GS-NS0 cell line. When
cell line 6A1(100)3 was engineered to ex-
press p21
CIP1
from an inducible promoter,
the specific production rate increased by
up to 1.5- to 4.5-fold to 35 to 45 pg per
cell day [55]. However from the data pre-
sented, it can be inferred that, overall,
there was no increase in volumetric pro-
ductivity. The GS-NS0 cell line 6A1(100)3
has also been engineered to constitutively
express a mutant Bcl-2 with p21
CIP1
under
the control of an inducible promoter [53].
Again, an increase in specific production
was seen from about 10 pg per cellday,
which is similar to the parent 6A1 (100)3,
to about 50 pg per cellday. Examination
of the growth curve data for batch cultures
of these cell lines again suggests that the
loss of viable biomass outweighs the in-
crease in specific production rate so that
there was no benefit to volumetric produc-
tivity. Interestingly, the choice of Bcl-2
gene used to transfect the GS-NS0 parent
had a profound affect upon the degree of
protection against apoptosis. Previous
studies [44] used the wild-type protein and
saw an increase in IVC compared to the
non-transfected parent, 6A1(100)3. When a
mutant Bcl-2 protein that lacks any cell cy-
cle activity was introduced into cell line
6A1(100)3, no increase in the IVC was ob-
served [53].
The rational design approaches to im-
proving the phenotype are based upon the
direct manipulation of the transcriptome
through control of specific genes. The
problem with such approaches is that the
. . . profile of an ideal cell depends on a
multitude of genes that are rather poorly
understood, mostly unknown, and broadly
distributed throughout the genome [57].
Although we may be able to manipulate
genes (e.g., Bcl-2 or p21
CIP1
) that are
known to have a major role in regulating
complex pathways, the impact of these
changes upon other complex pathways
for example, the synthesis and secretion of
a recombinant antibody cannot be pre-
dicted. This can be seen when the impact
of over-expression of Bcl-2 is examined.
Improvements in volumetric productivity
were seen for some cell lines in some cul-
tures systems, but not others. Thus, there
still appears to be a role for the classical
strain improvement methods where cells
with desired phenotypes are isolated from
a mutagenized population.
Cell cycle mutants especially tempera-
ture-sensitive (ts) ones are a good source
of cell lines in which progression through
the cell cycle can be reversibly arrested. Typ-
ically, these cell lines have the potential to
maintain high viability for extended periods.
Jenkins and Hovey [58] isolated ts-mutants
from CHO-K1 and engineered these mu-
tants to express TIMP using a GS expression
vector. Optimization of temperature control
was investigated by repeatedly exposing the
culture to the non-permissive temperature
(398C), with recovery at the permissive tem-
perature (348C). The concentration of TIMP
increased from 200 to 300 mg L
1
due to a 3-
fold increase in specific production rate to
3.4 pg per cellday, with growth arrest and
no loss of culture viability.
4.4.3
Summary
In summary, a number of cell engineering
approaches have been evaluated with GS
cell lines to improve volumetric productiv-
ity. These approaches have included uncou-
pling cell growth from productivity and in-
creasing the space-time yield of viable bio-
mass by decreasing the death rate. Some
of these approaches resulted in an increase
in the specific production rate, but no in-
crease in the volumetric production rate
was seen: the latter is the key parameter
for a commercial manufacturing organiza-
tion. A few of these approaches have been
evaluated in state-of-the-art manufacturing
processes for biopharmaceutical proteins,
where different results to those obtained
in laboratory studies were obtained. The
reasons for this are not clear, but they may
be due to the elimination of apoptosis trig-
gers during process optimization.
4.5
Selection of Useful Cell Sub-populations
In addition to metabolic engineering, it is
sometimes possible to isolate useful sub-
populations of cell lines.
A limitation in the use of CHO cell lines
for producing biopharmaceutical proteins
has been the long time it can take to adapt
such cell lines to single cell suspension
culture in serum- or protein-free media. A
variant of the CHO-K1 cell line that grows
spontaneously in protein-free suspension
culture has been described for use with
the GS system [59]. The isolation of natu-
ral variants has also been exploited to iso-
late an NS0 clone which no longer re-
quires cholesterol [60]. This nutrient is in-
soluble and its addition to protein-free me-
dia is not straightforward.
4.6
Process Development
4.6.1
Media
In recent years there has been a drive to
remove serum, serum proteins and other
animal-derived materials from cell culture
media, motivated in large part by concerns
regarding the potential introduction of ad-
ventitious agents. The removal of complex
additions such as proteins offers other ad-
vantages; particularly cost reduction and
easier purification of product. In addition,
chemical definition of the medium greatly
assists process optimization.
Serum and serum proteins have diverse
functions which are now reasonably well
understood for the industrially important
cell lines, and which can generally be sub-
stituted by non-protein alternatives. Mam-
malian cells typically require a source of
fatty acids, which were historically sup-
plied by serum. To supply these, serum-
free media usually contain plasma lipopro-
tein fractions, free fatty acids complexed to
serum albumin or fatty acid/phospholipid
microemulsions [61]. A high-density lipo-
protein serum-fraction in medium contain-
ing bovine serum albumin was used by
Seamans et al. [62] to replace serum in
cultures of a recombinant antibody-produc-
ing GS-NS0 cell line. Further, they found
that the serum-fraction could be replaced
with a commercially available non-protein-
aceous lipid emulsion and a pluronic F-
68/cholesterol emulsion. This gave equiva-
lent growth and productivity (100 mg L
1
).
The requirement for cholesterol supple-
mentation for the serum-free culture of
NS0 cells is thought to be a function of their
ancestry as they are derived from the NS-1
cell line. The NS-1 cell line is deficient in
3-ketosteroid reductase activity, which is re-
sponsible for the conversion of lathosterol
to cholesterol, and leads to a requirement
for cholesterol [63]. However, the require-
ment for cholesterol can be circumvented.
Birch et al. [60] successfully isolated choles-
terol-independent variants of the NS0 host.
They achieved this by dilution cloning in a
medium that was free of both serum and
cholesterol. One of these variants was able
to grow in protein-free chemically defined
medium, without the addition of any lipids,
and with a population doubling time
equivalent to the parental cell line. In con-
trast, without cholesterol supplementation
the original NS0 cell line died within
24 hours. Keen and Hale [64] adapted an
antibody-producing GS-NS0 cell line to
grow in the absence of cholesterol. This re-
moved the final animal-derived raw materi-
al from their medium, which was further
improved by elevating the concentration of
glutamate, asparagine, ribonucleosides,
and choline chloride.
Iron delivery to cells in culture needs
careful consideration: transferrin has been
used successfully for many years, human
transferrin being more effective than bo-
vine [65]. However, this is still an animal-
derived raw material and thus undesirable
for biopharmaceutical manufacturing.
Some cells do not need an iron carrier and
can be supplied with soluble iron com-
pounds such as ferric ammonic citrate
[66]. In other cases an iron carrier may be
required such as the synthetic lipophilic
iron carrier tropolone [67, 68].
4.6.2
Glutamine-free Media
Aside from its use as a selectable marker,
there are physiological advantages in intro-
ducing GS to remove the glutamine de-
pendence of cells. Glutamine is relatively
unstable in culture media and degrades to
release ammonia, which can accumulate
to inhibitory levels. Several studies have
described the metabolic engineering of hy-
bridoma cell lines with GS to achieve glu-
tamine prototrophy [60, 69, 70]. Birch et
al. [60] demonstrated that a hybridoma
transfected with GS had increased anti-
body productivity when grown in the ab-
sence of glutamine.
4.6.3
Culture Conditions
It is usual to control pH, dissolved oxygen,
and temperature in bioreactors. Small
changes in pH can have dramatic effects
on process performance. Wayte et al. [71]
compared the effect of pH on a GS-NS0
and a hybridoma in shake-flask culture.
Using this approach, they found that the
specific growth rate was relatively constant
over the range pH 7.05 to 7.4, but that be-
low this pH culture growth was signifi-
cantly inhibited. In fed-batch bioreactor
cultures the response of different cell lines
to culture pH was variable. One hybrido-
ma had an increased IVC and a lower spe-
cific production rate at low culture pH,
whilst a second hybridoma showed an in-
creased IVC and no change in specific pro-
duction rate. However, in both cases, de-
creasing the culture pH from 7.2 to 7.1
caused an increase in the harvest antibody
concentration. The GS-NS0 cell line exam-
ined in this study was less sensitive to cul-
ture pH than the hybridomas, and larger
changes in culture pH were needed to af-
fect the culture. At pH 7.1, both the IVC
and the specific production rate were in-
creased compared to pH 7.4, resulting in
an increase in antibody concentration from
119 mg L
1
to 194 mg L
1
.
Osman et al. [72] investigated the effect of
pH shifts and perturbations in cultures of
the antibody-producing GS-NS0 cell line,
6A1(100)3, cultured in serum containing
batch culture. Cells growing at pH 7.3 were
able to continue growing after a shift in cul-
ture pH in the range of pH 7.0 to 8.0. A
shift in culture pH of greater than 0.2 pH
units caused a transient increase in the pro-
portion of apoptotic cell in the culture, but
the cultures were able to recover from this.
However, cultures were not able to recover
if the pH was decreased below 7.0 or in-
creased above 8.0. The culture pH affected
both growth and metabolism. The antibody
concentration was highest at pH 7.0 as a re-
sult of increased IVC, whilst the specific
production rate was constant over the rela-
tively wide pH range of 6.5 to 8.0. The
authors also investigated the effect of transi-
ent shifts in culture pH which could poten-
tially occur as a result of zoning in large-
scale reactors as a result of, for example, al-
kali addition to poorly mixed areas. Transi-
ent shifts had to be quite large to have an af-
fect on growth. Increases of culture pH to
above 8.5 for longer than 10 minutes in-
duced a lag and caused a reduction in the
maximum viable cell concentration. Similar
effects were seen at low pH (below pH 6.5),
but the perturbation needed to be for sev-
eral hours.
Using a GS-NS0 producing an IgG
1
in
protein-free fed-batch culture, Moran et al.
[73] investigated the effect of a range of
parameters on the growth rate, specific
production rate, IVC and antibody concen-
tration at harvest. In contrast to the results
of Wayte et al. [71] and Osman et al. [72],
there was no statistically significant effect
on any of these parameters within the
range of culture pH from 7.1 to 7.5. More
importantly, there was no detectable
change in the distribution of glycoforms of
the antibody. It is not clear why there are
such differences but it may depend upon
the particular cell line as well as the pro-
cess.
4.6.4
Fed-batch Cultures
The early mammalian cell processes were
typically batch. As culture media and pro-
cesses have developed over the years, ad-
vances in feeding strategies for fed-batch
processes have increased productivity to
several grams per liter for both GS-NS0 [4]
and GS-CHO cell lines [5].
One approach to implementing a fed-
batch strategy is to feed cultures with me-
dium concentrates. This can offer a rapid
approach to increasing productivity, and
can also be relatively simple to implement
[17, 74]. Using GS-NS0 cells producing an
antibody, Bibila et al. [17] fed cultures with
10 basal medium concentrates (Iscoves
Modified Dulbeccos medium) to increase
productivity. Sodium chloride, potassium
chloride and sodium bicarbonate were
omitted from the medium concentrates in
order to minimize the increases in osmo-
larity caused by feeding. In their system,
feeding basal medium concentrates did
not result in an increase in the maximum
viable cell concentration or the IVC. How-
ever, the final antibody concentration was
increased 1.9-fold as a result of an increase
in the specific production rate.
A further refinement of the fed-batch
method is to feed the supplements added
to the medium in addition to the basal me-
dium concentrates [17]. This approach was
shown to be more effective than concen-
trates alone, and led to increases in the max-
imum viable cell concentration (1.7- to 2-
fold), the IVC (2.3- to 3.3-fold) and the spe-
cific production rate (2-fold). These effects
combined to produce an up to a 7-fold in-
crease in antibody concentration. Further
increases might be expected by feeding
more nutrients, though above a certain vol-
ume of additions a decrease in process per-
formance was observed. This was thought
to be the result of increases in osmolarity
caused by the medium components. Infor-
mation on metabolism gained from the me-
dium concentrate experiments was then
used to develop an optimized fed-batch pro-
cess. The strategy chosen for this was to
maintain nutrient homeostasis, where the
amino acid concentrations were maintained
at their original concentrations and the cul-
ture was supplemented with glucose, lipids
and proteins. Further development of the
fed-batch process required significant pro-
cess development time and effort. How-
ever, this led to product concentration at
harvest of 1.8 and 1.2 g L
1
.
The effect of medium osmolarity on the
growth of GS-NS0 cells was investigated
by Bibila et al. [17]. Cell growth was re-
duced when the osmolarity was increased
to 400 mOsm and completely inhibited
above 500 mOsm. The specific production
rate increased as the osmolarity was in-
creased from the baseline of 270 mOsm to
300 and 400 mOsm. However, as a direct
result of reduced growth, the cultures at
400 mOsm reached a lower product con-
centration than the controls. Zhou et al.
[2] noted that increases in osmolarity be-
low 450 mOsm had little impact upon pro-
ductivity, but above this level there was a
rapid increase in the specific production
rate. However, growth cessation occurred
at this elevated osmolarity.
Zhou et al. [2] refined the nutrient home-
ostasis approach further by feeding cultures
based on the IVC, with the aim of keeping
nutrient concentrations around their origi-
nal concentrations. However, this assumes
that the consumption and yields of these
nutrients are constant throughout the cul-
ture, which may not be correct. On-line
measurement of the oxygen uptake rate
(OUR) was used to infer nutrient depletion.
Rapid decreases in OUR were observed that
could be reversed by addition of amino
acids. This did not result in an increased
cell concentration, indicating that another
nutrient was limiting or that some factor
had accumulated to a growth-inhibitory lev-
el. The addition of an increased amount of a
cholesterol complex in conjunction with the
amino acid feed was able to restore growth
and increased the product concentration to
2.7 g L
1
. Despite responsive feeding based
on the OUR, it was not possible to maintain
growth indefinitely. It was a reduction in
the cell death rate that resulted in a pro-
longed culture lifetime. During the decline
phase there was a much slower linear de-
crease in the viable cell concentration and
OUR, rather than the rapid decreases in
OUR observed previously. The authors sug-
gested that this indicated that cell death
Fig. 4.6 Changes in process parameters during optimization of a GS-CHO process
producing an IgG
4
antibody using chemically defined animal-component free me-
dia in 10-L laboratory-scale airlift bioreactors: (a) growth parameters; (b) productiv-
ity parameters.
might be caused by environmental condi-
tions such as high osmolarity rather than
nutrient limitation. One drawback noted
was that although no base was used for
pH control, because the amino acid solution
had a pH of 9.5, feeding ultimately resulted
in a rise in the CO
2
concentration in the re-
actor as a result of maintaining the culture
pH set point. As nutrient metabolism
changes through the different growth
phases, they proposed a two-feed strategy
where one feed is used to extend the cell
growth phase, after which a different feed
is used to prolong culture longevity.
deZengotita et al. [3], using the GS-NS0
cell line described by Zhou et al. [2], found
that feeding phosphate prolonged the cell
growth phase and delayed the onset of
apoptosis, resulting in a doubling of the
maximum viable cell concentration. An in-
creased IVC resulted in an increase in the
product concentration at harvest from 0.5
to 1.3 g L
1
. This also delayed the meta-
bolic shift from lactate production to lac-
tate consumption.
Sauer et al. [74] discussed the need for
high-yielding generic fed-batch processes
to decrease the amount of development
time prior to manufacturing. They used a
similar approach to that of Bibila et al.
[17], starting with partial media concen-
trates and initially controlled feed additions
based on glucose concentration. For an
Sp2/0 cell line producing an antibody, this
led to a 3-fold increase in product concen-
tration, from 70 to 220 mg L
1
. By changing
the glucose concentration at which feeds
were added, it was possible to demonstrate
the effect of underfeeding, comparable to
batch culture, and overfeeding: both condi-
tions showed a reduction in the final anti-
body concentration. That the feeding re-
gime was robust was demonstrated by the
range of glucose concentration over which
the process could operate without adversely
affecting process performance. To test the
general applicability of the process it was
tested on a panel of cell lines. In each case,
compared to batch culture, feeding in-
creased both the exponential growth phase
Table 4.2 Oligosaccharide profiles determined by MALDI-TOF MS
for a GS-NS0 IgG
4
antibody during process optimization in chem-
ically defined animal component-free fed-batch culture
Structure Relative Peak Intensity [%]
0.37 g L
1
0.48 g L
1
0.75 g L
1
1.0 g L
1
1.4 g L
1 a)
G2F+2 (a-Gal) 4.0 3.5 3.3 2.4 3.1
G2F+(a-Gal) 8.4 6.8 7.6 6.0 6.0
G2F 39.1 43.2 41.8 41.9 40.4
G1F 34.7 32.1 33.7 32.7 37.8
G0F 9.1 9.9 10.5 12.6 12.7
G1F-GN 1.5 1.6 1.2 1.7 ND
G0 0.7 0.5 0.8 0.6 ND
G0F-GN 1.5 1.5 1.2 1.4 ND
Man-5 1.0 0.9 0.0 0.9 ND
a) Analyzed in separate assay to other samples.
ND = Not detected.
and the culture duration. Most of the in-
crease in product concentration was a result
of an increased IVC rather than increased
specific production rate. Between the differ-
ent cell lines there were marked differences
in the specific glucose consumption rate, up
to a factor of 4-fold, whilst the apparent
yield of lactate on glucose was relatively un-
changed. Interestingly, there was an inverse
correlation between specific glucose utiliza-
tion rate and IVC.
Similar improvements in productivity
were obtained by Dempsey et al. [68], who
performed repeated rounds of nutrient
supplementation and analysis to develop
nutrient supplements for their GS-NS0
cultures. They tested these supplements
on three cell lines producing different anti-
bodies, and attained a 10-fold increase over
the original product concentrations.
Shaw et al. [24] showed that the chemi-
cally defined animal component-free pro-
cess they developed using the GS-NS0 cell
line 6A1(100)3 was applicable to other cell
lines. Using a different cell line that was
making 1 g L
1
in a serum-free process,
with no optimization for this second cell
line, an antibody concentration of 1.8 g L
1
was attained. This has subsequently been
confirmed with other cell lines producing
above 1 g L
1
(unpublished results).
Process optimization using our model
GS-CHO cell line (22H11) was achieved
using multiple rounds of fermentations in
chemically defined media (Fig. 4.6). The
initial optimization was performed by
changing the base medium, and the feeds
were modified using the approach of spent
medium analysis and re-supplementation.
This increased the yield from 139 mg L
1
to 585 mg L
1
a 4-fold increase in pro-
ductivity. This optimized process was then
used as the starting point for a new, non-
amplified GS-CHO cell line (LB01), and
this resulted in a 14-fold increase over the
original process, to 1917 mg L
1
. Further
process optimization was then performed
using the new cell line. The progress of
the optimization is shown in Fig. 4.6. For
iterations 4 and 5 (compare LB01 v4 and
LB01 v5 in Fig. 4.6), the pH control was
optimized, which resulted in a further im-
provement in productivity to 4301 mg L
1
,
a 31-fold increase over the original pro-
cess. It is apparent from the data shown
in Fig. 4.6 that it is possible to improve
productivity by optimising several parame-
ters, namely specific production rate, IVC,
and maximum viable cell concentration.
4.6.5
Process Optimization and Product Quality
One of the concerns with increasing the
product concentration is that the product
quality characteristics are maintained. We
have monitored the product quality of a
GS-NS0 cell line throughout the optimiza-
tion process. Through successive rounds
of optimization involving changes in the
composition of the feeds, culture pH and
extending culture duration, the product
concentration from the GS-NS0 process
was increased from 0.37 to 1.4 g L
1
. There
were no major changes observed in the oli-
gosaccharide profiles during this optimiza-
tion process (Table 4.2). It cannot however
be assumed that changes will not occur,
and it is essential to monitor product qual-
ity during process development. For exam-
ple, we found an increased proportion of
an aglycosyl variant of an antibody pro-
duced in GS-NS0 during process optimiza-
tion. This was shown to be a result of glu-
cose becoming limiting under revised
feeding conditions.
4.7
Summary
Significant progress has been made in re-
cent years in the development of high-
yielding processes for the production of
biopharmaceuticals. Highly efficient non-
amplified gene expression systems such as
that based on glutamine synthetase, in
combination with new approaches to
screening, have provided highly productive
cell lines. We can expect to see further im-
provements to cell lines resulting from de-
liberate engineering of desirable character-
istics. Improved understanding of cell
physiology using modern omics tools
will contribute significantly to these ef-
forts, and we are already seeing the first
indications of this [75]. In parallel with
these developments in the design of cell
lines, we have also seen impressive pro-
gress in the optimization of culture pro-
cesses, particularly through the use of so-
phisticated feeding strategies for fed-batch
culture. For recombinant antibodies it is
now possible regularly to achieve yields in
excess of 1 g L
1
in completely chemically
defined media, and it is probable that
yields for modern biopharmaceuticals of at
least 10 g L
1
will be achieved in the fore-
seeable future.
References
1 C. R. Bebbington, G. Renner, S. Thomson, D.
King, D. Abrams, G. T. Yarranton, Bio/Technol.
1992, 10, 169175.
2 W. Zhou, C-C. Chen, B. C. Buckland, J. G. Au-
nins, Biotechnol. Bioeng. 1997, 55, 783792.
3 V. M. deZengotita, W. M. Miller, J. G. Aunins,
W. Zhou, Biotechnol. Bioeng. 2000, 69, 566
576.
4 Varma, A. E. Schmelzer, S. White, M. Patel, C.
DiGirolamo, S. Case. The Waterside Conference,
8
th
Annual Conference, San Francisco (Abst).
2003.
5 D. O. Mainwaring, 227
th
American Chemical
Society Meeting, Anaheim (Abst). 2004.
6 P. G. Sanders, R. H. Wilson, EMBO J. 1984, 3,
6571.
7 C. R. Bebbington, Methods: A Companion to
Methods in Enzymology. 1991, 2, 136145.
8 M. I. Cockett, C. R. Bebbington, G. T. Yarran-
ton, Bio/Technology 1990, 8, 662667.
9 D. R. Burton, J. Pyati, R. Koduri, S. J. Sharp,
G. B. Thornton, P. W. H. I. Parren, L. S. W. Saw-
yer, R. M. Hendry, N. Dunlop, P. L. Nata, M.
Lanacchia, E. Garratty, E. R. Stiehm, Y. J. Bry-
son, Y. Cao, J. P. Moore, D. D. Ho, C. F. Bar-
bas, Science 1994, 266, 10241027.
10 V. E. Laubach, E. P. Garvey, P. Sherman, Bio-
chem. Biophys. Res. Commun. 1996, 218, 802
807.
11 J. McKnight, B. J. Classon, Immunology 1992,
75, 286292.
12 D. Guerini, S. Schroder, D. Foletti, E. Carafoli,
J. Biol. Chem. 1995, 270, 1464314650.
13 R. Flesher, J. Marzowski, W.-C. Wang, H. V.
Raff, Biotechnol. Bioeng. 1995, 46, 399407.
14 K. N. Baker, M. H. Rendall, A. E. Hills, M.
Hoare, R. B. Freedman, D. C. James, Biotech-
nol. Bioeng. 2001, 73, 188202.
15 M. R. Lifely, C. Hale, S. Boyce, M. J. Keen, J.
Phillips, Glycobiology 1995, 5, 813822.
16 T. E. Hassell, H. N. Brand, G. L. Renner, A. J.
Westlake, R. P. Field, In: Animal Cell Technol-
ogy: Developments, Processes and Products (R. E.
Spier, J. B. Griffiths, C. MacDonald, Eds.), But-
terworth-Heinemann, UK, pp. 4247, 1992.
17 T. A. Bibila, C. Ranucci, K. Glazomitsky, B. C.
Buckland, J. G. Aunins, Ann. N. Y. Acad. Sci.
1994, 745, 277284.
18 T. C. Peakman, J. Worden, R. H. Harris, H.
Cooper, J. Tite, M. J. Page, D. R. Gewert, M.
Bartholomew, J. S. Crowe, S. Brett, Hum. Anti-
bod. Hybrid. 1994, 5, 6574.
19 G. J. Pendse, S. Karkare, J. E. Bailey, Biotech-
nol. Bioeng. 1992, 40, 119129.
20 M. B. Gu, P. Todd, D. S. Kompala, Cytotech-
nology. 1996, 18, 159166.
21 U. H. Weidle, P. Buckel, J. Wienberg, Gene
1988, 66, 193203.
22 M. I. Cockett, C. R. Bebbington, G. T. Yarran-
ton, Nucleic Acids Res. 1991, 19, 319325.
23 M. E. Reff, Internl. Pat. WO 94/11523, 1994.
24 M. Shaw, D. O. Mainwaring, T. S. Root, E. E.
Allen, Cell Culture Engineering XIV, Cancn
(Abst). 2004.
25 R. S. Barnett, K. L. Limoli, T. B. Huynh, E. A.
Ople, M. E. Reff, In: Antibody Expression and
Engineering (H. Y. Wang, Ed.), American
Chemical Society, USA, pp. 2740, 1995.
26 G. F. Hollis, G. E. Mark, Internl. Pat. WO95/
17516, 1995.
27 T. Benton, T. Chen, M. McEntee, B. Fox,
D. King, R. Crombie, T. C. Thomas, C. R. Beb-
bington, Cytotechnology 2002, 38, 4346.
28 J. Running Deer, D. S. Allison, Biotechnol.
Prog. 2004, 20, 880889.
29 K. T. Powell, J. C. Weaver, Bio/Technology 1990,
8, 333337.
30 S. C. G. Brezinsky, G. G. Chiang, A. Szilvasi, S.
Mohan, R. I. Shapiro, A. MacLean, W. Sisk, G.
Thill, J. Immunol. Methods 2003, 277, 41155.
31 R. Manz, M. Assenmacher, E. Pflger, S. Mil-
tenyi, A. Radbruch, Proc. Natl. Acad. Sci. USA
1995, 92, 19211925.
32 P. Holmes, M. Al-Rubeai, J. Immunol. Methods
1999, 230, 141147.
33 A. J. Racher, Third Annual Biological Production
Forum, Edinburgh (Abst.). 2004.
34 H. N. Brand, S. J. Froud, H. K. Metcalfe, A. O.
Onadipe, A. Shaw, A. J. Westlake, In: Animal
Cell Technology: Products for Today, Prospects for
Tomorrow (R. E. Spier, J. B. Griffiths, W. Ber-
thold, Eds.), Butterworth-Heinemann Ltd, UK,
pp. 5560, 1994.
35 L. Cosgrove, G. O. Lovrecz, A. Verkuylen, L.
Cavaleri, L. A. Black, J. D. Bentley, G. J. How-
lett, P. P. Gray, C. W. Ward, N. M. McKern,
Prot. Purificat. 1995, 6, 789798.
36 P. Bird, E. Bolam, L. Castell, O. Obeid, N.
Darton, G. Hale, In: New Developments and
New Applications in Animal Cell Technology
(O.-W. Merten, P. Perrin, J. B. Griffiths, Eds.),
Kluwer Academic Publishers, The Nether-
lands, pp. 4349, 1998.
37 L. M. Barnes, C. M. Bentley, A. J. Dickson, Bio-
technol. Bioeng. 2001, 73, 261270.
38 L. M. Barnes, C. M. Bentley, A. J. Dickson, Bio-
technol. Bioeng. 2003, 85, 115121.
39 M. Al-Rubeai, D. Mills, A. N. Emery, Cytotech-
nology 1990, 4, 1328.
40 N. Arden, M. J. Betenbaugh, Trends Biotechnol.
2004, 22, 174180.
41 D. J. DiStefano, G. E. Mark, D. K. Robinson,
Biotechnol. Lett. 1996, 18, 1071072.
42 H. Zou, W. J. Henzel, X. Liu, A. Lutschg,
X. Wang, Cell. 1997, 90, 405413.
43 D. G. Kirsch, A. Doseff, B. N. Chau, D.-S. Lim,
N. C. de Souza-Pinto, M. B. Kastan, Y. A. La-
zebnik, J. M. Hardwick, J. Biol. Chem. 1999,
274, 2115521161.
44 B. T. Tey, R. P. Singh, L. Piredda, M. Piacenti-
ni, M. Al-Rubeai, Biotechnol. Bioeng. 2000, 68,
3143.
45 B. T. Tey, R. P. Singh, L. Piredda, M. Piacenti-
ni, M. Al-Rubeai, J. Biotechnol. 2000, 79, 147
159.
46 K. Murray, C.-E. Ang, K. Gull, J. A. Hickman,
A. J. Dickson, Biotechnol. Bioeng. 1996, 51,
298304.
47 S. Terada, T. Komatsu, T. Fujita, A. Terakawa,
T. Nagamune, S. Takayama, J. C. Reed, E. Su-
zuki, Cytotechnology 1999, 31, 141149.
48 N. S. Kim, G. M. Lee, Biotechnol. Bioeng. 2001,
71, 184193.
49 S. L. McKenna, T. G. Cotter, Biotechnol. Bioeng.
2000, 67, 165176.
50 Perani, J. Kenworthy, I. Alam, G. Gilchrist,
H. K. Metcalfe, J. R. Birch, A. J. Racher, Second
European Meeting on Cell Engineering, Costa
Brava, Spain (Abst), 2001.
51 W. M. Miller, H. W. Blanch, C. R. Wile, Biotech-
nol. Bioeng. 1988, 32, 947965.
52 E. Suzuki, D. F. Ollis, Biotechnol. Prog. 1990,
6, 231236.
53 N. Ibarra, S. Watanabe, J.-X. Bi, J. Shuttle-
worth, M. Al-Rubeai, Biotechnol. Prog. 2003,
19, 224228.
54 M. Fussenegger, J. E. Bailey, Biotechnol. Prog.
1998, 14, 807833.
55 S. Watanabe, J. Shuttleworth, M. Al-Rubeai,
Biotechnol. Bioeng. 2002, 77, 17.
56 J.-X. Bi, J. Shuttleworth, M. Al-Rubeai, Bio-
technol. Bioeng. 2004, 85:741749.
57 G. Stephanopoulos, Nature Biotechnol. 2002,
20, 666668.
58 N. Jenkins, A. Hovey, In: Animal Cell Technol-
ogy: Basic and Applied Aspects. Vol. 5. Kluwer
Academic Publishers, The Netherlands,
pp. 267272, 1993.
59 M. H. Rendall, A. Maxwell, D. Tatham, P.
Khan, R. D. Gay, R. C. Kallmeier, J. R. T. Wayte,
A. J. Racher, In: Animal Cell Technology
Meets Genomics (F. Gdia, M. Fussenegger,
Eds.) Springer, The Netherlands, pp. 701704,
2005.
60 J. R. Birch, R. C. Boraston, H. K. Metcalfe,
M. E. Brown, C. R. Bebbington, R. P. Field,
Cytotechnology. 1994, 15, 1116.
61 S. Schmidt, In: Production of Biologicals from
Animal Cells in Culture (R. E. Spier, J. B. Grif-
References 831
fiths, B. Meigner, Eds.) Butterworth-Heine-
mann, UK, pp. 6769, 1991.
62 T. C. Seamans, S. L. Gould, D. J. DiStefano, M.
Silberklang, D. K. Robinson, Ann. N. Y. Acad.
Sci. 1994, 745, 240243.
63 J. D. Sato, T. Kawamoto, T. Okamoto, J. Exp.
Med. 1987, 165, 17611766.
64 M. J. Keen, C. Hale Cytotechnology 1996, 18,
207217.
65 T. O. Messmer, Biochim. Biophys. Acta 1973,
320, 663670.
66 M. Titeux, U. Testa, F. Louache, P. Thomo-
poulos, H. Rochant, J. Breton-Gorius, J. Cell
Physiol. 1984, 121, 251256.
67 H. K. Metcalfe, R. P. Field, S. J. Froud, In: Pro-
duction of Biologicals from Animal Cells in Cul-
ture (R. E. Spier, J. B. Griffiths, B. Meigner,
Eds.), Butterworth-Heinemann, UK, pp. 88
90, 1994.
68 J. Dempsey, S. Ruddock, M. Osborne, A. Rid-
ley, S. Sturt, R. P. Field, Biotechnol. Prog. 2003,
19, 175178.
69 S. L. Bell, C. R. Bebbington, M. F. Scott, J. N.
Wardell, R. E. Spier, M. E. Bushell, P. G. San-
ders, Enzyme Microb. Technol. 1995, 17, 98
106.
70 C. Paredes, E. Prats, J.aJ. Cair, F. Azorn, L.
Cornudella, F. Gdia, Cytotechnology 1999, 30,
8593.
71 J. R. T. Wayte, R. C. Boraston, H. Bland, J. Var-
ley, M. E. Brown, Genet. Engineer. Biotechnol.
1997, 17, 125132.
72 J. J. Osman, J. R. Birch, J. Varley, Biotechnol.
Bioeng. 2001, 75, 6373.
73 E. B. Moran, S. T. McGowan, J. M. McGuire,
J. E. Frankland, I. A. Oyebade, W. Waller, L. C.
Archer, L. O. Morris, J. Pandya, S. R. Nathan,
L. Smith, M. L. Cadette, J. T. Michalowski, Bio-
technol. Bioeng. 2000, 69, 242255.
74 P. W. Sauer, J. E. Burky, M. C. Wesson, H. D.
Sternard, L. Qu, Biotechnol. Bioeng. 2000, 67,
585597.
75 C. M. Smales, D. M. Dinnis, S. H. Stansfield,
D. Alete, E. A. Sage, J. R. Birch, A. J. Racher,
C. T. Marshall, D. C. James, Biotechnol. Bioeng.
2004, 88, 474488.
Abstract
Technical advances made during the past 20
years have enabled the genetic transforma-
tion and regeneration of transgenic plants
and animals for the tissue-specific accumu-
lation of recombinant human proteins.
These transgenic systems provide produc-
tion technology for biopharmaceuticals re-
quiring complex multi-subunit assembly
(e.g., vaccines and secretory antibodies)
and for proteins that can not be efficiently
synthesized by current commercial blood
fractionation microbial mammalian cell cul-
ture systems. The manufacture of biothera-
peutics in transgenic animals and plants
grown using conventional agronomic and
farming practices also offers the opportu-
nity to produce practically unlimited sup-
plies of life-saving products at low cost. In
addition, the production of biotherapeutics
using transgenic systems (e.g., milk) offers
the highest accumulation level of heterolo-
gous protein accumulation obtained from
a recombinant production system. Trans-
genic plants allow the production of bio-
pharmaceuticals free of potential animal-de-
rived contaminants and pathogens such as
prions in a matrix that can be used for oral
delivery, without additional purification and
with no requirement for refrigeration. Seeds
provide a stable matrix for handling and
storing biopharmaceuticals for years after
harvest decoupling downstream processing
from biosynthesis. In addition to these at-
tractive advantages, the implementation of
transgenic systems for biopharmaceutical
production offers the opportunity to im-
prove agricultural efficiency and profitabil-
ity whilst reinforcing the public perception
of biotechnology as an important tool to en-
hance both agriculture and healthcare.
Transgenic systems can deliver innovative
biotherapeutics to treat cancer, infectious
diseases, inflammation, organ rejection,
skin conditions, genetic deficiencies, and
respiratory ailments. These biopharmaceu-
ticals will be both affordable and accessible
to broad segments of the population and de-
veloping regions of the world that currently
do not have access to these treatments. This
chapter will focus on biopharmaceuticals
833
ISBN: 3-527-31184-X
5
Biopharmaceuticals Derived from Transgenic Plants and Animals
Julio Baez
Vivat, Crescat, Floreat
A Ripe and Blooming Market for Transgenic Animals and Plants
derived from transgenic animals and plants
that are currently commercialized, or that
have human clinical experience. Exploring
the properties and performance of these
transgenic-derived biopharmaceuticals used
as diagnostics, protein replacement therapy,
cancer therapeutics, immunoprophylactics,
as anti-infectives, nutraceuticals, excipients,
and in medical devices will provide under-
standing of the status and of the potential
for transgenic-based production systems.
Abbreviations
AAT alpha-1-antitrypsin
AIDS aquired immunodeficiency
syndrome
AT-III antithrombin III
BSSL bile salt-stimulated lipase
DMF drug master file
EpCAM epithelial cellular-adhesion mol-
ecule
HAE hereditary angioedema
hBChE human butyrylcholinesterase
HIV human immunodeficiency
LTB labile toxin B
MPS-I mucopolysaccharidosis
rhAAT recombinant human alpha-1-
antitrypsin
rhAGLU recombinant human alpha-glu-
cosidase
rhC1I recombinant human C1 inhibi-
tor
rhFIB recombinant human fibrinogen
rhLF recombinant human lactoferrin
rhLZ recombinant human lysozyme
TMV tobacco mosaic virus
USDA US Department of Agriculture
USDA/ US Department of Agricul-
APHIS ture/Animal and Plants Health
Inspection Service
5.1
Introduction
During the past 20 years, the application
of recombinant DNA technology to health-
care has enabled the introduction of more
than 140 biopharmaceutical products pro-
viding innovative diagnostic, preventive,
and therapeutic treatment for cancer, car-
diovascular disease, diabetes, sepsis, infec-
tious diseases, inflammation, organ rejec-
tion, skin ailments, autoimmune condi-
tions, respiratory ailments, genetic defi-
ciencies, and asthma [1]. About 370
biopharmaceuticals are currently under-
going clinical trials, targeting more than
200 diseases [1]. Recombinant human pro-
teins used as biotherapeutics are derived
from mammalian cell culture and micro-
bial fermentation. The application of these
recombinant production technologies dur-
ing the past 25 years has delivered many
innovative products that have provided
new opportunities for growth to the phar-
maceutical industry, while making avail-
able life-saving diagnostic, prophylactic,
and therapeutic approaches to health pro-
viders and to patients. Many of these re-
combinant human products have replaced
biologics prepared from animal/human
tissues, whilst others have been made
available for the first time, as they could
not be recovered from natural sources. Mi-
crobial and mammalian culture-based re-
combinant production technologies, sup-
plemented with insect cell culture and sol-
id-phase protein synthesis, have also pro-
vided valuable reagents for the discovery,
development, and analysis of proteins and
non-protein-based new chemical entities
used as drugs [2].
In parallel to these efforts to develop re-
combinant systems for the production of
innovative drugs, the same recombinant
DNA technology has been applied to im-
provements in agricultural systems for the
successful commercialization of transgenic
crop plants and farm animal recombinant
products. Taken together, this has resulted
in enhanced agronomic performance and
productivity. The first wave of agricultural
biotechnology products herbicide/pest-
tolerant plants and bovine growth hor-
mone are currently providing higher
profits to farmers and agricultural compa-
nies whilst minimizing the negative im-
pact of agricultural activities on the envi-
ronment. Farmers have quickly embraced
recombinant technology when it is avail-
able, as illustrated by its successful imple-
mentation in the US. By 2004 just eight
years after the introduction of the first
commercial pest-resistant crops 45% of
the corn, 85% of soy and 76% cotton fields
planted in the US will have genetically en-
hanced plants [3]. A second wave of geneti-
cally enhanced crop plants are on the hori-
zon, these being designed to deliver im-
proved shelf-life and quality food products
with higher concentrations of designes
health-enhancing oils, proteins, and vita-
mins. Alternatively, crop plants have been
designed for improved agronomic perfor-
mance to further enhance the value of
agricultural biotechnology to society.
The use of transgenic animals and
plants as factories for valuable pharmaceu-
tical and industrial products represents the
third wave of agricultural biotechnology
products derived from genetically en-
hanced organisms. Recombinant DNA
technology allows the accumulation of re-
combinant human proteins in all tissues
of a transgenic organism, or selectively in
a particular tissue. Biopharmaceutical ac-
cumulation can be directed into conven-
tional agricultural products such as milk,
eggs, foliage, fruits, stems, and seeds that
are normally harvested from farm animals
and crop plants. Directed accumulation of
biopharmaceuticals into these familiar
agricultural products facilitates the imple-
mentation of transgenic production tech-
nology using established agricultural prac-
tices. The use of recombinant DNA tech-
nology to generate transgenic-derived bio-
pharmaceuticals is a continuation to the
historical use of animal and plant tissues
and derived products to provide valuable
health-enhancing agents. Since the begin-
ning of medicine, human/animal blood,
animal tissues, and plants had been the
source of many oral, topical, and injectable
therapeutics, such as Factor VIII for hemo-
philiacs, serum albumin used as plasma
expander, porcine insulin for diabetes
treatment, egg viral vaccines for immuni-
zation, therapeutic polyclonal antibodies,
steroids, and plant morphine for pain
treatment. Of the 100 most frequently pre-
scribed US drugs, about one-fifth are ob-
tained directly from plants; representing
products such as birth control pills (Mexi-
can yam), digitalis (foxglove), and recent
anticancer therapeutics such as taxol from
the Pacific yew tree. Transgenic animals
and plants are simply providing innovative
ways to enhance the use of animal and
plant tissues and derived products to pro-
vide new biopharmaceuticals.
The initial implementation of recombi-
nant DNA technology in agriculture to de-
liver improved agronomic performance to
crops and to animals had limited direct
impact upon the efficiency and profitabil-
ity of food production, and did not deliver
value-added products to consumers. This
created the perception in the food industry
and consumers that agricultural biotech-
nology has no direct value to them. These
groups instead focused on the perceived
high risk associated with the implementa-
tion of agricultural biotechnology, and ig-
nored its potential to improve nutrition
and healthcare. This perception of low val-
ue/high risk has created significant contro-
versy related to the implementation of
products derived from agricultural biotech-
nology. Valid technical concerns related
with the unknown long-term impact of ge-
netic modifications on food safety and the
environment combined with questions
about the actual cost benefits to the farm-
ing community, food industry, consumers
and governments providing farming subsi-
dies resulted in the generation of signifi-
cant opposition to the implementation of
agricultural biotechnology, especially out-
side the US. Extensive testing has been
conducted to certify the safety to both con-
sumers and the environment of products
derived from genetically enhanced plants
and animals. However, no long-term expo-
sure data are available, and this has cre-
ated the demand by many organizations to
slow down the implementation of geneti-
cally enhanced food and to provide proper
labeling for these products. The use of
crop plants and farm animals for the pro-
duction of biopharmaceuticals will have to
be implemented in this controversial envi-
ronment. Those involved in developing the
technology, the food industry, and the reg-
ulatory agencies are working together to
ensure that the production of biopharma-
ceuticals using transgenic animals and
plants can be implemented without con-
taminating the food supply or the environ-
ment [4].
The first successful therapeutic protein
made in a transgenic system was human
tissue plasminogen activator regulated by
a milk-directed promoter for accumulation
in mouse milk [5]. Human growth hor-
mone, which was one of the first proteins
produced using recombinant microbial
systems in the early 1980s, became the
first human protein expressed in plants
(tobacco cells) in 1986 [6]. Since then, over
200 biotherapeutics of diverse origin,
structure, and function such as antibodies,
enzymes, antigens for vaccines, and hor-
mones have been successfully expressed in
tissues capable of being regenerated into
transgenic animals and plants. Today,
transgenic-derived recombinant human
proteins are commercially available for
non-human use in research, processing,
and as diagnostics. There is no biothera-
peutic derived from a transgenic plant or
animal approved for human therapeutic
use. One product, goat milk-derived inject-
able recombinant human antithrombin,
was submitted for European approval (see
Part IV, Chapter 11). Transgenic produc-
tion strategies, material used for extracting
the recombinant product based on these
strategies, host organisms, commercial or
academic institutions developing these
strategies, and product examples are listed
in Table 5.1. For transgenic animals, pro-
duction is conducted in specially built
barns designed and managed exclusively
for recombinant protein manufacturing.
Using transgenic plants, there are two ba-
sic strategies for manufacturing plant-de-
rived biopharmaceuticals: 1) production in
open fields, greenhouses or underground;
and 2) production in bioreactors contain-
ing transgenic aquatic plants, cells, or tis-
sues in suspension such as roots, me-
chanically chopped foliage, or germinated
seeds. Biopharmaceuticals are recovered
from milk, semen, whole organisms,
blood, and eggs using transgenic animals;
from foliage (transgenic or using viral in-
fection), tubers, stems, fruits, and seeds
using plants. Most farming animals and
commercial crops have been used for the
production of biopharmaceuticals by over
120 institutions (commercial and aca-
demic; see Table 5.1), though several of
the commercial establishments listed are
no longer operational. Table 5.1 also lists
the large number of recombinant proteins
Table 5.1 Production strategies, material used for extraction, host organ-
isms, institutions, and products from transgenic systems.
Production
strategy
Material used
for extraction
Host
organism
Institutions Product examples Reference
Transgenic
animals housed
in specialized
barns
Milk Mouse
Rabbit
Cow
Goat
Sheep
Pig
Astra, Sweden
INRA, France
Human extracellular
superoxide dismutase
36
BioProtein
Technologies
Antibodies, vaccines,
human C1 inhibitor, ery-
thropoetin, superoxide
dismutase
37, 38
Gala Biotechnology Not available 39
GTC Biotherapeutics Human antithrombin,
Human serum albumin,
HIV vaccine, malaria vac-
cine, Monoclonal antibod-
ies, Peptides, Fusion
proteins, Beta-interferon,
Interferon alpha, Glutamic
acid decarboxylase, Hu-
man growth hormone,
Insulin, Tissue plasmino-
gen activator
9, 4049
Infigen human collagen type I,
Human fibrinogen, alpha
glucosidase, gelatin
50, 51
Institut fr Tierzucht
Tierverhalten,
Neustadt, Germany &
Fraunhofer,
Hannover, Germany
Human factor VIII 52
Korea Institute of
Science and
Technology, Taejon
Human granulocyte
colony-stimulating factor
53
Nexia human butyrylcholinester-
ase, spider silk
54, 55
Pharming C1 esterase inhibitor,
fibrinogen, collagen I &
II, Lactoferrin, Factor VII,
Factor IX
5658
PPL Therapeutics Bile human gastric Lipase,
Fibrinogen, thrombin,
Factor VII, Factor IX, al-
pha-antitrypsin, calcitonin
(salmon), collagen, super-
oxide dismutase, Glucagon
lipopeptide, Human se-
rum albumin, Protein C
5963
Production
strategy
Material used
for extraction
Host
organism
Virginia Tech,
Blacksburg, VA
American Red Cross
Human protein C 64, 65
Virtanen Institute,
University of Kuopio
Finland
Human granulocyte-
macrophage colony-
stimulating factor, human
erythropoietin
66
Blood Cow
Rabbit
Hematech Polyclonal antibodies 67, 68
Therapeutic Human
Polyclonal Inc.
Polyclonal antibodies 69
Semen Pig TGN Biotech Human follicle-
stimulating hormone
70
Urine Mice Catholic University
of Korea, Seoul
Human granulocyte-
macrophage colony-
stimulating factor
71
NYU/USDA/U.
Vermont
Human growth hormone 72, 73
Whole animal Caterpillars
Shrimp
Advanced
Bionutrition
Not available 74
Chesapeake PERL Not available 75
Organs/Cells Pig
Cow
Advanced Cell
Technology
Cell transplantation 76
Nextran (Baxter) Organ xenotransplantation 77
Eggs Chicken Avigenics Interferon, antibodies 78, 79
BioAgri alpha-1 antitrypsin,
biogeneric
8082
GeneWorks Inc. Human growth factor,
antibodies
83
GenWay Biotech Antibodies 84
Origen Therapeutics Not available 85
TransGenRx Proinsulin 86
TransXenoGen Anti-Neoplastic Urinary
Protein, Insulin, Human
Serum Albumin
87
Viragen Vaccines 88
Vivalis Vaccines 89, 90
Production
strategy
Material used
for extraction
Host
organism
Open or
contained
growth in
greenhouses or
underground
cultivation
Foliage Alfalfa
Potato
Tobacco
Melon
Brassica carinata
Brassica napus
Lettuce
Sunflower
Turnip
Agriculture and
Agri-Food Canada
IL-10 91
Battelle, Pacific
Northwest National
Laboratories
hEGF, Factor VIII, IX,
XIII, Thrombin
9298
Boyce Thompson In-
stitute/Texas A&M/
Axis Genetics
hepatitis B surface anti-
gen, enterotoxigenic E. coli
fusion protein, Norwalk
virus antigen
99103
Center for Genetic
Engineering and
Biotechnology,
Havana, Cuba
sFv Anti-Hepatitis B virus
surface antigen, coat
protein potato leaf roll
virus
104, 105
CICV, INTA- Buenos
Aires, Argentina
INIA, Madrid, Spain
Structural protein VP1 of
foot-and-mouth disease
virus, spike protein from
swine-transmissible
gastroenteritis coronavirus
106110
Chlorogen Plastid accumulation of
vaccines, proinsulin,
antibodies, human serum
albumin
111114
Chonbuk National
University, Jeonju
Korea
Plastid accumulation of B
subunit of E. coli
enterotoxin
115
Cobeto Human Intrinsic Factor,
human transcobalamin
116
Copenhagen
University, Denmark
Monoclonal antibody 117
ENEA, Rome, Italy Antibodies 118, 119
EpiCyte/Scripps Re-
search Institute
Antibodies, secretory
antibodies
120124
ERA Plantech Calcitonin 125
Farmacule Not available 126, 127
Fraunhofer Antibodies, vaccines 128130
Friederich Miescher-
Institut, Basel,
Switzerland
Human interferon 131
Gent University Antibodies 132137
Hebrew University of
Jerusalem
Interferon beta 138, 139
Hokkaido University,
Japan
Human interferon-al-
pha2b, IL 8, Human tu-
mor necrosis factor
140, 141
Production
strategy
Material used
for extraction
Host
organism
Icon Interferon-a, b,
somatotropin
Restriction enzyme
Single-chain antibodies
Monoclonal antibodies
Antigens, Glucocerebrosi-
dase, Thaumatin. Albumin
DNAse, RNAse inhibitor,
Insulin
142145
Institute of Plant
Genetics and
Cultivated Plant Re-
search in Gatersleben
Human papillomavirus
(HPV), Type 16 virus-like
particles, spider silk,
single chain antibodies
146151
Jefferson Medical
College, Philadelphia
Antibodies 438
KIST, S. Korea IL-6 152
Kyoto University,
Japan
Erythropoietin 153, 154
Medicago IgGs, Thrombin,
Aprotinin, tPa, Superoxide
dismutase, Protease
inhibitor, Collagen frag-
ment, Enzyme for CO
2
solution, hemoglobin
155158
Meristem Gastric lipase, Human
serum albumin, Lactofer-
rin, collagen, MAbs,
hemoglobin, Beta-
interferon
159168
Mogen International Human serum albumin 169
Monsanto/
Agracetus
Monoclonal antibodies,
single chain antibodies,
Human growth hormone
(plastid), collagen
170177
Monash University
Victoria, Australia
Measles virus
hemagglutinin protein
178180
National Institute of
Agrobiological
Tsukuba and Ibaraki,
Japan
Lactoferrin, lactoalbumin,
human epidermal growth
factor
181183
Nexgen/Guardian Not available 184
North Carolina State
University
Canine oral papilloma-
virus protein
185
Phylogix Lectin-based proteins 186, 187
Planet Biotechnology Secretory antibodies
CaroRX, RhinoRX,
DoxoRX
188194
Production
strategy
Material used
for extraction
Host
organism
Plantigen GAD (glutamic acid de-
carboxylase) and cytokines,
Interleukin-10, Interleu-
kin-4, MHC (major histo-
compatibility complex)
and cytokines
91, 195
199
Roswell Park
Institute, Buffalo,
New York
Potato vaccine booster
with injected hepatitis B
vaccine
200
St. George London
John Innes Centre
Vaccines, secretory
antibodies
201
Universit degli Studi
di Verona, Italy
Diabetes-inducing auto-
antigen glutamic acid
decarboxylase
202
University of Guelph,
Canada
Porcine epidermal growth
factor, Swine Viral Epitope
Fusion
203, 204
University of Ken-
tucky
Engineered Antimicrobial
Peptides
205, 206
University of Milan,
Italy
E. coli toxin B subunit
tuberculosis antigen
207209
University of Western
Ontario
Diabetes-inducing auto-
antigen glutamic acid
decarboxylase
196
UTS Biotech, Rome,
Italy
Human papillomavirus 16
E7 protein
210, 211
Wageningen
University,
The Netherlands
Antibody subunits,
glycosylation research
212216
York University,
Toronto
HIV antigen 217
Foliage infected
by recombinant
virus
Tobacco
Brassica
Agrenvec Not available 218220
CNR, Turin, Italy Single-chain Fv antibody
fragment
221
Fraunhofer Vaccines, antibodies 222230
Icon Genetics Interferon-a, b, Somatotro-
pin, Restriction enzyme,
Single-chain antibodies,
Antigens
142, 144,
233
Large Scale Biology Antigen from cancer cells
as personalized cancer
vaccines, Aprotinin
Alpha-galactosidase,
Hematopoietic factors,
lysosomal acid lipase
234238,
239
Production
strategy
Material used
for extraction
Host
organism
Tuber Potato
Carrot
Arizona State
University/Boyce
Thompson I.
Human papillomavirus
like particles, Norwalk
virus antigen
101, 240
Battelle, Pacific
Northwest National
Laboratories
hEGF, Factor VIII, IX,
XIII, Thrombin
92, 94
98, 241
Institute of Agrobio-
technology CSIC,
Pamplona, Spain
Human serum albumin 242
Institute of Plant
Genetics and
Cultivated Plant Re-
search in Gatersleben
Spider silk, viral particles 146, 147
Loma Linda
University
Lactoferrin, diabetes-in-
ducing autoantigen fusion
proteins of insulin and
glutamic acid decarboxy-
lase to cholera toxin
243247
MPB Cologne scFv antibodies 248, 249
New Zealands Crop
and Food Research
Atrial natriuretic factor 250
Novoplant Oral-delivered MAbs 251
Planton (Kiel) Human antimicrobial
proteins
252
Stems Sugarcane
Rubber Tree
Rubber Research In-
stitute of Malaysia
Human Serum Albumin 253255
Texas A&M/Procane Collagen 256
Fruit Banana
Tomato
Melon
Arizona State
Boyce Thompson I.
E. coli endotoxin fusion
protein
102, 257
University of
Colorado, Boulder
Respiratory syncytial virus
fusion protein
258
University of Delhi,
New Delhi, India
Cholera toxin B subunit 259, 260
ViroGene Vaccines 261
Production
strategy
Material used
for extraction
Host
organism
Seed Barley
Phaseolus v.
Corn
Rice
Sawflower-
derived Oil
Brassica napus
derived oil
Peas
Soybean
Tobacco
Cropdesign Not available 262
Dow Antibodies, peptides 263
Epicyte Anti-herpes and anti-
sperm secretory antibodies
for topical gels
120, 121,
194, 264,
265
Fraunhofer Antibody fragments 266, 267
Gent University Enkephalins 268, 269
Helsinki University Gelatin 270
Institute of Plant
Genetics and Culti-
vated Plant Research
in Gatersleben
Antibodies 148
Iowa State University E. coli enterotoxin B sub-
unit, porcine alpha-
lactalbumin
271, 272
Lethbridge Research
Centre, Alberta,
Canada
Bovine virus protein 273
Maltagene Not available 274
Meristem Gastric Lipase, Human
Serum Albumin, Lactofer-
rin
159161,
163165,
275
Monsanto Protein
Technologies/
Agracetus/Calgene
human growth hormone
171176,
276
Novoplant Oral-delivered MAbs 277
Orf Genetics GM-CSF, Interleukin-3,
Stem cell factor, Erythro-
poietin, Beta-interferon
278
ProdiGene Beta-glucuronidase, avidin,
trypsin, vaccines,
aprotinin, laccase
23, 24,
279291
Saint George London
John Innes Centre
Single chain antibodies,
secretory antibodies
267, 292,
293
Sembiosys (oil) Insulin, ApoA1, hirudin,
somatotropin
294297
Sungene Not available 12, 298
Syngenta Not available 299
Universidade de Sao
Paulo, Brazil
Human Growth Hormone 300
University of Ottawa Human insulin-like
growth factor, Human
granulocyte-macrophage
colony stimulating factor,
glycoprotein B from hu-
man cytomegalovirus
27, 301
306
Production
strategy
Material used
for extraction
Host
organism
Ventria Alpha antitrypsin,
Lactoferrin, lysozyme
307314
Washington State
University
Human lactoferrin,
lysozyme, gelatin
315, 316
Transgenic
aquatic plants,
cells in liquid
suspension, or
tissues express-
ing recombinant
protein in a
bioreactor
Growth media
or harvested
cells/tissue
Moss Greenovation Monoclonal antibody 317
Duckweed
(Lamna)
Biolex Plasmin, Human growth
hormone, Monoclonal
antibodies, alpha-
interferon
318
Algae Phycotransgenics Viral antigens 319
Chlamydomo-
nas reinhardti
Scripps Research
Institute/Rincon
Pharmaceuticals
Monoclonal antibodies 232, 320,
321
Roots Phytomedics Human placental alkaline
phosphatase
322324
Mechanically
injured-induced
promoter for
foliage secretion
CropTech Glucocerebrosidase, alpha-
iduronidase, serum
proteins and monoclonal
antibodies, vaccines
325328
Germinated oil
seed
UniCrop Monoclonal antibodies 329, 330
Plant cell
culture
Flanders University Antibodies 331
Fraunhofer Antibody fragments 310
John Innes Centre,
Norwich, UK
Antibodies 310
National Institute of
Public Health, Tokyo,
Japan
Human monoclonal
antibody anti-hepatitis B
virus surface antigen
332
Phytoprotein Vaccines 333
Protalix (Metabogal) Glucocerebrosidase,
334
ROOTec Not available 335
University New
South Wales,
Sydney, Australia
Antibodies 336338
used as biopharmaceuticals, including
monoclonal antibodies, hormones, en-
zymes, inhibitors, fusion proteins, vac-
cines, and structural proteins that have
been successfully expressed in crop plants
and farm animals such as corn, tobacco,
alfalfa, tomato, potato, barley, rice, cows,
goats, pigs, and chicken.
In this review we will first discuss why
transgenic technology is being considered
for the production of biopharmaceuticals,
and then focus on commercialized prod-
ucts derived from transgenic systems and
on those products with clinical experience
to illustrate the potential of the technology
to impact the healthcare industry. There
are many publications and conference pre-
sentations describing the production tech-
nologies available for the manufacture of
recombinant human proteins in transgenic
systems and criteria for selection of these
production systems [731, 233]. Some ex-
cellent reviews by Knblein also compare
transgenic plant systems on the basis of
recent data [3235]. The main point to re-
member is that for each particular bio-
pharmaceutical the production technolo-
gies must be carefully analyzed in order to
match specific quality, amount, marketing,
regulatory, level of containment, and sell-
ing price requirements related with each
product, and its medical use. Transgenic
systems meeting provide new opportunities
to commercialize some biopharmaceuticals,
and to make others more accessible to
healthcare providers.
5.2
Advantages and Disadvantages
of Transgenic Systems for the Production
Today, transgenic systems are being con-
sidered for the production of many diverse
biopharmaceuticals, including monoclonal
antibodies, hormones, therapeutic en-
zymes, structural proteins, and vaccines.
Table 5.1 illustrates that there are about
120 academic institutions and commercial
enterprises considering transgenic systems
as attractive production technologies for
the commercialization of more than 130
biopharmaceuticals.
There are many reasons why transgenics
should be considered for biopharmaceuti-
cal production. First, crop plants and farm
animals are capable of producing human
proteins with similar complex post-transla-
tional modifications, folding and assembly
as native human proteins. Biochemical
and structural equivalency with human
proteins facilitates the development, regu-
latory approval, and commercialization of
recombinant-derived biopharmaceuticals,
thereby increasing the probability of ob-
taining satisfactory human-like safety and
efficacy profiles in clinical trials. Second,
transgenic systems provide improved ma-
terial traceability and source reproducibili-
ty compared with what is available for bio-
logics obtained from natural sources such
as human blood or from animals by-prod-
ucts. The use of transgenic plants provides
improved safety by avoiding animal-de-
rived pathogens and immunogenic con-
taminants. Third, transgenics offer a reli-
able and cost-advantageous alternative to
mammalian cell culture and eukaryotic
microbial systems, particularly for biophar-
maceuticals required in large quantities
(more than tens of metric tons) and at low
cost (<$5 g
1
). Significant operational and
capital cost savings can result from repla-
cing the bioreactor-based biosynthesis step
with farming [284]. The downstream recov-
ery/purification facility cost remains un-
changed or may be lower with transgenics
as feed streams especially in the case of
milk, eggs, and seeds are more reprodu-
5.2 Advantages and Disadvantages of Transgenic Systems for the Production of Biopharmaceuticals 845
cible and may contain fewer contaminat-
ing proteins than broth obtained from cell
culture or microbial fermentation pro-
cesses. It has been reported that the pro-
duction cost for a recombinant protein at
1 metric ton scale is in the range of $20 to
$40 per gram for transgenic animals, and
$10 to $20 per gram for transgenic plants
[339]. The same report estimates that, at
the same scale, bioreactor-based systems
using yeast results in production costs in
the range of $50 to $100 per gram, while
using mammalian systems costs are in the
$3005000 per gram range. Bacterial-based
production systems can deliver recombi-
nant proteins in the $15 per gram cost
range at 1 metric ton scale because of the
high volumetric productivity of these sys-
tems and the use of large bioreactors.
Plants and animals provide a readily scal-
able production system based on accepted
agronomic and farming practices that re-
quire minimal capital investment. Manu-
facturing capital avoidance enabled by the
use of transgenic systems could enable
companies with limited funding to retain
their independence from large companies
to commercialize biopharmaceuticals al-
lowing them to direct their resources to
support clinical trials rather than building
manufacturing facilities. The impact on ca-
pacity of different transgenic systems
based on optimal productivity reported in
the literature is shown in Table 5.2. As
these requirements are based on optimal
productivity levels, capacity requirements
can be significantly higher at early devel-
opment stages or with particular recombi-
nant proteins having lower accumulation
levels than the optimal productivity re-
ported for these systems.
Transgenic systems provide enabling
technology for the commercialization of
some biopharmaceuticals that cannot be
produced efficiently by mammalian cell
culture or by microbial fermentation due
to the complex multimeric assembly, or
proteolysis sensitivity of these products.
Transgenic systems also provide enabling
technology for biopharmaceuticals to be
delivered in edible form in the expression
matrix, usually of plant origin, allowing
their commercialization without any addi-
tional purification. Edible vaccines ex-
pressed in plants are examples of the ap-
plication of this cost-effective drug delivery
option. If desired, the protein in the accu-
mulation matrix can be stabilized by
freeze-drying the plant or the animal tis-
sue containing the desired product. This
results in biopharmaceuticals that do not
require cold-chain storage, and is a signifi-
cant advantage for the successful introduc-
tion of biopharmaceuticals into developing
Table 5.2 Comparison of capacity of production systems based
on highest reported crude protein expression levels (acreage/ani-
mals/bioreactor capacity adjusted to 50% purification yield)
Production system Optimal productivity Requirement for 1 metric ton per yr
Potato foliage 250 kg protein acre yr
1
[29] 8 acres
Alfalfa/tobacco foliage 40 kg protein acre yr
1
[29] 50 acres
Rice/barley seed 40 kg protein acre yr
1
[314] 50 acres
Chicken egg 40 g protein hen yr
1
[31] 50000 hens, 14 million eggs
Goat milk 12 kg protein goat yr
1
[340] 160 goats, 87000 L milk (@20 g L
1
)
Mammalian cell culture 900 kg/15000 L yr
1
2 bioreactors (20 runs/year @ 3 g L
1
)
Microbial fermentation 2400 kg/40000 L yr
1
1 bioreactor (20 runs/year @ 3 g L
1
)
regions of the world. In addition, the raw
materials used to produce biopharmaceuti-
cals namely milk, eggs, and seeds have
well-characterized and reproducible com-
positions that facilitate downstream pro-
cess design and performance. Usually,
plant extracts contain few proteins that are
similar to human-derived proteins. For ex-
ample, eggs contain only 12 proteins, com-
pared with over 20000 in traditional fer-
mentation systems.
Manufacturing flexibility is another im-
portant advantage of using transgenic sys-
tems. The long term storage stability (e.g.,
seeds) can uncouples biological production
from downstream processing. Costly bior-
eactor-based facilities need to be operated
at maximum capacity to be profitable. On
occasion, there has been a demand to
build new facilities simply to meet pro-
jected peak demands which sometimes
do not materialize. The multi-ton scale
production capability enabled by trans-
genics can be quickly adjusted to meet
lower market demands, as producing the
material in milk or seeds is very inexpen-
sive. Grains and oil seeds are readily
stored for long periods, without refrigera-
tion, and without loss of protein activity as
a result of degradation or enzyme hydroly-
sis. Obtaining more animals or land to
plant crops in order to meet peak-projected
demand is relatively simple and inexpen-
sive. For milk, eggs, cereal grains (e.g.,
corn, rice, barley), for oilseeds (e.g., soy-
beans, canola, safflower), for some foliage
(e.g., tobacco, alfalfa), and for tubers there
is an efficient commercial infrastructure
for their growth, harvest, transportation
and storage, and this can easily be adapted
to the production of biopharmaceuticals al-
most to any volume.
However, there is no ideal production
technology suitable for all biopharmaceuti-
cals. The authors qualitative ranking of
transgenic systems relative to conventional
production systems for biopharmaceuticals
microbial fermentation, baculovirus-in-
duced insect cell culture, and mammalian
cell culture is listed in Table 5.3.
Table 5.3 illustrates three categories of
parameters comparing transgenic systems
with other recombinant protein production
systems. The first category (scalability,
cost, product storage and stability) illus-
trates the significant advantages of trans-
genic systems already discussed above.
The second category (endogenous patho-
genicity, accumulation level, biosynthesis)
illustrates that animals and plants offer
different advantages and benefits that
must be carefully considered based on the
product to be made, its use, required pur-
ity, demand, targeted cost, and other com-
mercial considerations. The third category
(speed for system selection and develop-
ment, regulatory risk, containment) shows
the potential disadvantages of transgenic
systems when compared with bioreactor-
based production technologies.
5.2.1
Cost and Capacity Advantages
During the past five years, the cost and ca-
pacity advantage of transgenic systems
compared with mammalian cell culture
has been reduced by significant recent im-
provements in the productivity of cell cul-
ture systems and improved culture devel-
opment technologies [341]. Increased pro-
ductivity, coupled with increased capacity
derived from the construction of several
large facilities, will increase the annual
mammalian cell culture the estimated ca-
pacity to 14000 kg by 2006 at a targeted
production cost below $200 g
1
. Currently,
most commercially available human thera-
peutic proteins sell at prices higher than
$10000 g
1
, whilst the annual requirement
is <10 kg. Very few biopharmaceuticals
have a market demand exceeding
100 kg yr
1
or require a selling price
<$50 g
1
when transgenic systems become
attractive.
5.2.2
Development and Implementation Times
Most transgenic systems require consider-
ably longer development and implementa-
tion times, and this results in a need for
higher levels of development resources
compared to other production systems.
The selection and optimization activities
for high productivity, stable, and fertile
founder animals or master seeds require
years rather than months, as is the case
for other production systems. The low rate
of live birth for transgenic animals trans-
formed using microinjection technology,
combined with a low rate of trait transfer
to the offspring, limits the attractiveness
of using transgenic animals for biophar-
maceutical production. The possibility that
the target protein might leak into the
blood circulation can result in physiologi-
cal perturbations and poor growth perfor-
mance. For plants, the low rate of healthy
generation after transformation results in
a long development and productivity opti-
mization timeline. As mentioned earlier,
microbial and mammalian cell culture sys-
tems are improving productivity, cost com-
petitiveness, and the time advantage over
transgenics. The long development time
and relatively high cost required for trans-
genic animal development is mitigated by
the use of tissue-directed synthesis based
on high expression promoters (e.g., casein
for milk expression); this allows the high-
est protein accumulation ever seen in a re-
combinant protein production system,
without affecting the overall physiological
state of the productive organism. This is
usually not possible in cell culture, and
Table 5.3 Qualitative ranking to illustrate strengths and weakness
of recombinant expression systems
Parameter Worst ? ? ? Best
Scalability >100 kg yr
1
products
Mammalian Insect cells Bacteria Yeast Animals Plants
Cost for <$50 g
1
products
Mammalian Insect cells Yeast Bacteria Animals Plants
Product storage stability
Mammalian Insect cells Yeast Animals Bacteria Plants (seeds)
Endogenous pathogenicity
Animals Mammalian Insect cells Yeast Bacteria Plants
Accumulation level
Plants Mammalian Insect cells Yeast Bacteria Animals (milk)
Folding and processing
Bacteria Yeast Plants Insect cells Animals Mammalian
Speed to production organism selection and development
Animals Plants Yeast Mammalian Insect cells Bacteria
Regulatory risk
Plants Animals Insect cells Yeast Mammalian Bacteria
Containment
Plants Animals Yeast Bacteria Insect cells Mammalian
microbial systems that can be physiologi-
cally perturbed by high heterologous pro-
tein accumulation levels. The average ac-
cumulation of biopharmaceuticals seen in
milk is 210 g of product per liter, com-
pared with 0.22 g L
1
accumulation that
is typically seen in microbial and mamma-
lian systems. The implementation of nu-
clear transplantation leads to a more rapid
development of founder animals. Some
animals mature quickly. For example,
chickens mature in 27 weeks, compared to
48 weeks for first-generation corn seed, 64
weeks for goats, and 140 weeks for cows
[86]. Rabbits have short gestation (1
month) and sexual maturity times, and
typically a rabbit can have eight to ten lit-
ters each year, thereby allowing for a quick
production scale-up [38]. There is no
known prion disease in rabbit, and no
known serious disease transmission to hu-
mans, which makes the rabbit safer than
other milk-producing animals [70]. Reports
have been made that the expression of
some human recombinant proteins in
milk can induce a lactational phenotype
resulting in the abnormal morphological,
biochemical, and functional differentiation
of mammary gland cells [342]. This sug-
gests that, especially in animals, it is im-
portant to monitor the effect of the ex-
pressed protein on the host organism. The
rapid reproduction cycle of hogs allows
production scale-up for biopharmaceuticals
from seminal fluid in 23 months. The use
of seminal fluid offers a closed secretory
system, so that the host is unaffected by
the proteins it produces [70]. For plants,
the use of fast-growing aquatic plants and
viral induction can deliver products as fast
as or faster than mammalian cell cul-
ture or microbial systems. The Lemna
aquatic plant system allows the generation
of transformed plants in 6 weeks, and has
the fastest growth rate of all higher plants,
doubling its biomass every 1.5 days [318].
Lemna grows in aqueous solutions of sim-
ple inexpensive mineral salts and requires
minimal energy inputs. Biopharmaceutical
productivity is enhanced by the high pro-
tein content (35% based on a dry weight
basis) and their ability to secrete biologi-
cally active, recombinant protein to the in-
organic media surrounding the aquatic
plants.
5.2.3
Post-translational Modification
The lack of identical human-like post-
translational modification (glycosylation,
sulfation, phosphorylation) constitutes an
obstacle to implement the use of trans-
genic systems for the production of bio-
pharmaceuticals. Transgenic systems, as
with all other recombinant production sys-
tems, are not capable of identical human-
like post-translational modification, specifi-
cally glycosylation [216, 343]. For example,
human antithrombin III, as isolated from
blood plasma, contains mainly biantennary
disialylated glycosylation. The same recom-
binant human protein produced in goat
milk contains a mixture of biantennary
mono- and disialylated glycosylation with
low levels of altered fucosylation, N-glyco-
lylneuraminic acid, N-acetylgalactosamine
for galactose substitution, terminal galac-
tose, and high-mannose structures. Hu-
man proteins expressed in plants can con-
tain a sugar not found in mammalian sys-
tems (xylose), a different fucose linkage,
terminal N-acetylglucosamine, and plants
are not capable of adding terminal galac-
tose or sialic acid [153]. A comparison of
the structures of the N-linked glycans at-
tached to the heavy chains of a tobacco-de-
rived monoclonal antibody with the corre-
sponding antibody of murine origin indi-
cated that all glycosylation sites are N-gly-
cosylated as in mouse, but the number of
glycoforms was higher in the plant than in
the mammalian expression system [343].
In addition to high-mannose-type N-gly-
cans, 60% of the oligosaccharides that are
N-linked to the plant-derived antibody have
b(1,2)-xylose and a(1,3)-fucose residues
linked to the core Man
3
GlcNAc
2
. Differ-
ences in glycosylation can alter the func-
tion, stability, pharmacokinetics, immuno-
genicity, protease sensitivity, and consis-
tency of biopharmaceuticals, and this re-
sults in the need for extensive bio-
equivalency studies if human-derived prod-
ucts were used originally for therapeutic
indications [344, 345]. The presence of
plant-specific oligosaccharide structures
may not be a limitation to the use of
plant-derived antibodies for topical immu-
notherapy, as plant-derived antibodies were
shown to be biologically active. However,
their immunogenic potential may raise
concerns for systemic applications of
plant-glycosylated antibodies in humans.
Variations in immunogenicity are very dif-
ficult to detect, as animal models are not
predictive of immunogenicity in humans.
The induction of antibodies against the
biopharmaceutical can limit its effective-
ness, alter its pharmacokinetics perfor-
mance, and may also lead to the genera-
tion of antibodies against the native hu-
man protein, resulting in the loss of a crit-
ical human physiological function. One
study indicated that having plant-like gly-
cosylation in a tobacco-derived murine
monoclonal antibody did not elicit an
immunogenic response in mice [346].
Methionine oxidation, tryptophan photo-
oxidation, aspartate isomerization, aspara-
gine and glutamine deamidation, proteoly-
sis, and protein aggregation are other
modifications that are influenced by the
production system and downstream pro-
cessing. Tissue-directed genetic manipula-
tion of plants and animals to clone human
genes responsible for post-translational
modifications can result in the develop-
ment of host organisms designed to con-
duct human-like post-translational modifi-
cations. Research has demonstrated the in-
duction of human-like glycosylation in
plants (addition of terminal galactose) and
human-like proline hydroxylation for col-
lagen in plants and animals by expression
of the corresponding human modification
enzyme [212]. These results illustrate the
possibility of having mammalian-like post-
translational modifications in transgenic
systems in a near future (see Part IV,
Chapter 7).
5.2.4
Regulatory Experience
The lack of regulatory experience since
no product derived from a transgenic sys-
tem has been approved for human use
is another significant limitation for trans-
genic systems to gain acceptance as a pro-
duction technology for biopharmaceuticals,
although as discussed above this will
change very soon if antithrombin from
goat milk is approved. The eventual
approval by regulatory agencies (one has
already been submitted for approval) of
products from transgenic systems will be
facilitated by the familiarity of the regula-
tory agencies and of the public with the
safe use of crop plants and farm animals
and their derived products (milk, eggs,
seeds) used for biopharmaceutical accumu-
lation. Crop plants and farm animals have
genetic and breeding properties that are
well-understood, and their derived prod-
ucts have well-characterized inherent tox-
ins, anti-nutrients, and exogenous con-
taminant profiles. Technologies to be em-
ployed for producing biopharmaceuticals
from transgenic systems will not differ
from current agricultural, farming, and
biopharmaceutical production practices.
Dedicated, custom-designed equipment
and facilities will be developed with proce-
dures equivalent to the Good Manufactur-
ing Practices currently in place for the
production of biopharmaceuticals. Once
the protein is extracted from the trans-
genic tissue, similar purification technolo-
gies as those used for cell culture and mi-
crobial recombinant systems can be ap-
plied. In many cases, the host-contaminat-
ing proteins and toxic metabolites such as
endotoxins are present at lower concentra-
tion and diversity in transgenic systems
compared to cell culture, non-transgenic
tissues, or microbial production systems.
Transgenic animal production will be con-
ducted minimizing virus- and prion-trans-
fer risks, providing a safer alternative to
the extraction of these products from hu-
man or non-transgenic animal tissues.
The use of plants will practically eliminate
this risk. The FDA, in collaboration with
the USDA/APHIS (US Department of
Agriculture/Animal and Plants Health In-
spection Service), published the GUID-
ANCE FOR INDUSTRY: Drugs, Biolog-
ics, and Medical Devices Derived From
Bioengineering Plants For Use in Humans
and Animals containing suggested norms
for field growth of plants and their proper
handling and transportation [347]. A simi-
lar document is available for transgenic
animals [348].
5.2.5
Containment
Containment is another critical issue,
especially when using crop plants such as
corn that can transfer its genes to other re-
lated plants via pollen or insect-pollinating
crops (e.g., alfalfa). The cost of procedures
associated with containment together
with subsequent testing and monitoring
and the need for dedicated equipment,
waste disposal, training, security, regula-
tory documentation and liability insurance
for the potential of mingling the trans-
genic-derived product with food products
can significantly reduce the cost advantage
of transgenics compared with other sys-
tems. Increased production costs combine
with the perception of risk could hinder
the implementation of transgenic systems.
The possibility of contamination of the
food supply with transgenic-derived bio-
pharmaceuticals results in the opposition
to the use of transgenic systems for bio-
pharmaceutical production from the food
industry, public advocates, and environ-
mentalists concerned about potential
health and safety risks of transgenic-de-
rived products in the environment. Animal
activists oppose the use of animals for
manufacturing as many animals are de-
stroyed as part of the selection process.
There have been recent incidents of corn
contamination of a cloned insecticidal pro-
tein not approved for human consump-
tion, and of soy seeds with corn residues
possibly containing a biopharmaceutical.
The food production industry, the public,
the media, and government agencies re-
acted strongly and negatively to these inci-
dents, even when the contamination could
not be detected or correlated with any
health threat. Industry groups such as the
Grocery Manufacturers of America and
grain processors are asking that the use of
food crops for pharmaceutical production
should be banned, while others are sug-
gesting that crops grown to produce bio-
pharmaceuticals should only be grown in
regions where that crop is not grown for
food or feed [349]. This latter practice
could eliminate the attraction of trans-
genics, namely that existing infrastructure
can be used to achieve low cost and farm-
ing flexibility. In response, several compa-
nies are developing technology for produc-
tion in closed systems such as green-
houses and underground mines, or in
bioreactors using transgenic-derived tis-
sues. Gene transfer via pollen can also be
avoided by using chloroplast expression
systems, by growing male sterile plants, by
mechanical pollen removal, or by using
self-pollinating crops such as barley and
rice.
The initial relatively small number of
animals and acreage (see Table 5.2) re-
quired for biopharmaceutical production
facilitates the implementation of pollen
control, containment practices, quality as-
surance procedures, and monitoring sys-
tems for all aspects of animal growth and
crop production, harvesting, post-harvest
handling, storage, transportation, and bio-
processing [350].
5.3
Commercial Biopharmaceuticals with
Human Clinical Experience for Therapeutic,
Immunoprophylactic, and Medical Device
Use derived from Transgenic Systems
As shown in Table 5.1, there are over 130
of products for medical applications that
have been or are being considered as
candidates for production using transgenic
systems. In this section, we will discuss
those biopharmaceuticals that are cur-
rently commercially available for non-hu-
man use (plant-derived avidin, trypsin, and
aprotinin), and one product that has been
submitted for regulatory approval, namely
milk-derived antithrombin. Ten therapeutic
proteins derived from transgenic systems
that were (or are) included in human clini-
cal testing will be also discussed. These
products are being considered for the
treatment of inherited diseases (milk-de-
rived alpha-1-antitrypsin, milk-derived C1
inhibitor, corn-derived gastric lipase, milk-
derived human bile salt stimulated lipase,
milk-derived acid alpha-glucosidase), auto-
immune conditions (milk-derived alpha-fe-
toprotein), cancer (corn-derived monoclo-
nal antibody, tobacco-derived single chain
antibodies), and as anti-infective agents
(milk-derived lactoferrin, tobacco-derived
secretory antibodies). We will also discuss
the development of oral vaccines and re-
combinant structural proteins used for
medical devices. An understanding of
these products will provide a good perspec-
tive as to how agricultural biotechnology
companies and the pharmaceutical indus-
try are considering the implementation of
transgenic systems, and highlight the fu-
ture impact of this technology on health-
care systems.
5.3.1
Commercial Products Derived from
Transgenic Systems
The three commercial products available
for medical applications derived from
transgenic systems are all available from
the Sigma Chemical Company, and are de-
rived from plants.
5.3.1.1 Corn Seed-derived Recombinant
Chicken Egg White Avidin
to Replace Egg-derived Avidin
Avidin was the first commercially available
recombinant protein (1997) to be derived
from a transgenic system (corn seed) for
use in medical applications [285, 286, 291].
As it forms strong, non-covalent bonds
with biotin, avidin is used in medical and
biochemical diagnostic kits for the detec-
tion of biotin-containing proteins and nu-
cleic acids. Transgenic corn which accu-
mulated avidin was developed jointly by
ProdiGene, Pioneer Hi-Bred International
and the US Department of Agriculture
(USDA) with its first field trials reported
in 1993 [25, 285]. The product was devel-
oped to improve the quality and the con-
sistency of this reagent at a lower produc-
tion cost compared with egg-derived avi-
din. It was estimated that a single $2.50
bushel of corn yields the same amount of
avidin as 1 ton of eggs, which costs about
$1000 to produce. Avidin is available from
the Sigma Chemical Company (#A8706)
and is sold as a homogeneous protein of
consistent quality, providing better perfor-
mance and reproducibility compared to
avidin produced by extraction from egg
white. Avidin is purified from corn by af-
finity chromatography using a 2-iminobio-
tin agarose resin. The native source of avi-
din is avian, reptilian or amphibian egg
white. Avidin is naturally induced by pro-
gesterone, tissue trauma, toxic agents, and
infections. It is a glycoprotein with four
identical subunits each of 16 kDa and 128
amino acids. There are many avidin-based
analytical methods available to quantify
DNA and proteins in biological samples
[351].
Recombinant avidin expressed in food
or in feed grain can be used as a non-
selective biopesticide against a spectrum of
storage-generated insect pests [352]. Ex-
pressing avidin with other traits such as
herbicide resistance or co-production with
another industrial protein illustrates the
concept of gene stacking for developing
agricultural crops with higher value and
improved performance. Insect control ef-
fected by avidin results from the created
biotin deficiency that is toxic to pests. The
recombinant avidin in corn seed is not
toxic to mice when administered as the
sole component of their diet for 21 days.
Avidins insecticidal activity is different
from other transgenic alternatives such as
Bt and from conventional insecticides. As
avidin is a natural food protein found in
eggs, it may be easier to register than
other insecticide proteins that are not pres-
ent in the food supply [352]. Beta-glucuro-
nidase was also produced in corn seed by
ProdiGene for use as a diagnostic enzyme,
but it is not commercially available [285,
286].
5.3.1.2 Corn Seed-derived Recombinant
Human Trypsin to Replace Bovine
Pancreas-derived Trypsin
Trypsin was the first commercially recom-
binant enzyme for medical applications
that was produced from a transgenic sys-
tem (corn seed) and available in kilogram
quantities [280]. Trypsin is a pancreatic en-
zyme which catalyzes the breakdown of
proteins at specific sites as part of the di-
gestion process. The enzyme has a molec-
ular weight of 2328 kDa, and in protein
molecules is active only against peptide
bonds adjacent to arginine and lysine.
Trypsin is one of the most site-selective of
all commercially available proteolytic en-
zymes, which makes it an ideal regent for
the analysis and specific processing of pro-
teins during manufacture. The most im-
portant industrial and medical use of tryp-
sin is in the manufacture of insulin, as it
cleaves the precursor protein into an active
form by removing peptide C from proinsu-
lin. Another bioprocessing application is
for the cleavage of undesired antigens
from cells in both human and veterinary
vaccines manufacture. Trypsin is used in
wound care as debridement treatment to
reduce inflammatory edema, hematoma
and pain associated with internal and ex-
ternal wounds [353, 354]. Proteases, in-
cluding trypsin, are used in therapy for
autoimmune diseases such as Type I dia-
betes [355]. Trypsin can also be used as a
5.3 Commercial Biopharmaceuticals with Human Clinical Experience 853
food additive to baby formula, assisting
the digestion of difficult proteins and lead-
ing to improved absorption in the digestive
tract. These therapeutic and nutraceutical
applications could be expanded by the
availability of a low-cost, abundant, and
safe source of recombinant trypsin.
Corn seed-derived recombinant human
trypsin was developed as an animal com-
ponent-free alternative to the currently
available bovine pancreas-derived trypsin
for use in bioprocessing and as a cell cul-
ture media component for growth en-
hancement. Corn-derived recombinant
trypsin is not approved for human thera-
peutic use. Recombinant trypsin has been
produced by cell culture, bacteria and
yeast, but not in quantities sufficient to
meet the needs of biopharmaceutical man-
ufacturers requiring kilogram quantities at
relatively low cost [280]. The corn seed-de-
rived recombinant protease is functionally
equivalent to the native bovine pancreatic
trypsin used in most current applications
[286]. TrypZean
, developed by Prodi-
Gene, is available from Sigma Chemical
Company (#T3568) for use in the dissocia-
tion of cultured mammalian cells from
plastic-ware, thereby eliminating the intro-
duction of potential animal-derived con-
taminants found in bovine and porcine
trypsin (see also Part IV, Chapter 1). Tryp-
Zean allows the use of a consistent quality
product which delivers the same kinetics
for cell detachment as the current product;
this results in a need for minimal protocol
changes during cell preparation. Soybean
trypsin inhibitors and other inhibitors
have a similar inhibitory performance to
TrypZean as they do with native trypsins.
Recombinant trypsin is manufactured by
Sigma by utilizing ProdiGenes transgenic
corn protein expression system.
5.3.1.3 Cornseed and Viral-induced Tobacco
Foliage-derived Recombinant Bovine
Aprotinin to Replace Bovine Lung-
derived Aprotinin
Tobacco-produced recombinant research-
grade bovine aprotinin (Apronexin NP)
[238] is also available from Sigma Aldrich.
Aprotinin is a protease inhibitor which is
used as a research reagent in biomanufac-
turing for several therapeutic applications.
It has been traditionally extracted from bo-
vine lung tissue. Aprotinin is a single, 58
amino-acid polypeptide with three di-
sulfide bonds, and inhibits several serine
proteases such as trypsin, chymotrypsin,
plasmin, and kallikrein.
Bovine lung-derived aprotinin (Trasy-
lol
, Bayer) is used in therapeutic applica-

tions as a clotting agent; its mode of ac-
tion is to inhibit the degradation of fibri-
nogen by plasmin, thereby reducing blood
loss during heart surgery and the need for
blood transfusions. Aprotinin also protect
platelets from damage in circulation ma-
chines during surgery [356, 357]. It can be
used to treat acute pancreatitis [358]. Use
of bovine-derived aprotinin can cause se-
vere anaphylaxis in 0.51% of patients. In
patients who receive aprotinin a second
time this percentage is greater [359]. The
recombinant product should have an im-
proved safety profile when it becomes
available for human use.
Apronexin NP (Sigma, #A6103) is sold
as a research reagent, for cell culture, and
for other biomanufacturing applications.
Large Scale Biology uses its biomanufac-
turing facility and its Geneware gene ex-
pression technology for Apronexin NP pro-
duction based on the use of recombinant
tobacco mosaic viruses containing the bo-
vine aprotinin gene to express the recom-
binant protein in non-transgenic tobacco
plants. This process consists of growing
non-transgenic tobacco in open fields or
greenhouses, infecting the leaves with the
recombinant plant virus to induce expres-
sion of a aprotinin in the leaves, and trans-
porting the harvested leaves (fresh or fro-
zen) to a closed facility for protein extrac-
tion.
Recombinant aprotinin has been pre-
pared by yeast fermentation [360] and
from corn seed, developed by ProdiGene
[283] under the tradename AproliZean
.
Corn seed-derived aprotinin has been
grown in field trials since 1998. In addi-
tion to research and bioprocessing uses,
corn-grown aprotinin can be used as avi-
din as a built-in insecticide based on re-
combinant aprotinins trypsin-inhibiting
activity. For its recovery, recombinant apro-
tinin together with a native corn trypsin
inhibitor is purified using a trypsin-agar-
ose column. The corn inhibitor binds to
the column, but the recombinant human
aprotinin does not. As the corn inhibitor
can also be purified, co-production of valu-
able recombinant and non-recombinant
materials from the same crop will clearly
increase the value of a crop.
5.3.2
Products Derived from Transgenic Systems
Submitted for Regulatory Approval
Until now, only one product derived from
transgenic technology, namely goat milk-
derived recombinant human antithrombin
III, has been submitted for approval in Eu-
rope.
5.3.2.1 Goat Milk-derived Recombinant
Human Antithrombin III to Replace
Human Blood Plasma-derived
THROMBATE III
Human antithrombin III (ATryn
) was
the first therapeutic recombinant protein
to be produced using a transgenic system
(goat milk), tested in clinical trials (1994),
and submitted for approval to a regulatory
agency (Marketing Authorization Applica-
tion to European regulatory agencies) in
2004. GTC Biotherapeutics, the developer
of ATryn
, plans a market launch of the

product in Europe by mid-2005 (see Part
IV, Chapter 11).
Antithrombin III (AT-III), a single-chain
glycoprotein of 58 kDa and 480 amino
acids, is synthesized in the liver. It is a ser-
ine protease inhibitor, and acts as the most
important inhibitor in the coagulation cas-
cade to avoid blood clot formation. AT-III
inhibits a wide spectrum of serine pro-
teases including thrombin, factors IXa, Xa
and XIa, kallikrein, plasmin, urokinase,
C1-esterase, and trypsin. AT-III interacts
with heparin by binding to specific sul-
fated and non-sulfated monosaccharide
units on heparin. The binding of AT-III to
heparin enhances the inhibition of factors
IXa, Xa, and thrombin.
THROMBATE III
(Bayer) is a plasma-
derived product that is used to treat AT-III
deficiency, the first inherited trait discov-
ered which was identified in 1965. The
condition is associated with thrombophilia
caused by low levels of AT-III or the pres-
ence of altered AT-III activity. Both condi-
tions can result in excessive blood clotting.
Acquired AT-III deficiency is another con-
dition that occurs in situations with high
risk of thrombosis such as trauma, burns,
and sepsis. GTC Biotherapeutics con-
ducted clinical trials with ATryn
in high-
risk situations such as surgery or child de-
livery to prevent deep-vein thrombosis in
hereditary AT-III-deficient patients and
heparin-resistant patients with acquired
AT-III deficiency undergoing coronary by-
pass. A pharmacokinetic study in patients
with hereditary AT-III deficiency indicated
that the administration of the recombinant
product resulted in an increase in blood
AT-III activity which was similar to that
seen with the plasma-derived product.
Clinical use of ATryn
in patients who
had surgical procedures indicated that it
was well tolerated in all cases. Shortages
in THROMBATE III
supplier resulted in
the use of ATryn
on compassionate
grounds when the plasma product was not
available. No clinical evidence of thrombo-
sis has been reported in any of these com-
passionate-use patients. The routine use of
THROMBATE III
as replacement thera-
py is not generally recommended due to
the high cost, the risk of infection, and the
need for frequent intravenous administra-
tion. The recombinant product derived
from milk should have improved safety
and reduced cost, resulting in a product
that is suitable for chronic use.
ATryn
production in goat milk is the

only recombinant production system
shown able to deliver AT-III at costs com-
mercially acceptable. GTC has several goat
production lines accumulating human AT-
III in milk at 26 g L
1
concentration. Pur-
ification from milk includes tangential
flow filtration to remove casein, micelles,
fat globules, bacteria, somatic cells, and
viruses followed by a heparin-based af-
finity chromatography to remove DNA,
lactose, mineral salts, vitamin, hormones,
and most of the milk proteins. Two addi-
tional chromatographic steps anion-ex-
change to remove additional milk proteins
including lactoferrin and hydrophobic in-
teraction to remove goat AT-III and milk
proteins complete the process. From
300 L of milk, 300 g of human AT-III are
purified to 99.9% purity with an overall
55% recovery yield. The effectiveness of
the purification process to remove viruses
and prions was demonstrated, with the
high degree of purity of ATryn
being
further illustrated by the observation that
no immunogenic events were detected in
clinical trials after examining 180 subjects,
including healthy volunteers and patients
with hereditary or acquired AT-III deficien-
cy, at 1 day after administration and 28
days later.
In addition to ATryn
, GTC Biopharma-
ceuticals has several other products under
development in collaboration with Cento-
cor, Abbott, Elan, Bristol-Myers Squibbs,
Alexion, Progenics Pharmaceuticals, Im-
munoGen, and Merrimak Pharmaceuticals.
These collaborations and internal programs
involve the production of several monoclo-
nal antibodies, recombinant human serum
albumin, a malaria vaccine, fusion proteins,
and a protein for the treatment of autoim-
mune conditions. Some of the products that
are currently undergoing clinical trials and
will be discussed later [361].
The commercial success of ATryn
could stimulate the use of transgenic sys-

tems for the commercialization of other
plasma-derived biotherapeutics [27, 52].
Milk-derived antitrypsin and human C1 in-
hibitor currently in clinical trials will be
discussed later. The development of tobac-
co foliage-derived factor VIII and factor IX
for the treatment of hemophilia, and factor
XIII and thrombin for wound healing
therapies as clot promoters is currently in
progress at Battelle [9497]. For factor VIII
it was estimated that, based on the current
non-optimized process, the needs of an
average 80-kg patient could be satisfied
with 23.5 kg of tobacco foliage, or two
plants. It was estimated that fewer than
4000 plants, suitable for greenhouse
growth, would be needed to supply the
worldwide demand for factor VIII. Throm-
bin, currently derived from bovine blood,
is used to control bleeding associated with
surgical procedures. Each year thrombic is
used in over 500000 operations in the
United States. Bovine thrombin is fre-
quently applied topically during surgery to
achieve hemostasis. It also is used as a
component of commercial fibrinogen pre-
parations. The use of Bovine-derived
thrombin has been associated with the de-
velopment of antibodies that may cross-re-
act with human blood proteins [362].
These antibodies could lead to serious
bleeding complications. A transgenic prod-
uct will have both improved safety and re-
duced cost.
Another important plasma-derived prod-
uct considered for production using trans-
genics systems by several organizations
(see Table 5.1) is human serum albumin
(hSA). This $1.3-billion, 400-tonne (in
2000) product is currently produced from
donor human blood. Its role is to maintain
fluid balance in the blood and to transport
amino acids, fatty acids, and hormones
[169, 254, 363, 364]. Current clinical uses
of hSA include blood volume replacement
during shock, the treatment of serious
burns, emergency surgeries, and as an ex-
cipient to maintain structural stability and
activity in biological drug formulations. As
with AT-III, hSA is difficult to express in
conventional recombinant production sys-
tems at amounts and cost-competitively
with the current blood-derived product.
GTC Biotherapeutics produces hSA de-
rived from a transgenic system based in
milk expression. GTC is producing qualifi-
cation batches for evaluation by potential
clients for use as an excipient and to pre-
pare a Drug Master File (DMF) allowing
data related to rhSA to be used in the
manufacture of therapeutic products. GTC
considers that the commercialization of
rhSA is possible by 2007, and estimates
that the annual total production volume
required to meet the needs of the excipient
market will be 12 tonnes [40].
The success of ATryn
may also stimu-

late the use of transgenic technology for
similar replacement therapeutic enzymes
and protein-based inhibitors to treat inher-
ited or acquired conditions caused by re-
duced amounts or variant activity of a criti-
cal enzyme or inhibitor. In the next sec-
tion, we will discuss in detail five such
products (milk-derived alpha-1-antitrypsin,
milk-derived C1 inhibitor, corn-derived
gastric lipase, milk-derived human bile salt
stimulated lipase, and milk-derived acid al-
pha-glucosidase) with human clinical trial
experience. Transgenic technology is at-
tractive for the category of therapeutic re-
placement products that are used for life-
long enzyme/inhibitor replacement thera-
py in small numbers of patients who are
also in need of other costly supplemental
drugs and expensive medical services re-
lated to their condition. These replacement
enzymes and inhibitors are expensive be-
cause they tend to be complex proteins
which are difficult to prepare by recombi-
nant technology and available only in natu-
ral form from human tissue or mamma-
lian cell culture. The preparation of a com-
plex product in small amounts and at a
relatively low cost is difficult using conven-
tional recombinant systems, and this cre-
ates a good opportunity for transgenic sys-
tems. Another advantage for transgenic
producers is that these products can be
commercialized as orphan drugs because
the limited number of patients allows sig-
nificant market protection to the devel-
oper.
5.3.3
Products Derived from Transgenic Systems
with Human Clinical Experience
for Therapeutic Use
At least 10 proteins developed for thera-
peutic use have been derived from trans-
genic systems, and have been used in hu-
man clinical trials according to published
data. Five of these proteins (milk-derived
antitrypsin, milk-derived C1 inhibitor,
corn-derived gastric lipase, milk-derived
bile salt-stimulated lipase, milk-derived
acid alpha-glucosidase) were developed to
treat inherited diseases, while the others
were developed to treat autoimmune con-
ditions (milk-derived alpha-fetoprotein)
and cancer (corn-derived monoclonal anti-
body, tobacco-derived cancer cell antigens),
or to serve as anti-infective agents (milk-
derived lactoferrin, tobacco-derived secre-
tory antibodies).
5.3.3.1 Transgenic-derived
Biopharmaceuticals to Treat
Inherited Conditions
As noted earlier, this is a therapeutic cate-
gory with a significant need for products
that need to be accessible to a small group
of patients for life-long treatment at low
cost. Moreover, they must be free of hu-
man/animal-derived infectious agents and
of immunogenic animal components be-
cause of chronic use, ATryn
belongs to
this category.
Sheep milk-derived recombinant human
alpha-1-antitrypsin to replace human blood-
derived Prolastin
PPL Therapeutics, in
conjunction with Bayer [who manufacture
human plasma-derived alpha-1-antitrypsin
(AAT); Prolastin
], sought to develop
sheep milk-derived recombinant human
alpha-1-antitrypsin (rhAAT), also known as
alpha-1-protease inhibitor. AAT is a human
blood glycoprotein which is produced in
the liver and inhibits several proteases, in-
cluding neutrophil elastase, a lung enzyme
that digests phagocyte cells and bacteria,
thereby promoting tissue healing [365].
Without ATT, lung elastase is not properly
regulated and can destroy the lung tissue.
Many respiratory diseases, including cystic
fibrosis and chronic obstructive pulmonary
disease, occur as a result of the imbalance
of AAT and elastase in the lungs. AAT de-
ficiency is an inherited condition which
occurs in approximately 1 in 5000 live
births. Severe AAT deficiency (hereditary
emphysema) affects around 150000
200000 people in the US and Europe. Pro-
lastin
can be used to alleviate this condi-

tion, but its availability fluctuates as it is
produced by only one company. Indeed,
Prolastin
is the only available product for

chronic replacement of AAT in congenital
AAT-deficient patients with emphysema.
The efficient production of hrATT in
milk was achieved by PPL Therapeutics,
who reported the highest ever accumula-
tion levels for a recombinant product,
namely 35 g L
1
in mouse milk and
20 g L
1
in goat milk. PPL Therapeutics
had a flock of 4000 AAT-producing trans-
genic sheep in New Zealand, and subse-
quently conducted successful clinical trials
using sheep milk-derived rhATT in pa-
tients with cystic fibrosis. Data were also
available from Phase I/II clinical studies
in patients with congenital AAT deficiency.
Studies of the treatment of emphysema
and acute respiratory distress syndrome
were also planned. In 2000, PPL and Bayer
announced a placebo-controlled Phase III
efficacy study for the treatment of AAT de-
ficiency designed to commercialize an
aerosol formulation of rhAAT. However, in
2003, it was announced by PPL that the
rhAAT development program had been put
on hold because of financial risk related to
construction of the manufacturing facility.
There were also questions regarding the
purity of the AAT produced in sheep milk,
as well as some minor (but unexplained)
side effects that included mild coughing.
Overall, the suggestion was that this prod-
uct would be too expensive to develop [60].
The production of rhAAT in rice cell cul-
ture and rice plants was reported by Ven-
tria [311, 366] (see Part IV, Chapter 8). As
part of the program to produce human
proteins as nutritional supplements to in-
fant formula. Human milk contains rela-
tively high concentrations of AAT. The sta-
bility of rice-derived rhAAT was examined
by biochemical and functional assays such
as elastase and trypsin inhibition, follow-
ing exposure to heat, low pH, and in vitro
digestion. Native AAT can withstand in vi-
tro digestion modeled after conditions in
the infant gut. Studies show that rhAAT
may survive the conditions of the infant
stomach and duodenum and effect protein
digestion in the infant small intestine. An
rhAAT expression equivalent to 20% of the
total secreted proteins was achieved, and
purification yielded active rhAAT with pur-
ity greater than 95% [311].
In addition to the development of rhAAT
products based on transgenic technology,
clinical studies to investigate the use of
gene therapy based on a recombinant ade-
no-associated virus with the hAAT gene
had been proposed to treat AAT-deficient
human subjects [367]. Yeast-derived rhAAT
developed for aerosol administration to treat
AAT-deficient individuals was shown to be
safe and resulted in increased lung anti-
neutrophil elastase defenses [368].
Milk-derived recombinant human C1 inhibi-
tor Pharming is conducting human clini-
cal trials for recombinant human C1 in-
hibitor (rhC1I) in patients suffering hered-
itary angioedema (HAE), a condition
caused by human C1 inhibitor deficiency
[58]. Human C1 inhibitor is a 105 kDa pro-
tein which has a critical role in the inhibi-
tion of proteases involved in fibrinolytic
and complement pathways. Deficiency of
this inhibitor is manifested in acute swel-
ling of soft tissues in the hands, feet, limbs,
face, intestinal tract, mouth, or airway (lar-
ynx or trachea). If swelling closes the air-
way, it can be fatal. Attacks of swelling can
become more severe in late childhood and
adolescence. In the US and Europe, there
are some 22000 patients suffering from
HAE, acute attacks of which can be treated
by intravenous injection of C1 inhibitor pu-
rified from human blood plasma (C1I-Im-
muno; ImmunoAG, Austria). The condition
may be prevented in some instances by
treatment with attenuated androgenic male
hormones. As with many other blood plas-
ma products, the supply of human C1 in-
hibitor is limited and is prone to viral con-
tamination, while the use of male hor-
mones can cause severe adverse side effects.
The results from a Phase II clinical trial es-
tablished the efficacy, safety and tolerability
of rhC1I in the treatment of acute attacks of
HAE [58]. All patients treated with rhC1I
showed rapid onset of relief and time to
complete resolution of the attack, accompa-
nied with acceptable pharmacokinetic and
pharmacodynamic performance. After treat-
ment, patients were monitored for 90 days
to evaluate the safety of rhC1I. Treatment
with milk-derived rhC1I did not elicit any
allergic or clinically relevant immune re-
sponses. None of the patients treated with
milk-derived rhC1I in the clinical trial ex-
perienced relapse of the initial HAE attack.
Corn seed-derived dog gastric lipase as re-
placement of porcine-derived Pancrelipase
Corn seed-derived dog gastric lipase is un-
der development by Meristem Therapeu-
tics for the treatment of gastric lipase defi-
ciency (steatorrhea), neonatal deficient
pancreatic function, or when pancreatic
function is compromised by diseases such
as cystic fibrosis or chronic alcoholism
[159]. The absence of gastric lipase in stea-
torrhea patients prevents the digestion of
food lipids. Pre-term infants have difficulty
digesting lipids and, unless they are
breast-fed, have no access to this enzyme.
Cystic fibrosis a genetic disease which af-
fects approximately 30000 children and
adults in the US can lead to reduced
levels of gastric lipase. For all of these pa-
tients, oral delivery of active gastric lipase
can significantly improve the quality of life
and nutrition.
Gastric lipase could also be used as an
industrial enzyme for the digestion of fats
to make soaps, the tanning of leather, and
the preparation of animal feedstuffs.
Currently, gastric lipase is available for
oral delivery as porcine or bovine pancreat-
ic extract capsules or tablets (Pancrelipase;
Ortho/McNeil Pharmaceuticals, Axcan
Pharma, Organon, and Digestive Care).
The pancreatic enzyme concentrate is pre-
dominantly steapsin (pancreatic lipase),
but it also contains amylases and pro-
teases. These products are unpalatable and
have low activity; hence patients are re-
quired to take large numbers of tablets
daily (perhaps 3050), resulting in poor
compliance. These products are also un-
suitable for pre-term infants.
Meristem selected dog lipase (a 379-ami-
no acid enzyme with three cysteines) for
expression in tobacco foliage and corn
seed as it was the most active of the ani-
mal gastric lipases when tested using
long-chain triacylglycerols which represent
the major components of human dietary
fats [159, 166]. Dog lipase is a glycoprotein
with 13% carbohydrate content by weight.
The carbohydrate moiety is not required
for activity, but it does increase the solubil-
ity and availability of the enzyme. Meri-
stem produced dog gastric lipase using
vacuolar retention and secretion to obtain
active glycosylated enzyme. The secreted
enzyme had improved proteolytic matura-
tion compared with the vacuolar retained
enzyme, and this resulted in a product
with higher specific activity [166]. Corn
seed-derived gastric lipase has similar ac-
tivity, acid resistance, and acidic optimum
pH, as the native animal-derived enzyme.
The transgenic-produced lipase is more ac-
tive and more palatable than current lipase
products.
Corn seed-derived dog gastric lipase also
has improved quality and consistency, and
will be significantly cheaper than other pu-
rified gastric lipases. Meristem claims a
75% reduction in production costs, and
has already successfully manufactured
kilogram quantities of pharmaceutical-
grade gastric lipase for clinical trials. Mer-
istem can purify 350 g of gastric lipase per
week from 700 kg of corn seed in their
manufacturing facility, which is capable of
producing up to 20 kg of purified lipase
each year.
Other recombinant versions of gastric li-
pases are available from yeast, Pseudomo-
nas and filamentous fungi [369]. Sheep
milk-derived human bile salt-stimulated li-
pase (BSSL) an enzyme produced in the
pancreas and in human milk was devel-
oped by PPL Therapeutics for similar indi-
cations, and it is discussed below [63].
Sheep milk-derived human bile salt stimulat-
ed lipase to replace porcine-derived lipase
PPL Therapeutics reported positive results
of its Phase II clinical trial using trans-
genic BSSL to treat patients with pancrea-
tic insufficiency [63]. The milk-derived
product was equally as effective as
Creon
, the current market leader used to

improve digestion and restore fat absorp-
tion to normal levels. Creon
capsules are
administered orally and contain delayed-re-
lease microencapsulated porcine pancreatic
pancrelipase. Six patients with chronic
pancreatic insufficiency were successfully
treated, four patients having chronic pan-
creatitis and one suffering from cystic fi-
brosis. The larger commercial opportunity
for this product is for premature infants
who do not receive their mothers milk,
which naturally contains BSSL. PPL devel-
oped BSSL in collaboration with AstraZe-
neca.
Rabbit milk-derived human acid alpha-gluco-
sidase for the treatment of Pompe disease
Pharming conducted clinical trials to test
the long-term safety and efficacy of rabbit
milk-derived recombinant human alpha-
glucosidase (rhAGLU) used to treat a lyso-
somal storage disorder: Pompe disease
[370]. This disease, which has a frequency
of 1 in 40000, usually results in infant
death at a median age of 68 months.
Pompe disease is caused by the absence of
alpha-glucosidase activity and/or presence
of deleterious mutations in the alpha-glu-
cosidase gene. This enzyme is used to
breakdown glycogen within cellular lyso-
somes. The absence of glycogen catabo-
lism results in loss of muscle strength, in-
cluding the heart, preventing infants from
developing. Milder forms of the disease
can lead to life-long metabolic and physi-
cal disorders. For a clinical trial, four pa-
tients with infantile Pompe disease show-
ing a lack of the processed forms of alpha-
glucosidase were treated. The rhAGLU
was well tolerated by the patients during
more than 3 years of treatment. Anti-
rhAGLU immunoglobulin G titers initially
increased during the first 2048 weeks of
therapy, but declined thereafter. All four
patients had mature forms of alpha-gluco-
sidase as shown by Western blot analysis,
indicating that the enzyme was properly
targeted to lysosomes. Pre-clinical testing
using a knockout mouse model (see Part
III, Chapter 4) indicated full correction of
acid alpha-glucosidase deficiency in all tis-
sues except brain after a single dose of
milk-derived enzyme administration [371,
372]. Weekly enzyme infusions over a peri-
od of 6 months resulted in the degradation
of lysosomal glycogen in heart, and skele-
tal and smooth muscle.
Products for the treatment of life-threat-
ening lysosomal storage enzyme deficien-
cies provide good targets for production
using transgenic systems. Large Scale Biol-
ogy Corporation (LSBC) developed viral-in-
duced tobacco foliage-derived alpha-galac-
tosidase [239], while CropTech produced
two human lysosomal enzymes, beta-glu-
cocerebrosidase [328] and iduronidase, ex-
pressed in tobacco foliage using a me-
chanically induced promoter to induce se-
cretion of the recombinant enzyme in a
bioreactor. The LSBC product, alpha-galac-
tosidase, catalyzes the hydrolysis of globo-
triaosylceramides and related glycosphin-
golipids in lysosomes [239]. This enzyme
is missing or insufficient in patients suf-
fering Fabry disease, a genetic disorder re-
sulting in the accumulation of these glyco-
lipids in cells within the kidneys, heart,
skin, and cells lining the blood vessels, the
result being early death. There are an esti-
mated 5000 Fabry patients in Europe, and
7000 in the US. The recombinant product
Fabrazyme was recently approved for use
in these patients to reduce cellular lipid
build-up and related neuropathic pain.
Pre-clinical data with the tobacco-derived
enzyme showed positive results using an
animal model of Fabry disease [239]. The
study showed that reductions of excess lip-
ids characteristic of Fabry disease can be
achieved in target organs with proper tar-
geting of the enzyme to affected organs
and no apparent toxicity or tissue damage.
No neutralizing antibodies were observed
to the glycosylated plant-derived enzyme
in extended infusion studies, indicating
that plant glycosylation in this protein may
be non-immunogenic.
CropTech developed beta-glucocerebrosi-
dase which catalyzes hydrolysis of the glu-
cocerebrosides in lysosomes. This enzyme
is either missing or insufficient in Gau-
chers disease, the result being pain, fatigue,
jaundice, bone damage, nerve damage, or
even death. This condition is present in
1000020000 patients in the US. Beta-glu-
cocerebrosidase is available as Ceredase
derived from human placenta, or as Cerezy-

me
that is a mammalian cell culture re-

combinant product. The protein is a mono-
meric glycoprotein of 497 amino acids with
modified carbohydrates (6% content by
weight). The recombinant product requires
in vitro altering of the sugar residues at
the non-reducing ends of the oligosacchar-
ide to obtain efficient enzymatic activity
and to be recognized by carbohydrate recep-
tors on macrophage cells. The recombinant
version is delivered to patients every 2
weeks for life at an annual cost of
$160000, making it one of the most expen-
sive drugs available. This high price is
mostly due to manufacturing costs. Using
the CropTech process, a single tobacco plant
could make enough glucocerebrosidase for
one dose, resulting in significant cost reduc-
tion. Studies are in progress on carbohy-
drate modification of recombinant glycopro-
teins produced in plants to reduce or elimi-
nate the need and cost of carbohydrate re-
modeling required for this enzyme [212].
CropTech also produced the lysosomal en-
zyme alpha-l-iduronidase responsible for
mucopolysaccharidosis (MPS-I), a condition
that severely affects normal growth and de-
velopment.
5.3.3.2 Transgenic-derived Biopharma-
ceuticals in Clinical Trials to Treat
Autoimmune Conditions
Goat-milk derived recombinant human al-
pha-fetoprotein In 2003, Merrimack Phar-
maceuticals began a clinical trial to study
MM-093, a recombinant version of human
alpha-fetoprotein, to develop a novel thera-
peutic option for treatment of autoim-
mune diseases such as rheumatoid arthri-
tis and multiple sclerosis. GTC Biothera-
peutics provided purified MM-093 using
its transgenic goat milk production tech-
nology. A transgenic system was selected,
since recombinant human alpha-fetopro-
tein has been difficult to express and puri-
fy using traditional recombinant produc-
tion systems. Merrimack recently com-
pleted dosing in its Phase I trial to deter-
mine the safety, tolerability, and phar-
macokinetics of MM-093 in healthy volun-
teers [49]. Merrimack is planning a clinical
study to assess the safety and tolerability
of MM-093 in patients with rheumatoid
arthritis, evaluating the relationship be-
tween pharmacokinetics and pharmacody-
namic markers. This study will be even-
tually expanded to other autoimmune pa-
tient populations. MM-093 is the second
protein using GTC Biopharmaceuticals
production technology to enter human
clinical studies.
Biopharmaceuticals in Clinical Trials
for Cancer Treatment
Corn seed-derived huNR-LU-10 monoclonal
antibody The NeoRx Corporation, in col-
laboration with Monsantos Integrated Pro-
tein Technologies, conducted the first clini-
cal trial of a transgenic (corn seed)-derived
monoclonal antibody (MAb), huNR-LU-10,
in 1998 [171, 373].
This full-size immunoglobulin G recog-
nizes an epithelial cellular-adhesion mole-
cule (EpCAM) present in colorectal cancer
and other solid-tumor cancers. Serum anti-
bodies such as immunoglobulin G are tet-
rameric glycoproteins composed of two
identical heavy chains and two identical
light chains. Antibodies are designed to
recognize and bind target antigens with
great specificity, making them great tools
for the diagnosis, prevention, and treat-
ment of disease. Antigen recognition de-
pends on having the correct folding and
assembly that needs to take place in the
transgenic system during protein synthesis
in order for antibodies to correctly recog-
nize their antigens.
Corn seed-derived huNR-LU-10 was pre-
pared from seeds harvested from multiple
grooving sites and seasons followed by
grinding in buffer, extraction, and purifica-
tion based on two or three chromatographic
steps. The initial chromatographic step con-
sisted of Protein A affinity chromatography
that is usually used for antibody purifica-
tion. The purified antibody purity was
>99% using a process with >50% recovery
yield from corn seed to purified product.
The process was shown to remove contami-
nants such as endotoxins, corn proteins and
corn DNA to undetectable levels. The puri-
fied MAb was then conjugated with a radio-
isotope-binding complex (chelator), fol-
lowed by in vivo conjugation with radioiso-
topes to allow its use for tumor imaging
and for delivery of high doses of radiation
directly to tumors with less exposure to nor-
mal tissue. The use of tumor-directed radio-
therapy requires lower radiation exposure,
potentially avoiding the need for marrow
transplantation. Corn seed-derived huNR-
LU-10 was genetically altered to remove
the glycosylation site in the constant region
of the heavy chain, thus avoiding plant-like
glycosylation. Plant-like glycosylation is a
potential source of structural variability
and of human immunogenicity. Removing
carbohydrates in MAbs such as huNR-LU-
10 was shown to be possible because anti-
gen binding is not affected by glycosylation,
and the glycosylation-dependent effector
functions are not required for the intended
radiotherapy use of this MAb. Using bind-
ing studies, blood analysis, whole-body
imaging, and cell-based assays it was deter-
mined that removal of glycosylation had no
effect on antigen binding, pharmacoki-
netics, and in vivo targeting properties in
both mice and humans, but resulted in re-
duced cellular-mediated complement activa-
tion and complete elimination of the anti-
body-dependent complement activation
[373]. It was shown that accumulation of
MAbs in corn seed allows long-term stable
storage (at least 2 years) at ambient tem-
peratures, without notable degradation or
loss of activity [171]. The high stability of
the product in corn seed allows the initial
processing to be conducted at room tem-
perature, which simplifies facility design
and further reduces costs. Viral clearance
steps are not needed, and this results in
further cost savings.
The purity, consistency, potency, antigen
binding, serum and urine clearance, tar-
get-tissue binding and stability for the puri-
fied antibody were reproducible and accept-
able for conducting FDA-approved clinical
trials. Pre-clinical studies and human clini-
cal comparability studies with the analo-
gous glycosylated mammalian cell culture-
derived huNR-LU-10 indicated that the corn
seed-derived product was as non-immuno-
genic and as effective as a anti-cancer agent
as the mammalian-derived MAb. However,
both molecules were withdrawn from devel-
opment because of diarrhea and other side
effects in Phase II trial patients, probably
as a result of a cross-reaction of the MAb
with related epitopes on the digestive sys-
tem. These effects were not specific to the
corn-derived antibody.
In addition to immunoglobulins, plants
have been shown to be capable of produc-
ing many other antibody-related mole-
cules. Two of those will be discussed in
the next sections: antibody fragments used
as antigens for cancer vaccines; and Car-
oRx, a secretory antibody for topical use.
Secretory antibodies are dimers of serum
antibodies linked by a joining peptide. In
order properly to assemble a secretory
antibody, plants need to express four dif-
ferent cloned genes. Plants have also suc-
cessfully accumulated single chain and di-
meric antibody fragments, bispecific frag-
ments, diabodies, and antibody fusion pro-
teins [121, 122]. Antibodies have been
expressed successfully in many different
plant systems in yields in excess of 1% of
the total soluble protein. Antibodies have
also been expressed in transgenic animals,
and using aquatic plants, plant cell cul-
ture, and virus-infected plants (see Table
5.1).
MAbs were initially thought to be the
ideal product category to implement the
use of transgenic systems for the produc-
tion of biopharmaceuticals. Several reasons
made the use of transgenic systems attrac-
tive for MAb production. First, a eukaryot-
ic production is needed for MAb produc-
tion because of the MAbs structural com-
plexity. Second, MAbs have demonstrated
good accumulation levels in many trans-
genic systems with acceptable clinical per-
formance, as discussed above. Third,
MAbs are designated as well-characterized
biologics facilitating the implementation of
manufacturing changes expected to occur
when a new technology is implemented.
Finally, there are a large number of MAb
products in the development pipeline, and
those approved had rapid commercial ac-
ceptance in important therapeutic areas
(cancer, arthritis, organ rejection, infec-
tious disease treatment) leading to the po-
tential requirement for large quantities of
some of these MAbs. There are more than
200 MAbs under development, with nearly
all of them being produced via mamma-
lian cell culture. The need to develop
transgenic technology for MAb production
was further justified by information in the
late 1990s that their commercial success
depended in having a low-cost, high-capac-
ity production technology as enabled by
transgenic systems. At that time, it was
projected that there would be a significant
shortage of biopharmaceutical production
capacity. Table 5.1 illustrates that 43 com-
panies and institutes had (or have) pro-
grams for the production of MAbs using
transgenic technology. The leaders in the
industry GTC Biotherapeutics, Dow, and
Monsanto have (or had) several MAb
production programs for milk or corn ex-
pression of several important commercial
MAbs. GTC Biotherapeutics has programs
to evaluate the production of important
commercial MAbs such as Remicade
and
Humira
that could be required in large

quantities, and an additional six MAbs
(IgG1, IgG2, IgG4) and three MAb-fusion
biotherapeutics. Monsanto produced the
glycosylated form of huNR-LU-10 in corn
and several other immunoglobulins G in
non-glycosylated (e.g., BR96, developed
with Bristol-Myers Squibb Co. and Seattle
Genetics) and glycosylated forms. How-
ever, recent increases in MAb accumula-
tion in mammalian cell culture, coupled
with the use of protein-free media and the
expansion of the biopharmaceutical pro-
duction capacity, had resulted in the reali-
zation that conventional mammalian tech-
nology can produce most of the antibody-
related molecules considered for therapeu-
tic use (see Part IV, Chapter 1).
Tobacco foliage-derived tumor-specific anti-
gens used as individualized cancer vaccines
Large Scale Biology Corporation (LSBC) is
using recombinant tobacco mosaic virus
(TMV) containing genes for antigenic can-
cer markers to infect non-transgenic tobac-
co plants to deliver large amounts of indi-
vidualized patient-specific antigenic pro-
tein in few weeks after cloning (see Part I,
Chapter 2). These antigenic proteins are
produced to manufacture personalized vac-
cines for use in cancer treatment [238].
These antigens are tumor-specific proteins
containing the characteristics of each indi-
vidual patients cancer. LSBC has been able
to manufacture these vaccines for 90% of
patients qualified for clinical treatment in
as little as 6 weeks from the receipt of
biopsy materials.
The first vaccines currently in clinical
trial are used for the treatment of non-
Hodgkins lymphoma, the most prevalent
form of lymphoma and the sixth leading
cause of cancer-related deaths in the US.
This treatment works for lymphoma be-
cause all the cancer cells have the same
surface protein, the unique antibody made
by the original B cell that became malig-
nant. The recombinant viral-infected tobac-
co plants are used to make a fragment
from this unique antibody.
A mouse model of lymphoma was used
in pre-clinical studies to validate the vacci-
nation procedure using a tobacco-derived
idiotype-specific single-chain variable re-
gion fragment of the immunoglobulin
from the cancerous mouse B-cell lympho-
ma [236]. Non-vaccinated mice died within
3 weeks of tumor injection, while 80% of
the vaccinated mice were protected from
the cancer and survived.
So far, 16 individual vaccines from 16 pa-
tients have been produced from tobacco, 15
are glycosylated, and one is a non-glycosy-
lated vaccine [374]. These vaccines were ap-
plied as 6-monthly subcutaneous injections
in studies conducted at Stanford University.
Excellent safety profile in all patients at all
immunization times was observed, with sig-
nificant cellular and humoral responses ob-
served in 8/16 and 7/16 patients, respec-
tively, equivalent to the response seen in
previous cancer vaccine trials.
Biopharmaceuticals in Clinical Trials
as Anti-infective Agents
Tobacco foliage-derived recombinant secre-
tory antibody CaroRx
Planet Biotechnol-
ogys CaroRx
is a tobacco-derived secretory
antibody in clinical trials since 1998 to pre-
vent the adhesion of tooth decay-causing
bacteria to the tooth surface [122, 191,
194]. The antibody recognizes the main
adhesion protein of Streptococcus mutans,
the oral pathogen responsible for tooth
decay in humans. Tooth decay caused by
bacterial infection results in about 70% of
US dental expenditures, or $50 billion an-
nually. The treatable population in the US
and Europe is estimated at approximately
115 million people. CaroRx
has com-
pleted Phase I clinical trials under an ap-
proved US FDA Investigational New Drug
(IND) application. A Phase II clinical trial
indicated that CaroRx
on a topical appli-
cation after the bacteria have been re-
moved from the mouth helps to prevent
recolonization by S. mutans for several
months. CaroRx
can be applied either by

dental hygienists or by the patients them-
selves after tooth cleaning. CaroRx
is ex-
pected to eliminate the decay-causing bac-
teria in 2 years.
CaroRx
is purified from tobacco foliage

and then applied topically to the teeth.
CaroRx
is a chimeric secretory immuno-

globulin A/G that is produced in trans-
genic tobacco plants through the expres-
sion of four separate cloned genes. These
genes were stacked by the sequential
crossing of independent transgenic plants,
each expressing a different component. Se-
cretory antibodies such as CaroRx
consist
of secreted immunoglobulin A dimers
the most abundant form of immunoglobu-
lin in mucosal secretions, present in sali-
va, sweat, colostrum, and the mucosal
epithelia of the human body. Secretory an-
tibodies protect a vast surface of perma-
nently exposed areas of the body against
being attacked by exogenous pathogens,
and they play a major role in host defense
at mucosal surfaces by inhibiting coloniza-
tion of pathogenic microorganisms. The
immunoglobulin A dimers are associated
with the joining J-chain that is added dur-
ing secretion. Engineering plants to gener-
ate functional secretory antibodies allows
the development of mucosal passive immu-
nization. All other recombinant systems
tested for the production of secretory anti-
bodies had resulted in poor productivity
and unacceptable yields of fully assembled
antibodies.
Planet Biotechnology selected tobacco as
the production platform for CaroRx
as
the accumulation of heterologous proteins
in tobacco is a well-established technology
with high productivity due to tobaccos
high biomass yield. In 2004, Large Scale
Biology and Planet Biotechnology entered
into a biomanufacturing agreement to ex-
tract and purify CaroRx
[375]. Tobacco
plants expressing CaroRx
will be ex-
tracted at the Owensboro, Kentucky, LSBC
manufacturing facility.
The production of secretory antibodies
in plants represents an important opportu-
nity for the commercialization of plant-de-
rived biopharmaceuticals. Planet Biotech-
nology is developing two additional secre-
tory antibodies. RhinoRx is under develop-
ment for the treatment of colds due to
rhinovirus, which represents about half of
all common colds and over 20 million doc-
tors office visits a year. For the prevention
of doxorubicin-induced hair loss (alopecia)
a disturbing side effect for cancer pa-
tients undergoing chemotherapy Planet
Biotechnology is developing DoxoRx. Each
year in the US, over 250000 patients re-
ceive chemotherapy that results in hair loss.
EpicCyte was developing secretory anti-
bodies to provide products for unmet needs
for sexual health (genital herpes, 45 million
US patients), contraception (spermicidal, 42
million US potential users), HIV/AIDS
(500000 US potential users), respiratory
conditions such as pneumonia (5 million
US patients), and gastrointestinal condi-
tions such as intestinal infections [191, 265].
Cows milk-derived recombinant human lac-
toferrin Pharming completed Phase I hu-
man clinical studies using cows milk-de-
rived human lactoferrin as an anti-infec-
tive agent and for ophthalmic indications
[376]. Pharming was the first to breed a
transgenic bovine the bull Herman
(1990) and Hermans daughters produced
milk containing lactoferrin. Lactoferrin is
an iron-binding glycoprotein which is se-
creted into colostrum, milk and tears, and
protects the new-born baby from potential
infections [377]. Lactoferrin is also present
in secondary granules of neutrophils de-
posited by these circulating cells in septic
sites to attack infection and inflammation
[378]. Its principal function is to act as a
scavenging agent for non-protein-bound
iron in body fluids and inflamed areas in
order to suppress free radical-mediated
damage and decrease accessibility of the
metal to invading bacterial, fungal, and
neoplastic cells. Potential therapeutic indi-
cations include the treatment of iron-defi-
ciency anemia, gastrointestinal infections,
dry-eye syndrome, and for use as an anti-
inflammatory agent, anti-oxidant, and neu-
tralizing agent for heparin. Iron-deficiency
anemia is the most common nutritional
deficiency in the world, affecting almost
25% of the world population, mostly
young children and women of childbear-
ing age [379]. Studies have shown a re-
duced frequency of diarrhea in breast-fed
children, this being attributed to the anti-
microbial action of the human milk lacto-
ferrin and lysozyme by inhibiting growth
of diarrhea-associated organisms such as
rotavirus, Cholera, Salmonella, and Shigella.
Lactoferrin could also be a nutritional sup-
plement aimed at the prevention and treat-
ment of gastrointestinal tract infections,
especially for patients under immunosup-
pressed conditions after chemotherapy or
radiotherapy. Lactoferrin might also be ef-
fective in ophthalmic and pulmonary ap-
plications, as the protein is naturally pres-
ent in tears and lung secretions. Lactofer-
rin has been demonstrated to inhibit ma-
lignant tumor growth, presumably
through immunomodulation [380]. In ad-
dition to these anti-infective and anti-in-
flammatory uses, lactoferrin was found to
increase osteoblast differentiation, reduce
osteoblast apoptosis, and increase prolif-
eration of primary chondrocytes, thereby
indicating a role in new bone formation
and a potential therapeutic use for the
treatment of bone disorders [381]. Lactofer-
rin could also be used in food preserva-
tion, fish farming, and oral hygiene [378].
Phase I studies conducted by Pharming
indicated that cows milk-derived recombi-
nant human lactoferrin (rhLF) is well tol-
erated at high doses. Volunteers were in-
jected intravenously with rhLF without
negative side effects. An oral biodistribu-
tion study in human volunteers has shown
that natural and recombinant hLF behave
in a similar fashion in the digestive tract.
Ventria is producing rhLF and recombi-
nant human lysozyme (rhLZ) accumulat-
ing in rice seed at a level of 5 g kg
1
flour
weight [382]. Rice-derived rhLF and rhLZ
were shown to be identical to the human
proteins, and to have stability similar to
native human lactoferrin and lysozyme
when exposed to heat, pH changes, and in
vitro digestion [308, 314, 383]. Both lacto-
ferrin and lysozyme are multifunctional
proteins and play key roles in many as-
pects of human health. The initial applica-
tion of rhLF and rhLZ will be for the de-
velopment of an oral rehydration solution
to prevent and treat diarrhea in infants,
babies and travelers [382]. Further applica-
tions of rhLF and rhLZ include use in
functional foods for health maintenance of
individuals with a compromised immune
system due to medical treatment of cancer,
aging or HIV/AIDS, as both rhLF and
rhLZ have reported immunostimulatory
activity [384, 385]. Recent findings on lac-
toferrin prevention of biofilm formation by
microbial isolates from lung [386] and the
critical role of lysozyme in pulmonary
health [387] indicate that rhLF and rhLZ
might be developed to treat patients with
lung disease, for example those with cystic
fibrosis.
Meristem is producing lactoferrin in corn
seed for the treatment of dry-eye syndrome
and gastrointestinal infections [164]. Lacto-
ferrin is a complementary product to gastric
lipase, Meristems lead development prod-
uct. Correct N-glycosylation has been deter-
mined to be important to maintain lactofer-
rins stability. An analysis of corn-derived
lactoferrin indicated that both N-glycosyla-
tion sites are mainly substituted by typical
plant-type glycans, with beta-1,2-xylose and
alpha-1,3-linked fucose at the proximal N-
acetylglucosamine. As expected, the com-
plex-type glycans typical of human proteins
are not present in maize recombinant lacto-
ferrin [164]. Lactoferrin has also been accu-
mulated in transgenic potatoes [243].
5.3.5
Transgenic-derived Oral Vaccines
Vaccines available today are injectable anti-
gens made from killed or weakened ver-
sions of a pathogen or of some material
(usually proteins or protein fragments) de-
rived from a pathogen. Injection of these
antigens stimulates the immune system to
behave as if the pathogen had infected the
body; this results in a response to elimi-
nate the pathogen and the creation of mem-
ory cells that later will repel the pathogen.
Oral vaccines would be the preferred way
to provide certain antigens, particularly for
those pathogens that infect by oral routes
of entry. Oral vaccines work by stimulating
the digestive, respiratory, and/or reproduc-
tive mucosal immune system by delivering
antigens in a stable matrix (usually the ed-
ible part of a plant) as described in many
publications [10, 26, 112, 113, 115, 146,
192, 209, 226, 259, 260, 271, 273, 388
398]. Antigens expressed within the matrix
of transgenic plants can assemble into com-
plex structures and be stored in plant tis-
sues; this allows them to act as effective
antigens when released from the plant cell
in the lower intestine. These antigenic ma-
terials are resistant to digestion and capable
of reaching lymphoid tissues [146].
There are several challenges to develop
effective oral vaccines that can be met by
producing oral antigens in transgenic
plants. Oral immunization requires larger
amounts of antigen compared to injectable
antigens typically milligram rather than
microgram quantities. Oral antigen cannot
be produced economically in such large
quantities using current vaccine produc-
tion technologies (microbial fermentation,
cell culture, eggs). Transgenic plants pro-
vide advantages in addition to cost and
convenience, because these plants can pro-
duce significant quantities of protein and
deliver oral antigens in an acceptable ma-
trix for administration without purifica-
tion. The use of plant derivatives with a
long shelf-life or processed by freeze-dry-
ing can achieve antigen preservation at
room temperature, avoiding cold chain
transportation and storage. This represents
a significant advantage of plant-derived
vaccines over current products, especially
when they are to be delivered in develop-
ing regions, and due to lower costs com-
pared to current vaccines and the reduc-
tion of hazards associated with injection.
For the successful implementation of
transgenic systems producing oral vac-
cines, the system must accumulate stable,
fully assembled, orally available antigens
at high, consistent accumulation levels to
enable controlled dosing. Plants that pro-
duce edible leaves, roots and fruits are the
best choices for oral vaccines. Potato was
selected as the first oral vaccine host be-
cause transformation and cultivation tech-
nologies were available in the late 1980s
[399, 400], but today bananas, lettuce, lu-
pine, spinach, sweet potato, corn, and to-
mato are all being developed for human
vaccines, and alfalfa, corn and beans for
animal vaccines. Emphasis is given to
foods that are well-liked, consumed raw,
and have a long shelf-life so that the ac-
ceptability and effectiveness of the technol-
ogy are enhanced.
In 1990, plants were shown capable of
expressing biologically active antigens, as
demonstrated by the expression of the sur-
face protein antigen of the dental bacteri-
um Streptococcus mutans in tobacco [401
403]. When fed to mice, biologically active
antibodies were induced to inhibit growth
of these bacteria.
The heat-labile toxin B subunit of E. coli
(LTB) [398], hepatitis B surface antigen
[404], respiratory syncytial virus F protein
[258], measles virus hemagglutinin [180],
and Norwalk virus capsid protein [240,
405, 406] have each been successfully ex-
pressed in plants and delivered orally in
animals or humans to determine their im-
munoprophylactic activity. The first ac-
count of a human clinical trial of oral vac-
cine based on an E. coli enterotoxin as
antigen delivered by consuming raw pota-
toes was published in 1998 [407]. Ten of
the 11 test subjects produced specific anti-
bodies to the toxin used, whilst no specific
antibodies were produced in the control
subjects. Immunity level was comparable
to that measured in volunteers exposed to
live organisms. The study demonstrated
that oral vaccines could survive digestion
delivered in a plant matrix and effectively
stimulate an immune response. Several
other animal and human studies based on
E. coli enterotoxin for the prevention of
travelers diarrhea have been reported,
these having used potato, tobacco, and
corn as the delivery matrix [102, 115, 207,
271, 408, 409]. In addition to these two
anti-bacterial plant-derived vaccines, other
immunoprophylactics delivered in trans-
genic plants have been tested in clinical
trials, including the hepatitis B surface
antigen, analogous yeast-derived Recombi-
vax
by Merck, expressed in potato or let-

tuce. This antigen has been used as a pro-
totype for vaccines and also expressed in
banana, tobacco, and lupin [100, 104, 200,
226, 332, 404, 410414]. Other examples of
antigens expressed in transgenic plants for
the prevention of human diseases include
tomato-derived rabies glycoprotein [222],
Helicobacter pylori antigen for preventing
peptic ulcer/cancer [415], rotavirus antigen
for preventing severe diarrhea [415], tobac-
co-derived human cytomegalovirus glyco-
protein [301], tobacco and potato-derived
cholera antigen [244, 246, 247], and Nor-
walk virus capsid protein expressed in po-
tato, tobacco, or tomato [240, 257, 400,
405, 406, 416]. Research is also under way
to create oral vaccine candidates in trans-
genic corn for AIDS prevention and for
treatments using antigens derived from
the simian immunodeficiency virus [288].
Genetically engineered spinach has been
used to express HIV-suppressing proteins
in an attempt to develop a safe and inex-
pensive AIDS vaccine [417]. The Institut
fr Pflanzengenetik und Kulturpflanzen-
forschung has genetically engineered
plants to produce human papillomavirus
Type 16 virus-like particles to develop plant
vaccines against cervical cancer, the third
most common cancer among women
worldwide [101, 146, 185, 210, 418, 419].
Dow Pharma (Midlands, IL, USA) is work-
ing with the Fraunhofer Institute and the
National Institutes of Health to develop
technology based on viral particle produc-
tion in foliage for the rapid development
of vaccines against infectious diseases, in-
cluding biowarfare agents, to be delivered
by capsule or nasal spray [420].
Many animal vaccines are also under de-
velopment such as those based on potato-
derived rabbit hemorrhagic disease virus
antigen, mink enteric virus antigen, Arabi-
dopsis and alfalfa-derived foot-and-mouth
disease antigens for agricultural domestic
animals [421, 422], and Arabidopsis, tobac-
co, corn-derived transmissible gastroenteri-
tis coronavirus antigens for pigs [287].
Transgenic animals have also been con-
sidered for the production of vaccines.
GTC Biotherapeutics is currently working
with the National Health Institute to devel-
op a malaria vaccine based on the produc-
tion of the viral surface protein antigen
MSP-1 in goat milk. This antigen is diffi-
cult to express in conventional recombi-
nant production systems.
One challenge related to the use of oral
vaccines is the potential risk of inducing
oral tolerance to antigens, and thus the
use of oral vaccines requires control and
monitoring of the administration of these
products as any other biopharmaceutical.
However, oral tolerance may also be used
for the development of orally delivered
treatments for autoimmune diseases.
Plant-derived autoantigens may be able
to suppress immune activity related to
autoimmune diseases. Autoimmune re-
sponse is involved in several diseases, in-
cluding insulin-dependent diabetes, psoria-
sis, systemic lupus erythematosus, Graves
disease and rheumatoid arthritis. The oral
administration of disease-specific autoanti-
gens may either prevent or delay the onset
of autoimmune disease symptoms. Studies
are in progress to determine if plant-de-
rived oral autoantigens can suppress auto-
immunity related to type I diabetes [244
247]. Insulin-producing beta-cell proteins
can elicit autoimmunity in people predis-
posed to type I diabetes. A potato-based
diabetes vaccine candidate was developed
based on using insulin or glutamic acid
decarboxylase linked to the innocuous B
subunit of the cholera toxin to enhance
uptake of the antigenic fusion protein by
the gut-associated lymphoid tissues [244
247]. Studies showed that repeated feeding
of these antigens to mice with a tendency
to become diabetic was effective in sup-
pressing autoimmune responses, and de-
laying the onset of high blood sugar levels,
an indication of diabetes [244].
Plant-derived vaccines and toxin deacti-
vation enzymes can also play a critical role
in biodefense by enabling the availability
of unlimited, low-cost proteins and other
biological agents to generate stable stock-
piles of preventive and therapeutic agents
to protect military and civilian populations.
Research is currently in progress to ad-
dress the danger of smallpox, plague, Ebo-
la virus, and anthrax virus when used as
bioterrorist weapons, with those nations at
risk developing programs rapidly to vacci-
nate part or all of their population, though
this requires stockpiles to be prepared,
stored, and renewed. Plant-derived vac-
cines can become ideal substitutes for tra-
ditional vaccines for biodefense applica-
tions. A plant-derived vaccine for anthrax,
based on the anthrax protective antigen, is
under development using tobacco trans-
formed for expression of the antigen in
the chloroplast [423]. The current anthrax
vaccine, which was designed in 1950,
causes edema and has other lethal facets
that lead to harmful side effects. The use
of chloroplast expression minimizes the
risk of gene spread, and can lead to higher
accumulation levels than with nuclear ex-
pression. Spinach infected with a recombi-
nant virus containing the protective anti-
gen is also being considered for the pro-
duction of a vaccine against anthrax in-
fection [424].
In a program related to biodefense, Nex-
ia is developing jointly with US and Cana-
dian defense agencies a recombinant ver-
sion in goat milk of human butyrylcholi-
nesterase (hBChE) enzyme (Protexia
)
[15]. This agent can act as an enzymatic
bioscavenger for nerve agents, such as so-
man, sarin, VX and tabun, to absorb and
degrade organophosphate poisons before
they cause neurological damage. Studies
using plasma-derived hBChE have shown
that increasing hBChE concentrations in
the blood protects laboratory animals from
the toxic effects of nerve agents. Protexia
is being developed for post-exposure (res-

cue) therapy and military prophylaxis to
prevent against the toxic effects of nerve
agents. Pharmacokinetics studies in ani-
mals have shown that a single injection of
Protexia
resulted in a sustained elevation

of blood BChE levels for many hours.
5.3.6
Transgenic-derived Proteins for Use
in Medical Devices and Drug Delivery
Many animal- and human-derived protein
including silk, elastin, fibrinogen, collagen,
and gelatin, are used in medical devices
and for drug delivery. Transgenic systems
can provide cost effective technology in the
production of structural proteins for medi-
cal use that are required to be biocompati-
ble and stable in biological tissues.
The most advanced structural protein
production program based on using a
transgenic system is the production of
goat milk-derived spider silk (BioSteel
)
by Nexia [55]. Spider silk is the strongest
fiber known, holding up to 400000 lb per
square inch (281 10
6
kg m
2
) without
breaking. Dragline spider silk proteins
contain iterated alanine-rich crystal-form-
ing blocks with mechanical strength and
glycine-rich amorphous blocks that provide
elasticity. Recombinant production is the
only alternative for producing this fiber
commercially as it cannot be harvested
from spiders [425]. Transgenic technology
is required for its production because con-
ventional recombinant microbial or cell
culture production systems have been not
been successful in expressing silk genes.
These genes are large and contain many
repetitive units which stress the protein
synthetic machinery when these organ-
isms are grown in bioreactors. Nexia is
producing soluble recombinant spider silk
using traditional goat dairy techniques for
milk collection, after which the protein is
extracted from milk and then spun into fi-
bers [426]. In 2000, two transgenic goats
Peter and Webster were born with the spi-
der silk gene incorporated into their genet-
ic composition, and used to generate the
milking herd. BioSteel
is being devel-
oped for use in medical device products
used in surgery, ophthalmic applications
and as prostheses. BioSteel
can also be
used in industrial applications such as in
lightweight, flexible bullet-resistant body
armor for the military and law enforce-
ment agencies, as well as high-perfor-
mance sporting equipment such as biode-
gradable fishing lines and nets. Nexia have
worked with the US Army Soldier and
Biological Chemical Command to develop
techniques for making fibers from soluble
recombinant spider silk proteins, and have
demonstrated that wet spinning of fibers
is possible from a concentrated aqueous so-
lution of mammalian cell culture and goat
milk-derived spider silk monofilaments.
Nexia also collaborated with Acordis Speci-
alty Fibers to develop spider silk fibers and
specialty materials for industrial, textile,
medical, and hygiene applications.
Synthetic spider silk has also been pro-
duced in transgenic tobacco and potato ex-
pressing the endogenous silk protein
genes of the spider Nephila clavipes [147].
Proteins of up to 100 kDa in size, and with
90% identity to silk protein, were pro-
duced in tobacco leaves, potato leaves and
potato tubers at up to 2% of the total solu-
ble protein accumulation level [151].
Pharming and Infigen are each develop-
ing milk-derived recombinant human fibri-
nogen (rhFIB) to be used as tissue sealant
to stop internal or external bleeding during
surgery or after traumatic injury [50, 51,
427]. Human fibrinogen is a soluble blood
protein that can form insoluble fibrin poly-
mers after activation by thrombin. In 2000,
the fibrinogen market was estimated at
US$284 million. Commercially available fi-
brin sealants use fibrinogen purified from
human donor plasma. As with other plasma
products, there are safety, availability, quali-
ty, and reproducibility concerns related to
the use of plasma-derived fibrinogen.
Furthermore, a substantial shortage of hu-
man fibrinogen is anticipated as market de-
mand is expected to increase to 500 kg per
year and require >10
6
L of donor blood.
Pharming has established a production line
with high expression of rhFIB that is vir-
tually identical to plasma fibrinogen, whilst
Infigen claims an accumulation level of
2.4 g L
1
rhFIB in cows milk.
Several groups have demonstrated an ac-
cumulation of recombinant human col-
lagen-related proteins in tobacco, milk,
and silkworms [50, 51, 56, 177, 376, 428
431]. Collagen available today for medical
uses is mostly a by-product of the meat in-
dustry, produced mainly from bovine
hooves and porcine/bovine bones. There is
a mammalian cell culture-derived collagen
which is available only for selected, high-
value applications due to its limited avail-
ability and high cost. Current uses of col-
lagen as a component of medical devices
include hemostats, vascular/tissue sea-
lants, implant coatings, artificial skin,
bone graft substitutes, dental implants,
and wound dressings. Injectable collagen
solutions are used for dermal augmenta-
tion and the treatment of incontinence.
Many additional applications are under de-
velopment, such as a component for the
tissue engineering of cartilage, bone, skin,
artificial tendons, blood vessels, nerve re-
generation, and for drug delivery [177].
Gelatin, made from hydrolyzed collagen,
also has many medical applications such
as a vaccine/biologic stabilizer and as a
plasma expander. Gelatin is also used to
manufacture hemostat sponges, hard cap-
sules, soft capsules and gel tablets. Most
of the collagen and gelatin available is a
variable mixture of several collagens and is
not highly purified; this leads to the possi-
bility of causing inflammatory reactions in
those individuals who are sensitive to ani-
mal-derived components. In addition, over
the past few years there has been growing
concern about the potential for contamina-
tion of bovine products with mad cow dis-
ease and its human variant causing
CreutzfeldtJakob disease.
Recombinant collagen and gelatin prod-
ucts provide a consistent and reliable hu-
man, animal or engineered amino acid
composition material which is compatible
with current pharmaceutical manufactur-
ing processes and potentially free of ani-
mal-derived components and pathogens
[177]. FibroGen developed a production
technology that demonstrated the use of re-
combinant insect cells, yeast, transgenic
plants, and transgenic animals, accumulat-
ing stable recombinant human collagen
and gelatin. This technology is based on ex-
pressing the genes for collagen or collagen
fragments simultaneously with prolyl hy-
droxylase, resulting in recombinant col-
lagen which is stable at biologically relevant
temperatures. Yeast was selected for the
production of recombinant collagen and
gelatin for most medical applications as it
accumulates fully assembled, stable col-
lagens and gelatin fragments at high levels.
Transgenic systems could be used for
the cost-effective, large-scale production of
recombinant human collagens and gela-
tins for selected medical applications re-
quiring large quantities at low cost. Hu-
man types I and III collagen homotrimers
have been expressed in transgenic tobacco
plants [177], while transgenic mice have
been engineered to produce full-length
type I procollagen homotrimer in milk
[428, 429]. Most recently, a transgenic silk-
worm system was used to produce a fu-
sion protein containing a collagenous se-
quence [430]. As seen in other recombi-
nant expression systems, these transgenic
systems lack sufficient endogenous prolyl
hydroxylase activity to produce fully hy-
droxylated collagen. In mice and tobacco,
this deficiency was overcome by over-ex-
pression of human prolyl hydroxylase, ana-
logous to the procedures conducted in
yeast and insect cell culture [177]. Pharm-
ing and Infigen have each demonstrated
an accumulation of recombinant collagen
related proteins in milk, with cows at Infi-
gen accumulating recombinant human
collagen-related molecules at a concentra-
tion of 8 g L
1
in milk [50]. Meristem has
shown that human collagen can be pro-
duced in transgenic tobacco plants, and that
the protein is spontaneously processed and
assembled into its typical triple-helix confor-
mation [160, 167]. The plant-derived col-
lagen had a low thermal stability owing to
the lack of hydroxyproline residues, but this
was remedied by co-expressing with animal
proline-4-hydroxylase [163].
Recombinant elastin has not been pro-
duced in transgenic systems as it is not
currently a commercial product with high-
volume demand. Elastin is used for the
production of vascular grafts [432436]
which today are frequently used to replace
a damaged artery or to create a new artery
for improved blood flow. Whilst use of
autologous vessels is preferable, synthetic
grafts made from expanded polytetra-
fluoroethylene are used on many occasions
when autologous vessels are not available.
However, this synthetic material may be
thrombogenic and result in smooth mus-
cle cell hyperplasia. It has been shown that
recombinant elastin fibers can form fibril-
lar structures and can also be cross-linked
to form stable, insoluble structures similar
to natural elastin [432]. These self-assem-
bly and cross-linking abilities combined
with the biological characteristics of low
platelet activation and inhibitory effects on
smooth muscle cell growth make recom-
binant elastin-based polypeptides ideal
candidates for coating vascular grafts.
5.4
Conclusions
Transgenic systems offer production alter-
natives for biopharmaceuticals, and have
significant advantages when compared
with bioreactor-based microbial and cell
culture-based recombinant production sys-
tems. However, transgenic systems remain
largely untested for production biopharma-
ceuticals, as products derived from trans-
genic systems are not yet commercially
available for human use. By the time this
book has been published however, this sit-
uation may have changed, as antithrombin
III derived from goat milk may be mar-
keted. As discussed earlier, there are three
transgenic plant-derived commercial prod-
ucts available for non-human use, and 10
plant- and milk-derived biotherapeutics
with human clinical experience. Many vac-
cines are also under development from
transgenic systems, some with limited hu-
man clinical data, and several structural
proteins under development for potential
medical uses. These diverse products repre-
sent what hopefully will be by 2010 the first
wave of products from transgenic systems
that will facilitate the implementation and
acceptance of transgenic technology by the
pharmaceutical industry, regulatory agen-
cies, the medical community, and patients.
The benefits on healthcare of biophar-
maceuticals derived from transgenic tech-
nology will be reflected by the commercial
availability of new products and therapies,
while facilitating the delivery, availability,
and accessibility of existing biopharmaceu-
ticals that cannot be produced using cur-
rent production approaches. Likewise, the
implementation of transgenic systems will
enable the commercialization of complex
multimeric proteins (e.g., viral particles for
oral vaccines and secretory antibodies) that
are difficult if not impossible to create
using current methods. Transgenic sys-
tems will also allow the production of fu-
sion proteins, metabolic toxic proteins,
and unstable peptides that cannot be pro-
duced efficiently in bioreactor-based pro-
duction systems.
The implementation of human-like post-
translational processing in specific tissues
5.4 Conclusions 873
of transgenic systems generating human-
like processed biopharmaceuticals may re-
sult in products with improved homogeneity
and engineered therapeutic performance.
By making biopharmaceuticals available
practically in unlimited quantities and at
low cost, transgenic systems will improve
the accessibility of biopharmaceuticals to
patients with life-long needs for such prod-
ucts. Likewise, transgenics may enable bio-
generics to allow pharmaceutical compa-
nies to maintain the profit margins re-
quired to sustain product development.
Unlimited, high-quality, low-cost biophar-
maceuticals will also facilitate the imple-
mentation of products with multiple indi-
cations and the administration of biophar-
maceuticals using non-injectable delivery
routes (oral, transdermal, pulmonary).
Non-injectable biopharmaceutical delivery
often requires high doses due to poor bio-
availability and degradation associated with
these routes of administration. The thera-
peutic use of blood proteins derived from
transgenic systems can expand the use of
blood-derived factors as the availability,
economics and safety parameters asso-
ciated with many such current biopharma-
ceuticals are unattractive.
Transgenic production will facilitate the
use of biopharmaceuticals and nutraceuti-
cals in oral formats for the prevention of
infectious and autoimmune diseases, and
also for biodefense. The availability of bio-
pharmaceuticals and nutraceuticals in
stable matrixes that do not require refrig-
eration during transportation and storage
will undoubtedly impact on the accessibil-
ity of these products to healthcare and nu-
trition crises in developing countries.
Finally, transgenic systems may permit
the commercialization of functionalized,
consistent quality, low-cost, protein-based
biomaterials resulting in improved bio-
compatibility, accessibility, quality and per-
formance for medical devices and for ad-
vanced drug delivery systems.
We can expect the implementation of
transgenic technology for the production
of biopharmaceuticals to take place in the
future as this technology is suited to meet
several critical needs confronting the
healthcare industry during the next few
decades. First, there is a worldwide in-
crease in the health-conscious and physi-
cally active aging population that requires
innovative high-performance health main-
tenance products, together with nutrition
products in larger quantities and at lower
cost that potentially only transgenic sys-
tems will be able to deliver. The economic
improvements in many currently develop-
ing regions of the world such as Eastern
Europe, China, and India will significantly
increase the demand for biopharmaceuti-
cals, as well as the growth resulting from
an increasingly aging population in the
US, Japan, and Western Europe. Second,
the worldwide growth of nationalized
healthcare systems and of managed care
organizations will result in significant
pressure to reduce the costs of biopharma-
ceuticals. In the US, major efforts are cur-
rently being made to obtain drugs from
outside the country to reduce costs, and
this is being encouraged by some govern-
ment officials. It is clear that keeping drug
prices at high levels in particular regions
will be difficult to achieve in the future.
The increasing use of combination thera-
pies for diseases such as arthritis, diabetes
and cancer further aggravates the need to
control overall therapy costs, as only lim-
ited resources are available for each pa-
tient. Third, the pharmaceutical market is
fractionated into many companies, with no
single company controlling more than
15% of the market. Patent expirations
leading to the eventual introduction of bio-
generics (see Part VIII, Chapter 3), re-
duced R&D productivity, increased develop-
ment costs ($800 million for a new thera-
peutic agent), and outsourcing activities to
regions with lower costs will place increas-
ing pressure on the pharmaceutical industry
to improve its efficiency (see Part IV, Chap-
ter 16). Transgenic technology can reduce
the production cost of biopharmaceuticals,
thereby allowing the industry to maintain
its traditionally high profit margins that
are required in a difficult business and inno-
vative drug development environment.
Fourth, the regulatory agencies are be-
coming harmonized worldwide and sub-
ject to economic and political pressures to
improve the diversity, cost benefits, and ac-
cessibility of drugs. Once transgenic sys-
tems become accepted by these organiza-
tions, there will be significant pressure to
accelerate their implementation.
Finally, new technologies related to the
rapid advancements in genomics, proteo-
mics, bioanalytics, high-throughput screen-
ing, and protein engineering will improve
the therapeutic ratio of current biotherapeu-
tics. This in turn will create new therapies
that should generate many new biopharma-
ceuticals requiring transgenic systems for
their successful commercialization.
The success of any innovative technol-
ogy resides in providing benefits to all of
those involved in its implementation and
commercialization. Transgenic technology
will be implemented if its products meet
the needs of the many entities involved in
their commercialization: agricultural bio-
technology companies, seed producers,
farmers, processors, food producers, phar-
maceutical industry, medical personnel, in-
surance companies, regulatory agencies,
and patients. The acceptance of transgenic
systems for the production of biologics
could be negatively impacted by another
accident such as the Starlink incident
[437], which involved the inadvertent mix-
ing of plant or animal materials contain-
ing a biopharmaceutical into the food sup-
ply, or into the environment. The govern-
ment, activists, and food producers will de-
mand that farmers and others involved in
the processing of agricultural materials
each test for the presence of biopharma-
ceuticals to very low detection levels,
though no scientific evidence of harm to
humans or animals may be demonstrated.
The producers would have to pay for the
testing, and also could lose sales abroad
and domestically, even forcing food pro-
ducers to reformulate their products to re-
move any use of potentially contaminated
material. At present, producers are work-
ing on containment strategies to insure
that no agricultural material from a trans-
genic-derived biopharmaceutical will enter
the food chain or the environment (see
Procedures required for transgenic tech-
nology implementation and currently un-
defined regulatory requirements will affect
the cost of biopharmaceuticals from trans-
genic systems. There are many un-
knowns in the regulatory environment,
and these pose risks for those entities
whose participation downstream of the
transgenic production technology specifi-
cally the pharmaceutical industry is nec-
essary to implement these production sys-
tems. As with any new technology, risks
will have to be in balance with the benefits
to patients, farmers, processors, food pro-
ducers, governments, insurers, seed com-
panies, and to the healthcare industry of
the commercialization of transgenic-de-
rived biopharmaceuticals. The success of
the technology developers, processors, and
farmers in managing these hazards com-
bined with public/industry/regulatory ac-
ceptance of the first wave of transgenic-de-
rived biopharmaceuticals will allow
transgenic technologies to deliver signifi-
5.4 Conclusions 875
cant benefits to the healthcare and agricul-
tural industries, thus illustrating the sig-
nificant value of the agricultural biotech-
nology as applied to human health.
References
1 Guide to Biotechnology. 2004. Biotechnology
Industry Association www.bio.org.
2 DePalma A. Is a Green Plant in Your Manu-
facturing Future? BioPharm Int., 2003; 16(11):
2433.
3 USDA. US Biotech Crop Plantings Fact Sheet
A Precis From The USDA Report. http://
usda.mannlib.cornell.edu/reports/nassr/field/
pcp-bba/acrg0604.txt 2004.
4 Peterson R, Arntzen CJ. On risk and plant-
based biopharmaceuticals. Trends Biotechnol.,
2004; 22(2).
5 Denman J, Hayes M, ODay C, et al. Trans-
genic expression of a variant of human tissue-
type plasminogen activator in goat milk: puri-
fication and characterization of the recombi-
nant enzyme. Biotechnology (NY), 1991; 9(9):
839843.
6 Barta A. The expression of a nopaline synthase
human growth hormone chimaeric gene in
transformed tobacco and sunflower callus tis-
sue. Plant Mol. Biol., 1986; 6: 347357.
7 Steiner U. Business Case for Plant Factories.
2003; Plant-Made Pharmaceuticals Confer-
ence, Quebec City, Quebec, Canada (on line
copy available at http://www.cpmp2003.org/).
8 Baez J. State of the Science: Role of Trans-
genic Technology in the Biosynthesis of Bio-
Pharmaceuticals and Industrial Protein. 2004
(Corn-Produced Pharmaceuticals and Indus-
trials Risk Assessment Symposium, sponsored
by the Biosafety Institute for Genetically Mod-
ified Agricultural Products Symposium,
Ames, Iowa).
9 Behboodi E, Chen M, Destrempes L, et al.
Principles of Cloning. Elsevier Science (USA),
2002.
10 Larrick JW, Thomas DW. Producing proteins
in transgenic plants and animals. Curr. Opin.
Biotechnol., 2001; 12(4): 411418.
11 Giddings G, Allison G, Brooks D, et al. Trans-
genic plants as factories for biopharmaceuti-
cals. Nature Biotechnol., 2000; 18(11): 1151
1155.
12 Herbers K, Sonnewald U. Production of new/
modified proteins in transgenic plants. Curr.
Opin. Biotechnol., 1999; 10(2): 163168.
13 Okada Y. [Transgenic plants as medicine pro-
duction systems]. Nippon Yakurigaku Zasshi,
1997; 110 Suppl 1: 1P6P.
14 Schillberg S, Fischer R, Emans N. Molecular
farming of recombinant antibodies in plants.
Cell. Mol. Life Sci., 2003; 60(3): 433445.
15 Twyman RM, Stoger E, Schillberg S, et al.
Molecular farming in plants: host systems
and expression technology. Trends Biotechnol.,
2003; 21(12): 570578.
16 Erickson L, Yu W, Brandle J, et al. (Eds), Mo-
lecular farming of plants and animals for hu-
man and veterinary medicine. Dordrecht, The
Netherlands: Kluwer Academic Publishers,
2002.
17 Cunningham C, Porter A (Eds), Methods in
molecular Biology the production of clini-
cally important recombinant proteins in
plants. Totowa, New Jersey, U.S.: Humana
Press, 1998.
18 Rohricht P, Ayares D. Transgenic Protein Pro-
duction: Part 3, Nuclear Transfer Technology.
BioPharm, 1999; 5660.
19 Fischer R, Stoger E, Schillberg S, et al. Plant-
based production of biopharmaceuticals. Curr.
Opin. Plant Biol., 2004; 7(2): 152158.
20 Fischer R, Emans N. Molecular farming of
pharmaceutical proteins. Transgenic Res.,
2000; 9(4/5): 279299.
21 Houdebine L. Transgenic animal bioreactors.
Transgenic Res., 2000; 9(4/5): 305320.
22 Giddings G. Transgenic plants as protein fac-
tories. Curr. Opin. Biotechnol., 2001; 12(5):
450454.
23 Kusnadi A, Nikolov Z, Howard JA. Production
of recombinant proteins in transgenic plants:
practical considerations. Biotechnol. Bioeng.,
1997; 56: 473484.
24 Horn ME, Woodard SL, Howard JA. Plant mo-
lecular farming: systems and products. Plant
Cell. Rep., 2004; 22(10): 711720.
25 Hood EE, Kusnadi A, Nikolov Z, et al. Molec-
ular farming of industrial proteins from trans-
genic maize. Adv. Exp. Med. Biol., 1999; 464:
127147.
26 Langridge WH. Edible vaccines. Sci. Am.,
2000; 283(3): 6671.
27 Ganz P, Sardana R, Dudani A, et al. In: Owen
M, Pen J (Eds), Transgenic Plants: A produc-
tion system for pharmaceutical proteins. New
York: John Wiley and Sons Ltd.; 1996: 281
297.
28 Molecular farming may be next wave for bio-
tech drug manufacturers. January 31, 2003.
The Food and Drug Letter Issue, No. 668.
29 Van Brunt J. Molecular Farming Factories.
Signals Magazine, Online Magazine of
Biotechnology Industry Analysis 2002;
http://www.signalsmag.com.
30 Raskin I, Ribnicky DM, Komarnytsky S, et al.
Plants and human health in the twenty-first
century. Trends Biotechnol., 2002; 20(12):
522531.
31 Meni Y, Kovacs J. In: Erickson L, Yu W, Bran-
dle J, et al. (Eds), Molecular Farming of Plants
and Animals for Human and Veterinary Medi-
cine. Kluwer Academic; 2002: 287317.
32 Knblein J, McCaman M. Modern Biopharma-
ceuticals Recombinant Protein Expression
in Transgenic Plants. Screening. Trends Drug
Discov., 2003; 6: 3335.
33 Knblein J. Biotech: A New Era In The New
Millennium? Fermentation and Expression of
Biopharmaceuticals in Plants. Screening.
Trends Drug Discov., 2003; 4: 1416.
34 Knblein J. Biopharmaceuticals expressed in
plants a new era in the new Millennium.
In: Mller R, Kayser O (Eds), Applications in
Pharmaceutical Biotechnology. Wiley-VCH
2004.
35 Knblein J. Plant-based Expression of Biophar-
maceuticals. In: Meyers RE (Ed), Encyclopedia
of Molecular Cell Biology and Molecular Med-
icine. Wiley & Sons, 2005; 2nd edition, Vol.
10: p. 385410.
36 Stromqvist M, Houdebine L, Andersson J, et
al. Recombinant human extracellular superox-
ide dismutase produced in milk of transgenic
rabbits. Transgenic Res., 1997; 6(4): 271278.
37 Galet C, Le Bourhis CM, Chopineau M, et al.
Expression of a single beta-alpha chain pro-
tein of equine LH/CG in milk of transgenic
rabbits and its biological activity. Mol. Cell.
Endocrinol., 2001; 174(1/2): 3140.
38 BioProtein Technologies Scientific Publica-
tions. http://www.bioprotein.com/gb/
news_events.htm#Squence_1 2004.
39 Transgametic
Process Gala Transgenics.

http://www.gala.com/transgametic.html 2004.
40 GTC Biotherapeutics Scientific Publications.
http://www.transgenics.com/publications.html
2004.
41 Echelard Y, Meade H. Toward a new cash cow.
Nature Biotechnol., 2002; 20(9): 881882.
42 Stowers AW, Chen LH, Zhang Y, et al. A re-
combinant vaccine expressed in the milk of
transgenic mice protects Aotus monkeys from
a lethal challenge with Plasmodium falciparum.
Proc Natl Acad Sci USA, 2002; 99(1): 339
344.
43 Pollock DP, Kutzko JP, Birck-Wilson E, et al.
Transgenic milk as a method for the produc-
tion of recombinant antibodies. J. Immunol.
Methods, 1999; 231(1/2): 147157.
44 Young MW, Meade H, Curling JM, et al. Pro-
duction of recombinant antibodies in the milk
of transgenic animals. Res. Immunol., 1998;
149(6): 609610.
45 Edmunds T, Van Patten SM, Pollock J, et al.
Transgenically produced human antithrombin:
structural and functional comparison to hu-
man plasma-derived antithrombin. Blood,
1998; 91(12): 45614571.
46 Ebert KM, Ditullio P, Barry CA, et al. Induc-
tion of human tissue plasminogen activator in
the mammary gland of transgenic goats. Bio-
technology (NY), 1994; 12(7): 699702.
47 Ditullio P, Cheng SH, Marshall J, et al. Pro-
duction of cystic fibrosis transmembrane con-
ductance regulator in the milk of transgenic
mice. Biotechnology (NY), 1992; 10(1): 7477.
48 Konkle BA, Bauer KA, Weinstein R, et al. Use
of recombinant human antithrombin in pa-
tients with congenital antithrombin deficiency
undergoing surgical procedures. Transfusion,
2003; 43(3): 390394.
49 Merrimack Pharmaceuticals Completes Dos-
ing in Phase I Study of Immunomodulator,
MM-093. http://www.merrimackpharma.com./
news.archive.html#11_19_03 2003.
50 Infigen Transgenic Production Technology.
http://www.infigen.com/sci_tech_prot.html
2004.
51 Infigen Terminates Agreement With Pharm-
ing Holding, N.V. http://www.biospace.com/
ccis/news_story.cfm?StoryID=6578115&full=1
2001.
52 Niemann H, Halter R, Carnwath J, et al. Ex-
pression of human blood clotting factor VIII
in the mammary gland of transgenic sheep.
Transgenic Res., 1999; 8(3): 237347.
53 Ko J, Lee C, Kim K, et al. Production of bio-
logically active human granulocyte colony
stimulating factor in the milk of transgenic
goat. Transgenic Res., 2000; 9(3): 215222.
References 877
54 Nexia Biotechnology Fundamental Biology.
http://www.nexiabiotech.com/en/03_bio/
index.php 2004.
55 Lazaris A, Arcidiacono S, Huang Y, et al. Spi-
der silk fibers spun from soluble recombinant
silk produced in mammalian cells. Science,
2002; 295(5554): 472476.
56 Pharming Products. http://
www.pharming.com/index.php?act=prod 2004.
57 van Berkel PH, Welling MM, Geerts M, et al.
Large scale production of recombinant human
lactoferrin in the milk of transgenic cows. Na-
ture Biotechnol., 2002; 20(5): 484487.
58 Pharming Features Clinical Results Of C1 In-
hibitor At Investigator Meeting. http://
www.pharming.com/
index.php?act=show&pg=5 2004.
59 PPL Therapeutics Products. http://www.ppl-
therapeutics.com/products/products.html
2004.
60 Vogel G. Biotechnology. Sheep fail to produce
golden fleece. Science, 2003; 300(5628): 2015
2016.
61 Harris DP, Andrews AT, Wright G, et al. The
application of aqueous two-phase systems to
the purification of pharmaceutical proteins
from transgenic sheep milk. Bioseparation,
1997; 7(1): 3137.
62 Chen SH, Vaught TD, Monahan JA, et al. Effi-
cient production of transgenic cloned calves
using preimplantation screening. Biol. Re-
prod., 2002; 67(5): 14881492.
63 Positive clinical trial result on BSSL. http://
www.ppl-therapeutics.com/news/
news_2_content_15.asp 2001.
64 Van Cott K, Lubon H, Russell C, et al. Pheno-
typic and genotypic stability of multiple lines
of transgenic pigs expressing recombinant hu-
man protein C. Transgenic Res., 1997; 6(3):
203212.
65 Van Cott K, Lubon H, Gwazdauskas F, et al.
Recombinant human protein C expression in
the milk of transgenic pigs and the effect on
endogenous milk immunoglobulin and trans-
ferrin levels. Transgenic Res., 2001; 10(1): 43
51.
66 Uusi M, Hyttinen J, Korhonen V, et al. Bovine
as1-casein gene sequences direct high level ex-
pression of human granulocyte-macrophage
colony-stimulating factor in the milk of trans-
genic mice. Transgenic Res., 1997; 6(1): 75
84.
67 Hematech Scientific Publications. http://
www.hematech.com/hematech/publications/
default.asp 2004.
68 Cloned Cows Generate Human Antibodies.
Nature Biotechnol., 2002; 10: 1038.
69 Therapeutic Human Polyclonals Products.
http://www.polyclonals.com/index.html 2004.
70 Birse D, Lacroix D, Dyck M, et al. Biothera-
peutics from Transgenic Porcine Sources: Bio-
processing Approaches and Challenges. Bio-
Processing J., 2004; March/April: 3744.
71 Ryoo Z, Kim M, Kim K, et al. Expression of
recombinant human granulocyte macrophage-
colony stimulating factor (hGM-CSF) in
mouse urine. Transgenic Res., 2001; 10(3):
193200.
72 Utilizing the Bladder as a Bioreactor in Trans-
genic Animals. http://www.med.nyu.edu/OIL/
Biotech/BSun.html 2004.
73 Kerr DE, Liang F, Bondioli KR, et al. The
bladder as a bioreactor: urothelium production
and secretion of growth hormone into urine.
Nat Biotechnol. 1998; 16(1): 7579.
74 Advanced Bionutrition Crustacean Expression
System. http://www.advancedbionutrition.-
com/html/prod_platforms.html#crustacean
2004.
75 Chesapeake PERL Technology. http://
www.c-perl.com/2004.
76 Advanced Cell Technology Scientific Publica-
tions. http://www.advancedcell.com/
Scientific-papers.htm 2004.
77 Nextran Information. http://www.baxter.ca/
htdocs/en/doctors/renal_therapies/
nextran.html 2004.
78 AviGenics Technology Overview. http://
www.avigenics.com/services.htm 2004.
79 Rapp JC, Harvey AJ, Speksnijder GL, et al.
Biologically active human interferon alpha-2b
produced in the egg white of transgenic hens.
Transgenic Res., 2003; 12(5): 569575.
80 BioAgri Products. http://
www.bioagricorp.com/strategic.aspx 2004.
81 Alper J. Biotechnology. Hatching the golden
egg: a new way to make drugs. Science, 2003;
300(5620): 729730.
82 Chang K, Qian J, Jiang M, et al. Effective gen-
eration of transgenic pigs and mice by linker
based sperm-mediated gene transfer. BMC
Biotechnol., 2002; 2(1): 5.
83 Coghlan A. Big breakfast Crack open an egg
and cure a disease. New Scientist Magazine,
1999; 11.
84 GenWay Biotech Products. http://
www.genwaybio.com/
DesktopPage.aspx?TabID=3329&Lang=en-US
2004.
85 Etches RJ, Verrinder Gibbins AM. Strategies
for the production of transgenic chicken.
Methods Mol. Biol., 1997; 62: 433450.
86 Linda Foster Benedict. Transforming Chickens
to Lay Golden Eggs. Louisiana Agriculture
Magazine OnLine 2003; 46(4).
87 TranXenoGen Technology. http://www.tranxe-
nogen.com/NewFiles/background.html 2004.
88 Viragen Avian Technology. http://
www.viragen.com/avian_intro.htm 2004.
89 Vivalis Transgenics Information. http://
www.vivalis.com/ 2004.
90 Pain B, Chenevier P, Samarut J. Chicken em-
bryonic stem cells and transgenic strategies.
Cells Tissues Organs, 1999; 165(3/4): 212219.
91 Menassa R, Kennette W, Nguyen V, et al. Sub-
cellular targeting of human interleukin-10 in
plants. J. Biotechnol., 2004; 108(2): 179183.
92 Pacific NorthWest Laboratories Transgenic
Plants Patents. http://availabletechnolo-
gies.pnl.gov/biomedical/trans.stm 2004.
93 Dai Z, Hooker BS, Anderson DB, et al. Ex-
pression of Acidothermus cellulolyticus endoglu-
canase E1 in transgenic tobacco: biochemical
characteristics and physiological effects. Trans-
genic Res., 2000; 9(1): 4354.
94 Hooker B, Dai Z, Gao J, et al., inventors;
Transgenic Plant-derived Human Blood
Coagulation Factors. 1999. PCT-World Intellec-
tual Property Organization WO 99/58699.
95 Plants make blood coagulants. Chemical &
Engineering News 1999; 44.
96 Hooker B, Dai Z, Kingsley M, et al., inventors;
Method for Producing Human Growth Factors
from Whole Plants or Plant Cell Culture.
1998. PCT-World Intellectual Property Organi-
zation WO98/21348.
97 Gao J, Hooker BS, Anderson DB. Expression
of functional human coagulation factor XIII
A-domain in plant cell suspensions and whole
plants. Protein Expr. Purif., 2004; 37(1): 89
96.
98 Hooker B. Two Crops, One Plant. BioPharm.,
1999; 2930.
99 Warzecha H, Mason HS. Benefits and risks
of antibody and vaccine production in trans-
genic plants. J. Plant Physiol., 2003; 160(7):
755764.
100 Smith ML, Richter L, Arntzen CJ, et al.
Structural characterization of plant-derived
hepatitis B surface antigen employed in oral
immunization studies. Vaccine, 2003;
21(25/26): 40114021.
101 Warzecha H, Mason HS, Lane C, et al. Oral
immunogenicity of human papillomavirus-
like particles expressed in potato. J. Virol.,
2003; 77(16): 87028711.
102 Walmsley AM, Alvarez ML, Jin Y, et al. Ex-
pression of the B subunit of Escherichia coli
heat-labile enterotoxin as a fusion protein in
transgenic tomato. Plant Cell Rep., 2003;
21(10): 10201026.
103 Axis Genetics and Biomira sign research col-
laboration for the development of a cancer
vaccine. http://www.biomira.com/news/
detailNewsRelease/89/:November 9, 1998.
104 Ramirez N, Ayala M, Lorenzo D, et al. Ex-
pression of a single-chain Fv antibody frag-
ment specific for the hepatitis B surface
antigen in transgenic tobacco plants. Trans-
genic Res., 2002; 11(1): 6164.
105 Lopez L, Muller R, Balmori E, et al. Molecu-
lar cloning and nucleotide sequence of the
coat protein gene of a Cuban isolate of pota-
to leafroll virus and its expression in Escheri-
chia coli. Virus Genes, 1994; 9(1): 7783.
106 Carrillo C, Wigdorovitz A, Trono K, et al. In-
duction of a virus-specific antibody response
to foot and mouth disease virus using the
structural protein VP1 expressed in trans-
genic potato plants. Viral Immunol., 2001;
14(1): 4957.
107 Gil F, Brun A, Wigdorovitz A, et al. High-
yield expression of a viral peptide vaccine in
transgenic plants. FEBS Lett., 2001; 488(1/2):
1317.
108 Gomez N, Wigdorovitz A, Castanon S, et al.
Oral immunogenicity of the plant derived
spike protein from swine-transmissible gas-
troenteritis coronavirus. Arch. Virol., 2000;
145(8): 17251732.
109 Wigdorovitz A, Carrillo C, Dus Santos MJ, et
al. Induction of a protective antibody re-
sponse to foot and mouth disease virus in
mice following oral or parenteral immuniza-
tion with alfalfa transgenic plants expressing
the viral structural protein VP1 Virology,
1999; 255(2): 347353.
110 Carrillo C, Wigdorovitz A, Oliveros JC, et al.
Protective immune response to foot-and-
mouth disease virus with VP1 expressed in
References 879
transgenic plants. J. Virol., 1998; 72(2):
16881690.
111 Chlorogen Technology. http://www.chloro-
gen.com/technology.htm 2004.
112 Daniell H, Khan MS, Allison L. Milestones
in chloroplast genetic engineering: an envi-
ronmentally friendly era in biotechnology.
Trends Plant Sci., 2002; 7(2): 8491.
113 Daniell H, Lee SB, Panchal T, et al. Expres-
sion of the native cholera toxin B subunit
gene and assembly as functional oligomers
in transgenic tobacco chloroplasts. J. Mol.
Biol. 2001; 311(5): 10011009.
114 Rachel Melcer. Creve Coeur Startup Devel-
ops Human Plasma from Tobacco. St. Louis
Post-Dispatch. June 20, 2003.
115 Kang TJ, Loc NH, Jang MO, et al. Expres-
sion of the B subunit of E. coli heat-labile
enterotoxin in the chloroplasts of plants and
its characterization. Transgenic Res., 2003;
12(6): 683691.
116 Cobento Biotech products. http://
www.cobento.com/2004.
117 Bouquin T, Thomsen M, Nielsen L, et al.
Human Anti-Rhesus D IgG1 Antibody Pro-
duced in Transgenic Plants. Transgenic Res.,
2002; 11(2): 115122.
118 Tavladoraki P, Benvenuto E, Trinca S, et al.
Transgenic plants expressing a functional
single-chain Fv antibody are specifically pro-
tected from virus attack. Nature, 1993;
366(6454): 469472.
119 Benvenuto E, Ordas RJ, Tavazza R, et al.
Phytoantibodies: a general vector for the ex-
pression of immunoglobulin domains in
transgenic plants. Plant Mol. Biol., 1991;
17(4): 865874.
120 Vine ND, Drake P, Hiatt A, et al. Assembly
and plasma membrane targeting of recombi-
nant immunoglobulin chains in plants with
a murine immunoglobulin transmembrane
sequence. Plant Mol. Biol., 2001; 45(2): 159
167.
121 Ma JK, Hiatt A, Hein M, et al. Generation
and assembly of secretory antibodies in
plants. Science, 1995; 268(5211): 716719.
122 Ma JK, Lehner T, Stabila P, et al. Assembly
of monoclonal antibodies with IgG1 and IgA
heavy chain domains in transgenic tobacco
plants. Eur. J. Immunol., 1994; 24(1): 131
138.
123 Hiatt A, Ma JK. Monoclonal antibody engi-
neering in plants. FEBS Lett., 1992; 307(1):
7175.
124 Hiatt A, Cafferkey R, Bowdish K. Production
of antibodies in transgenic plants. Nature,
1989; 342(6245): 7678.
125 ERA Biotech Technology. http://
www.eraplantech.com/ang/zera.html 2004.
126 Yang IC, Iommarini JP, Becker DK, et al. A
promoter derived from taro bacilliform bad-
navirus drives strong expression in trans-
genic banana and tobacco plants. Plant Cell
Rep., 2003; 21(12): 11991206.
127 Farmacule Technology. http://
www.farmacule.com/technology.htm#T2
2004.
128 Fraunhofer Center for Molecular Biology
Technology. http://www.fraunhofer-cmb.org/
index.cfm?act=technology 2004.
129 Stoger E, Schillberg S, Twyman RM, et al.
Antibody production in transgenic plants.
Methods Mol. Biol., 2004; 248: 301318.
130 Fischer R, Schumann D, Zimmermann S, et
al. Expression and characterization of bispe-
cific single-chain Fv fragments produced in
transgenic plants. Eur. J. Biochem., 1999;
262(3): 810816.
131 de Zoeten GA, Penswick JR, Horisberger
MA, et al. The expression, localization, and
effect of a human interferon in plants. Virol-
ogy, 1989; 172(1): 213222.
132 Peeters K, De Wilde C, De Jaeger G, et al.
Production of antibodies and antibody frag-
ments in plants. Vaccine, 2001; 19(1719):
27562761.
133 Peeters K, De Wilde C, Depicker A. Highly
efficient targeting and accumulation of a
F(ab) fragment within the secretory pathway
and apoplast of Arabidopsis thaliana. Eur. J.
Biochem., 2001; 268(15): 42514260.
134 De Jaeger G, De Wilde C, Eeckhout D, et al.
The plantibody approach: expression of anti-
body genes in plants to modulate plant me-
tabolism or to obtain pathogen resistance.
Plant Mol. Biol., 2000; 43(4): 419428.
135 De Jaeger G, Buys E, Eeckhout D, et al.
High level accumulation of single-chain vari-
able fragments in the cytosol of transgenic
Petunia hybrida. Eur. J. Biochem., 1999;
259(1/2): 426434.
136 Bruyns AM, De Jaeger G, De Neve M, et al.
Bacterial and plant-produced scFv proteins
have similar antigen-binding properties.
FEBS Lett., 1996; 386(1): 510.
137 De Neve M, De Loose M, Jacobs A, et al. As-
sembly of an antibody and its derived anti-
body fragment in Nicotiana and Arabidopsis.
Transgenic Res., 1993; 2(4): 227237.
138 Edelbaum O, Stein D, Holland N, et al. Ex-
pression of active human interferon-beta in
transgenic plants. J. Interferon Res., 1992;
12(6): 449453.
139 Rosenberg N, Reichman M, Gera A, et al.
Antiviral activity of natural and recombinant
human leukocyte interferons in tobacco pro-
toplasts. Virology, 1985; 140(1): 173178.
140 Ohya K, Itchoda N, Ohashi K, et al. Expres-
sion of biologically active human tumor ne-
crosis factor-alpha in transgenic potato plant.
J. Interferon Cytokine Res., 2002; 22(3):
371378.
141 Ohya K, Matsumura T, Ohashi K, et al. Ex-
pression of two subtypes of human IFN-al-
pha in transgenic potato plants. J. Interferon
Cytokine Res., 2001; 21(8): 595602.
142 Icon Genetics Products. http://
www.icongenetics.com/html/tech3_6.htm
2004.
143 Marillonnet S, Giritch A, Gils M, et al. In
planta engineering of viral RNA replicons:
efficient assembly by recombination of DNA
modules delivered by Agrobacterium. Proc.
Natl. Acad. Sci. USA, 2004; 101(18): 6852
6857.
144 Gleba Y, Marillonnet S, Klimyuk V. Engi-
neering viral expression vectors for plants:
the full virus and the deconstructed virus
strategies. Curr. Opin. Plant Biol., 2004;
7(2): 182188.
145 Borisjuk NV, Borisjuk LG, Logendra S, et al.
Production of recombinant proteins in plant
root exudates. Nature Biotechnol., 1999;
17(5): 466469.
146 Biemelt S, Sonnewald U, Galmbacher P, et
al. Production of human papillomavirus type
16 virus-like particles in transgenic plants. J.
Virol., 2003; 77(17): 92119220.
147 Scheller J, Guhrs KH, Grosse F, et al. Pro-
duction of spider silk proteins in tobacco
and potato. Nature Biotechnol., 2001; 19(6):
573577.
148 Fiedler U, Conrad U. High-level production
and long-term storage of engineered anti-
bodies in transgenic tobacco seeds. Biotech-
nology (NY), 1995; 13(10): 10901093.
149 Artsaenko O, Peisker M, Zur NU, et al. Ex-
pression of a single-chain Fv antibody
against abscisic acid creates a wilty pheno-
type in transgenic tobacco. Plant J., 1995;
8(5): 745750.
150 Conrad U, Fiedler U. Expression of engi-
neered antibodies in plant cells. Plant Mol.
Biol., 1994; 26(4): 10231030.
151 Scheller J, Henggeler D, Viviani A, et al.
Purification of spider silk-elastin from trans-
genic plants and application for human
chondrocyte proliferation. Transgenic Res.,
2004; 13(1): 5157.
152 Byun K, Hong J, Lee C, et al., inventors;
KIST SK, assignee. Human interleukin-6
producing tobacco. 1996.
153 Matsumoto S, Ikura K, Ueda M, et al. Char-
acterization of a human glycoprotein (ery-
thropoietin) produced in cultured tobacco
cells. Plant Mol. Biol., 1995; 27(6): 1163
1172.
154 Matsumoto S, Ishii A, Ikura K, et al. Expres-
sion of human erythropoietin in cultured to-
bacco cells. Biosci. Biotechnol. Biochem.,
1993; 57(8): 12491252.
155 DAoust M, Busse U, et al. In: Fischer R,
Schillberg S (Eds), Molecular Farming:
Plant-made Pharmaceuticals and Technical
Proteins. Wiley-VCH; 2004.
156 Medicago Technology. http://www2.medica-
go.com/en/tech_platform/proficia/ 2004.
157 Busse ULVTSeVLP. In: L. Erickson W-JYJB,
Rymerson R (Eds), Molecular Farming of
Plants and Animals for Human and Veterin-
ary Medicine. Kluwer Academic Publishers;
2002.
158 DAoust M, Busse U, et al. In: Christou P,
Klee H (Eds), Handbook of Plant Biotech-
nology. Wiley & Sons; 2004.
159 Meristem Products. http://
www.meristem-therapeutics.com/ 2004.
160 Perret S, Merle C, Bernocco S, et al. Unhy-
droxylated triple helical collagen I produced
in transgenic plants provides new clues on
the role of hydroxyproline in collagen fold-
ing and fibril formation. J. Biol. Chem.,
2001; 276(47): 4369343698.
161 Samyn-Petit B, Wajda Dubos JP, Chirat F, et
al. Comparative analysis of the site-specific
N-glycosylation of human lactoferrin pro-
duced in maize and tobacco plants. Eur. J.
Biochem., 2003; 270(15): 32353242.
References 881
162 Mokrzycki-Issartel N, Bouchon B, Farrer S,
et al. A transient tobacco expression system
coupled to MALDI-TOF-MS allows validation
of the impact of differential targeting on
structure and activity of a recombinant ther-
apeutic glycoprotein produced in plants.
FEBS Lett., 2003; 552(2/3): 170176.
163 Merle C, Perret S, Lacour T, et al. Hydroxy-
lated human homotrimeric collagen I in
Agrobacterium tumefaciens-mediated transient
expression and in transgenic tobacco plant.
FEBS Lett., 2002; 515(13): 114118.
164 Samyn-Petit B, Gruber V, Flahaut C, et al.
N-glycosylation potential of maize: the hu-
man lactoferrin used as a model. Glycoconj
J., 2001; 18(7): 519527.
165 Theisen M. Production of recombinant
blood factors in transgenic plants. Adv. Exp.
Med. Biol., 1999; 464: 211220.
166 Gruber V, Berna P, Arnaud T, et al. Large-
scale production of a therapeutic protein in
transgenic tobacco plants: effect of subcellu-
lar targeting on quality of a recombinant
dog gastric lipase. Molecular Breeding, 2001;
7(4): 329340.
167 Ruggiero F, Exposito JY, Bournat P, et al.
Triple helix assembly and processing of hu-
man collagen produced in transgenic tobac-
co plants. FEBS Lett., 2000; 469(1): 132136.
168 Dieryck W, Pagnier J, Poyart C, et al. Hu-
man haemoglobin from transgenic tobacco.
Nature, 1997; 386(6620): 2930.
169 Sijmons PC, Dekker BM, Schrammeijer B,
et al. Production of correctly processed hu-
man serum albumin in transgenic plants.
Biotechnology (NY), 1990; 8(3): 217221.
170 Schlittler MR, Taylor DW, Russell DA, et al.
Production of human somatotropin from
corn seed. Am. Chem. Soc., 221st BIOT-019,
2001.
171 Baez J, Russell D, Craig J. Corn Seed Pro-
duction of Therapeutic Proteins Moves For-
ward. BioPharm., 2000; 3.
172 Baez J, Russell D. Characterization of hu-
man monoclonal antibodies produced in
plants. 219th ACS National Meeting, San
Francisco, CA, March 2630, 2000, BIOT-
018, 2000.
173 Staub JM, Garcia B, Graves J, et al. High-
yield production of a human therapeutic
protein in tobacco chloroplasts. Nature Bio-
technol., 2000; 18(3): 333338.
174 Russell DA. Feasibility of antibody produc-
tion in plants for human therapeutic use.
Curr. Top. Microbiol. Immunol., 1999; 240:
119138.
175 Zeitlin L, Olmsted SS, Moench TR, et al. A
humanized monoclonal antibody produced
in transgenic plants for immunoprotection
of the vagina against genital herpes. Nature
Biotechnol., 1998; 16(13): 13611364.
176 Francisco JA, Gawlak SL, Miller M, et al. Ex-
pression and characterization of bryodin 1
and a bryodin 1-based single-chain immuno-
toxin from tobacco cell culture. Bioconjug.
Chem., 1997; 8(5): 708713.
177 Olsen D, Yang C, Bodo M, et al. Recombi-
nant collagen and gelatin for drug delivery.
Adv. Drug Deliv. Rev., 2003; 55(12): 1547
1567.
178 Webster DE, Thomas MC, Strugnell RA, et
al. Appetising solutions: an edible vaccine
for measles. Med. J. Aust., 2002; 176(9):
434437.
179 Webster DE, Cooney ML, Huang Z, et al.
Successful boosting of a DNA measles im-
munization with an oral plant-derived
measles virus vaccine. J. Virol., 2002; 76(15):
79107912.
180 Huang Z, Dry I, Webster D, et al. Plant-de-
rived measles virus hemagglutinin protein
induces neutralizing antibodies in mice.
Vaccine, 2001; 19(15/16): 21632171.
181 Tawaiwa F, Katsumata K, Anzai H. Produc-
tion of human lactoferrin in transgenic
plants. 2nd International Molecular Farming
Conference; London, Ontario, Canada, 1999.
182 Takase K, Hagiwara K. Expression of human
alpha-lactalbumin in transgenic tobacco. J.
Biochem. (Tokyo), 1998; 123(3): 440444.
183 Higo K, Saito Y, Higo H. Expression of a
chemically synthesized gene for human epi-
dermal growth factor under the control of
cauliflower mosaic virus 35S promoter in
transgenic tobacco. Biosci. Biotechnol. Bio-
chem., 1993; 57(9): 14771481.
184 NexGen Molecular Pharming Program.
http://www.campnexgen.com/english/mf/
mf.htm 2004.
185 Allina S, Heng Lui Y, Him S, et al. Produc-
tion of a Canine oral papillomavirus vaccine
in Tobacco. 2nd International Molecular
Farming Conference; London, Ontario, Ca-
nada, 1999.
186 Phylogix Information. http://
www.phylogix.com/about_us.htm 2004.
187 Moore JG, Fuchs CA, Hata YS, et al. A new
lectin in red kidney beans called PvFRIL
stimulates proliferation of NIH 3T3 cells ex-
pressing the Flt3 receptor. Biochim. Biophys.
Acta, 2000; 1475(3): 216224.
188 Planet Biotechnology Publications. http://
www.planetbiotechnology.com/publica-
tions.html 2004.
189 Larrick JW, Thomas DW. Producing proteins
in transgenic plants and animals. Curr.
Opin. Biotechnol., 2001; 12: 411418.
190 Daniell H, Streatfield SJ, Wycoff K. Medical
molecular farming: production of antibodies,
biopharmaceuticals and edible vaccines in
plants. Trends Plant Sci., 2001; 6(5): 219
226.
191 Larrick JW, Yu L, Naftzger C, et al. Produc-
tion of secretory IgA antibodies in plants.
Biomol. Eng., 2001; 18(3): 8794.
226.
193 Larrick JW, Yu L, Chen J, et al. Production
of antibodies in transgenic plants. Res. Im-
munol., 1998; 149(6): 603608.
194 Ma JK, Hikmat BY, Wycoff K, et al. Charac-
terization of a recombinant plant monoclo-
nal secretory antibody and preventive immu-
notherapy in humans. Nature Med., 1998;
4(5): 601606.
195 Plantigen technology. http://www.lhsc.on.ca/
plantigen/science_tech.html 2004.
196 Ma S, Huang Y, Yin Z, et al. Induction of
oral tolerance to prevent diabetes with trans-
genic plants requires glutamic acid decar-
boxylase (GAD) and IL-4. Proc. Natl. Acad.
Sci. USA, 2004; 101(15): 56805685.
197 Menassa R, Kennette W, Nguyen V, et al.
Subcellular targeting of human interleukin-
10 in plants. J. Biotechnol., 2004; 108(2):
179183.
198 Ma S, Jevnikar AM. Autoantigens produced
in plants for oral tolerance therapy of auto-
immune diseases. Adv. Exp. Med. Biol.,
1999; 464: 179194.
199 Ma SW, Zhao DL, Yin ZQ, et al. Transgenic
plants expressing autoantigens fed to mice
to induce oral immune tolerance. Nature
Med., 1997; 3(7): 793796.
200 Kong Q, Richter L, Yang YF, et al. Oral im-
munization with hepatitis B surface antigen
expressed in transgenic plants. Proc. Natl.
Acad. Sci. USA, 2001; 98(20): 1153911544.
201 Ma JK, Drake PM, Christou P. The produc-
tion of recombinant pharmaceutical proteins
in plants. Nature Rev. Genet., 2003; 4(10):
794805.
202 Avesani L, Falorni A, Battista G, et al. Im-
proved in planta expression of the human is-
let autoantigen glutamic acid decarboxylase
(GAD65). Transgenic Res., 2003; 12(2): 203
212.
203 Tuboly T, Yu W, Bailey A, et al. Immuno-
genicity of porcine transmissible gastroente-
ritis virus spike protein expressed in plants.
Vaccine, 2000; 18(19): 20232028.
204 Du S, Yu W, DeLange C, et al. Expression of
Porcine Epidermal Growth Factor in Plants.
2nd International Molecular Farming Con-
ference; London, Ontario, Canada, 1999.
205 Desai UA, Sur G, Daunert S, et al. Expres-
sion and affinity purification of recombinant
proteins from plants. Protein Expr. Purif.,
2002; 25(1): 195202.
206 Li Q, Lawrence CB, Xing HY, et al. En-
hanced disease resistance conferred by ex-
pression of an antimicrobial magainin ana-
log in transgenic tobacco. Planta, 2001;
212(4): 635639.
207 Rigano MM, Alvarez ML, Pinkhasov J, et al.
Production of a fusion protein consisting of
the enterotoxigenic Escherichia coli heat-labile
toxin B subunit and a tuberculosis antigen
in Arabidopsis thaliana. Plant Cell Rep.,
2004; 22(7): 502508.
208 Rigano MM, Sala F, Arntzen CJ, et al. Tar-
geting of plant-derived vaccine antigens to
immunoresponsive mucosal sites. Vaccine,
2003; 21(7/8): 809811.
209 Sala F, Manuela RM, Barbante A, et al. Vac-
cine antigen production in transgenic plants:
strategies, gene constructs and perspectives.
Vaccine, 2003; 21(7/8): 803808.
210 Franconi R, Di Bonito P, Dibello F, et al.
Plant-derived human papillomavirus 16 E7
oncoprotein induces immune response and
specific tumor protection. Cancer Res., 2002;
62(13): 36543658.
211 Franconi R, Roggero P, Pirazzi P, et al.
Functional expression in bacteria and plants
of an scFv antibody fragment against topo-
References 883
viruses. Immunotechnology, 1999; 4(3/4):
189201.
212 Bakker H, Bardor M, Molthoff JW, et al. Ga-
lactose-extended glycans of antibodies pro-
duced by transgenic plants. Proc. Natl. Acad.
Sci. USA, 2001; 98(5): 28992904.
213 Schouten A, Roosien J, de Boer JM, et al.
Improving scFv antibody expression levels in
the plant cytosol. FEBS Lett., 1997; 415(2):
235241.
214 Schouten A, Roosien J, van Engelen FA, et al.
The C-terminal KDEL sequence increases the
expression level of a single-chain antibody
designed to be targeted to both the cytosol
and the secretory pathway in transgenic to-
bacco. Plant Mol. Biol., 1996; 30(4): 781793.
215 van Engelen FA, Schouten A, Molthoff JW,
et al. Coordinate expression of antibody sub-
unit genes yields high levels of functional
antibodies in roots of transgenic tobacco.
Plant Mol. Biol., 1994; 26(6): 17011710.
216 Elbers I, Stoopen G, Bakker H, et al. Influ-
ence of growth conditions and developmen-
tal stage on N-glycan heterogeneity of trans-
genic immunoglobulin G and endogenous
proteins in tobacco leaves. Plant Physiol.,
2001; 126: 13141322.
217 Zhang G, Leung C, Murdin L, et al. In plan-
ta expression of HIV-1 p24 protein using an
RNA plant virus-based expression vector.
Mol. Biotechnol., 2000; 14(2): 99107.
218 Agrenvec Information. http://agrenvec.com/
eng/ 2004.
219 Sanchez F, Wang X, Jenner CE, et al.
Strains of Turnip mosaic potyvirus as de-
fined by the molecular analysis of the coat
protein gene of the virus. Virus Res., 2003;
94(1): 3343.
220 Jenner CE, Sanchez F, Nettleship SB, et al.
The cylindrical inclusion gene of Turnip mo-
saic virus encodes a pathogenic determinant
to the Brassica resistance gene TuRB01. Mol.
Plant Microbe Interact., 2000; 13(10): 1102
1108.
221 Roggero P, Ciuffo M, Benvenuto E, et al.
The expression of a single-chain Fv antibody
fragment in different plant hosts and tissues
by using Potato virus X as a vector. Protein
Expr. Purif., 2001; 22(1): 7074.
222 Yusibov V, Hooper DC, Spitsin SV, et al. Ex-
pression in plants and immunogenicity of
plant virus-based experimental rabies vac-
cine. Vaccine, 2002; 20(25/26): 31553164.
223 Koprowski H, Yusibov V. The green revolu-
tion: plants as heterologous expression vec-
tors. Vaccine, 2001; 19(1719): 27352741.
224 Belanger H, Fleysh N, Cox S, et al. Human
respiratory syncytial virus vaccine antigen
produced in plants. FASEB J., 2000; 14(14):
23232328.
225 Yusibov V, Shivprasad S, Turpen TH, et al.
Plant viral vectors based on tobamoviruses.
8194.
226 Kapusta J, Modelska A, Figlerowicz M, et al.
A plant-derived edible vaccine against hepati-
tis B virus. FASEB J., 1999; 13(13): 1796
1799.
227 Yusibov V, Shivprasad S, Turpen TH, et al.
Plant viral vectors based on tobamoviruses.
8194.
228 Verch T, Yusibov V, Koprowski H. Expres-
sion and assembly of a full-length monoclo-
nal antibody in plants using a plant virus
vector. J. Immunol. Methods, 1998; 220(1/2):
6975.
229 Modelska A, Dietzschold B, Sleysh N, et al.
Immunization against rabies with plant-de-
rived antigen. Proc. Natl. Acad. Sci. USA,
1998; 95(5): 24812485.
230 Fraunhofer Vaccine Development Technol-
ogy. http://www.fraunhofer-cmb.org/
index.cfm?act=tech_VaccineDevelopment
2004.
231 Marillonnet S, Giritch A, Gils M, et al. In
planta engineering of viral RNA replicons:
efficient assembly by recombination of DNA
modules delivered by Agrobacterium. Proc.
Natl. Acad. Sci. USA, 2004; 101(18): 6852
6857.
232 Crabtree P. Rincon Pharmaceuticals. San
Diego, Union-Tribune, May 20, 2005.
233 Fischer R, Schillberg S. Molecular Farming
Plant made pharamaceuticals and techni-
cal proteins. Weinheim, Wiley-VCH, 2004.
234 Gelderman M, Oliver K, Yazdani A, et al.
Preclinical Studies with Plant-Produced
alpha-Galactosidase A in Fabry Mice Show
Potential for Replacement Therapy. NA. Pre-
clinica, 2004; 2(1): 6774.
235 Kumagai MH, Donson J, Della-Cioppa G, et
al. Rapid, high-level expression of glycosy-
lated rice alpha-amylase in transfected plants
by an RNA viral vector. Gene, 2000; 245(1):
169174.
236 McCormick AA, Kumagai MH, Hanley K, et
al. Rapid production of specific vaccines for
lymphoma by expression of the tumor-de-
rived single-chain Fv epitopes in tobacco
plants. Proc. Natl. Acad. Sci. USA, 1999;
96(2): 703708.
237 Turpen TH, Reinl SJ, Charoenvit Y, et al.
Malarial epitopes expressed on the surface
of recombinant tobacco mosaic virus. Bio-
technology (NY), 1995; 13(1): 5357.
238 Large Scale Biology Corporation Announces
Biomanufacturing and Commercial Distribu-
tion Agreement with Sigma-Aldrich Cor-
poration. March 6, 04 Wall Street Reporter
2004.
239 Joint NIH And Large Scale Biology Corpora-
tion (LSBC) Studies With Plant-Produced Al-
pha-Galactosidase Show Its Potential For En-
zyme Replacement Therapy In Fabry Disease.
http://www.biospace.com/news_story.cfm?
StoryID=14899320&full=1 2004.
240 Tacket CO, Mason HS, Losonsky G, et al.
Human immune responses to a novel Nor-
walk virus vaccine delivered in transgenic
potatoes. J. Infect. Dis., 2000; 182(1): 302
305.
241 Dai Z, Hooker BS, Anderson DB, et al. Ex-
pression of Acidothermus cellulolyticus endo-
glucanase E1 in transgenic tobacco: bio-
chemical characteristics and physiological ef-
fects. Transgenic Res., 2000; 9(1): 4354.
242 Farran I, Sanchez-Serrano JJ, Medina JF, et
al. Targeted expression of human serum al-
bumin to potato tubers. Transgenic Res.,
2002; 11(4): 337346.
243 Chong DK, Langridge WH. Expression of
full-length bioactive antimicrobial human
lactoferrin in potato plants. Transgenic Res.,
2000; 9(1): 7178.
244 Arakawa T, Yu J, Chong DK, et al. A plant-
based cholera toxin B subunit-insulin fusion
protein protects against the development of
autoimmune diabetes. Nature Biotechnol.,
1998; 16(10): 934938.
245 Arakawa T, Langridge WH. Plants are not
just passive creatures! Nature Med., 1998;
4(5): 550551.
246 Arakawa T, Chong DK, Langridge WH. Effi-
cacy of a food plant-based oral cholera toxin
B subunit vaccine. Nature Biotechnol., 1998;
16(3): 292297.
247 Arakawa T, Chong DK, Merritt JL, et al. Ex-
pression of cholera toxin B subunit oligo-
mers in transgenic potato plants. Transgenic
Res., 1997; 6(6): 403413.
248 During K, Hippe S, Kreuzaler F, et al. Syn-
thesis and self-assembly of a functional
monoclonal antibody in transgenic Nicotiana
tabacum. Plant Mol. Biol., 1990; 15(2): 281
293.
249 Klaus During. Engineered storage organs as
bioreactors for protein production. 2nd Inter-
national Molecular Farming Conference;
London, Ontario, Canada, 1999.
250 Collins S. Potato the medical factory of to-
morrow. New Zealand Herald. April 23, 2003.
251 Novoplant Technology. http://
www.novoplant.de/ 2004.
252 Planton technology. http://www.planton.de/
en/index.html 2004.
253 Malaysia Rubber Institute Biotechnology and
Strategic Research Unit. http://
www.lgm.gov.my/r&d/bsru/transgenic.html
2004.
254 Arokiaraj P, Rueker F, Carter D, et al. Ex-
pression of Human Serum Albumin in the
laticifers of Transgenic Hevea brasiliensis.
255 Yeang, H. Engineering Crop Plants for In-
dustrial End Uses. London, UK: Portland
Press; 1998: 5564.
256 Rod Santa Ana III. Medical proteins from
sugarcane. Farm Press online, USA 2003.
257 Michael Smith. Norwalk virus vaccine grown
in tomatoes. United Press International.
February 16, 2003.
258 Sandhu JS, Krasnyanski SF, Domier LL, et
al. Oral immunization of mice with trans-
genic tomato fruit expressing respiratory
syncytial virus-F protein induces a systemic
immune response. Transgenic Res., 2000;
9(2): 127135.
259 Jani D, Singh NK, Bhattacharya S, et al.
Studies on the immunogenic potential of
plant-expressed cholera toxin B subunit.
Plant Cell Rep., 2004; 22(7): 471477.
260 Jani D, Meena LS, Rizwan-ul-Haq QM, et al.
Expression of cholera toxin B subunit in
transgenic tomato plants. Transgenic Res.,
2002; 11(5): 447454.
261 Arazi T, Lee HP, Huang PL, et al. Production
of antiviral and antitumor proteins MAP30
and GAP31 in cucurbits using the plant virus
vector ZYMV-AGII. Biochem. Biophys. Res.
Commun., 2002; 292(2): 441448.
References 885
262 Cropdesign Technology. http://
www.cropdesign.com 2004.
263 Dow Plant Technology. http://
www.dow.com/plantbio 2004.
264 Frigerio L, Vine ND, Pedrazzini E, et al. As-
sembly, secretion, and vacuolar delivery of a
hybrid immunoglobulin in plants. Plant
Physiol., 2000; 123(4): 14831494.
265 Company of the Month: Epicyte. http://
www.biotechjournal.com/Journal/Jun2001/
juneartA2001.pdf 2001; San Diego Biotech
Journal.
266 Stoger E, Vaquero C, Torres E, et al. Cereal
crops as viable production and storage sys-
tems for pharmaceutical scFv antibodies.
Plant Mol. Biol., 2000; 42(4): 583590.
267 Perrin Y, Vasquero C, Gerrard I, et al. Trans-
genic pea seeds as bioreactors for the pro-
duction of a single-chain Fv fragment anti-
bodies used in cancer diagnosis and therapy.
Molecular Breeding, 2000; 6: 345352.
268 Vandekerckhove J, Van Damme J, Van Lijse-
betten M, et al. Enkephalins produced in
transgenic plants using modified 2S seed
storage proteins. Bio/Technology, 1989; 7:
929932.
269 De Jaeger G, Scheffer S, Jacobs A, et al.
Evaluation of Phaseolus vulgaris arcelin-5I
gene for single-chain antibody production in
transgenic plants. European Society Plant
Physiology Annual Meeting, Helsinki, Fin-
land, 2001.
270 Mkinen K. Production of recombinant gela-
tin in barley. http://www.honeybee.helsinki.
fi/mmsbl/Biokemlab/makinen/research.html
2004.
271 Chikwamba R, Cunnick J, Hathaway D, et
al. A functional antigen in a practical crop:
LT-B producing maize protects mice against
Escherichia coli heat labile enterotoxin (LT)
and cholera toxin (CT). Transgenic Res.,
2002; 11(5): 479493.
272 Yang SH, Moran DL, Jia HW, et al. Expres-
sion of a synthetic porcine alpha-lactalbumin
gene in the kernels of transgenic maize.
Transgenic Res., 2002; 11(1): 1120.
273 Eudes F, Gilbert S, Acharya S, et al. Expres-
sion of a Bovine Virus Protein in Barley: a
Potential Edible Vaccine. 2nd International
Molecular Farming Conference; London,
Ontario, Canada, 1999.
274 Maltagene Technology. http://www.malta-
gen.de/index-english.htm 2004.
275 Mokrzycki-Issartel N, Bouchon B, Farrer S,
et al. A transient tobacco expression system
coupled to MALDI-TOF-MS allows validation
of the impact of differential targeting on
structure and activity of a recombinant ther-
apeutic glycoprotein produced in plants.
FEBS Lett., 2003; 552(2/3): 170176.
276 Schlittler MR, Taylor DW, Russell DA, et al.
Production of human somatotropin from
corn seed. Am. Chem. Soc., 221st BIOT-019
2001.
277 Novoplant Technology. http://www.novo-
plant.de/ 2004.
278 Orf Genetics Technology. http://
www.orfgenetics.com/orf/wgorf.nsf/key2/
Technology 2004.
279 Bailey MR, Woodard SL, Callaway E, et al.
Improved recovery of active recombinant lac-
case from maize seed. Appl. Microbiol. Bio-
technol., 2004; 63(4): 390397.
280 Woodard SL, Mayor JM, Bailey MR, et al.
Maize (Zea mays)-derived bovine trypsin:
characterization of the first large-scale, com-
mercial protein product from transgenic
plants. Biotechnol. Appl. Biochem., 2003;
38(Pt 2): 123130.
281 Lamphear BJ, Streatfield SJ, Jilka JM, et al.
Delivery of subunit vaccines in maize seed.
J. Control. Release, 2002; 85(13): 169180.
282 Streatfield SJ, Jilka JM, Hood EE, et al.
Plant-based vaccines: unique advantages.
Vaccine, 2001; 19(1719): 27422748.
283 Azzoni AR, Kusnadi AR, Miranda EA, et al.
Recombinant aprotinin produced in trans-
genic corn seed: extraction and purification
studies. Biotechnol. Bioeng., 2002; 80(3):
268276.
284 Evangelista RL, Kusnadi AR, Howard JA, et
al. Process and economic evaluation of the
extraction and purification of recombinant
beta-glucuronidase from transgenic corn.
Biotechnol. Prog., 1998; 14(4): 607614.
285 Kusnadi AR, Hood EE, Witcher DR, et al.
Production and purification of two recombi-
nant proteins from transgenic corn. Biotech-
nol. Prog., 1998; 14(1): 149155.
286 Hood E. From green plants to industrial en-
zymes. Enzyme Microbial Technol., 2002;
30(3): 279283.
287 Lamphear BJ, Jilka JM, Kesl L, et al. A corn-
based delivery system for animal vaccines:
an oral transmissible gastroenteritis virus
vaccine boosts lactogenic immunity in
swine. Vaccine, 2004; 22(19): 24202424.
288 Horn ME, Pappu KM, Bailey MR, et al. Ad-
vantageous features of plant-based systems
for the development of HIV vaccines. J Drug
Target., 2003; 11(810): 539545.
289 Streatfield SJ, Lane JR, Brooks CA, et al. Corn
as a production system for human and ani-
mal vaccines. Vaccine, 2003; 21(7/8): 812815.
290 Kusnadi AR, Evangelista RL, Hood EE, et al.
Processing of transgenic corn seed and its
effect on the recovery of recombinant beta-
glucuronidase. Biotechnol. Bioeng., 1998;
60(1): 4452.
291 Hood E, Witcher DR, Maddock S, et al.
Commercial production of avidin from
transgenic maize: characterization of trans-
formant, production, processing, extraction,
and purification. Mol. Breed., 1997; 3: 291
306.
292 Nicholson L, Christou P, Ma J, et al. Expres-
sion of recombinant SIgA in cereals. 2nd In-
ternational Molecular Farming Conference;
London, Ontario, Canada, 1999.
293 Stoger E, Vaquero C, Torres E, et al. Cereal
crops as viable production and storage sys-
tems for pharmaceutical scFv antibodies.
Plant Mol. Biol., 2000; 42(4): 583590.
294 Moloney MM, Holbrook LA. Subcellular tar-
geting and purification of recombinant pro-
teins in plant production systems. Biotech-
nol. Genet. Eng. Rev., 1997; 14: 321336.
295 van Rooijen GJ, Moloney MM. Plant seed
oil-bodies as carriers for foreign proteins.
Biotechnology (NY), 1995; 13(1): 7277.
296 Parmenter DL, Boothe JG, van Rooijen GJ,
et al. Production of biologically active hiru-
din in plant seeds using oleosin partitioning.
Plant Mol. Biol., 1995; 29(6): 11671180.
297 Hogge L, Boothe J, Maloney M, et al. Struc-
ture Confirmation of Recombinant Hirudin
Using MALDI with C-Terminal Sequencing.
298 SunGene Technology. http://
www.sungene.de/ 2004.
299 Syngenta BioPharma. http://
www.syngenta.com/en/biopharma/2004.
300 Leite A, Kemper E, Bocaccorsi E, et al. Cor-
rectly Processed Human Growth Hormone
is Expressed in Seeds of Transgenic Tobacco
Plants. 2nd International Molecular Farming
Conference; London, Ontario, Canada, 1999.
301 Tackaberry ES, Prior F, Bell M, et al. In-
creased yield of heterologous viral glycopro-
tein in the seeds of homozygous transgenic
tobacco plants cultivated underground. Ge-
nome, 2003; 46(3): 521526.
302 Sardana RK, Alli Z, Dudani A, et al. Biologi-
cal activity of human granulocyte-macro-
phage colony stimulating factor is main-
tained in a fusion with seed glutelin peptide.
Transgenic Res., 2002; 11(5): 521531.
303 Wright KE, Prior F, Sardana R, et al. Sorting
of glycoprotein B from human cytomegalo-
virus to protein storage vesicles in seeds of
transgenic tobacco. Transgenic Res., 2001;
10(2): 177181.
304 Tackaberry ES, Dudani AK, Prior F, et al.
Development of biopharmaceuticals in plant
expression systems: cloning, expression and
immunological reactivity of human cytome-
galovirus glycoprotein B (UL55) in seeds of
transgenic tobacco. Vaccine, 1999; 17(23/24):
30203029.
305 Panahi M, Cheng X, Callagham M, et al.
Expression of human insulin-like growth
factor (IGF-I) in plants. 2nd International
Molecular Farming Conference; London,
Ontario, Canada, 1999.
306 Panahi M, Alli Z, Cheng X, et al. Recombi-
nant protein expression plasmids optimized
for industrial E. coli fermentation and plant
systems produce biologically active human
insulin-like growth factor-1 in transgenic
rice and tobacco plants. Transgenic Res.,
2004; 13(3): 245259.
307 Dalton R. California edges towards farming
drug-producing rice. Nature, 2004;
428(6983): 591.
308 Suzuki YA, Kelleher SL, Yalda D, et al. Ex-
pression, characterization, and biologic activ-
ity of recombinant human lactoferrin in rice.
J. Pediatr. Gastroenterol. Nutr., 2003; 36(2):
190199.
309 Humphrey BD, Huang N, Klasing KC. Rice
expressing lactoferrin and lysozyme has anti-
biotic-like properties when fed to chicks. J.
Nutr., 2002; 132(6): 12141218.
310 Torres E, Vaquero C, Nicholson L, et al. Rice
cell culture as an alternative production sys-
tem for functional diagnostic and therapeu-
tic antibodies. Transgenic Res., 1999; 8(6):
441449.
311 Huang J, Sutliff TD, Wu L, et al. Expression
and purification of functional human alpha-
References 887
1-Antitrypsin from cultured plant cells. Bio-
technol. Prog., 2001; 17(1): 126133.
312 Nandi S, Suzukib Y, Huang J, et al. Expres-
sion of human lactoferrin in transgenic rice
grains for the application in infant formula.
Plant Sci. 163(4): 713722.
313 Huang J, Nandi S, Wu L, et al. Expression
of natural antimicrobial human lysozyme in
rice grains. Mol. Breeding, 2004; 10(1/2):
8394.
314 Huang N. High-level Protein Expression Sys-
tem Uses Self-Pollinating Crops as Hosts.
BioProcess Int., 2004; 5459.
315 WSU Dieter Von Wettstein Research Inter-
est. http://molecular.biosciences.wsu.edu/
faculty/vonwettstein.html 2004.
316 Horvath H, Huang J, Wong O, et al. The
production of recombinant proteins in trans-
genic barley grains. Proc. Natl. Acad. Sci.
USA, 2000; 97(4): 19141919.
317 Greenovation Technology. http://
www.greenovation.com/Projects_Produc-
tion.htm 2004.
318 Gasdaska J, Spencer D, Dickey L. Advan-
tages of Therapeutic Protein Production in
the Aquatic Plant Lemna. BioProcessing J.,
2003; 4953.
319 Phycotransgenic Technology. http://
www.phycotransgenics.com/ 2004.
320 Franklin SE, Mayfield SP. Prospects for mo-
lecular farming in the green alga Chlamydo-
monas. Curr Opin. Plant Biol., 2004; 7(2):
159165.
321 Mayfield SP, Franklin SE, Lerner RA. Ex-
pression and assembly of a fully active anti-
body in algae. Proc. Natl. Acad. Sci. USA,
2003; 100(2): 438442.
322 Borisjuk NV, Borisjuk LG, Logendra S, et al.
Production of recombinant proteins in plant
root exudates. Nature Biotechnol., 1999;
17(5): 466469.
323 Gleba D, Borisjuk NV, Borisjuk LG, et al.
Use of plant roots for phytoremediation and
molecular farming. Proc. Natl. Acad. Sci.
USA, 1999; 96(11): 59735977.
324 Gaume A, Komarnytsky S, Borisjuk N, et al.
Rhizosecretion of recombinant proteins
from plant hairy roots. Plant Cell Rep.,
2003; 21(12): 11881193.
325 Medina-Bolivar F, Cramer C. Production of
recombinant proteins by hairy roots cultured
in plastic sleeve bioreactors. Methods Mol.
Biol., 2004; 267: 351364.
326 Medina-Bolivar F, Wright R, Funk V, et al.
A non-toxic lectin for antigen delivery of
plant-based mucosal vaccines. Vaccine, 2003;
21(9/10): 9971005.
327 Cramer CL, Boothe JG, Oishi KK. Trans-
genic plants for therapeutic proteins: linking
upstream and downstream strategies. Curr.
Top. Microbiol. Immunol., 1999; 240: 95
118.
328 Cramer CL, Weissenborn DL, Oishi KK,
et al. Bioproduction of human enzymes in
transgenic tobacco. Ann. N.Y. Acad. Sci.,
1996; 792: 6271.
329 Koivu K. Contained Plant Bioreactor. 4th In-
ternational Conference, Production & Eco-
nomics of Biopharmaceuticals. San Diego,
California, 2001.
330 Unicrop Publications. http://www.unicrop.fi/
index.jsp?p=74, 79, 88 2004.
331 Geert De Jaeger Research Team Plant Cell
Culture Program. http://www.vib.be/
Research/EN/Research+Departments/
Department+of+Plant+Systems+Biology/
Geert+De+Jaeger/2004.
332 Yano A, Maeda F, Takekoshi M. Transgenic
tobacco cells producing the human monoclo-
nal antibody to hepatitis B virus surface anti-
gen. J. Med. Virol., 2004; 73(2): 208215.
333 Information update on Phytoprotein Biotech
Pte Ltd. http://www.agenix.com/news/
biotechnews11apr1.htm 2001.
334 Protalix Technology. http://www.protalix.-
com/html/technology_technology.htm 2004.
335 ROOTec Technology. http://
www.pharmaceutical-map.com/
non-member/storefront_view.php?id=26
2004.
336 Tsoi BM, Doran PM. Effect of medium prop-
erties and additives on antibody stability and
accumulation in suspended plant cell cul-
tures. Biotechnol. Appl. Biochem., 2002;
35(Pt 3): 171180.
337 Sharp JM, Doran PM. Characterization of
monoclonal antibody fragments produced by
plant cells. Biotechnol. Bioeng., 2001; 73(5):
338346.
338 Doran PM. Foreign protein production in
plant tissue cultures. Curr. Opin. Biotech-
nol., 2000; 11: 199204.
339 Mison D, Curling J. The industrial produc-
tion costs of recombinant therapeutic pro-
teins expressed in transgenic corn. Bio-
Pharm, 2000; 4854.
340 Young M. Production of biopharmaceutical
proteins in the milk of transgenic animals.
BioPharm, 1997; 10(6): 3438.
341 Price B. The Capacity and Capability Crunch:
Myth or Reality. Plant-Made Pharmaceuticals
conference, Quebec, Canada, 2003.
342 Palmer C, Lubon H, McManaman J. Trans-
genic mice expressing recombinant human
protein C exhibit defects in lactation and im-
paired mammary gland development. Trans-
genic Res., 2003; 12(3): 283292.
343 Cabanes-Macheteau M, Fitchette-Laine AC,
Loutelier-Bourhis C, et al. N-Glycosylation of
a mouse IgG expressed in transgenic tobac-
co plants. Glycobiology, 1999; 9(4): 365372.
344 Gala FA, Morrison SL. V region carbohy-
drate and antibody expression. J. Immunol.,
2004; 172(9): 54895494.
345 Rifai A, Fadden K, Morrison SL, et al. The
N-glycans determine the differential blood
clearance and hepatic uptake of human im-
munoglobulin (Ig)A1 and IgA2 isotypes. J.
Exp. Med., 2000; 191(12): 21712182.
346 Chargelegue D, Vine N, van Dolleweerd C,
et al. A murine monoclonal antibody pro-
duced in transgenic plants with plant-specif-
ic glycans is not immunogenic in mice.
Transgenic Res., 2000; 9(3): 187194.
347 United States Department of Agriculture.
Field testing of plants engineered to produce
pharmaceutical and industrial compounds.
2003, pp. 1133711340. United States Feder-
al Register 68.
348 FDA-CBER. Points to Consider in the Manu-
facture and Testing of Therapeutic Proteins
for Human Use Derived from Transgenic
Animals. http://mbcr.bcm.tmc.edu/BEP/
ERMB/ptc_tga.html 1995.
349 Bair J. Food Value Chain Perspective. Corn-
Produced Pharmaceuticals and Industrials
Risk Assessment Symposium, sponsored by
the Biosafety Institute for Genetically Modi-
fied Agricultural Products Symposium,
Ames, Iowa, 2004.
350 Wall R, Kerr D, Bondioli K. Transgenic dairy
cattle: genetic engineering on a large scale.
J. Dairy Sci., 1997; 80: 22132224.
351 Masarik M, Kizek R, Kramer KJ, et al. Appli-
cation of avidin-biotin technology and ad-
sorptive transfer stripping square-wave vol-
tammetry for detection of DNA hybridiza-
tion and avidin in transgenic avidin maize.
Anal. Chem., 2003; 75(11): 26632669.
352 Kramer KJ, Morgan TD, Throne JE, et al.
Transgenic avidin maize is resistant to stor-
age insect pests. Nature Biotechnol., 2000;
18(6): 670674.
353 Grigorian AV, Shekhter AB, Tolstykh PI, et
al. [Proteolytic enzymes in the overall treat-
ment of wounds]. Khirurgiia (Mosk), 1979;
(8): 1923.
354 Yucel VE, Basmajian JV. Decubitus ulcers:
healing effect of an enzymatic spray. Arch.
Phys. Med. Rehabil., 1974; 55(11): 517519.
355 Roep BO, van den Engel NK, van Halteren
AG, et al. Modulation of autoimmunity to
beta-cell antigens by proteases. Diabetologia,
2002; 45(5): 686692.
356 Landis RC, Asimakopoulos G, Poullis M, et
al. The antithrombotic and anti-inflamma-
tory mechanisms of action of aprotinin.
Ann. Thorac. Surg., 2001; 72(6): 21692175.
357 Landis RC, Haskard DO, Taylor KM. New
anti-inflammatory and platelet-preserving ef-
fects of aprotinin. Ann. Thorac. Surg., 2001;
72(5): S1808S1813.
358 Belorgey D, Dirrig S, Amouric M, et al. In-
hibition of human pancreatic proteinases by
mucus proteinase inhibitor, eglin c and
aprotinin. Biochem. J., 1996; 313 (Pt 2):
555560.
359 Laxenaire MC, Dewachter P, Pecquet C.
[Allergic risk of aprotinin]. Ann. Fr. Anesth.
Reanim., 2000; 19(2): 96104.
360 Barthel T, Kula MR. Studies on the extrac-
tion of DesPro(2)-Val15-Leu17-aprotinin
from the culture broth of a recombinant
Saccharomyces cerevisiae. Bioseparation,
1992; 3(6): 365372.
361 GTC Biotherapeutics Reports Second Quar-
ter 2004 Financial Results. http://
www.transgenics.com/pressreleases/
pr080504.html 2004.
362 Winterbottom N, Kuo J, Nguyen K, et al.
Antigenic Responses to Bovine Thrombin
Exposure During Surgery: A Prospective
Study of 309 Patients. J. Appl. Res. Clin.
Exp. Ther., 2002; 2(1).
363 GTC Biotherapeutics Reports Fourth Quar-
ter and Year End 2003 Financial Results.
http://www.transgenics.com/pressreleases/
pr030304.html 2004.
364 Melcer R. Creve Coeur Startup Develops
Human Plasma from Tobacco. St. Louis
Post-Dispatch, June 20, 2003.
References 889
365 Lomas DA, Parfrey H. Alpha1-antitrypsin
deficiency. 4: Molecular pathophysiology.
Thorax, 2004; 59(6): 529535.
366 Chowanadisai W, Huang J, Huang N, et al.
Stability of recombinant human alpha-1-anti-
trypsin produced in rice in infant formula.
J. Nutr. Biochem., 2003; 14(7): 386393.
367 Flotte TR, Brantly ML, Spencer LT, et al.
Phase I trial of intramuscular injection of a
recombinant adeno-associated virus alpha 1-
antitrypsin (rAAV2-CB-hAAT) gene vector to
AAT-deficient adults. Hum. Gene Ther.,
2004; 15(1): 93128.
368 Hubbard RC, McElvaney NG, Sellers SE, et
al. Recombinant DNA-produced alpha 1-anti-
trypsin administered by aerosol augments
lower respiratory tract antineutrophil elas-
tase defenses in individuals with alpha 1-
antitrypsin deficiency. J. Clin. Invest., 1989;
84(4): 13491354.
369 Smerdon GR, Aves SJ, Walton EF. Produc-
tion of human gastric lipase in the fission
yeast Schizosaccharomyces pombe. Gene,
1995; 165(2): 313318.
370 Van den Hout JM, Kamphoven JH, Winkel
LP, et al. Long-term intravenous treatment
of Pompe disease with recombinant human
alpha-glucosidase from milk. Pediatrics,
2004; 113(5): e448e457.
371 Bijvoet AG, Van Hirtum H, Kroos MA, et al.
Human acid alpha-glucosidase from rabbit
milk has therapeutic effect in mice with gly-
cogen storage disease type II. Hum. Mol.
Genet., 1999; 8(12): 21452153.
372 Bijvoet AG, Kroos MA, Pieper FR, et al. Re-
combinant human acid alpha-glucosidase:
high level production in mouse milk, bio-
chemical characteristics, correction of en-
zyme deficiency in GSDII KO mice. Hum.
Mol. Genet., 1998; 7(11): 18151824.
373 Reno J. Performance of corn-seed derived
huNR-LU-10. 1998; IBC conference on
Transgenic Production, San Francisco,
September 1998.
374 Pogue G. Clinical and Pre-Clinical Applica-
tion of Plant-Derived Glycoproteins.
BIO2005 International Conference, San
Francisco, 2004.
375 Large Scale Biology and Planet Biotechnol-
ogy Announce Biomanufacturing Agree-
ment. http://www.lsbc.com/pdfs/Planet-
Bio.pdf 2004.
376 Pharming Group 2002 annual report. http://
www.pharming.com/downloads/Annual_Re-
port_2002.pdf 2004.
377 van Berkel PH, Welling MM, Geerts M, et
al. Large scale production of recombinant
human lactoferrin in the milk of transgenic
cows. Nature Biotechnol., 2002; 20(5): 484
487.
378 Weinberg ED. The therapeutic potential of
lactoferrin. Expert Opin. Investig. Drugs,
2003; 12(5): 841851.
379 Bethell DR, Huang J. Recombinant human
lactoferrin treatment for global health is-
sues: iron deficiency and acute diarrhea. Bio-
metals, 2004; 17(3): 337342.
380 Wolf JS, Li D, Taylor RJ, et al. Lactoferrin in-
hibits growth of malignant tumors of the
head and neck. ORL J. Otorhinolaryngol. Re-
lat. Spec., 2003; 65(5): 245249.
381 Cornish J, Callon KE, Naot D, et al. Lactofer-
rin is a potent regulator of bone cell activity
and increases bone formation in vivo. Endo-
crinology, 2004; 145(9): 43664374.
382 Huang J, Nandi S, Wu L, et al. Expression
of natural antimicrobial human lysozyme in
rice grains. Mol. Breeding, 2002; 10(1/2):
8394.
383 Nandi S, Suzukib Y, Huang J, et al. Expres-
sion of human lactoferrin in transgenic rice
grains for the application in infant formula.
Plant Sci., 2002; 163(4): 713722.
384 Sava G. Pharmacological aspects and thera-
peutic applications of lysozymes. EXS, 1996;
75: 433449.
385 Brock JH. The physiology of lactoferrin. Bio-
chem. Cell. Biol., 2002; 80(1): 16.
386 Singh PK, Parsek MR, Greenberg EP, et al.
A component of innate immunity prevents
bacterial biofilm development. Nature, 2002;
417(6888): 552555.
387 Markart P, Korfhagen TR, Weaver TE, et al.
Mouse lysozyme M is important in pulmo-
nary host defense against Klebsiella pneumo-
niae infection. Am. J. Respir. Crit. Care
Med., 2004; 169(4): 454458.
388 A plethora of hi-tech vaccines genetic, ed-
ible, sugar glass, and more. CVI Forum,
1999; (18): 523.
389 Arntzen CJ. Edible vaccines. Public Health
Rep., 1997; 112(3): 190197.
390 Azhar AM, Singh S, Anand KP, et al. Ex-
pression of protective antigen in transgenic
plants: a step towards edible vaccine against
anthrax. Biochem. Biophys. Res. Commun.,
2002; 299(3): 345351.
391 Bonetta L. Edible vaccines: not quite ready
for prime time. Nature Med., 2002; 8(2): 94.
392 Bonn D. Edible vaccines tackle mucosal in-
fections head on. Lancet Infect Dis., 2002;
2(5): 263.
393 Clough J. Edible vaccines against human pa-
pilloma virus. Drug Discov. Today, 2002;
7(17): 886887.
226.
395 Fooks AR. Development of oral vaccines for
human use. Curr. Opin. Mol. Ther., 2000;
2(1): 8086.
396 Keusch GT. The National Institutes of
Health agenda for international research in
micronutrient nutrition and infection inter-
actions. J. Infect. Dis., 2000; 182 Suppl 1:
S139S142.
397 Lauterslager TG, Hilgers LA. Efficacy of oral
administration and oral intake of edible vac-
cines. Immunol. Lett., 2002; 84(3): 185190.
398 Lauterslager TG, Florack DE, van der Wal
TJ, et al. Oral immunisation of naive and
primed animals with transgenic potato tu-
bers expressing LT-B. Vaccine, 2001;
19(1719): 27492755.
399 Wu YZ, Li JT, Mou ZR, et al. Oral immuni-
zation with rotavirus VP7 expressed in trans-
genic potatoes induced high titers of muco-
sal neutralizing IgA. Virology, 2003; 313(2):
337342.
400 Mason HS, Ball JM, Shi JJ, et al. Expression
of Norwalk virus capsid protein in trans-
genic tobacco and potato and its oral immu-
nogenicity in mice. Proc. Natl. Acad. Sci.
USA, 1996; 93(11): 53355340.
401 Curtiss R, Cardineau G, inventors; Washing-
ton University, Mycogen, assignees. Oral im-
munization by transgenic plants. US Patent
US 5, 654, 184; US 5679880; US 5686079.
1997.
402 Koga T, Oho T, Shimazaki Y, et al. Immuni-
zation against dental caries. Vaccine, 2002;
20(16): 20272044.
403 Ma JK. The caries vaccine: a growing pro-
spect. Dent. Update, 1999; 26(9): 374380.
404 Richter LJ, Thanavala Y, Arntzen CJ, et al.
Production of hepatitis B surface antigen in
transgenic plants for oral immunization. Na-
ture Biotechnol., 2000; 18(11): 11671171.
405 Kirk DD, Rempel R, Pinkhasov J, et al.
Application of Quillaja saponaria extracts as
adjuvants for plant-made vaccines. Expert
Opin. Biol. Ther., 2004; 4(6): 947958.
406 Kirk DD, Vonhof W, Eibner J, et al. Model
Production of a Potent Plant-Made Vaccine.
NFID Sixth Annual Conference on Vaccine
Research, 58 May, Arlington, VA, 2003.
407 Tacket CO, Mason HS, Losonsky G, et al.
Immunogenicity in humans of a recombinant
bacterial antigen delivered in a transgenic
potato. Nature Med., 1998; 4(5): 607609.
408 Mason HS, Haq TA, Clements JD, et al. Ed-
ible vaccine protects mice against Escherichia
coli heat-labile enterotoxin (LT): potatoes ex-
pressing a synthetic LT-B gene. Vaccine,
1998; 16(13): 13361343.
409 Lam D, Arntzen C, Mason H, inventors;
ProdiGene, assignee. Vaccines expressed in
plants. US Patent US 6034298. 2000.
410 Stiefelhagen P. [Oral vaccines a future pro-
spect! Hepatitis vaccination in the vegetable
market?]. MMW Fortschr. Med., 2003;
145(47): 10.
411 Smith ML, Mason HS, Shuler ML. Hepatitis
B surface antigen (HBsAg) expression in
plant cell culture: Kinetics of antigen accu-
mulation in batch culture and its intracellu-
lar form. Biotechnol. Bioeng., 2002; 80(7):
812822.
412 Smith ML, Keegan ME, Mason HS, et al.
Factors important in the extraction, stability
and in vitro assembly of the hepatitis B sur-
face antigen derived from recombinant plant
systems. Biotechnol. Prog., 2002; 18(3): 538
550.
413 Thanavala Y, Yang YF, Lyons P, et al. Immu-
nogenicity of transgenic plant-derived hepa-
titis B surface antigen. Proc. Natl. Acad. Sci.
USA, 1995; 92(8): 33583361.
414 Mason HS, Lam DM, Arntzen CJ. Expres-
sion of hepatitis B surface antigen in trans-
genic plants. Proc. Natl. Acad. Sci. USA,
1992; 89(24): 1174511749.
415 Edible vaccines for Helicobacter pylori and ro-
taviruses: Deprohealth project. http://
www.flair-flow.com/consumer-docs/
ffe63503.html 2004.
416 Nakata S. [Vaccine development for Norwalk
virus]. Nippon Rinsho, 2002; 60(6): 1222
1227.
References 891
417 Koprowski H. [Old and new prescriptions
for infectious diseases and the newest re-
cipes for biomedical products in plants].
Arch. Immunol. Ther. Exp. (Warsz), 2002;
50(6): 365369.
418 Bosch X. Tobacco fights back. Lancet Oncol.,
2004; 5(5): 264.
419 Varsani A, Williamson AL, Rose RC, et al.
Expression of Human papillomavirus type
16 major capsid protein in transgenic Nicoti-
ana tabacum cv. Xanthi. Arch. Virol., 2003;
148(9): 17711786.
420 Dow enters NIH research agreement to
develop rapid vaccine production system.
http://www.fraunhofer-cmb.org/index.cfm?
act=news 2004.
421 Hood EE, Jilka JM. Plant-based production
of xenogenic proteins. Curr. Opin. Biotech-
nol., 1999; 10(4): 382386.
422 Walmsley AM, Arntzen CJ. Plants for deliv-
ery of edible vaccines. Curr. Opin. Biotech-
nol., 2000; 11(2): 126129.
423 Touchette N. New Company to Make Drugs
from Plants. GNN News Alerts 2003.
424 Beck E. Researchers at Thomas Jefferson
University developed a way to make a safer
anthrax vaccine. United Press International.
March 12, 2003.
425 Huang JK, Li M. [Silk protein fiber biomate-
rials and tissue engineering]. Zhongguo Xiu
Fu Chong Jian Wai Ke Za Zhi, 2004; 18(2):
127130.
426 Williams D. Sows ears, silk purses and
goats milk: new production methods and
medical applications for silk. Med. Device
Technol., 2003; 14(5): 911.
427 Prunkard D, Cottingham I, Garner I, et al.
High-level expression of recombinant human
fibrinogen in the milk of transgenic mice.
Nature Biotechnol., 1996; 14(7): 867871.
428 Toman PD, Pieper F, Sakai N, et al. Produc-
tion of recombinant human type I procolla-
gen homotrimer in the mammary gland of
transgenic mice. Transgenic Res., 1999; 8(6):
415427.
429 John DC, Watson R, Kind AJ, et al. Expres-
sion of an engineered form of recombinant
procollagen in mouse milk. Nature Biotech-
nol., 1999; 17(4): 385389.
430 Tomita M, Munetsuna H, Sato T, et al.
Transgenic silkworms produce recombinant
human type III procollagen in cocoons. Na-
ture Biotechnol., 2003; 21(1): 5256.
431 Ruggiero F, Chanut H, Fichard A. Produc-
tion of Recombinant Collagen for Biomedi-
cal Devices. BioPharm, 2000; 3237.
432 Bellingham CM, Lillie MA, Gosline JM, et
al. Recombinant human elastin polypeptides
self-assemble into biomaterials with elastin-
like properties. Biopolymers, 2003; 70(4):
445455.
433 Bellingham CM, Woodhouse KA, Robson P,
et al. Self-aggregation characteristics of re-
combinantly expressed human elastin poly-
peptides. Biochim. Biophys. Acta, 2001;
1550(1): 619.
434 Keeley FW, Bellingham CM, Woodhouse
KA. Elastin as a self-organizing biomaterial:
use of recombinantly expressed human elas-
tin polypeptides as a model for investiga-
tions of structure and self-assembly of elas-
tin. Philos. Trans. R. Soc. Lond. B Biol. Sci.,
2002; 357(1418): 185189.
435 Woodhouse KA, Klement P, Chen V, et al.
Investigation of recombinant human elastin
polypeptides as non-thrombogenic coatings.
Biomaterials, 2004; 25(19): 45434553.
436 Yang G, Woodhouse KA, Yip CM. Substrate-
facilitated assembly of elastin-like peptides:
studies by variable-temperature in situ atom-
ic force microscopy. J. Am. Chem. Soc.,
2002; 124(36): 1064810649.
437 Bucchini L, Goldman LR. Starlink corn:
a risk analysis. Environ. Health Perspect.,
2002; 110(1): 513.
438 Benowitz S. Jefferson Scientistis Create
Plant Factories churning out Antibodies
against Tumor Cells (Press Release). Jeffer-
son University Hospital, Philadelphia, May
3, 2005.
Abstract
This chapter reviews progress and chal-
lenges in the area of production of recom-
binant proteins, in particular biopharma-
ceuticals, in plants. Different expression
platforms are summarized, including
those based on the use of transgenic,
transplastomic or transfected plants as
production hosts. The quality and yield of
recombinant proteins produced in and pu-
rified from plants, as well as progress in
clinical trials with plant-made pharmaceu-
tical proteins are described. The advan-
tages, limitations and biological safety as-
pects of plant-based production of biophar-
maceuticals are discussed.
Abbreviations
BSE bovine spongiform
encephalopathy
Bt Bacillus thuringiensis
CJD CreutzfeldJakob disease
EPSPS 5-enolpyruvylshikimate-3-
phosphate synthase
GAD glutamic acid decarboxylase
GD Gaucher disease
GM genetically modified
GUS b-glucuronidase
hVEGF human vascular endothelial
growth factor
IL interleukin
Mabs monoclonal antibodies
MHC major histocompatible complex
NPT neomycin phosphotransferase
PAT phosphinothricin acetyltransferase
PTGS post-transcriptional gene silencing
TMV tobacco mosaic virus
TSP total soluble protein
vCJD variant Creutzfeld-Jakob disease
6.1
Introduction
Numerous reviews concerning plant mo-
lecular farming have been published in
recent years [16]. Analysis of these re-
views and of recent research publications
shows a change of priorities in the per-
ceived advantages of plants as production
hosts for recombinant proteins. Initially,
the emphasis was on unlimited scalability
and low cost of plant-based production,
whereas yield and biosafety issues were
not properly addressed. However, the last
two parameters are crucial for determining
the economics and, consequently the
chances for commercial success of each
specific plant-based system.
893
6
Production of Recombinant Proteins in Plants
Victor Klimyuk, Sylvestre Marillonnet, Jrg Knblein, Michael McCaman, and Yuri Gleba
ISBN: 3-527-31184-X
A fresh boost was given to plant-based
molecular farming in recent years, as the
biopharmaceutical industry is trying to
eliminate manufacturing processes that
rely on production in animal cells due to
the possible contamination of these prod-
ucts by human pathogens such as bovine
spongiform encephalopathy (BSE) or
CreutzfeldJakob disease (CJD, vCJD) (see-
Part IV, Chapters 1, 2, 3, and 4).
In this chapter, we discuss different ex-
pression systems that are being developed.
We consider the potential of each system
by taking into account the impact of sev-
eral parameters on economics and regula-
tory acceptability of the system: productiv-
ity (absolute and relative yield), biological
safety (in particular transgene contain-
ment), scalability, versatility (ability to ac-
commodate diverse proteins and to express
a recombinant protein identical to the nat-
ural one), speed of research, development
and commercial scalability provided by
each of these systems.
6.2
Plant-based Expression Systems
Numerous plant-based expression systems
that differ by the type of expression cas-
sette and the location of such expression
cassettes within a plant host cell have been
developed. In the simplest way, the sys-
tems can be classified depending on how
the foreign genetic material is incorpo-
rated into a plant cell. For example, sys-
tems can be based either on expression of
heterologous sequences stably incorporated
into nuclear (Section 6.2.1) or plastid (Sec-
tion 6.2.2) genome, or on sequences tran-
siently expressed within plant cells (Sec-
tion 6.2.3).
6.2.1
Vectors Introduced via Stable
Nuclear Transformation
The majority of genetically modified (GM)
plants in use today have been produced by
nuclear transformation. The vectors de-
signed for such stable transformation
usually include an expression cassette un-
der control of a strong constitutive, induc-
ible or tissue-specific promoter. The use of
such promoters for expression of the pro-
tein of interest has been described in
many publications (Table 6.1). The most
commonly used promoters to drive tran-
scription in dicotyledonous plants are the
constitutive cauliflower mosaic virus
(CaMV) 35S promoter [7] and the tuber-
specific patatin promoter from potato [8
10]. Recently, a highly efficient system
based on the use of the promoter-termina-
tor of Chrysanthemum morifolium rbcS1
was reported, providing up to 10% of total
soluble protein (TSP) in tobacco leaf for
gusA expression [11]. The stress-inducible
peroxidase (SWPA2) promoter was suc-
cessfully used for expressing human lacto-
ferrin in cultured ginseng cells [12]. An-
other group of frequently used promoters
are those with seed-specific expression pat-
terns. Such specificity is usually conserved
in a heterologous host, as seed-specific
promoters from monocotyledonous plants
can also drive seed-specific expression in
dicotyledonous plants. For example, the
rice glutelin 3 promoter was used to ex-
press the recombinant glycoprotein b gene
of the human cytomegalovirus in tobacco
seeds [13]. Such promoters, like the hor-
dein gene promoter, maize ubiquitin pro-
moter and patatin promoter, were used for
expression of genes of interest in barley
grains [14], maize seeds [15], and potato
tubers [16], respectively. However, the pro-
moters of maize ubiquitin 1 [17] and rice
Table 6.1 Expression levels of selected recombinant proteins from nuclear transgenes
Recombinant protein Expression level Plant tissue Refer-
ence(s)
b-Glucuronidase
b-Glucuronidase
Avidin
Lysozyme
Spider silk proteins
Spider silk proteins
Human somatotropin
Human serum albumin
Human serum albumin
Human collagen
Human a-lactalbumin
B subunit of E. coli enterotoxin
Thermostable b-glucanase
Human interleukin-2
Aprotinin
Bovine trypsin
b-Casein
Cholera toxin B subunit
HIV p-24 casid protein
Hepatitis B surface antigen
Human lactoferrin
Human lactoferrin
Human lactoferrin
Human lactoferrin
Human lactoferrin
Human cytomegalovirus
glycoprotein B
scFv
scFv
scFv
scFv
Diabody
Human interferon-alpha
HPV major capsid protein L1
Human interleukin-18
Human glucocerebrosidase
Human a and b hemoglobin
Human a1 antitrypsin
Human placental alkaline
phosphatase
Human vascular endothelial growth
factor
0.40.7% TSP
10% TSP
5.7% TSP
545% TSP
2% TSP
2% TSP
0.16% TSP
0.2% TSP
0.2% TSP
0.1 mg g
1
5 lg g
1
13 lg g
1
0.15.4% TSP
115 U g
1
0.17% TSP
3.3% TSP
0.01% TSP
0.3% TSP
0.35% TSP
0.3316 lg g
1
0.1% TSP
0.5%
a)
Unknown
4.3% TSP
3% TSP
1% TSP
0.010.25% TSP
0.5 mg g
1
1% TSP
11.7 lg g
1 b)
0.016.8% TSP
30 lg g
1
1% TSP
36.5% TSP
0.5 mg kg
1
560 IU g
1
0.20.5% TSP
0.05% TSP
1 mg g
1
0.05% TSP
18.224 mg g
1 c)
2028 lg g
1 b)
30 mg L
1 d)
seeds (corn)
leaves (tobacco)
seeds (corn)
seeds (rice)
tuber (potato), leaves (tobacco)
seeds (tobacco)
tuber (potato), leaves,
cell culture
(tobacco)
tobacco (leaves)
tobacco (leaves)
tuber (potato)
seeds (barley)
microtuber (potato)
seeds (corn)
seeds (corn)
leaves (potato)
leaf, tuber (potato)
leaves (tobacco)
tuber (potato)
tuber(potato)
seeds (rice)
seeds (corn)
cell culture (tobacco)
cell culture (ginseng)
seeds (tobacco)
leaves (tobacco)
leaves (tobacco)
alfalfa
rhizosecretion (tobacco)
leaves, seeds (tobacco)
leaves, seeds (wheat, rice)
Petunia hybrida
seeds (Arabidopsis)
leaves (tobacco)
potato
tuber (potato), leaves (tobacco)
leaves (tobacco)
leaves (tobacco)
seeds (tobacco)
cell culture (rice)
rhizosecretion (tobacco)
cell culture (moss Ph. patens)
3436
11
36
32, 33
10
10, 45
46
47, 48
47, 48
49
50
51
14
16
52
53
54
55
56
57, 58
59
60
61
62
12
13
6365
66, 67
68
69
7072
37
73
38
74
75
76
75
77
78
79
80,
81
82
a) Relative yield per gram of dry seed weight.
b) Rhizosecretion of protein per gram dry roots in 24 h.
c) Accumulation of protein in the medium per gram dry cells
biomass in 5070 h.
d) Secreted amount of protein per 1 L of culture in 24 h.
actin 1 [18] genes are the most frequently
used for monocotyledonous plants.
In general, yields of recombinant pro-
teins under control of constitutive promo-
ters are low (ca 0.1% of TSP). Even though
some high-yield expression systems have
also been reported (e.g., [11]), such sys-
tems do not appear to represent broadly
applicable solutions for the production of
pharmaceutical proteins in plants, because
constitutive expression of many biologi-
cally active proteins at high levels often
compromises plant growth and develop-
ment. It probably also can trigger trans-
gene silencing.
Interesting alternatives to constitutive
expression are expression systems based
on inducible promoters. Such systems pro-
vide for separation of the growth and pro-
duction phases, thus theoretically allowing
for improved yield with proteins that are
toxic or interfere with plant physiology/de-
velopment. Several inducible systems
based on application of small molecules-
inducers have been described: the tetracy-
cline-inducible system [1921], the copper-
inducible [22], steroid-inducible systems
[23, 24], ethanol- [25, 26] or acetaldehyde-
inducible [27] systems, and insecticide
methoxyfenozide-inducible system [28]. A
hybrid system representing a combination
of two different inducers was also de-
scribed. It consists of a chimeric promoter
that can be switched on by the glucocorti-
coid hormone dexamethasone and
switched off by tetracycline [29]. For the
latest review on chemically inducible sys-
tems, see Ref. [30]. At present, it is still
not known whether inducible systems will
provide yields higher than those generated
by transgenes expressed under control of
constitutive or tissue-specific promoters.
Also, the leakiness of such systems might
be a serious problem for expression of cy-
totoxic proteins. Additionally, the use of
some chemical inducers such as steroids
and antibiotics is not desirable for large-
scale application. Potentially, the ethanol
switch proposed by Caddick and colleagues
could perhaps satisfy requirements for the
commercial use of a chemical gene switch
[25]. The use of vectors that can potentially
produce high yields, such as plant viral re-
plicons stably incorporated into plant nu-
clear DNA, does not provide a solution
either, because such vectors are subject to
transgene silencing. However, a promising
approach was found by using post-tran-
scriptional gene silencing (PTGS) suppres-
sors to boost expression of genes of inter-
est in plants. By crossing a transgenic to-
bacco plant carrying the potato virus X-
based replicon with a transgenic plant pro-
viding for the viral PTGS suppressor HC-
Pro, progeny expressing high level of a
gene of interest (GUS) could be obtained
[31]. It is clear however, that such an
approach is not suitable for the expression
of cytotoxic proteins, and that further im-
provements of the system are required.
One possible improvement is the develop-
ment of a tightly regulated inducible sys-
tem that does not allow leaky expression
of the transgene and the PTGS suppres-
sor.
The use of seed-specific promoters ap-
pears to be the most promising (see Table
6.1). Seeds are also an attractive choice for
molecular farming because they can be
transported and stored for downstream
processing, without any significant loss of
yield or quality of the recombinant protein.
Several seed-specific promoters from dicot-
yledonous and monocotyledonous plants
have been isolated and used for expression
of recombinant proteins in rice [32, 33],
corn [3436], barley [14], wheat [37], tobac-
co [13], and even Arabidopsis [38]. The yield
of recombinant proteins in seeds can
reach up to 45% of TSP, as was shown in
the case of seed-specific expression of the
human lysozyme gene in rice [33].
The use of major food/feed crops (rice,
corn, wheat, barley, etc.) for molecular farm-
ing and especially for the production of
pharmaceutical proteins carries the risk
of uncontrolled release into the environ-
ment and of entering the food chain by con-
tamination of non-transgenic crops. In such
cases, biological safety issues should be
carefully considered and given the highest
priority for each specific recombinant pro-
tein. Potential problems for companies
ignoring or miscalculating the problem
can range from costly delays in field trials
(such as one that Ventria Bioscience faced
this year with transgenic rice expressing hu-
man lactoferrin and lysozyme [3941]), to
high financial liabilities once a permit for
commercial production has been obtained.
ProdiGene Inc. was ordered to pay US$ 3
million penalty by the United States Depart-
ment of Agriculture (USDA) for contami-
nating soybean fields with transgenic vol-
unteer corn plants [42]. The unpaid portion
of this penalty was paid by Stine Seed Com-
pany as a part of ProdiGenes rescue pur-
chase [43]. The choice of corn as a produc-
tion host for monoclonal antibodies (Mabs)
is one of the major reasons behind the re-
cent failure of Epicyte Pharmaceuticals to
raise new capital [44]. In contrast, several
small plant biotechnology companies have
chosen to use plants grown in a closed envi-
ronment as the basis of their production
platform. Among them, Biolex uses a small
aquatic plant Lemna (duckweed) grown in
closed environment, and Medicago Inc.
uses alfalfa grown in contained glasshouses.
Other companies, such as Plantigen, Planet
Biotechnology, and Large Scale Biology rely
on open-field cultivation, but using tobacco,
a production host that cannot contaminate
feed and food stocks. SemBioSys, a Cana-
dian biotechnology company, has chosen
safflower (Carthamus tinctrius L.), a minor
crop plant with medicinal, industrial, and
food applications, as an alternative produc-
tion host. The choice of a minor crop plant
reduces the risks of cross-contaminating
non-transgenic crops, while benefiting the
developer who can rely on well-established
agro-cultivation technologies.
In brief, it appears very unlikely that
high yields of recombinant proteins in
transgenic plants can be achieved with the
use of constitutive promoters, especially in
the case of recombinant proteins that have
deleterious effects on plant growth and de-
velopment. The use of inducible systems
is a more promising approach, though this
requires the development of leakage-
proof systems for the production of cyto-
toxic proteins. Today, production in the
seeds of transgenic plants using seed-spe-
cific promoters is the most obvious choice
for those interested, and a viable solution
for many recombinant proteins.
6.2.2
Plastid-based Vectors
Since the first successful transformation of
tobacco chloroplasts [83], expression sys-
tems based on the transformation of plant
plastids has attracted the attention of plant
biotechnologists. The features that make
this technology so attractive are its poten-
tial for high protein yield, along with in-
herent biosafety features such as limited
plastid transfer via pollen (due to maternal
inheritance of plastid-encoded genes) [84]
and the relatively low probability of trans-
gene movement from the chloroplast to
the nucleus [85, 86].
Many other heterologous proteins have
been expressed in tobacco chloroplast, and
in many cases yields exceeding 5% of TSP
were reported (Table 6.2). Examples of such
proteins are neomycin phosphotransferase
(NPTII) (23% of TSP) [87], 5-enolpyruvyl-
shikimate-3-phosphate synthase (EPSPS)
(10% of TSP) [87], phosphinothricin acetyl-
transferase (PAT) (7% of TSP) [88], Bacillus
thuringiensis (Bt) toxin (41% of TSP). As ex-
pected, another large group of proteins suc-
cessfully expressed in plastids are different
antigens without any obvious cytotoxicity
that are derived from human and animal
pathogens of prokaryotic nature (Table
6.2). These data suggest that expression in
plastids can be a good choice for an expres-
sion platform for these proteins.
There are numerous reviews describing
the state of the art in plastid-based expres-
sion systems [8994], so there is no need
to go into their detailed description. The
main point worth mentioning is that, at
present, only tobacco plastids and plastids
of the unicellular alga Chlamydomonas re-
inghardtii [95] can be transformed routine-
ly. However, this situation is changing. In
addition to tobacco, successful plastid
transformation was reported for Brassica-
ceae [96, 97], potato [98], tomato [99], and
cotton (H. Daniell, personal communica-
tion). Transplastomic potato plants, tomato
cell lines as well as other Nicotiana species
(N. benthamiana, N. excelsiana and N. ex-
celsior), were also produced (H.-U. Koop,
personal communication).
There are limitations on the choice of the
recombinant protein to be expressed in
transplastomic hosts due to the prokaryotic
nature of the translational and post-transla-
tional machinery of plastids [100] (see Part
Table 6.2 Expression levels of selected recombinant proteins in plastids
Recombinant protein Expression level
[% TSP]
Host Reference(s)
b-Glucuronidase
Neomycin phosphotransferase (NPTII)
Green fluorescent protein
EPSP synthase
Phosphinothricin acetyltransferase
Bacillus thuringiensis toxin
Human somatotropin
Human serum albumin
Interferon-alpha
Interferon-gamma
Tetanus toxin fragment C
B subunit of E. coli enterotoxin
Native cholera toxin B subunit
Rotavirus VP6 protein
2L21 peptide (virulent canine parvovirus)
Xylanase
Phenylalanine ammonia lyase
Bacillus anthracis protective antigen
lsc antibody
Green fluorescent protein
b-Glucuronidase
1.38.8
0.1623
5
0.00110
7
246.1
0.27
0.0211.1
1
0.16
1025
2.5
4.1
0.63
2331
6
1.01.5
1.718.1
1
0.5
0.01
Tobacco
Tobacco
Tobacco
Tobacco
Tobacco
Tobacco
Tobacco
Tobacco
Tobacco
Tobacco
Tobacco
Tobacco
Tobacco
Tobacco
Tobacco
Tobacco
Tobacco
C. reinhardtii
C. reinhardtii
C. reinhardtii
C. reinhardtii
111, Icon
a)
112114
98
87
88
115118
102
119
Icon
a)
104
120
121
122
123
124
125
Icon
a)
126
127
128
129
a) Icon Genetics, unpublished data.
IV, Chapters 5 and 7; and Part V, Chapter 1).
Whereas some constraints of translation are
more easily addressed through vector de-
sign, lack of some essential post-transla-
tional capabilities (in particular N-glycosyla-
tion) cannot be easily corrected. In some
cases, the latter can be an advantage if gly-
cosylation is not essential for recombinant
protein function, since N-linked oligosac-
charides of plant origin might result in
new immunogenic structures [101]. Also,
achievement of the correct amino acid se-
quence at the N-terminal end of proteins
that do not start by a methionine might be
a problem for the expression of such pro-
teins in plastids. For example, human so-
matotropin is a secreted protein with a N-
terminal signal peptide cleaved off after se-
cretion into the endoplasmic reticulum
(ER), resulting in a processed protein with
a N-terminal amino acid that is not a
methionine. As a solution for expression
of this protein in the chloroplast, an N-ter-
minal fusion with the ubiquitin gene was
made, and the fusion protein processed by
cleavage of ubiquitin during the extraction
from plant tissue [102]. The protein was ex-
pressed at a high level (more than 7% of
TSP), but removal of N-terminal ubiquitin
from the ubiquitin-human somatotropin fu-
sion resulted in only 3080% (depending on
extraction conditions) of all molecules being
correctly processed. In contrast, human so-
matotropin expression in tobacco seeds
based on stable nuclear transformation vec-
tors yielded the correctly processed protein
with the correct amino acid sequence, but
at a very low level (0.16% of seed TSP) [103].
In general, expression levels of recombi-
nant proteins in transplastomic plants can
be significantly higher than those in nucle-
ar transformants, subject to optimal design
of the expression cassette and depending
on the nature of the recombinant protein
of interest. Clear evidence of an impact of
the expression cassette is the differences
in expression levels obtained using various
vectors for Bt toxin (25-fold difference),
human serum albumin (510
2
-fold differ-
ence), EPSPS (10
4
-fold difference) (see Ta-
ble 6.2). A significant increase in yield can
also be achieved by using translational fu-
sion of a gene of interest to another one
(usually a reporter gene) which is easy to
express. An example of such an increase
in yield is the expression level of the inter-
feron-a gene as translational fusion with
the GUS reporter gene [104].
Another remarkable advantage of plastid
transformation technology is the availabil-
ity of efficient methods, similar to those
used in bacterial genetics for example,
the use of homologous or site-specific re-
combination, for complete removal of
transformation markers and other un-
necessary sequences from plastid transfor-
mants. There are several approaches to
achieve this result. Some approaches, also
used with nuclear transformants [105], are
based on site-specific recombination using
the CRE/lox system [106, 107] (see Part
III, Chapter 2 and Part IV, Chapters 2 and
3). Two other approaches are based on ho-
mologous recombination between two
short direct repeats flanking the selectable
marker [108, 109], and do not require
crosses with plants that provide for the
site-specific recombinase. An approach re-
cently developed by Icon Genetics [109]
has the advantage that it allows the easy
selection of homoplastomic lines after the
antibiotic resistance gene has been lost.
Removal of the antibiotic resistance gene
improves the biosafety parameters of the
system, since the presence of the antibiotic
resistance gene in plants grown in an
open field is not desirable.
An expression system based on the plas-
tids of the green algae Chlamydomonas rein-
hardtii represents a special case. Despite ex-
hibiting relative yields of recombinant pro-
tein generally lower than those achieved
with transplastomic tobacco, the system al-
lows for very fast scale-up, producing gram
quantities of recombinant protein within 2
months after transformation, with the pos-
sibility of further increase in geometric pro-
gression, reaching yields of up to 150 kg of
recombinant protein per acre of fully con-
tained growth area annually (S. Mayfield,
personal communication). For comparison,
the generation of a homoplastomic tobacco
plant with a transgene stably incorporated
into plastid DNA requires at least 6 months.
Therefore in terms of speed but not yield
the Chlamydomonas system is comparable
to many known microbial systems. For the
most recent review on molecular farming
in plastids of C. reinhardtii, see Ref. [110].
In conclusion, plastid-based expression
systems provide for potentially higher ex-
pression levels than the majority of nuclear
expression systems, and with a higher level
of transgene containment, but the platform
provides limited post-translational proces-
sing choices. Also, despite reports of suc-
cessful plastid transformation of new spe-
cies beyond Nicotiana, the choice of produc-
tion host is still predominantly restricted to
the Nicotiana family. However, this limita-
tion does not present a serious drawback
for production of pharmaceutical proteins
(except perhaps for edible vaccines) since
tobacco is a non-feed/non-food crop with a
well-established agriculture.
6.2.3
Vectors for Transient Transformation
Transient expression [130] is a fast and
convenient alternative to stable transforma-
tion because of the speed it provides for
research and development and, to some
extent, because of less complex regulatory
hurdles, as no stably transformed plants
are involved in the production process.
Transient expression can be achieved by
transfecting plants with viral vectors, or,
on a small scale, by agroinfiltrating plant
tissue with a standard expression cassette
under control of a constitutive promoter,
for example the 35S promoter [131].
Usually, agroinfiltration itself does not pro-
vide for high yield, but in combination
with PTGS suppressors such as p19 or
HcPro, protein expression levels can be in-
creased up to 50-fold [132]. However, even
such significant improvement is still well
below the levels that can be achieved with
some viral vectors. The two approaches
(agroinfiltration and the use of viral vec-
tors) can be combined, and we recently
demonstrated that tobacco mosaic virus
(TMV)-based vectors can be delivered via
agroinfiltration as separate structural
blocks and assembled in planta with the
help of a site-specific recombinase [133].
Viruses first began to attract attention as
a potential basis for developing expression
systems almost 20 years ago. Since then,
tremendous progress has been achieved in
the development of viral vectors. There are
numerous reviews describing different vi-
ral vectors and strategies for their use
[134139]. Comparison of the expression
levels provided by different vectors (Table
6.3) makes it clear that TMV-based expres-
sion systems achieve the highest yield. In
a recently published report [133], it was
shown that TMV-based expression of a re-
porter gene (GFP or DsRed) could reach
biological limits of the plant leaf system,
producing up to 5 mg of recombinant pro-
tein per gram fresh leaf biomass. The rela-
tive yield with such system can approach
80% of total soluble protein. Such high ab-
solute and relative yield translates into
much more efficient and cost-effective up-
stream and downstream processes. Also,
high relative yield is possible because of
virus-controlled gene amplification and be-
cause of the relatively less understood
mechanism of virus-induced shut-off of
host protein biosynthesis. Other features
of viral vectors are their ability for cell-to-
cell and systemic movement, which allows
the vector to spread from the infection
zone to most of the plant tissue. Fully
functional viral vectors can tolerate only
relatively modest heterologous inserts with
an upper size limit of around 1 Kb. Signif-
icantly shorter inserts practically do not af-
fect viral functions, and there are numer-
ous examples of using such inserts as fu-
sions with the viral coat protein for anti-
genic epitopes production [140147]. In
contrast, larger inserts affect vector stabili-
ty, resulting in wild-type revertants that
successfully outcompete the original re-
combinant vector for all viral functions
amplification rate, cell-to-cell and systemic
movement. These problems have been ad-
dressed by applying the so-called decon-
structed virus strategy for example, by
removing the missing or undesired func-
tions of a viral vector and complementing
the necessary functions in trans (for a re-
view, see Ref. [139]). Such approach allows
not only for the payload capacity of the vir-
al vector to be increased, but also provides
for significantly better containment of the
recombinant viral vector. Using viral vec-
tors devoid of the coat protein gene, we
have achieved very high yield (up to 40%
of TSP) with inserts encoding different
proteins and protein fusions that are as
large as 2.2 Kb (see Fig. 6.3).
Systemic viral vectors are restricted to cer-
tain parts of infected plants (usually the
newly developing leaves, stem and roots),
and are excluded from a significant part of
plant biomass, including the mature leaves.
Clearly, the creation of transgenic plants
with a viral replicon precursor stably in-
Table 6.3 Transient expression levels of different recombinant proteins
Recombinant protein Expression level Vector Production host Refer-
ence(s)
Dihydrofolate reductase
Metallothionein II
Interferon aD
Interferon c
ScFv
ScFv
ScFv
ScFv
Monoclonal antibody
Glycoprotein D of BHV-1
a-amylase
Pollen allergen Bet v1
GFP, DsRed
Human somatotropin
Diabody
Human lactoferrin
8 lg g
1
0.5% TSP
2 lg g
1
0.5% TSP
5 lg g
1
1230 lg g
1
0.020.8 mg mL
1
IF
b)
0.251.2 mg g
1
Unknown
20 lg g
1
5% TSP
0.2 mg g
1
5 mg g
1
1.2 mg g
1
1.5 mg kg
1
0.6% TSP
CaMV
CaMV
CaMV
BMV
p35S
a)
TMV
TMV
TMV
TMV
TMV
TMV
TMV
TMV
TMV
Unknown
PVX
Turnip
Turnip
Turnip
Protoplasts (tobacco)
Leaves (tobacco)
N. benthamiana
N. benthamiana
N. benthamiana
N. benthamiana
N. benthamiana
N. benthamiana
N. benthamiana
N. benthamiana
N. benthamiana
Leaves (tobacco)
N. benthamiana
149
150
151
152
131
153
154
Icon
c)
155
156
157
158
133
Icon
c)
74
159
a) CaMV 35S promoter.
b) Interstitial fluid.
c) Icon Genetics, unpublished data.
serted on a chromosome could be a solu-
tion, but it was found that such plants pro-
duce very low amount of recombinant pro-
tein because of silencing [148]. As already
mentioned in Section 6.2.1, the use of PTGS
suppressors allows this problem to be ad-
dressed [31]. The problem of separating
the plant growth phase from the production
phase can also be resolved by developing
either a tightly regulated inducible system
or a virus-based transient expression sys-
tem, where tight control over transgene ex-
pression is not an issue. However, viral vec-
tor delivery to each plant cell in a transient
expression system is not a trivial task. Agro-
bacterium-mediated delivery of T-DNAs en-
coding RNA viral replicons provides only a
small fraction of plant cells with active viral
replicons due to low ability of the primary
transcripts to leave a nucleus. The RNAs
of plant plus-sense RNA viruses replicate
in the cytoplasm and normally never enter
the cell nucleus, and contain multiple se-
quence features that are likely to be impro-
Fig. 6.1 Comparative analysis of recombinant pro-
tein expression systems: state-of-the-art versus
magnICON
. Background: magnICON
-based ex-
pression of green fluorescent protein (GFP) in Ni-
cotiana benthamiana (Plants are exposed to UV
light). Left side insert: Coomassie blue-stained
polyacrylamide gel after SDS-electrophoretic sepa-
ration of total soluble protein (TSP) extracted
from N. benthamiana leaves. The framed area con-
tains GFP bands. w.t.: TSP extracted from the
leaves of wild-type N. benthamiana. Right side in-
sert: time-course of GFP expression levels for
magnICON
-based (ICON) and other state-of-

the-art systems.
perly recognized by the nuclear RNA proces-
sing machinery if delivered to the nucleus.
We recently developed highly active syn-
thetic templates for delivery of RNA viral
vectors as DNA precursors using Agrobac-
terium, and found that Agrobacterium deliv-
ery of such templates can be used to start
gene amplification and obtain high-level
expression in all mature leaves of a plant,
simultaneously. Such a transfection route
can be performed on an industrial scale by
vacuum-infiltration of batches of multiple
plants (Marillonnet et al., unpublished re-
sults). In this process, the bacteria assume
the (formerly viral) functions of primary
infection and systemic movement, whereas
the viral vector itself provides for cell-to-
cell (short distance) spread, amplification
and high-level expression. A comparative
analysis of state-of-the-art systems with
our approach is illustrated in Fig. 6.1.
Assuming protein yields as mentioned
above, and based on realistic yields of
100 tons of plant leaf biomass per hectare
(ha) of greenhouse per year, a 1-ha facility
should be capable of producing 280
400 kg of recombinant protein each year.
This means that for the vast majority of
pharmaceutical proteins, industrial-scale
production can be carried out entirely in a
partially or fully contained greenhouse fa-
cility. The whole process is a straightfor-
ward protocol, similar to existing industrial
microbial technologies; it requires, in addi-
tion to well-established industrial up-
stream (greenhouse plant cultivation) and
downstream (protein extraction and purifi-
cation) components, a contained technol-
ogy block that includes an apparatus for
vacuum-infiltration of batches of plants
and a chamber/greenhouse for short-term
subsequent incubation, as well as a small
bacterial fermentation apparatus. Such a
block would of course require certain
safety locks, so as to eliminate the re-
lease of agrobacteria into the open envi-
ronment and to protect the operating per-
sonnel.
6.3
Plant-made Recombinant Proteins available
Commercially, and under Development
A relatively small number of biotechnology
companies operate in the field of recombi-
nant proteins production in plants (Table
6.4). This small number represents roughly
one-third of approximately 60 dedicated
plant biotechnology companies an unfa-
vorable comparison to the several hundred
small and medium-sized businesses that
use bacterial, yeast or animal-based expres-
sion platforms and operate in the sector of
red biotechnology. The number of plant-
made recombinant proteins that have
reached the market is also very limited,
but has a healthy tendency to grow.
Although not a biopharmaceutical, one of
the first plant-derived recombinant proteins
of potential importance to reach the market,
trypsin, has been produced in transgenic
maize by ProdiGene, and is marketed by
Sigma-Aldrich Fine Chemicals as Tryp-
Zean
TM
. ProdiGene also established pro-
duction in maize of recombinant protease
inhibitor aprotinin (called AproliZean
TM
).
In 2002, ProdiGene planted 400 acres of
transgenic corn expressing trypsinogen, an
amount sufficient (according to ProdiGene)
to meet 5% of the market demand. The
company planned to scale-up the cultivation
and to meet full market demand in 2003,
but this did not materialize. The company
faced problems caused by field contamina-
tion with transgenic volunteer corn plants
expressing recombinant proteins, and this
forced the company to down-size its opera-
tions. Subsequently, ProdiGene was sold to
H. Stine Seeds [43]. Based on ProdiGenes
6.3 Plant-made Recombinant Proteins available Commercially, and under Development 903
data, planting 400 acres of transgenic corn
yielded ca. 1000 tons of seed expressing
trypsin at a level of 58 mg kg
1
seed, which
corresponds to a total harvest (assuming
50% recovery of 90% pure enzyme) of 11
13 kg of enzyme. Although the price of
plant-made trypsin (TrypZean) in the Sigma
catalogue (20042005) is approximately 20-
fold higher than that for trypsin of animal
origin, plant-made trypsin is bona fide hu-
man pathogen-free, and is safer for use in
many applications. However important, this
advantage does not justify such a high price,
but likely reflects the very high downstream
processing cost of low-expressing seed; cer-
tainly, a more efficient plant expression sys-
tem is needed in order to lower the cost of
plant-made trypsin to make it a commercial
success.
At the start of 2004, Large Scale Biology
Corporation (LSBC) began to ship test
quantities of its plant-produced recombi-
nant bovine-sequence aprotinin (rAproti-
nin) to customers for R&D and manufac-
turing applications. LSBC also has several
other products in development; for exam-
ple, plant-produced human therapeutic en-
zyme alpha-galactosidase A for enzyme re-
placement therapy showed positive results
in preclinical studies using an animal
model of Fabry disease. This enzyme is
currently undergoing clinical trials. LSBC
are also developing vaccines for animal
health in collaboration with Schering-
Plough Animal Health Corporation
(SPAH), and a human papillomavirus vac-
cine is currently under development in col-
laboration with the University of Louisville
Table 6.4 Companies using plant-based recombinant proteins production platforms
Company Internet link Production host
Biolex
Chlorogen
CropTech
a
Dow Chemical
Epicyte Pharmaceutical
a
Greenovation
Icon Genetics
Large Scale Biology
LemnaGene
Meristem Therapeutics
Medicago
Monsanto Protein Tech.
a)
MPB Cologne
a)
Novoplant
SemBioSys
Syngenta Biopharma
Phytomedics
Planet Biotechnology
Plantigen
ProdiGene
Protalix
Ventria
http://www.biolex.com
http://www.chlorogen.com
http://www.dow.com/plantbio/index.htm
http://www.biolex.com
http://www.greenovation.com
http://www.icongenetics.com
http://www.lsbc.com
http://www.lemnagene.com
http://www.meristem-therapeutics.com
http://www2.medicago.com
http://www.mpt.monsanto.com
http://www.novoplant.com/
http://www.sembiosys.ca
http://www.syngenta.com/en/biopharma/
http://www.phytomedics.com
http://www.planetbiotechnology.com
http://www.plantigen.com/
http://www.prodigene.com
http://www.protalix.com
http://www.ventriabio.com
Duckweed
Tobacco
Tobacco
Maize
Maize
Moss Physcomitrella
Tobacco
Tobacco
Duckweed
Maize, Tobacco
Alfalfa
Maize
Potato
Various
Safflower
Various
Tobacco
Tobacco
Tobacco
Maize
Cell culture
Rice
a) Operations terminated.
(Kentucky). The company uses a TMV-
based transient expression platform that is
fast and easy to apply. The platform is
built on viral vectors capable of systemic
movement.
California-based Ventria Bioscience (Sa-
cramento) is at the stage of field trials of
transgenic rice plants expressing recombi-
nant lactoferrin and lysozyme in rice seed
(see Part IV, Chapter 8). It is worth men-
tioning, that these products might reach
market later than planned considering dif-
ficulties with obtaining the necessary field
trial permits [3941].
Meristem Therapeutics (Clermont-Fer-
rand, France) has a gastric lipase, produced
in grains of corn, for the treatment of cystic
fibrosis in Phase II clinical trials. The com-
pany also has a recombinant lactoferrin at
the stage of preclinical trials, in addition
to human serum albumin, collagen and
monoclonal antibodies at R&D stage.
Several other companies have products
at different stages of development (see
Part IV, Chapter 5). We provide here a
brief update of their performance using
publicly available information, obtained
predominantly from the web pages of the
companies mentioned below.
SemBioSys Genetics Inc. (Calgary, Cana-
da), in addition to insulin and apolipopro-
tein AI, demonstrated a proof-of-concept
for proteins addressing osteoporosis, pul-
monary and liver fibrosis, psoriasis and
gastrointestinal disorders. Plantigen Inc.
(London, Canada) has proof of principle
for several products under development.
Those include glutamic acid decarboxylase
(GAD) for the treatment of type 1 diabetes;
interleukin-10 (IL-10) for treatment of in-
flammatory bowel disease; interleukin-4
(IL-4) for use as adjuvant to enhance im-
mune response; and major histocompati-
ble complex (MHC) and cytokines for use
in organ transplants.
Greenovation (Freiburg, Germany) uses
the advantage of highly efficient homolo-
gous recombination in the moss Physcomi-
trella to develop a new production system
for humanized antibodies with required
glycosylation pattern. For a more detailed
description of production of biopharma-
ceuticals with humanized glycosylation
(see Part IV, Chapter 7).
The productivity of moss bioreactor for
human vascular endothelial growth factor
(hVEGF) reached 30 mg L
1
per day [82].
The company claims that further increases
in yield can be achieved by optimization of
production strains for specific proteins.
Protalix (formerly Metabogal, Israel) be-
gan preclinical tests on enzyme therapy
for Gaucher disease (GD), a genetically
linked condition that affects the metabo-
lism of people deficient in the enzyme b-
glucocerebrosidase. Without this enzyme,
fatty deposits build up in the spleen, brain,
liver, and bone marrow, and this leads to
extreme pain and may even prove fatal. In
2000, the annual world-wide market for
GD was US$ 620 million.
Among pharmaceutical proteins, the
antibody market is potentially the most in-
teresting (see Part IV, Chapter 16 and Part
V, Chapter 1). Antibody production is tra-
ditionally based on animal cell culture,
commanding high manufacturing costs
($5001000 g
1
purified protein). To date,
no plant-made antibodies are available
commercially, and some analysts foresee
potential problems with production capaci-
ties within next few years. Most of the an-
tibodies available or under development
are immunoglobulins that are complex
glycosylated proteins, and cannot be ex-
pressed in bacterial cells. Expression in eu-
karyotic organisms (animal cells, insect
cells, yeast, or plants) (see Part IV, Chap-
ters 1, 2, 12, 13, and 14) results in glyco-
proteins that differ considerably from hu-
6.3 Plant-made Recombinant Proteins available Commercially, and under Development 905
man antibodies because of the differences
in protein glycosylation among different
organisms. In many cases, glycosylation
does not affect the pharmacological prop-
erties of the resultant protein, and it is not
required for proper protein folding. Thus,
some developers mutate glycosylation sites
in antibody genes in order to avoid glyco-
sylation altogether. Alternatively, the plant
host can be engineered to provide for a re-
quired humanized glycosylation pattern,
and this approach is currently being used
by Greenovation (Freiburg, Germany), as
mentioned above. A smaller proportion of
immunopharmaceuticals are proteins that
are less complex. Alternatively and prob-
ably more easily these proteins can be
manufactured in non-animal cells.
Starting from the pioneering work of
scientists at the Scripps Research Institute
[160, 161] and the Max-Planck-Institute
(Cologne, Germany) [162], numerous anti-
bodies have been expressed in plants and
shown to be properly processed and as-
sembled into fully functional molecules
with full immunological activity. Today, al-
most every plant biotechnology company
involved in the molecular pharming of
pharmaceutical proteins has antibodies in-
cluded in their product portfolio. The level
of expression in transgenic plants express-
ing heavy- and light-chain Mab polypep-
tides under a strong constitutive promoter
are rather low, amounting usually to 10
50 mg of antibody per kg leaf biomass (to-
bacco) or up to 1 g kg
1
seed (maize), and
may require up to 3 years before gram
quantities of the protein of interest can be
obtained for research or clinical studies.
The level of expression in plants in-
fected using viral vectors can be higher
(up to 200 mg kg
1
leaf biomass), and re-
search quantities can be obtained within
46 weeks, although the current versions
of viral vectors allow expression only of
single-chain antibodies. The technology
can be used immediately for applications
such as the production of individualized
antibody vaccines, for example vaccines for
non-Hodgkins lymphoma made by Large
Scale Biology. Phase I clinical trials were
successfully completed with these indivi-
dualized vaccines, and subsequently Large
Scale Biology Corporation built the worlds
first commercial-scale biopharmaceutical
production facility, based at Owensboro,
Kentucky.
Icon Genetics has conducted a feasibility
study aimed at the expression of single-
chain antibodies in plants using its viral
expression technology built on a TMV-
based transient expression system. Using
different antibodies currently at a preclini-
cal stage, it was possible to express them
at a level of 0.251.2 g kg
1
fresh leaf bio-
mass, with relative recombinant protein
amounts reaching up to 35% of TSP.
Among other recombinant proteins that
Icon is developing are included aprotinin,
trypsin, thrombin, several viral antigens,
and thaumatin.
Planet Biotechnology Inc. of Hayward,
California and LSBC announced recently a
biomanufacturing agreement to extract
and purify a plant-made antibody to con-
trol dental caries; Phase I clinical trials
were completed by Planet Biotechnology
Inc. The same company is also developing
a new approach for blocking infection by
rhinovirus, a major cause of the common
cold. In preclinical testing, the fusion pro-
tein proved highly protective against cellu-
lar damage caused by human rhinovirus
infection. Planet Biotechnology also suc-
cessfully completed pilot human clinical
trials with neutralizing toxicity of che-
motherapeutic drug doxorubicin by topical-
ly applied antibodies. The oral application
of antibodies can prevent chemotherapy-
induced gastrointestinal toxicity and, when
applied topically in liposomes, prevents
doxorubicin-induced hair loss.
Novoplant is developing, for the first
time, a comprehensive portfolio of anti-
bodies for use in veterinary medicines.
These monoclonal immunoglobulin pre-
parations are based on single-chain-anti-
bodies produced in plant seeds, and are
designed to combat typical diseases that
occur in swine, poultry, and calves in ani-
mal husbandry. They are designed to offer
protection against coccidiosis in chickens
and scour caused by ETEC Escherichia coli,
rotavirus and coronavirus in piglets and
calves.
Medicago currently has 10 Mabs and
plasma-proteins in early-stage develop-
ment, with preclinical studies planned to
start in 2005 for two products. Greenova-
tion has one antibody (ABC-48) in preclini-
cal development for the prevention of
deep-vein thrombosis. The secreted IgG
antibody was shown to be correctly as-
sembled, and displayed normal binding ac-
tivity to its natural ligand.
This brief account of currently available
products and under development (see Ta-
ble 6.4) provides evidence that, despite sev-
eral drawbacks, the production of recombi-
nant proteins in plants is steadily progres-
sing. However, it is also evident that all ex-
pression platforms in use have certain
limitations, and it appears that as yet there
is no universal platform to satisfy the re-
quirements for production of all biophar-
maceuticals of choice.
6.4
Comparative Analysis of the Expression
Systems and Production Platforms
In principle, all three expression platforms
(see Section 6.2) provide for expression
levels within essentially the same range.
Comparison of expression levels (see Ta-
bles 6.16.3) reveals that high yields (over
5% of TSP) are feasible using either nucle-
ar, plastid, or transient expression. How-
ever, some systems are more equal than
others when additional requirements are
introduced, such as the speed of R&D pro-
cess and biosafety parameters. The trans-
fection platform is by far the fastest: start-
ing with the DNA construct of the protein
of interest, milligram and gram quantities
of recombinant protein are available in 3
4 weeks. Thus, the platform supports the
highest possible speed of R&D in the in-
dustry, including the microbial and animal
cells systems (Fig. 6.2). The biosafety pa-
rameter shall not be underestimated, as
regulators, food producers, and almost
everybody else alerted by the mistakes of
Fig. 6.2 Research & Develop-
ment time course using differ-
ent recombinant protein pro-
duction platforms.
Prodigene are all calling for stronger
regulations to ensure that plants express-
ing pharmaceuticals are entirely and effec-
tively separated from the food supply. At-
tempts to derive certain assumptions by
monitoring the progress of plant biotech-
nology companies involved in pharmaceu-
tical proteins production can be fruitful,
despite the limited number of such com-
panies (see Table 6.4).
It is already clear that the early adoption
of food or feed crops especially cereals
with large production areas such as corn
and rice as production hosts for biophar-
maceuticals has been a fatal choice for
some players. However, we believe that
there are sufficient technological achieve-
ments in the fields of plant molecular biol-
ogy, genetics and biotechnology to make
any technology very safe, and a more de-
tailed discussion of these topics can be
found in a recent review [163]. Neverthe-
less, the main issue in biosafety is cross-
contamination of non-transgenic stock
an issue that can be simply and efficiently
resolved by the physical containment of
transgenic plants and/or by employing a
production host that is a non-food, non-
feed crop. Several biotech companies have
chosen these approaches. Medicago uses
alfalfa plants grown in high-tech glass-
houses, whilst other companies (e.g.,
Large Scale Biology) use both approaches
(glasshouse and open field) depending on
the volume of product required, with the
production host being tobacco, a non-food/
non-feed plant. It is worthy of note that in
the case of Large Scale Biology it is not
stably transformed but rather transiently
transformed tobacco plants that are being
used. Also, the choice of a plant host such
as duckweed (Biolex, Lemnagen), moss
Physcomitrella (Greenovation), cell suspen-
sion cultures (Protalix), or rhizosecretion
(Phytomedics) [80] leads to no other alter-
native but contained production. One Ca-
nadian company, Prairie Plant Systems
Inc., offers a new concept and facilities for
growing pharmaceutically active plants in
subterranean growth chambers. Undoubt-
edly, contained production shall be the so-
lution to many problems faced by plant
biotech companies using open-field and
food crops for molecular farming. How-
ever, growth in a closed environment is
substantially more expensive than in open
field and, may not be the solution for
products required in large (tons per year)
quantities at a competitive price, such as
human serum albumin.
Fortunately, the majority of recombinant
proteins for pharmaceutical use are re-
quired in quantities of, at most, hundreds
of kilograms per year, whilst in some cases
(e.g., glucocerebrosidase) even sub-kilo-
gram or gram quantities can satisfy mar-
ket needs. Although these recombinant
proteins can be produced in a closed envi-
ronment, in order to compensate for the
space limit and improve the economics of
the production process, high-yield expres-
sion systems are required. It is very un-
likely that the use of technology based on
seed-specific promoters and monocotyledo-
nous plants can be the solution for indoor
growth, as only a relatively small propor-
tion of total plant biomass (seed) can be
used for production. There is still no con-
sensus which plant host shall be used for
pharmaceutical protein production, but it
is evident that if we favor the safer choice
of glasshouse production to drastically re-
duce the danger of transgene escape, then
preference shall be given to crops allowing
for the highest possible yield of productive
plant biomass per year per hectare. These
crops shall be easy to transform, and shall
have well-established expression systems.
It is very natural that the most likely
choice of crop for such production is to-
bacco, as this can support several harvests
per year with yields of leaf biomass reach-
ing over 100 tons per hectare that is,
four to 30 times more than other crop can-
didates such as alfalfa, wheat, rice, and
corn [164]. Additionally, a transgenic tobac-
co line serving as a host for molecular
farming can be engineered to incorporate
other biosafety features, including male
sterility or competence for a specific ex-
pression vector (e.g., an ability to comple-
ment viral vector function such as cell-to-
cell or systemic movement), thereby
further increasing biosafety parameters
and drastically reducing the chance of
transgene escape.
6.5
It is clear that transient expression is sig-
nificantly faster for generating results, and
requires a significantly shorter time for op-
timization compared to systems requiring
stable transformation. For example, vectors
used in transient expression systems can
be evaluated within 47 days, whilst sys-
tems built on stable transformation re-
quire several months to obtain primary
transformants, which makes the process
of their optimization slower by at least an
order of magnitude. This parameter shall
not be underestimated in the highly com-
petitive environment of modern biotech-
nology. Surprisingly enough, no efforts
Fig. 6.3 magnICON
expression platform versatility: Coomas-

sie blue-stained SDSPAGE of total soluble protein extracted
from transfected Nicotiana leaves. The recombinant protein
bands are circled.
have been made within the biotech indus-
try to use transient expression systems at
an industrial scale, with the notable excep-
tion of Large Scale Biology Inc., that is
using TMV-based vectors and transfection
in tobacco as its main expression platform.
Scientists at Icon Genetics have further
improved the TMV-based system which, in
its present form, uses highly efficient agro-
bacterial delivery of optimized viral vectors
to practically every plant cell, thus remov-
ing the need for systemic and, to some ex-
tent, even cell-to-cell movement (unpub-
lished data). Moreover, such vectors pro-
vide for the highest possible yield, practi-
cally reaching the biological limits of the
system for many non-toxic proteins
(Fig. 6.3). The speed of the system is sup-
ported by the well-established protocol for
rapid optimization of recombinant protein
expression level [133].
Once the expression cassette has been
optimized, highly diluted agrobacterial sus-
pension carrying the cDNA of the TMV-
based vector is used for large-scale agroin-
filtration of whole tobacco plants. The bio-
mass can be harvested 710 days later and
used for isolation of the recombinant pro-
tein of interest. In comparison with other
systems, this protocol is extremely fast,
provides high yields, and has the potential
for unlimited scalability. In addition, our
Fig. 6.4 Comparison of expression systems.
transfection system has almost no restric-
tion on the size of biopharmaceutical (no
limitation in gene size) to be expressed
(see Section 6.2.3). However, improvement
is still required for the expression of het-
ero-oligomeric proteins, such as full-length
antibodies.
A comparison of the different expression
systems is summarized in Fig. 6.4.
Acknowledgments
The authors thank Dr. Leonid Shlumukov
for critical reading of this manuscript, and
Dr. Stefan Herz for providing the data in-
cluded in Table 6.2, and Dr. Romy Kandr-
zia for providing the picture for Fig. 6.3.
References
1 Horn, M. E., S. L. Woodard, and J. A. Howard,
Plant molecular farming: systems and prod-
ucts. Plant Cell Rep, 2004; 22(10):711720.
2 Fischer, R., et al., Plant-based production of
biopharmaceuticals. Curr Opin Plant Biol,
2004; 7(2):152158.
3 Stoger, E., et al., Antibody production in trans-
genic plants. Methods Mol Biol, 2004;
248:301318.
4 Ma, J. K., P. M. Drake, and P. Christou, The
production of recombinant pharmaceutical
proteins in plants. Nat Rev Genet, 2003;
4(10):794805.
5 Hellwig, S., et al., Plant cell cultures for the
production of recombinant proteins. Nat Bio-
technol, 2004; 22(11):14151422.
6 Knblein, J., Biopharmaceuticals expressed in
plants: a new era in the new Millennium. In:
Applications in Pharmaceutical Biotechnology. R.
Mller & O. Kayser (Eds.). Wiley-VCH.
7 Franck, A., et al., Nucleotide sequence of cau-
liflower mosaic virus DNA. Cell, 1980;
21(1):285294.
8 Liu, X. J., et al., Cis regulatory elements direct-
ing tuber-specific and sucrose-inducible ex-
pression of a chimeric class I patatin promo-
ter/GUS-gene fusion. Mol Gen Genet, 1990;
223(3):401406.
9 Jefferson, R., A. Goldsbrough, and M. Bevan,
Transcriptional regulation of a patatin-1 gene
in potato. Plant Mol Biol, 1990; 14(6):995
1006.
10 Scheller, J., et al., Production of spider silk
proteins in tobacco and potato. Nat Biotech-
nol, 2001; 19(6):573577.
11 Outchkourov, N. S., et al., The promoter-termi-
nator of chrysanthemum rbcS1 directs very
high expression levels in plants. Planta, 2003;
216(6):10031012.
12 Kwon, S. Y., et al., Transgenic ginseng cell
lines that produce high levels of a human lac-
toferrin. Planta Med, 2003; 69(11):10051008.
13 Tackaberry, E. S., et al., Increased yield of het-
erologous viral glycoprotein in the seeds of
homozygous transgenic tobacco plants culti-
vated underground. Genome, 2003; 46(3):521
526.
14 Horvath, H., et al., The production of recom-
binant proteins in transgenic barley grains.
Proc Natl Acad Sci USA, 2000; 97(4):1914
1919.
15 Hood, E. E., et al., Molecular farming of in-
dustrial proteins from transgenic maize. Adv
Exp Med Biol, 1999; 464:127147.
16 Park, Y. and H. Cheong, Expression and pro-
duction of recombinant human interleukin-2
in potato plants. Protein Expr Purif, 2002;
25(1):160165.
17 Christensen, A. H. and P. H. Quail, Ubiquitin
promoter-based vectors for high-level expres-
sion of selectable and/or screenable marker
genes in monocotyledonous plants. Transgenic
Res, 1996; 5(3):213218.
18 McElroy, D., et al., Isolation of an efficient ac-
tin promoter for use in rice transformation.
Plant Cell, 1990; 2(2):163171.
19 Gatz, C. and P. H. Quail, Tn10-encoded tet re-
pressor can regulate an operator-containing
plant promoter. Proc Natl Acad Sci USA,
1988; 85(5):13941397.
20 Gatz, C., C. Frohberg, and R. Wendenburg,
Stringent repression and homogeneous de-re-
pression by tetracycline of a modified CaMV
35S promoter in intact transgenic tobacco
plants. Plant J, 1992; 2(3):397404.
21 Weinmann, P., et al., A chimeric transactiva-
tor allows tetracycline-responsive gene expres-
sion in whole plants. Plant J, 1994; 5(4):559
569.
References 911
22 Mett, V. L., L. P. Lochhead, and P. H. Reynolds,
Copper-controllable gene expression system
for whole plants. Proc Natl Acad Sci USA,
1993; 90(10):45674571.
23 Aoyama, T. and N. H. Chua, A glucocorticoid-
mediated transcriptional induction system in
transgenic plants. Plant J, 1997; 11(3):605
612.
24 McNellis, T. W., et al., Glucocorticoid-inducible
expression of a bacterial avirulence gene in
transgenic Arabidopsis induces hypersensitive
cell death. Plant J, 1998; 14(2):247257.
25 Caddick, M. X., et al., An ethanol inducible
gene switch for plants used to manipulate car-
bon metabolism. Nat Biotechnol, 1998;
16(2):177180.
26 Roslan, H. A., et al., Characterization of the
ethanol-inducible alc gene-expression system
in Arabidopsis thaliana. Plant J, 2001;
28(2):225235.
27 Junker, B. H., et al., In plants the alc gene ex-
pression system responds more rapidly follow-
ing induction with acetaldehyde than with
ethanol. FEBS Lett, 2003; 535(13):136140.
28 Padidam, M., et al., Chemical-inducible, ecdy-
sone receptor-based gene expression system
for plants. Transgenic Res, 2003; 12(1):101
109.
29 Bohner, S., et al., Technical advance: transcrip-
tional activator TGV mediates dexamethasone-
inducible and tetracycline-inactivatable gene
expression. Plant J, 1999; 19(1):8795.
30 Padidam, M., Chemically regulated gene ex-
pression in plants. Curr Opin Plant Biol,
2003; 6(2):169177.
31 Mallory, A. C., et al., The amplicon-plus sys-
tem for high-level expression of transgenes in
plants. Nat Biotechnol, 2002; 20(6):622625.
32 Yang, D., et al., Expression and localization of
human lysozyme in the endosperm of trans-
genic rice. Planta, 2003; 216(4):597603.
33 Huang, J., et al., Expression of natural antimi-
crobial human lysozyme in rice grains. Molec-
ular Breeding, 2002; 10(12):8394.
34 Kusnadi, A. R., et al., Processing of transgenic
corn seed and its effect on the recovery of re-
combinant beta-glucuronidase. Biotechnol
Bioeng, 1998; 60(1):4452.
35 Evangelista, R. L., et al., Process and economic
evaluation of the extraction and purification of
recombinant beta-glucuronidase from trans-
genic corn. Biotechnol Prog, 1998; 14(4):607
614.
36 Kusnadi, A. R., et al., Production and purifica-
tion of two recombinant proteins from trans-
genic corn. Biotechnol Prog, 1998; 14(1):149
155.
37 Stoger, E., et al., Cereal crops as viable produc-
tion and storage systems for pharmaceutical
scFv antibodies. Plant Mol Biol, 2000;
42(4):583590.
38 De Jaeger, G., et al., Boosting heterologous
protein production in transgenic dicotyledo-
nous seeds using Phaseolus vulgaris regulatory
sequences. Nat Biotechnol, 2002; 20(12):1265
1268.
39 Dalton, R., California edges towards farming
drug-producing rice. Nature, 2004;
428(6983):591.
40 California rejects plan for drug-producing rice.
Nature, 2004; 428:790791.
41 Jacobs, P., Ventria Biosciences. http://
www.checkbiotech.org/root/index.cfm?fuseac-
tion=search&search=ventria&doc_id=
7582&start=1&fullsearch=0, 2004(16 April).
42 Fox, J. L., Puzzling industry response to Prodi-
Gene fiasco. Nat Biotechnol, 2003; 21(1):34.
43 Johnson, D., Stine Seed purchases biotech
company ProdiGene. http://www.
checkbiotech.org/root/index.cfm?fuseaction=
search&search=prodigene&doc_id=
5994&start=1&fullsearch=0, 2003.
44 Crabtree, P., Epicyte Pharmaceutical joins fail-
ing-biotech row. http://www.checkbiotech.org/
root/index.cfm?fuseaction=search&search=
%20epicyte&doc_id =7731&start=
1&fullsearch=1, 2004.
45 Scheller, J., et al., Purification of spider silk-
elastin from transgenic plants and application
for human chondrocyte proliferation. Trans-
genic Res, 2004; 13(1):5157.
46 Leite, A., et al., Expression of correctly pro-
cessed human growth hormone in seeds of
transgenic tobacco plants. Molecular Breeding,
2000; 6(1):4753.
47 Farran, I., et al., Targeted expression of hu-
man serum albumin to potato tubers. Trans-
genic Res, 2002; 11(4):337346.
48 Sijmons, P. C., et al., Production of correctly
processed human serum albumin in trans-
genic plants. Biotechnology (N Y), 1990
8(3):217221.
49 Ruggiero, F., et al., Triple helix assembly and
processing of human collagen produced in
transgenic tobacco plants. FEBS Lett, 2000;
469(1):132136.
50 Takase, K. and K. Hagiwara, Expression of hu-
man alpha-lactalbumin in transgenic tobacco.
J Biochem (Tokyo), 1998; 123(3):440444.
51 Lauterslager, T. G., et al., Oral immunisation
of naive and primed animals with transgenic
potato tubers expressing LT-B. Vaccine, 2001;
19(1719):27492755.
52 Azzoni, A. R., et al., Recombinant aprotinin
produced in transgenic corn seed: extraction
and purification studies. Biotechnol Bioeng,
2002; 80(3):268276.
53 Woodard, S. L., et al., Maize (Zea mays)-de-
rived bovine trypsin: characterization of the
first large-scale, commercial protein product
from transgenic plants. Biotechnol Appl Bio-
chem, 2003; 38(Pt 2):123130.
54 Chong, D. K., et al., Expression of the human
milk protein beta-casein in transgenic potato
plants. Transgenic Res, 1997; 6(4):289296.
55 Arakawa, T., et al., Expression of cholera toxin
B subunit oligomers in transgenic potato
plants. Transgenic Res, 1997; 6(6):403413.
56 Zhang, G. G., et al., Production of HIV-1 p24
protein in transgenic tobacco plants. Mol Bio-
technol, 2002; 20(2):131136.
57 Richter, L. J., et al., Production of hepatitis B
surface antigen in transgenic plants for oral
immunization. Nat Biotechnol, 2000;
18(11):11671171.
58 Joung, Y. H., et al., Expression of the hepatitis
B surface S and preS2 antigens in tubers of
Solanum tuberosum. Plant Cell Rep, 2004;
22(12):925930.
59 Chong, D. K. and W. H. Langridge, Expression
of full-length bioactive antimicrobial human
lactoferrin in potato plants. Transgenic Res,
2000; 9(1):7178.
60 Nandi, S., et al., Expression of human lactofer-
rin in transgenic rice grains for the applica-
tion in infant formula. 2002; 163:713722.
61 Samyn-Petit, V., et al., N-glycosylation poten-
tial of maize: the human lactoferrin used as a
model. Glycoconj J, 2001; 18:519527.
62 Choi, S. M., et al., High expression of a hu-
man lactoferrin in transgenic tobacco cell cul-
tures. Biotechnol Lett, 2003; 25(3):213218.
63 Stevens, L. H., et al., Effect of climate condi-
tions and plant developmental stage on the
stability of antibodies expressed in transgenic
tobacco. Plant Physiol, 2000; 124(1):173182.
64 Ko, K., et al., Function and glycosylation of
plant-derived antiviral monoclonal antibody.
8018.
65 Ko, K., et al., Elimination of alkaloids from
plant-derived human monoclonal antibody. J
Immunol Methods, 2004; 286(12):7985.
66 Ma, J. K., et al., Characterization of a recombi-
nant plant monoclonal secretory antibody and
preventive immunotherapy in humans. Nat
Med, 1998; 4(5):601606.
67 Ma, J. K., et al., Generation and assembly of
secretory antibodies in plants. Science, 1995;
268(5211):716719.
68 Khoudi, H., et al., Production of a diagnostic
monoclonal antibody in perennial alfalfa
plants. Biotechnol Bioeng, 1999; 64(2):135
143.
69 Drake, P. M., et al., Rhizosecretion of a mono-
clonal antibody protein complex from trans-
genic tobacco roots. Plant Mol Biol, 2003;
52(1):233241.
70 Fischer, R., et al., Expression and characteriza-
tion of bispecific single-chain Fv fragments
produced in transgenic plants. Eur J Biochem,
1999; 262(3):810816.
71 Fiedler, U., et al., Optimization of scFv anti-
body production in transgenic plants. Immu-
notechnology, 1997; 3(3):205216.
72 Schouten, A., et al., The C-terminal KDEL se-
quence increases the expression level of a sin-
gle-chain antibody designed to be targeted to
both the cytosol and the secretory pathway in
transgenic tobacco. Plant Mol Biol, 1996;
30(4):781793.
73 De Jaeger, G., et al., High level accumulation
of single-chain variable fragments in the cyto-
sol of transgenic Petunia hybrida. Eur J Bio-
chem, 1999; 259(12):426434.
74 Vaquero, C., et al., A carcinoembryonic anti-
gen-specific diabody produced in tobacco.
FASEB J, 2002; 16(3):408410.
75 Ohya, K., et al., Expression of two subtypes of
human IFN-alpha in transgenic potato plants.
J Interferon Cytokine Res, 2001; 21(8):595
602.
76 Biemelt, S., et al., Production of human papil-
lomavirus type 16 virus-like particles in trans-
genic plants. J Virol, 2003; 77(17):92119220.
77 Cramer, C. L., et al., Bioproduction of human
enzymes in transgenic tobacco. Ann NY Acad
Sci, 1996; 792:6271.
78 Dieryck, W., et al., Human haemoglobin from
transgenic tobacco. Nature, 1997;
386(6620):2930.
References 913
79 Terashima, M., et al., Utilization of an alterna-
tive carbon source for efficient production of
human alpha(1)-antitrypsin by genetically en-
gineered rice cell culture. Biotechnol Prog,
2001; 17(3):403406.
80 Borisjuk, N. V., et al., Production of recombi-
nant proteins in plant root exudates. Nat Bio-
technol, 1999; 17(5):466469.
81 Komarnytsky, S., et al., A quick and efficient
system for antibiotic-free expression of hetero-
logous genes in tobacco roots. Plant Cell Rep,
2004; 22(10):765773.
82 Decker, E. L. and R. Reski, The moss bioreac-
tor. Curr Opin Plant Biol, 2004; 7(2):166170.
83 Svab, Z., P. Hajdukiewicz, and P. Maliga,
Stable transformation of plastids in higher
plants. Proc Natl Acad Sci USA, 1990;
87(21):85268530.
84 Scott, S. E. and M. J. Wilkinson, Low probabil-
ity of chloroplast movement from oilseed rape
(Brassica napus) into wild Brassica rapa. Nat
Biotechnol, 1999; 17(4):390392.
85 Huang, C. Y., M. A. Ayliffe, and J. N. Timmis,
Direct measurement of the transfer rate of
chloroplast DNA into the nucleus. Nature,
2003; 422(6927):7276.
86 Stegemann, S., et al., High-frequency gene
transfer from the chloroplast genome to the
nucleus. Proc Natl Acad Sci USA, 2003;
100(15):88288833.
87 Ye, G. N., et al., Plastid-expressed 5-enolpyru-
vylshikimate-3-phosphate synthase genes pro-
vide high level glyphosate tolerance in tobac-
co. Plant J, 2001; 25(3):261270.
88 Lutz, K. A., J. E. Knapp, and P. Maliga, Expres-
sion of bar in the plastid genome confers her-
bicide resistance. Plant Physiol, 2001;
125(4):15851590.
89 Maliga, P., Plastid Transformation in Higher
Plants. Annu Rev Plant Physiol Plant Mol
Biol, 2004; 55:289313.
90 Maliga, P., Engineering the plastid genome of
higher plants. Curr Opin Plant Biol, 2002;
5(2):164172.
91 Maliga, P., Progress towards commercializa-
tion of plastid transformation technology.
Trends Biotechnol, 2003; 21(1):2028.
92 Heifetz, P. B. and A. M. Tuttle, Protein expres-
sion in plastids. Curr Opin Plant Biol, 2001;
4(2):157161.
93 Heifetz, P. B., Genetic engineering of the
chloroplast. Biochimie, 2000; 82(67):655666.
94 Bogorad, L., Engineering chloroplasts: an al-
ternative site for foreign genes, proteins, re-
actions and products. Trends Biotechnol,
2000; 18(6):257263.
95 Boynton, J. E., et al., Chloroplast transforma-
tion in Chlamydomonas with high-velocity
microprojectiles. Science, 1988;
240(4858):15341538.
96 Skarjinskaia, M., Z. Svab, and P. Maliga,
Plastid transformation in Lesquerella fendleri,
an oilseed Brassicacea. Transgenic Res,
2003; 12(1):115122.
97 Hou, B. K., et al., Chloroplast transformation
in oilseed rape. Transgenic Res, 2003;
12(1):111114.
98 Sidorov, V. A., et al., Technical Advance:
Stable chloroplast transformation in potato:
use of green fluorescent protein as a plastid
marker. Plant J, 1999; 19(2):209216.
99 Ruf, S., et al., Stable genetic transformation
of tomato plastids and expression of a for-
eign protein in fruit. Nat Biotechnol, 2001;
19(9):870875.
100 Harris, E. H., J. E. Boynton, and N. W. Gill-
ham, Chloroplast ribosomes and protein
synthesis. Microbiol Rev, 1994; 58(4):700
754.
101 Wilson, I. B., Glycosylation of proteins in
plants and invertebrates. Curr Opin Struct
Biol, 2002; 12(5):569577.
102 Staub, J. M., et al., High-yield production of
a human therapeutic protein in tobacco
chloroplasts. Nat Biotechnol, 2000;
18(3):333338.
103 Leite, A., et al., Expression of correctly pro-
cessed human growth hormone in seeds of
transgenic tobacco plants. Molecular Breed-
ing, 2000; 6:4753.
104 Leelavathi, S. and V. S. Reddy, Chloroplast
expression of His-tagged GUS-fusions: a
general strategy to overproduce and purify
foreign proteins using transplastomic plants
as bioreactors. Mol Breed, 2003; 11:4958.
105 Dale, E. C. and D. W. Ow, Gene transfer with
subsequent removal of the selection gene
from the host genome. Proc Natl Acad Sci
USA, 1991; 88(23):1055810562.
106 Corneille, S., et al., Efficient elimination of
selectable marker genes from the plastid ge-
nome by the CRE-lox site-specific recombi-
nation system. Plant J, 2001; 27(2):171178.
107 Hajdukiewicz, P. T., L. Gilbertson, and J. M.
Staub, Multiple pathways for Cre/lox-
mediated recombination in plastids. Plant J,
2001; 27(2):161170.
108 Iamtham, S. and A. Day, Removal of antibio-
tic resistance genes from transgenic tobacco
plastids. Nat Biotechnol, 2000; 18(11):1172
1176.
109 Klaus, S. M., et al., Generation of marker-
free plastid transformants using a transi-
ently cointegrated selection gene. Nat Bio-
technol, 2004; 22(2):225229.
110 Franklin, S. E. and S. P. Mayfield, Prospects
for molecular farming in the green alga
Chlamydomonas. Curr Opin Plant Biol, 2004;
7(2):159165.
111 Staub, J. M. and P. Maliga, Accumulation of
D1 polypeptide in tobacco plastids is regu-
lated via the untranslated region of the psbA
mRNA. EMBO J, 1993; 12(2):601606.
112 Carrer, H., et al., Kanamycin resistance as a
selectable marker for plastid transformation
in tobacco. Mol Gen Genet, 1993; 241(12):
4956.
113 Kuroda, H. and P. Maliga, Complementarity
of the 16S rRNA penultimate stem with se-
quences downstream of the AUG destabi-
lizes the plastid mRNAs. Nucleic Acids Res,
2001; 29(4):970975.
114 Kuroda, H. and P. Maliga, Sequences down-
stream of the translation initiation codon are
important determinants of translation effi-
ciency in chloroplasts. Plant Physiol, 2001;
125(1):430436.
115 Reddy, V. S., et al., Analysis of chloroplast
transformed tobacco plants with cry1Ia5 un-
der rice psbA transcriptional elements reveal
high level expression of Bt toxin without im-
posing yield penalty and stable inheritance of
transplastome. Mol Breed, 2002; 9:259269.
116 McBride, K. E., et al., Amplification of a chi-
meric Bacillus gene in chloroplasts leads to
an extraordinary level of an insecticidal pro-
tein in tobacco. Biotechnology (NY), 1995;
13(4):362365.
117 De Cosa, B., et al., Overexpression of the Bt
cry2Aa2 operon in chloroplasts leads to for-
mation of insecticidal crystals. Nat Biotech-
nol, 2001; 19(1):7174.
118 Kota, M., et al., Overexpression of the Bacil-
lus thuringiensis (Bt) Cry2Aa2 protein in
chloroplasts confers resistance to plants
against susceptible and Bt-resistant insects.
1845.
119 Fernndez-San Milln, A., et al., A chloro-
plast transgenic approach to hyper-express
and purify Human Serum Albumin, a pro-
tein highly susceptible to proteolytic degra-
dation. Plant Biotechnol J, 2003; 1:7179.
120 Tregoning, J. S., et al., Expression of tetanus
toxin Fragment C in tobacco chloroplasts.
Nucleic Acids Res, 2003; 31(4):11741179.
121 Kang, T. J., et al., Expression of the B sub-
unit of E. coli heat-labile enterotoxin in the
chloroplasts of plants and its characteriza-
tion. Transgenic Res, 2003; 12(6):683691.
122 Daniell, H., et al., Expression of the native
cholera toxin B subunit gene and assembly
as functional oligomers in transgenic tobac-
co chloroplasts. J Mol Biol, 2001;
311(5):10011009.
123 Birch-Machin, I., et al., Accumulation of ro-
tavirus VP6 protein in chloroplasts of trans-
plastomic tobacco is limited by protein sta-
bility. Plant Biotechnol J, 2004; 2(3): 261
270.
124 Molina, A., et al., High-yield expression of
viral peptide animal vaccine in transgenic to-
bacco chloroplasts. Plant Biotechnol J, 2004;
2:141153.
125 Leelavathi, S., et al., Overproduction of an
alkali- and thermo-stable xylanase in tobacco
chloroplasts and efficient recovery of the en-
zyme. Mol. Breed., 2003; 11:5967.
126 Watson, J., et al., Expression of Bacillus an-
thracis protective antigen in transgenic
chloroplasts of tobacco, a non-food/feed
crop. Vaccine, 2004; 22(3132):43744384.
127 Mayfield, S. P., S. E. Franklin, and R. A. Ler-
ner, Expression and assembly of a fully ac-
tive antibody in algae. Proc Natl Acad Sci
USA, 2003; 100(2):438442.
128 Franklin, S., et al., Development of a GFP
reporter gene for Chlamydomonas reinhardtii
chloroplast. Plant J, 2002; 30(6):733744.
129 Ishikura, K., et al., Expression of a foreign
gene in Chlamydomonas reinhardtii chloro-
plast. J Biosci Bioeng, 1999; 87:307314.
130 Fischer, R., et al., Towards molecular farm-
ing in the future: transient protein expres-
sion in plants. Biotechnol Appl Biochem,
1999; 30 (Pt 2):113116.
131 Vaquero, C., et al., Transient expression of a
tumor-specific single-chain fragment and a
chimeric antibody in tobacco leaves. Proc
Natl Acad Sci USA, 1999; 96(20):11128
11133.
References 915
132 Voinnet, O., et al., An enhanced transient
expression system in plants based on sup-
pression of gene silencing by the p19 pro-
tein of tomato bushy stunt virus. Plant J,
2003; 33(5):949956.
133 Marillonnet, S., et al., In planta engineering
of viral RNA replicons: efficient assembly by
recombination of DNA modules delivered by
Agrobacterium. Proc Natl Acad Sci USA,
2004; 101(18):68526857.
134 Awram, P., et al., The potential of plant viral
vectors and transgenic plants for subunit
vaccine production. Adv Virus Res, 2002;
58:81124.
135 Pogue, G. P., et al., Making an ally from an
enemy: plant virology and the new agricul-
ture. Annu Rev Phytopathol, 2002; 40:4574.
136 Porta, C. and G. P. Lomonossoff, Viruses as
vectors for the expression of foreign se-
quences in plants. Biotechnol Genet Eng
Rev, 2002; 19:245291.
137 Scholthof, K. B., T. E. Mirkov, and H. B.
Scholthof, Plant virus gene vectors: biotech-
nology applications in agriculture and medi-
cine. Genet Eng (NY), 2002; 24:6785.
138 Mor, T. S., et al., Geminivirus vectors for
high-level expression of foreign proteins in
plant cells. Biotechnol Bioeng, 2003;
81(4):430437.
139 Gleba, Y., S. Marillonnet, and V. Klimyuk,
Engineering viral expression vectors for
plants: the full virus and the deconstructed
virus strategies. Curr Opin Plant Biol, 2004;
7(2):182188.
140 Turpen, T. H., et al., Malarial epitopes ex-
pressed on the surface of recombinant tobac-
co mosaic virus. Biotechnology (NY), 1995;
13(1):5357.
141 Lomonossoff, G. P. and W. D. Hamilton,
Cowpea mosaic virus-based vaccines. Curr
Top Microbiol Immunol, 1999; 240:177189.
142 Brennan, F. R., T. D. Jones, and W. D. Hamil-
ton, Cowpea mosaic virus as a vaccine car-
rier of heterologous antigens. Mol Biotech-
nol, 2001; 17(1):1526.
143 Chatterji, A., et al., Cowpea mosaic virus:
from the presentation of antigenic peptides
to the display of active biomaterials. Intervi-
rology, 2002; 45(46):362370.
144 Nemchinov, L. G., et al., Development of a
plant-derived subunit vaccine candidate
against hepatitis C virus. Arch Virol, 2000;
145(12):25572573.
145 Bendahmane, M., et al., Display of epitopes
on the surface of tobacco mosaic virus: im-
pact of charge and isoelectric point of the
epitope on virushost interactions. J Mol
Biol, 1999; 290(1):920.
146 Wu, L., et al., Expression of foot-and-mouth
disease virus epitopes in tobacco by a tobac-
co mosaic virus-based vector. Vaccine, 2003;
21(2730):43904398.
147 Belanger, H., et al., Human respiratory syn-
cytial virus vaccine antigen produced in
plants. FASEB J, 2000; 14(14):23232328.
148 Angell, S. M. and D. C. Baulcombe, Consis-
tent gene silencing in transgenic plants ex-
pressing a replicating potato virus X RNA.
EMBO J, 1997; 16(12):36753684.
149 Brisson, N., et al., Expression of a bacterial
gene in plants by using a viral vector. Na-
ture, 1984; 310:511514.
150 Lefebvre, D., B. Miki, and J. Laliberte, Mam-
malian metallothionein functions in plants.
Bio-technology, 1987; 5:10531056.
151 De Zoeten, G. A., et al., The expression, lo-
calization, and effect of a human interferon
in plants. Virology, 1989; 172(1):213222.
152 Mori, M., et al., Efficient production of hu-
man gamma interferon in tobacco proto-
plasts by genetically engineered brome mo-
saic virus RNAs. J Gen Virol, 1993; 74 (Pt
7): 12551260.
153 McCormick, A. A., et al., Rapid production
of specific vaccines for lymphoma by expres-
sion of the tumor-derived single-chain Fv
epitopes in tobacco plants. Proc Natl Acad
Sci USA, 1999; 96(2):703708.
154 McCormick, A. A., et al., Individualized hu-
man scFv vaccines produced in plants: hu-
moral anti-idiotype responses in vaccinated
mice confirm relevance to the tumor Ig. J
Immunol Methods, 2003; 278(12):95104.
155 Verch, T., V. Yusibov, and H. Koprowski, Ex-
pression and assembly of a full-length
monoclonal antibody in plants using a plant
virus vector. J Immunol Methods, 1998;
220(12):6975.
156 Perez Filgueira, D. M., et al., Bovine herpes
virus gD protein produced in plants using a
recombinant tobacco mosaic virus (TMV)
vector possesses authentic antigenicity. Vac-
cine, 2003; 21(2730):42014209.
157 Kumagai, M. H., et al., Rapid, high-level ex-
pression of glycosylated rice alpha-amylase
in transfected plants by an RNA viral vector.
Gene, 2000; 245(1):169174.
158 Krebitz, M., et al., Rapid production of the
major birch pollen allergen Bet v 1 in Nicoti-
ana benthamiana plants and its immunologi-
cal in vitro and in vivo characterization.
FASEB J, 2000; 14(10):12791288.
159 Li, Y., et al., Expression of a human lactofer-
rin N-lobe in Nicotiana benthmiana with po-
tato virus X-based agroinfection. Biotechnol
Lett, 2004; 26(12):953957.
160 Hiatt, A., Antibodies produced in plants. Na-
ture, 1990; 344(6265):469470.
161 Hiatt, A., R. Cafferkey, and K. Bowdish, Pro-
duction of antibodies in transgenic plants.
Nature, 1989; 342(6245):7678.
162 During, K., et al., Synthesis and self-assem-
bly of a functional monoclonal antibody in
transgenic Nicotiana tabacum. Plant Mol
Biol, 1990; 15(2):281293.
163 Gleba, Y., S. Marillonnet, and V. Klimyuk,
Design of Safe and Biologically Contained
Transgenic Plants: Tools and Technologies
for Controlled Transgene Flow and Expres-
sion. Biotechnol Genet Eng Rev, 2004;
21:325367.
164 Daniell, H., S. J. Streatfield, and K. Wycoff,
Medical molecular farming: production of
antibodies, biopharmaceuticals and edible
vaccines in plants. Trends Plant Sci, 2001;
6(5):219226.
References 917
Abstract
Genetically engineered plants are promis-
ing systems for the production of biophar-
maceutical proteins. Among the different
plant-based systems, the moss bioreactor
shows unique properties. Mosses are culti-
vated as haploid, photo-autotrophically ac-
tive and fully differentiated gametophytic
tissue performed as suspension cultures in
bioreactors. In addition, moss is the only
known plant system which shows a high
frequency of homologous recombination
which allows for gene knock-outs, thereby
opening the possibility of genetic engineer-
ing of the glycosylation pathway. Here, we
present an overview of the biotechnologi-
cally relevant aspects of mosses, with a
special emphasis on glycoengineering per-
formed in Physcomitrella patens.
Abbreviations
ADCC antibody-dependent cell-
mediated cytotoxicity
CHO chinese hamster ovary
EST expressed sequence tag
FucT alpha 1,3-fucosyltransferase
GlcNAC N-acetylglucosaminyl resi-
due
GNT I N-acetylglucosaminyltrans-
ferase I
GNT II N-acetylglucosaminyltrans-
ferase II
GNT glucosaminyltransferase
hVEGF human vascular endothelial
growth factor
MALDI-TOF matrix assisted laser de-
sorption/ionization time-of-
flight
Mbp mega base pairs
rHSA human serum albumin
rhVEGF recombinant human vascu-
lar endothelial growth fac-
tor
XylT beta 1,2-xylosyltransferase
7.1
Introduction
Until now, biopharmaceuticals have been
produced either in microorganisms such
as E. coli or in animal cell cultures (e.g.,
Chinese hamster ovary (CHO) cells), if the
therapeutic protein requires complex post-
translational modification (see Part IV,
Chapters 1, 2, 3, 5, 12 and 13).
An expected shortage of manufacturing
capacities, as well as safety issues in terms
of virus load and contamination with TSE,
919
7
Humanized Glycosylation: Production of Biopharmaceuticals
in a Moss Bioreactor
ISBN: 3-527-31184-X
resulted in the search for alternative pro-
duction systems [1]. Plants are the most
promising production organisms for bio-
pharmaceuticals because of their cost-effi-
cient upstream processes [2] and excellent
safety aspects. In addition to molecular
pharming performed in greenhouses or on
the field, secretion-based plant systems
such as rhizosecretion from tobacco roots
have been developed [35; for a review, see
[6]]. Secretion of the target protein into the
medium is a major improvement for the
downstream process, because extraction
and purification of proteins from plant tis-
sues is a complex and costly process, as re-
cently reviewed by Knblein [7].
However, the major limitation for the use
of biopharmaceuticals produced by plants is
the glycosylation pattern. Animal and plant
N-linked glycosylation patterns are identical
in the core structure, but there are differ-
ences in the additional sugar residues. Plant
N-glycans contain beta 1,2-linked xylose and
alpha 1,3-linked fucose residues, whereas
the beta 1,4-linked terminal galactose resi-
due, which is typical for animal-derived gly-
coproteins (e.g., on antibodies) is not pres-
ent in plants. There is also some evidence
that plant-specific residues have immuno-
genic potential [8, 9].
Here, we present the moss bioreactor,
which is based on secretion of the target
protein into a simple medium [10] and, by
humanization of the N-glycans, avoids
plant-specific immunogenicity.
7.2
Mosses: Some General Aspects
In contrast to higher plants, the main
phase of the life cycle of mosses is not the
diploid sporophyte but the haploid, photo-
synthetically active gametophyte. The ga-
metophyte consists of the filamentous pro-
tonema, which shows apical growth, and
of the morphologically more distinct ga-
metophore (Fig. 7.1). The differentiation
steps in the development of the moss ga-
metophyte are clearly defined. Growth of
the first cell type, the chloronema, begins
after germination of spores or protoplasts.
The second cell-type in the branched pro-
tonema is the caulonema, on which buds
and later the complete gametophore are
developing. Chloronema and caulonema
cells can be distinguished not only by their
morphology but also by their predominant
cell-cycle phases [11]. Sporophyte develop-
ment occurs only under specific conditions
[12], and therefore mosses are propagated
in general vegetatively without sporulation
(for a review, see Refs. [13, 14]). The long-
term storage of mosses is performed on
solid medium, as well as by cryopreserva-
tion, with regrowth rates of 100% [15, 16].
Most investigations carried out with
Physcomitrella were based on a strain
which was collected during the 1960s by
H. L. K. Whitehouse. Engel [17] established
the in-vitro culture by subcultivation of
plant material grown from one spore. The
haploid nature of the gametophytic tissue,
the clearly defined differentiation pattern,
and the simple cultivation parameters have
provided an excellent basis for scientific
studies, and consequently many genetic
and physiological studies of Physcomitrella
have been performed during the past four
decades (e.g., [18]). The establishment and
optimization of protoplast isolation [19, 20]
opened the possibility for further develop-
ment of transformation methods. Interest-
ingly, agrobacteria-mediated transforma-
tion one of the main transformation
methods used in higher-plant technology
is not applicable to mosses due to the lack
of a useful agro strain. Biolistic transfor-
mation [21] and electroporation [22] can be
used, but neither method is particularly
7 Humanized Glycosylation: Production of Biopharmaceuticals in a Moss Bioreactor 920
sufficient. Polyethyleneglycol (PEG)-
mediated DNA transfer is the method of
choice for moss transformation (for a re-
view, see Ref. [23]), the protocol being
quite different to that used for higher
plants for example, there is no require-
ment for a cooling step [24, 25]. For re-
combinant expression in mosses, in most
cases constitutive active heterologous regu-
latory sequences such as the 35S promoter
or the rice actin 1 5' region were used [24,
26]. Inducible expression has also been de-
scribed for moss; for example, Zeidler et
al. [27] used the tetracycline-based Top 10
system successfully, whilst Knight et al.
[28] described expression driven by the
Em-promoter in Physcomitrella. Although
many expressed sequence tag (EST) data
were available for Physcomitrella (see be-
low), only recently the first endogenous
regulatory sequences from moss were
characterized [2931].
During the 1990s, Physcomitrella was dis-
covered as a new tool for functional geno-
mics and, for the first time in plants,
highly efficient targeted homologous re-
combination into the genome was de-
scribed for Physcomitrella [25, 32, 33]. Sub-
sequently, numerous molecular data have
been generated in this respect. The ge-
nome size was determined as 511 Mbp,
which is three- to four-fold that of Arabi-
dopsis thaliana [11]. Some 95% of the Phys-
comitrella transcriptome is known, based
on EST data from about 25000 protein-
coding genes [34, 35]. On the basis of
these data, a calculation of codon usage
was possible, and this resulted in there
being no significant preferences for Phys-
comitrella [36]. Thus, no codon optimiza-
tion is necessary for the recombinant ex-
pression of human proteins in moss. Tak-
ing these results together, it is clear that
Physcomitrella is indeed a well-character-
ized organism.
Fig. 7.1 Physcomitrella patens. Left: highly homogeneous protonema tissue. Right: a leafy gametophore.
7.3
Cell Culture
Mosses can be grown on solid medium as
well as in liquid suspension cultures. In
general, the moss tissue is cultivated un-
der photoautotrophic conditions. Light, air,
and a simple medium based on a mineral
salt composition without any sugars or
plant hormones are sufficient to cultivate
the fully differentiated tissue. Mosses can
be grown with ammonium or with nitrate
as nitrogen sources [37]. Although the op-
timal growth conditions are known for
many moss species, the composition of
medium components can be varied
broadly. Mosses can also be grown under
heterotrophic conditions, though the addi-
tion of sugar to the medium results in a
marked production of secondary metabo-
lites [38].
By utilizing medium supplementation
with plant hormones or additives such as
ammonium tartrate, the differentiation of
moss tissue can be influenced in a well-de-
fined manner. Whereas the addition of
auxin promotes the development of caulo-
nema cells, the addition of ammonium tar-
trate results in an arrest on the chlorone-
ma stage of the moss protonema [11]. The
latter effect is of interest for biotechnologi-
cal applications, because of the high
homogeneity of the cells. Although fila-
ments consisting of chloronema cells are
fully differentiated, only one type of cells
is present in such cultures. Moreover,
these cells are arrested in the G
2
/M phase
of the cell cycle [11]. In conclusion,
mosses can be grown as an extremely
homogeneous and well-defined suspension
culture, whilst the tissue material is both
fully differentiated and photosynthetically
active.
Liquid cultivation can be performed as
suspension culture not only in Erlenmeyer
flasks but also in bioreactors. Mosses were
cultivated in photo-bioreactors in stirred
glass tanks [39], as well as under airlift
conditions [26, 40]. Supplementation with
CO
2
and light are the major parameters
that influence the growth rates. Since light
is a limiting factor for the large-scale
photoautotrophic cultivation of mosses, a
glass tube reactor was developed. Based on
technology already established for the
large-scale photoautotrophic cultivation of
algae [41], the cultivation of the filamen-
tous protonema under sterile conditions
required significant adaptations [42]. The
prototype of the photobioreactor has a
working volume of 30 L (Fig. 7.2).
Down-scaling of the suspension cultures
to microtiter plates was successfully estab-
lished for testing culture supplementations
[43]. This allows efficient and rapid medi-
um optimization, and also opens the pos-
sibility of automating the upstream devel-
opment. In combination, the overall direct
transformation, clonal growth without any
crossing steps, and the technical ability to
speed the process development will mini-
mize the time-to-market for biopharma-
ceuticals produced in moss.
Fig. 7.2 A photobioreactor. The tubular photobio-
reactor design allows scaling of moss suspension
culture up to large volumes. (Illustration courtesy
of Prof. C. Posten, Karlsruhe.)
7.4
Recombinant Expression
Antibiotic resistance markers and reporter
genes were the first heterologous proteins
to be expressed in mosses [24, 26, 44]. The
production of human proteins in moss
was first shown by expression of human
vascular endothelial growth factor 121
(rhVEGF). rhVEGF was successfully tar-
geted to the secretory pathway, and this re-
sulted in an efficient secretion of the re-
combinant protein into the medium.
Moreover, the moss-derived rhVEGF was
shown to be biologically active [45, 46]. An
important criterion for the successful ex-
pression of a therapeutic protein from a
recombinant cell is to obtain a transgenic
plant that maintains stability, both of pro-
duction and at the molecular level. Several
transgenic moss strains which were aged
between 2 and 7 years were examined with
regard to expression of the target protein
rhVEGF and neomycin phosphotransferase
as an antibiotic resistance marker. Protein
levels of rhVEGF were measured using an
enzyme-linked immunosorbent assay (ELI-
SA), and found to be unchanged. Further-
more, 100% of the transgenic plant mate-
rial showed resistance to the antibiotic
G418, even after several years of cultiva-
tion without selection pressure. In addi-
tion, analysis at the molecular level con-
firmed the protein data [47]. Thus, the
moss appears to be an ideal production
system for biopharmaceuticals under strict
regulatory requirements. rhVEGF was
further used to optimize expression in
Physcomitrella, with molecular tools such
as endogenous promoters, 5' regulatory re-
gions and signal peptides being isolated
and characterized [30, 31, 48]. In the
meantime, as in other plant expression
systems, different human proteins were
successfully expressed in mosses, includ-
ing fully assembled antibodies [49] and hu-
man serum albumin (rHSA). Indeed, the
co-expression of rHSA simultaneously
with rhVEGF led to the development of a
new approach for enhanced protein recov-
ery [43].
Transient expression systems established
for higher plants rely on virus infection or
agroinfiltration [5052] (see Part IV, Chap-
ters 5 and 6; Part IV, Chapter 9). Due to
the lack of moss viruses and suitable agro-
bacteria strains, none of these methods is
applicable for mosses. Nevertheless, transi-
ent expression is a useful tool not only for
the rapid analysis of molecular tools but
also for the expression of recombinant pro-
tein for first characterization. Transient
transformation was developed and opti-
mized, and resulted in the expression of
up to 10 lg mL
1
rhVEGF [46]. In addi-
tion, this transient expression system was
used for promoter analysis [30, 31] and for
the analysis of secretion capacity [46].
Analysis of different expression vectors in
the transient system allows the selection
of the optimal combinations (e.g., promo-
ter, signal peptide) for each target protein
in a short time. Therefore, the time-con-
suming generation of stably transformed
plants can be focused on the optimal ex-
pression vectors at a very early stage of the
process. Overall, the transient system can
be used for further screening of molecular
tools to improve recombinant protein ex-
pression in mosses.
On the protein level, transient expres-
sion allows feasibility studies to be con-
ducted rapidly, as it enables production of
limited amounts of protein for first analy-
sis within weeks.
7.5
N-Glycosylation
The production of complex biopharmaceuti-
cals is closely associated with glycosylation
issues. N-glycosylation can be responsible
for protein folding [53] and prevention of
protein degradation in the cells, as well as
metabolism in the liver of mammals [54]
(see Part IV, Chapters 1 and 3). In addition,
effector functions such as antibody-depen-
dent cell-mediated cytotoxicity (ADCC) have
been discussed as being linked to N-glycan
structures [5557] (see Part I, Chapter 15
and Part V, Chapter 1). The process of N-gly-
cosylation seems to be highly conserved in
most eukaryotes, and in particular a mini-
mal core structure consisting of Man
3-
GlcNAc
2
is common for N-glycans (see Part
IV, Chapter 2 and Part VI, Chapter 2). In
glycoproteins, N-glycans are covalently
linked to the asparagine (Asn) residues of
the tripeptide Asn-X-Ser/Thr, where X can
be any amino acid except aspartic acid and
proline. N-glycosylation starts in the endo-
plasmic reticulum (ER). An oligosaccharide
precursor is transferred to the Asn residue
and further processed to a high-mannose
structure consisting of Man
9
GlcNAc
2
. The
next step in this process occurs in the Golgi
apparatus. Mannose residues are removed
by alpha-mannosidase I, followed by the ad-
dition of a GlcNAc residue to the terminal
mannose of one branch by the enzyme N-
acetyl glucosaminyltransferase I (GNTI).
The remaining mannose residues of the
high-mannose structure are processed by a
second, Golgi-located mannosidase (alpha-
ManII). The resulting structure, Man
3-
GlcNAc
3
, is used as a substrate by GNTII,
which transfers a second terminal GlcNAc
residue to the N-glycan resulting in com-
plex-type N-glycans.
Further modifications are different in
plants and mammals. In plants, an alpha
1,3-linked fucosylation to the first core
GlcNAc, and a beta 1,2-linked xylosylation
to the first core mannose residue, are
mediated by specific glycosyltransferases
(for a review, see Ref. [58]). These struc-
tures are common for plants, including
mosses [59, 60].
In contrast, mammalian N-glycans con-
tain alpha 1,6 fucosyl residues linked to
the first core GlcNAc. The terminal struc-
tures of mammalian N-glycans can be pro-
cessed in a much more complex manner
compared to plant structures. Galactose re-
sidues are attached to the terminal
GlcNAcs in 1,4-linkage. In mammals, sia-
lyltransferases use galactose-containing N-
glycans as substrates for further proces-
sing. Depending on the glycoproteins, the
N-glycan structures can be completely dif-
ferent. Whereas coagulation factors such
as factor IX are highly sialylated [61], anti-
bodies (if at all) contain only minor frac-
tions of sialylated N-glycans [55].
The major drawback for plant-derived
complex biopharmaceuticals is that the
plant-specific xylosyl and fucosyl residues
are attached to the core structure of N-gly-
cans. Although Chargelegue et al. [62] ob-
served no immunogenic effects of a plant-
derived murine monoclonal antibody in an
animal study based on a mouse model,
both residues are described in the litera-
ture as structures with high immunogenic
potential [8, 9].
Consequently, major efforts were made
to overcome this limitation. One approach
is based on an observation by von Schae-
wen et al. [63]. These authors isolated a
mutant strain of Arabidopsis thaliana (cgl)
which showed a loss of GNTI activity. All
N-glycan structures isolated from this mu-
tant strain were related to high mannose-
type, and no complex-type structures were
detected. The loss of complex-type N-gly-
cans was accompanied by the loss of the
plant-specific sugar residues on the core
structure. From these data it was known
that high-mannose structures, which are
processed in the ER, are not substrates for
the plant-specific alpha 1,3-fucosyltransfer-
ase (FucT) and the beta 1,2-xylosyltransfer-
ase (XylT). It was also shown that target-
ing of the recombinant protein to the ER
by attaching the ER retention signal KDEL
results in N-glycans of only the high-man-
nose type on these proteins, which showed
poor stability after injection into mice [64].
A second approach was based on anti-
sense technology (see Part I, Chapter 10;
Part III, Chapter 3; and the Introduction).
In this study, antisense constructs were de-
signed for the GNTI of Nicotiana benthami-
ana. Whereas reduction of GNTI activity to
2% was very successful, only minor changes
in N-glycan composition were observed.
The remaining very low GNTI activity
seemed to be sufficient for close to normal
N-glycan processing [65]. A completely
new approach was developed by Koprivova
et al. [66], who isolated the gene coding
for GNTI from Physcomitrella patens. Based
on the highly efficient homologous recom-
bination in mosses (see above), these
authors performed targeted disruption of
gntI. Although the gntI knock-out (ko) was
successful, the remaining N-glycan struc-
tures in the ko plants were similar to that
of the wild-type including the complex-
type structures. Again, the effect was com-
pensated, which illustrates the overall com-
plexity of the glycosylation pathway. Never-
theless, the putative genes coding for alpha
1,3-FucT and beta 1,2-XylT were isolated
from Physcomitrella [59], and knock-out con-
structs were designed. To remove the im-
munogenic potential of both plant-specific
N-glycans completely, plants containing
double knock-outs were generated. MAL-
DI-TOFanalysis of the remaining N-glycans
clearly demonstrated the absence of plant-
specific sugar residues in the double
knock-out, respectively [66]. Double knock-
outs were performed not only through gene
disruption but also through complete gene
replacement. All approaches resulted in
complete loss of the putative immunogenic
sugar residues, thus confirming the dys-
functional character of the genes and the
loss of corresponding glycosyltransferase
activity in all cases (Fig. 7.3). Although the
N-glycan structures with the xylosyl and fu-
cosyl residues are highly conserved over the
whole plant kingdom, surprisingly the dou-
ble knock-out plants showed no differences
in growth and differentiation. Regeneration
of the plants was also similar to that of the
wild-type, whilst the secretion capacity of
rhVEGF was as high as in the wild-type [46].
Human-like terminal beta 1,4-galactosy-
lation was recently described in tobacco
plants [67] and in tobacco-derived BY2 sus-
pension cell cultures [68] recombinantly
expressing the mammalian enzyme. More-
over, the extracellular proteins of the trans-
genic BY2 cell line GT6 recombinantly ex-
pressing the human beta 1,4-galactosyl-
transferase also contain galactose-extended
N-glycans [69]. Surprisingly, the same re-
sults were obtained by expression of the
human galactosyltransferase in the game-
tophytic tissue of mosses which already
lacks the plant-specific sugar residues due
to its double knock-out character, and
therefore shows extensive manipulation of
the glycosylation pathway [70].
Taken together, humanized glycosylation
at least sufficient for antibody production
was achieved in genetically engineered
moss strains, without any negative influ-
ence on growth rates in bioreactor cultures
or on their secretion capacity for recombi-
nant proteins.
7.6
Conclusions and Outlook
Physcomitrella is a well-established plant
system that, since the 1960s, has been cul-
tivated under axenic conditions in many
laboratories. It can be cultivated in liquid
cultures, and such photobioreactors are
scaleable. Transgenic moss strains show
high genetic stability. Different biopharma-
ceuticals, including human growth factor,
antibodies and non-glycosylated proteins
such as HSA were successfully expressed
and secreted into the medium.
Molecular tools such as promoters, other
regulatory sequences and signal peptides
were isolated and optimized for expression
in mosses. Together with its high secretion
capacity, the moss therefore allows the de-
velopment of high-production strains. By
using directed knock-out of plant-specific
glycosyltransferases via homologous re-
combination, and through the simulta-
neous introduction of human galactosyl-
transferase, the humanization of glycosyla-
tion was successfully achieved in Physcomi-
trella. This not only removes the
immunogenic potential of plant-produced
proteins for injection, but also opens per-
spectives for glycosylation design. With
further understanding of the function of
protein glycosylation, this will include on
the one hand improved effector function
of monoclonal antibodies such as ADCC.
On the other hand, glycosylation seems to
play a role in pharmacokinetics which is
only partly understood. The establishment
of sialylation in plants will be a major step
towards the production of authentic bio-
pharmaceuticals.
References
1 J. Knblein, Screening Trends in Drug Dis-
cov., 2003, 4, 1416.
2 J. Knblein, M. McCaman, Screening Trends
in Drug Discov., 2003, 6, 3335.
3 R. Wongsamuth, P. M. Doran, Biotechnol.
Bioeng., 1997, 54, 401415.
4 N. V. Borisjuk, L. G Borisjuk, S. Logendra, F.
Petersen, Y. Gleba, I. Raskin, Nat. Biotechnol.,
1999, 17, 466469.
5 P. M. Drake, D. M. Chargelegue, N. D. Vine,
C. J. van Dolleweerd, P. Obregon, J. K. Ma,
Plant Mol. Biol., 2003, 52, 233241.
6 J. K. Ma, P. M. Drake, P. Christou, Nat. Rev.
Genet. 2003, 4, 794805.
7 J. Knblein, Biopharmaceuticals expressed in
plants a new era in the new Millennium.
In: Applications in Pharmaceutical Biotechnol-
ogy. R. Mller & O. Kayser (Eds.). Wiley-VCH.
8 G. Garcia-Casado, R. Sanchez-Monge, M. J.
Chrispeels, A. Armentia, G. Salcedo, L. Go-
mez, Glycobiology, 1996, 6, 471477.
9 M. Bardor, C. Faveeuw, A. C. Fitchette, D. Gil-
bert, L. Galas, F. Trottein, L. Faye, P. Lerouge,
Glycobiology, 2003, 13, 427434.
10 E. L. Decker, G. Gorr, R. Reski, BIOforum Eu-
rope, 2003, 2, 9697.
11 G. Schween, G. Gorr, A. Hohe, R. Reski,
Plant Biol., 2003, 5, 5058.
12 A. Hohe, S. A. Rensing, M. Mildner, D. Lang,
R. Reski, Plant Biol., 2002, 4, 595602.
13 R. Reski, Bot. Acta, 1998, 111, 115.
14 R. Reski, D. J. Cove, Current Biol., 2004, 14,
R261R262.
References 927
Fig. 7.3 Glycoengineering in mosses. Mass spec-
troscopic analysis of N-glycans in mosses (Prof.
F. Altmann, Vienna) performed by matrix-assisted
laser desorption ionization time of flight (MALDI-
TOF) spectroscopy. Top: N-glycan analysis of wild-
type. Bottom: N-glycan analysis of a transgenic
moss strain lacking plant-specific sugar residues.
The mass shift of the major N-glycan structure in
wild-type to the complex-type GnGnXF to the ma-
jor structure GnGn [lacking the plant-specific 1,3-
linked fucosyl (F) and 1,2-linked xylosyl (X) resi-
dues] in the glycoengineered transgenic plant is
marked by an arrow.
3
15 N. H. Grimsley, L. A.Withers, Cryo-Letters,
1983, 4, 251258.
16 J. Schulte, R. Reski, Plant Biol., 2004, 6, 119
127.
17 P. P. Engel, Am. J. Bot. 1964, 55, 438446.
18 D. J. Cove, A. Schild, N. W. Ashton, E. Hart-
mann, Photochem. Photobiol., 1978, 27, 249
254.
19 S. Rother, B. Hadeler, J. M. Orsini, W. O. Abel,
R. Reski, J. Plant Physiol. 1994, 143, 7277.
20 A. Hohe, R. Reski, Plant Sci., 2002, 163, 69
74.
21 W. Sawahel, S. Onde, C. Knight, D. J. Cove,
Plant Mol. Biol. Rep., 1992, 10, 314315.
22 D. G. Schaefer, G. Bisztray, J.-P. Zryd, in: Bio-
technology in Agriculture and Forestry, Vol. 29,
Plant Protoplasts and Genetic Engineering V
(Y. P. S. Bajaj, Ed.). Springer-Verlag, Berlin
Heidelberg, 1994, Chapter II.11, 349364.
23 D. G. Schaefer, Annu. Rev. Plant Physiol.
Plant Mol. Biol., 2002, 477501.
24 M. Zeidler, E. Hartmann, J. Hughes, J. Plant
Physiol., 1999, 154, 641650.
25 R. Strepp, S. Scholz, S. Kruse, V. Speth, R.
Reski, Proc. Natl. Acad. Sci. USA, 1998, 95,
43684373.
26 K. Reutter, R. Reski, Pl. Tissue Cult. Biotech.,
1996, 2, 142147.
27 M. Zeidler, C. Gatz, E. Hartmann, J. Hughes,
Plant Mol. Biol., 1996, 30, 199205.
28 C. D. Knight, A. Sehgal, K. Atwal, J. C. Wal-
lace, D. J. Cove, D. Coates, R. S. Quatrano, S.
Bahadur, P. G. Stockley, A. C. Cuming, Plant
Cell, 1995, 7, 499506.
29 V. Horstmann, C. M. Huether, W. Jost, R. Re-
ski, E. L. Decker, BMC Biotechnol, 2004, 4,
13.
30 W. Jost, S. Link, V. Horstmann, E. L. Decker,
R. Reski, G. Gorr, Curr. Genet., 2005, 47,
111120.
31 A. Weise, M. Rodriguez-Franco, B. Timm, M.
Hermann, S. Link, W. Jost, G. Gorr, Isolation
of four members of a plant actin gene family
with remarkable gene structures and the use
of their 5 regions for high transgene expres-
sion. (submitted).
32 D. G. Schaefer, J. P. Zryd, Plant J., 1997, 11,
11951206.
33 T. Girke, H. Schmidt, U. Zahringer, R. Reski,
E. Heinz, Plant J., 1998, 15, 3948.
34 S. A. Rensing, S. Rombauts, Y. van de Peer, R.
Reski, Trends Plant Sci., 2002, 7, 535538.
35 T. Nishiyama, T. Fujita, T. Shin-I, M. Seki, H.
Nishide, I. Uchiyama, A. Kamiya, P. Carninci,
Y. Hayashizaki, K. Shinozaki, Y. Kohara, M.
Hasebe, Proc. Natl. Acad. Sci. USA, 2003,
100, 80078012.
36 E. L. Decker, R. Reski, Curr. Opin. Plant Biol.,
2004, 7, 166170.
37 M. Bopp, B. Knoop, Culture methods for
bryophytes, in: Cell culture and somatic cell ge-
netics of plants (J. K. Vasil, Ed.), 1984, 1, chap-
ter 12, 96105.
38 P. E. Simon, J. B. Naef, Physiol. Plant., 1981,
53, 1318.
39 A. Hohe, E. L. Decker, G. Gorr, G. Schween,
R. Reski, Plant Cell Rep., 2002, 20, 1135
1140.
40 P. J. Boyd, J. Hall, D. J. Cove, in: Methods in
bryology, Proc. Bryol. Meth. Workshop Mainz
(J. M. Glime, Ed.), Hattori Bot. Lab., Nichinan,
1988, 4145.
41 O. Pulz, Appl. Microbiol. Biotechnol., 2001,
57, 287293.
42 A. Lucumi, C. Posten, Improved in vitro cul-
ture of the moss Physcomitrella patens in a tu-
bular pilot photobioreactor. (submitted).
43 A. Baur, R. Reski, G. Gorr, Enhanced recovery
of a secreted recombinant human growth fac-
tor by stabilising additives and by coexpres-
sion of human serum albumin in the moss
Physcomitrella patens. Plant Biotechnol. J., in
press.
44 D. Schaefer, J.-P. Zryd, C. D. Knight, D. J.
Cove, Mol. Gen. Genet., 1991 226, 418424
45 G. Gorr, E. L. Decker, M. Kietzmann, R. Reski,
Naunyn-Schmiedebergs Arch. Pharmacol.,
2001, 363, Suppl: R 85.
46 A. Baur, F. Kaufmann, H. Rolli, A. Weise, R.
Luethje, B. Berg, M. Braun, W. Baeumer, M.
Kietzmann, R. Reski, G. Gorr A fast and flex-
ible PEG-mediated transient expression sys-
tem in plants for high level expression of se-
creted recombinant proteins. J. Biotechnol., in
press.
47 A. Baur, B. Timm, A. Weise, M. Rodriguez-
Franco, G. Gorr, Genetic stability of transgenic
Physcomitrella patens plants, Moss 2004, The
7th Annual Moss International Conference,
Freiburg, Germany.
48 A. Schaaf, R. Reski, E. L. Decker, Eur. J. Cell
Biol., 2004, 83, 145152.
49 W. Jost, E. Schulze, M. Rodriguez-Franco, A.
Weise, G. Gorr, Naunyn-Schmiedebergs Arch.
Pharmacol., 2004, 369, Suppl: R 80.
50 R. Fischer, C. Vaquero-Martin, M. Sack, J.
Drossard, N. Emans, U. Commandeur, Bio-
technol. Appl. Biochem., 1999, 30, 113116.
51 C. Vaquero, M. Sack, J. Chandler, J. Drossard,
F. Schuster, M. Monecke, S. Schillberg, R.
Fischer, Proc. Natl. Acad. Sci. USA, 1999, 96,
1112811133.
52 V. Klimyuk, S. Marillonnet, J. Knblein, M.
McCaman, Y. Gleba, Chapter 6.
53 A. Helenius, Mol. Biol. Cell, 1994, 5, 253265.
54 A. G. Morell, G. Gregoriadis, I. H. Scheinberg,
J. Hickman, G. Ashwell, J. Biol. Chem., 1971,
246, 14611467.
55 P. Umana, J. Jean-Mairet, R. Moudry, H. Am-
stutz, J. E. Bailey, Nat. Biotechnol., 1999, 17,
176180.
56 R. L. Shields, J. Rai, R. Keck, L. Y. OConnell,
K. Hong, Y. G. Meng, S. H. Weikert, L. G. Pre-
sta, J. Biol. Chem. 2002, 277, 2673326740.
57 T. Shinkawa, K. Nakamura, N. Yamane, E.
Shoji-Hosaka, Y. Kanda, M. Sakurada, K.
Uchida, H. Anazawa, M. Satoh, M. Yamasaki,
N. Hanai, K. Shitara, J. Biol. Chem., 2003,
278, 34663473.
58 P. Lerouge, M. Cabanes-Macheteau, C. Rayon,
A. C. Fischette-Laine, V. Gomord, L. Faye,
Plant Mol. Biol., 1998, 38, 3148.
59 A. Koprivova, F. Altmann, G. Gorr, S. Kopriva,
R. Reski, E. L. Decker, Plant Biol., 2003, 5,
582591.
60 R. Vietor, C. Loutelier-Bourhis, A. C. Fitchette,
P. Margerie, M. Gonneau, L. Faye, P. Lerouge,
Planta, 2003, 218, 269275.
61 Y. Makino, K. Omichi, N. Kuraya, H. Ogawa,
H. Nishimura, S. Iwanaga, S. Hase, J. Bio-
chem., 2000, 128, 175180.
62 D. Chargelegue, N. D. Vine, C. J. van Dolle-
weerd, P. M. Drake, J. K. Ma, Transgenic Res.,
2000, 9, 187194.
63 A. von Schaewen, A. Sturm, J. ONeill, M. J.
Chrispeels, Plant Physiol., 1993, 102, 1109
1118.
64 K. Ko, P. M. Rudd, D. J. Harvey, R. A. Dwek, S.
Spitsin, C. A. Hanlon, C. Rupprecht, B.
Dietzschold, M. Golovkin, H. Koprowski,
2003, 100, 80138018.
65 R. Strasser, F. Altmann, J. Glossl, H. Stein-
kellner, Glycoconj. J., 2004, 21, 275282.
66 A. Koprivova, C. Stemmer, F. Altmann, A.
Hoffmann, S. Kopriva, G. Gorr, R. Reski, E. L.
Decker, Plant Biotechnol. J., 2004, 2, 517523.
67 H. Bakker, M. Bardor, J. W. Molthoff, V. Go-
mord, I. Elbers, L. H. Stevens, W. Jordi, A.
Lommen, L. Faye, P. Lerouge, D. Bosch, Proc.
Natl. Acad. Sci. USA, 2000, 98, 28992904.
68 N. Q. Palacpac, S. Yoshida, H. Sakai, Y. Ki-
mura, K. Fujiyama, T. Yoshida, T. Seki, Proc.
Natl. Acad. Sci. USA, 1999, 96, 46924697.
69 R. Misaki, Y. Kimura, N. Q. Palacpac, S. Yoshi-
da, K. Fujiyama, T. Seki, Glycobiology, 2003,
13, 199205.
70 C. M. Huether, O. Lienhart, A. Baur, C. Stem-
mer, G. Gorr, R. Reski, E. L. Decker, Glyco-en-
gineering of moss lacking plant-specific sugar
residues. Plant Biol., in press.
References 929
Abstract
ExpressTec has been developed to produce
biopharmaceuticals both cost-effectively
and in large quantities. ExpressTec is suc-
cessful because it utilizes the latest devel-
opments in plant molecular biology with
the use of strong, endosperm-specific pro-
moters; signal peptides targeting the sub-
cellular compartments to prevent proteoly-
tic degradation of the recombinant protein;
optimized codons to maximize transla-
tional efficiency; and transcriptional activa-
tors that increase target gene transcription
and control of the expression of competi-
tive molecules. Several recombinant pro-
teins have been expressed using the Ex-
pressTec system. The expression level of
these proteins is between 0.1 to 1% of
brown rice weight, or 2560% of soluble
protein. Data shows that both the trans-
genes and their expression are stable over
5 years and 10 generations. The physical
and biochemical properties of the recombi-
nant proteins are the same as for native
proteins. Scale-up processing has shown
that recombinant proteins are easily ex-
tracted from cereal grains, and economical
analysis has placed the cost of biopharma-
ceuticals produced by ExpressTec at about
US$ 6 per gram.
Abbreviations
CFU colony-forming units
GC/MS gas chromatography/mass spec-
trometry
hITF human intestine trefoil factor
hLF human lactoferrin
hLZ human lysozyme
LPS lipopolysaccharides
nhLF native human lactoferrin
PBF prolamin-box binding factor
PMP plant made pharmaceuticals
PSV protein storage vacuole
rhLF recombinant human lactoferrin
rhLZ recombinant human lysozyme
8.1
Introduction
The 1990s and the beginning of the
twenty-first century mark a new era of
transgenic biopharmaceutical production
in plant and animal cells. Researchers all
over the world explore various ways to pro-
duce biopharmaceuticals in large volumes
at a low cost. Scientists at Ventria Bio-
science have developed a protein expres-
sion system, ExpressTec, which expresses
recombinant proteins, enzymes and sec-
931
8
ExpressTec: High-level Expression of Biopharmaceuticals
in Cereal Grains
ISBN: 3-527-31184-X
ondary metabolites in cereal grains. In Ex-
pressTec, target genes are codon-optimized
conforming to the codon preference of the
host genes. The codon-optimized target
genes are then linked to the strong endo-
sperm-specific promoters and the signal
peptides derived from the storage proteins.
The signal peptide leads the target protein
through the endoplasmic reticulum (ER)
where the signal peptides are cleaved and
the high level recombinant protein pro-
duced. The mature proteins are accumu-
lated in cell compartments such as protein
bodies in the endosperm. The target pro-
teins are isolated from these cereal grains
for various applications.
While ExpressTec can be used in all cer-
eal grains, Ventria has focused on the use
of rice and barley grains. There are several
advantages in using rice and barley grains
as the host to produce biopharmaceuticals:
1. Rice and barley grains are generally re-
garded as safe for consumption. In
many countries, rice flour is the first
solid food for infants. Rice-based infant
formulas are commercially available and
rice is considered hypoallergenic. Thus,
cereal grains such as rice are particularly
suitable for the production of recombi-
nant protein for oral applications.
2. The storage proteins in rice and barley
grains are synthesized during grain ma-
turation and stored in protein bodies for
use in the germination and seedling
growth of the next generation. Thus,
protein accumulation in rice and barley
grain is a natural process and suitable
for recombinant protein production.
3. Cereal grains can be produced in large
quantity at very low cost. There is essen-
tially no scale limitation.
4. Cereal grains can be stored for years
without loss of functionality, and there-
fore downstream processing can be con-
ducted independent from the growing
season.
5. Production of recombinant proteins in
cereal grain will be devoid of any animal
pathogen a risk present in transgenic
animal systems.
6. Rice and barley are both self-pollinating
crops. The pollen viability and out-cross-
ing rates are very low, reducing the seg-
regation requirement and the chance of
gene flow via pollen.
8.2
Development of ExpressTec
for High-level Expression of Recombinant
Proteins in Cereal Grains
Plant expression systems can be generally
categorized into three groups: 1) whole-
plant systems producing proteins in the
leaves or the entire plant; 2) cell culture
systems where the protein can be pro-
duced in the culture cells or secrete into
culture media; and 3) the seed/fruit or tu-
ber systems with proteins expressed in
storage organs. Regardless of which ex-
pression system is used to produce bio-
pharmaceuticals, the expression level of
the active protein is the foundation and
one of the most critical factors impacting
the commercialization of plant made phar-
maceuticals (PMP). In general, gene ex-
pression is regulated at four different lev-
els: transcription; post-transcription; trans-
lation; and post-translation. In order to
achieve high-level expression of recombi-
nant proteins, our strategy is to focus on
increasing transcription, enhancing trans-
lational efficiency, and improving the pro-
tein targeting and trafficking (Fig. 8.1).
8.2.1
Increasing Protein Expression
by Boosting Gene Transcription
Transcription is controlled by promoter ac-
tivity and regulated by the cis elements on
the promoters. The promoter activity is en-
hanced by transcriptional factors that inter-
act with cis elements. To search for the
strong promoters, we have examined and
screened various promoters from storage
protein genes via both transient expression
and transgenic analysis [1, 2]. GUS and hu-
man lysozyme genes are used as reporter
genes in transient and transgenic analysis,
respectively. Fig. 8.2 shows one such com-
parison study. Glutelin 1 (Gt1) promoter
from the rice glutelin gene and globulin
(Glb) promoter from the rice globulin gene
show the strongest promoter activities. The
expression level of lysozyme in R
1
seeds
reaches, on average, 10.5% (Glb) and 13%
(Gt1) total soluble protein (TSP). The best
line shows that the expression level is up
to 60% TSP which is derived from a Gt1-
based construct. On a Coomassie blue-
stained gel, the lysozyme band is the most
abundant band among all the rice protein
bands (Fig. 8.3B, lanes 1 to 3) indicating
the strength of the Gt1 promoter activity.
Transcription can be enhanced by tran-
scriptional factors that bind to cis elements
on the promoter. We tested the effects of
various transcriptional factors on recombi-
nant protein expression. The factors in-
clude REB binding to the rice globulin
8.2 Development of ExpressTec for High-level Expression of Recombinant Proteins in Cereal Grains 933
Fig. 8.1 Schematic diagram of ExpressTec development. The control points on gene expres-
sion regulation are illustrated; the technologies listed in the boxed panel represent areas
where major efforts have been made to increase recombinant protein expression.
promoter, PBF (Prolamin-box Binding Fac-
tor) from maize and Opaque 2 from
maize. Rice plants containing the human
lysozyme gene were generated both with
and without the transcription factor. The
results show a 3.7-fold increase of human
lysozyme expression when co-expressing a
Glb promoter-specific REB transcriptional
factor with the Glb-lys construct [3]. A sig-
nificant increase of human lysozyme was
observed when co-expressing PBF with the
Gt1-lys construct (Fig. 8.3A and B lane 4).
Furthermore, transient analysis shows that
PBF and O2 can act additively to enhance
the expression of the GUS reporter gene
in immature rice endosperm [4].
8.2.2
by Enhancing Translational Efficiency
Abundant transcripts would produce abun-
dant mRNA which sets the foundation of
effective protein translation. The untran-
slational leader sequence is one of the key
elements to translation initiation, a deter-
mining factor on the number of peptides
produced from each mRNA. We assume
that native 5' untranslational sequence of a
strongly expressed gene would be the best
for recombinant protein expression; thus,
the native 5' untranslational sequence was
used in our expression cassettes. These
cassettes were used in both transient ex-
pression analysis [2] and transgenic analy-
sis [5, 6]. High-level protein expression is
achieved with the use of native 5' untrans-
lational sequence.
The translational efficiency is one of the
other important elements that affect pro-
tein synthesis and accumulation. The
translational efficiency is highly impacted
by the use of genetic codons of the genes.
Due to genetic codon degeneracy, codon
usage has high diversity among different
organisms. In triple-letter genetic codons,
whilst the first and second positions are
largely conserved among organisms, the
third position is quite diverse. The pre-
ferred codons in rice genes at the third po-
sition are 100% G or C. This is, however,
not the case for other organisms. For ex-
Fig. 8.2 Comparison of the promoter activities
from the promoters of different storage protein
genes. Seven promoters were tested using a trans-
genic approach, and human lysozyme was used
as reporter gene. Gt1, Gt3, Glub-1 and Glub-2 are
from the rice glutelin gene family; Glb is from the
rice globulin gene; Rp6 is from the rice prolamin
gene; and Bx-7 is from the wheat high molecular-
weight glutenin gene. Human lysozyme was quan-
tified by turbidimetric assay.
ample, the preferred codons in Arabidopsis
genes are 15% G or C at the third position
of the codons. Therefore, when expressing
foreign genes, using the preferred codons
of the host can maximize the translational
efficiency. This has been confirmed in our
laboratory and in other laboratories [79].
In producing the human blood protein, al-
pha-1-antitrypsin, in rice culture cells, ex-
pression of the codon-optimized gene is
several folds higher than that of the native
gene [10, 11]. In expressing another pro-
tein, subtilisin, expression of the codon-op-
timized gene is over 100-fold higher than
that of native gene. All data point to a very
important conclusion that foreign genes
must be codon-optimized to match the co-
don preference of the host for high-level
expression.
8.2.3
by Managing the Space
for Recombinant Protein Deposition
8.2.3.1 Protein Bodies in Endosperm Cells
Provide a Safe Environment
for Foreign Protein Accumulation
Post-translational regulation mainly in-
cludes the signal peptide cleavage, glycosy-
lation, phosphorylation, and proper protein
folding while the protein is translocated
into the ER and transported to the Golgi
apparatus. Then, the protein is targeted
and trafficked to the protein storage va-
cuole (PSV), ER-derived protein bodies,
and other organelles, or secreted to cytosol
[1215]. In those steps, the proteins could
either be accumulated in high amounts or
rapidly turned over due to protease activity,
depending on the destination of the pro-
tein targeting. It has been shown that
PMPs expressed at very low levels in plant
Fig. 8.3 Human lysozyme expression level was
improved through using different strategies during
ExpressTec development. (A) Milestones of in-
creasing human lysozyme level in rice endosperm.
The lysozyme expression level was increased to
1% of dry weight by selection stronger promoters
(1), use of transcription factor (2), and space-
emptying strategy (3). (B) Human lysozyme pro-
tein in a Coomassie blue-stained gel from differ-
ent approaches. Lanes 13 shows the effects of
selecting the stronger promoters, Glb promoter/
Glb signal peptide (lane 1), Glb promoter/Gt1 sig-
nal peptide (lane 2) and Gt1 promoter/Gt1 signal
peptide (lane 3). Lane 4 shows the effects of
using the transcription factor. Lane 5 represents
the effects of using the space-emptying strategy.
Lane 6 is for non-transgenic TP309. Lane M is
molecular weight marker.
cells without a targeting destination [16
18] because of the protein(s) exposure to
protease(s) degradation. Therefore, a pro-
mising gene expression system for PMP
proteins should use a particular targeting
signal to deposit the recombinant proteins
to certain organelles or cell compartments
to prevent their degradation.
In rice, grain endosperm, two protein
bodies PBI and PBII are considered to
be the sink for protein storage during en-
dosperm development. We hypothesize that
protein bodies provide a safe environment
for the deposition and accumulation of
PMPs, because there is limited protease ac-
tivity within the protein bodies. Targeting
PMPs to protein bodies in cereal endo-
sperm cells can be achieved by attaching a
signal sequence to a mature peptide of the
PMP, which can guide PMPs through the
inner membrane system instead of to the
cytosol. As soon as the gene is transcribed
and processed, mRNA is bound to the sub-
domains of the ER, which determines
where the protein targets [1921]. Then,
the synthesized recombinant protein targets
to the protein bodies through the protein
trafficking pathway during endosperm de-
velopment [22]. Comparison studies with
and without signal peptide confirms this
hypothesis. In the expression of heat stable
beta-glucanase in barley grain, signal pep-
tide from hordien D, a barley storage pro-
tein, was used. The expression level of
beta-glucanase with hordein signal peptide
is several fold higher than the same con-
struct without the signal peptide [23]. In ex-
pressing recombinant protein in rice grain,
we use strong promoters and signal pep-
tides from two rice storage proteins, glute-
lin and globulin, to achieve high-level ex-
pression of the PMP [2]. The use of Gt1 sig-
nal peptide and its 5' untranslated sequence
promote higher expression than that from
Glb (Fig. 8.3B, lanes 1, 2 and 3).
To confirm that recombinant human ly-
sozyme (rhLZ) is present in protein body,
immature rice endosperm from LZ264
grain was harvested and sectioned. The
sections were incubated with anti-lysozyme
and anti-glutelin antibody. The anti-lyso-
zyme antibody derived from sheep was
specifically recognized by IgG conjugated
with green fluorescence; hence, the pres-
ence of rhLZ would appear green. The
anti-glutelin antibody derived from rabbit
was specifically recognized by IgG conju-
gated with red fluorescence; hence, the
presence of glutelin would appear red. A
yellow color would appear if rhLZ and glu-
telin were to be co-localized. As seen in
Fig. 8.4, panel A indicates the presence of
rhLZ shown in green, while panel B
shows the presence of glutelin shown in
red. Panel C shows yellow spots, indicat-
ing that rhLZ and glutelin are co-localized.
Panel D is a linear scan to show that the
green and red peak at the same position
rather than showing shadows from a near-
by color. Since glutelin is a storage protein
known to be stored in the protein body of
endosperm cells, Fig. 8.4 confirms that
rhLZ is targeted to the protein body in
LZ264 grain [24]. The deposit of rhLZ in
protein bodies is further confirmed by
electronic microscopic analysis [24].
8.2.3.2 Increasing Protein Expression using
Different Translational Machineries
To improve human lysozyme expression,
we attempted to improve heterologous re-
combinant protein expression levels by: 1)
co-transformation of Glb and Gt1 expres-
sion cassettes; and 2) crossing of two inde-
pendent transgenic lines expressing lyso-
zyme protein from Gt1 and Glb cassettes.
These experiments failed to improve lyso-
zyme expression. This implies that the
gene transcription and copy number are
not the bottle-neck for improvement of hu-
man lysozyme expression. Other possible
limiting factors include the translation effi-
ciency, protein trafficking, and targeting.
Therefore, we attempted to improve hu-
man lysozyme expression by using differ-
ent promoters/signal peptides which can
bind to different ER subdomains than that
of Gt1 and Glb promoters/signals would.
We hypothesized that we could achieve
higher expression by using different tar-
geting signals to use different ER subdo-
mains. A wheat puroindoline b promoter
and signal peptide have been tested. Co-ex-
pression of both constructs (Gt1 promoter,
its signal peptide plus the human lyso-
zyme gene and the puraindoline b promo-
ter, its signal peptide and the human lyso-
Fig. 8.4 Evidence of rhLZ targeting to protein
bodies in rice endosperm. The immature endo-
sperm at 14 days after pollination was used for
fluorescence microscopic studies. The endosperm
section was incubated with antibodies against hu-
man lysozyme (a) and rice glutelin (b). A merger
of the two images from (a) and (b) is shown in
(c). (d) Diagram of the fluorescence scans along
the white line shown in (c). Red line=glutelin;
green line=human lysozyme, showing both lyso-
zyme and glutelin co-localized in the same protein
bodies.
zyme gene), resulted in an increase in the
expression of human lysozyme by 79% to
8 mg kg
1
rice grain flour. Electron micro-
scopy studies show that the puroindoline-
based construct directed rhLZ to both pro-
tein body I and II.
8.2.3.3 Reduce Endogenous Protein
Expression to Reserve the Space
to Recombinant Protein Storage
When we examined the protein body struc-
ture of the lines expressing high levels of re-
combinant human lysozyme, we observed
that rice endosperm generated novel storage
vesicles or protein body variants for recom-
bination protein deposition [24]. It also indi-
cated that rice endosperm cells are capable
of generating novel storage vesicles for re-
combinant protein deposition when large
amounts of recombinant protein are ex-
pressed in the endosperm cells. Further-
more, we also observed a negative correla-
tion between native storage proteins and
the recombinant protein expression. We hy-
pothesized that the protein bodies are the
sink. When more source proteins are
available, they compete for the sink caus-
ing an imbalance between the sink and
the source. Reduced native storage protein
in high lysozyme-expressing lines indicates
that the recombinant protein can partially
compete for ER sub-domains with native
storage proteins and chaperones during
the trafficking. This implies that human ly-
sozyme expression could be further in-
creased by shutting down native storage pro-
tein expression, making more sink space
available to recombinant protein deposition.
Thus, we call this strategy space-empty-
ing. This concept was tested by reducing
the endogenous protein expression via anti-
sense technology. The antisense constructs
of glutelin and globulin were introduced
into the transgenic line that expressed high
levels of human lysozyme using gene stack-
ing. The expression of recombinant human
lysozyme in the best lines expressing the
antisense gene is increased from 5 to
10 mg g
1
rice flour (Fig. 8.3).
In summary, we conclude that ExpressTec
is a promising biomanufactory system for
expressing PMPs in rice grain as well as
other cereal grain endosperms. In addition
to all the advantages of other plant expres-
sion systems, it has a higher capacity for ob-
taining higher expression levels of PMPs.
Its core technology is to target the recombi-
nant proteins to protein bodies so that the
biopharmaceutical can be protected from
protease degradation and accumulated at a
very high level. Moreover, the technology
is improved by boosting transcription, en-
hancing translational efficiency, improving
protein trafficking and deposition to maxi-
mize PMP expression in cereal grains (Fig.
8.3). The recombinant protein expression
can be as high as 1% of flour weight.
8.3
High-level Expression of Biopharmaceuticals
in Cereal Grain using ExpressTec
The success of developing ExpressTec laid
the foundation for production of various
recombinant polypeptides, multipeptide
proteins and secondary metabolites. To ex-
press small peptide, a fusion strategy is
used to achieve high level expression.
8.3.1
Expression of Human Lysozyme
in Rice Grain
Human lysozyme (hLZ) hydrolyzes 1,4-
beta-linkage between N-acetylmuramic
acid and N-acetyl-d-glucosamine residue in
peptidoglycan. Human LZ exhibits antibac-
terial, antiviral, antifungal and antiparasitic
activities, and has also been implicated as
an anti-inflammatory/anti-oxidant agent or
direct binding to lipopolysaccharides (LPS)
for immunomodulation [25]. Human LZ is
found in human secretions, such as milk,
tears and saliva, and consists of an ungly-
cosylated polypeptide chain with 130 ami-
no acid residues, giving it a molecular
weight of 14.5 kDa.
To express hLZ in rice grain, the hLZ
gene was codon-optimized. A total of 92
codons out of 130 codons was modified,
resulting in the G+C content being raised
from 46 to 68%. The synthetic hLZ gene
was cloned to produce pAPI159, which
contains the Gt-1 promoter, Gt-1 signal se-
quence and nos terminator. After transfor-
mation, over 500 transgenic rice R
0
plants
were generated and seeds from fertile rice
plants were analyzed via LZ activity assay,
Coomassie blue-stained gel and Western
blot analysis to determine the amount of
recombinant human lysozyme (rhLZ) in
the endosperm. As shown in Fig. 8.3B, a
dominating band corresponding to the po-
sition of a protein with a molecular mass
of rhLZ was detected in the salt-soluble
fraction of crude extracts from the trans-
genic rice grains, while it is absent in un-
transformed rice. The identity of the pro-
tein was confirmed by Western blotting
analysis and verified further by N-terminal
analysis [5]. One of the lines, named
LZ159 was selected and advanced for 10
generations from 1999 to 2003. Both the
transgenes and expression levels are
stable. The expression level of LZ159 re-
mains at 5 g kg
1
rice flour, amounting to
60% total soluble protein. To determine
the bactericidal activity of rhLZ, an E. coli
strain, K12, was used. Bacterial culture
with the addition of rhLZ at 20 lg mL
1
re-
sulted in significantly fewer colony form-
ing units (CFU) than those where rhLZ
was not added (Fig. 8.5), thus proving that
rhLZ is biologically active. Further studies
8.3 High-level Expression of Biopharmaceuticals in Cereal Grain using ExpressTec 939
Fig. 8.5 Bactericidal effect of purified rhLZ. E. coli (10
5
CFU)
was incubated for 120 min with buffer plus 20 lg mL
1
puri-
fied rhLZ (A) or buffer only (B). At the end of the incubation
period, numbers of CFUs were determined by plating a sam-
ple of the incubation mixture.
show that rhLZ are thermal stable and ac-
tive in a wide range of pH [5].
8.3.2
Expression of Human Lactoferrin
in Rice Grain
Human lactoferrin (hLF) is an 80-kDa
iron-binding glycoprotein, and another
major component found in human milk
(average 12 mg mL
1
); lower concentra-
tions are present in the exocrine fluids of
glandular epithelium cells such as bile,
tears and saliva (0.10.3 mg mL
1
). LF has
been suggested to have several biological
activities, including antimicrobial, regula-
tion of iron absorption, immunomodula-
tion, protection from pathogen infection,
and cellular growth-promoting activity [26].
To express hLF, the gene was synthe-
sized based on the codon-preference of
rice genes. Of the 692 codons for the ma-
ture peptide of the hLF gene, 413 codons
were changed. The codon-optimized hLF
gene was expressed using ExpressTec. To-
tal soluble protein extracted from rice
grains is analyzed by SDS-PAGE (Fig.
8.6A, lane 4). Among the protein bands,
the recombinant human lactoferrin (rhLF)
band was the strongest, indicating that
rhLF is the most abundant soluble protein
extracted from rice grain. Quantitative
analysis by ELISA indicated that up to
25% of soluble protein or 0.5% flour weight
was rhLF. The expression level of rhLF in
the best line reaches 5.00.5 g kg
1
and is
stable through 10 generations (Fig. 8.6B).
In order to characterize the biochemical
properties of rhLF expressed in rice grain,
rhLF from transgenic rice grain was puri-
fied to homogeneity. The N-terminal se-
quence of rhLF was identical to the corre-
sponding region of hLF, indicating that
the rice signal peptidase recognized and
cleaved at the junction between the Gt1 sig-
nal peptide sequence and the mature pep-
tide of rhLF. The isoelectric point (pI) of
hLF and rhLF was similar, indicating that
both have similar surface charges. Both na-
tive human lactoferrin (nhLF) and rhLF can
reach iron-saturation by picking up iron
from a solution to form holo-LF. The stabil-
ity of iron-binding by rhLF toward low pH
was analyzed and compared to that of nhLF
(Fig. 8.6C). Iron release began at about pH
4, was completed around pH 2, and was
similar for both proteins.
The antimicrobial effect of rhLF was
tested against a Gram-negative strain of E.
coli, DH5a (Fig. 8.6D). E. coli at a concen-
tration of 10
5
CFU was mixed with and
without rhLF. After incubation at 378C for
120 min, the CFU after treatment with
rhLF were reduced by 90%, while CFU
without rhLF remained unchanged.
8.3.3
Expression of Fibrinogen in Rice Grain
Fibrinogenis a multi-chain proteininvolving
the assembly of three different polypeptides
with a molecular mass of 340 kDa. The mol-
ecule is arranged as a dimer with each half-
molecule containing a set of each of the three
different chains. The subunits and the
chains are linked together by three disulfide
bonds at the N-terminal portions of the poly-
peptides and form a symmetrical trinodular
structure. There are two symmetrical bonds
that are located between adjacent c chains
and another bond between a chains. In ad-
dition, there are 29 inter- and intra-chain dis-
ulfide bonds interspersed throughout the
molecule that are responsible for maintain-
ing proper structure. Fibrinogen is a blood
plasma protein that serves as one of the main
components in blood clotting.
Expression of multipolypeptide protein
in rice grain posed a new challenge to Ex-
pressTec. After gene codon optimization,
genes for individual chain of fibrinogen (a,
b, and c,) were delivered into the rice cell
by co-transformation. Transgenic plants
were obtained and rice grains analyzed for
the expression of the fibrinogen polypep-
tide via Western blot analysis. All three
polypeptides were expressed in the same
cell. Compared to the positive control, it is
estimated that the expression level of the
three polypeptide reaches about 0.4%
brown rice flour weight.
8.3.4
Expression of Intestine Trefoil Factor
in Rice Grains
Human intestine trefoil factor (hITF) con-
sists of 75 amino acids. After cleavage of a
signal peptide, the resulting mature hITF
contains 60 amino acids [27, 28]. Human
ITF is present in both monomer and di-
mer forms in gastrointestinal tissue [29].
Several biological functions of ITF have
Fig. 8.6 Biochemical properties and biological ac-
tivity of rhLF. (A) rhLF in Coomassie blue-stained
gel. Lanes 1 and 2 represent non-transgenics of
Golden Promise (barley) and TP309 (rice), respec-
tively; lanes 3 and 4 show rhLF from the endo-
sperm extracts of transgenic barley and rice, re-
spectively; lanes 57 indicate native human LF
standard, titrated to 6, 8 and 10 lg per lane, re-
spectively. M indicates molecular mass marker.
(B) Stable expression of rhLF over 10 generations
as determined by ELISA. (C) pH-dependent iron
release of recombinant and native human LF (see
Ref. [6]). (D, E) Bactericidal effect of purified rhLF.
E. coli (110
5
CFU) was incubated for 120 min
with buffer plus 1 mg mL
1
purified rhLF (D) or
buffer only (E). At the end of the incubation peri-
od, numbers of CFUs were determined by plating
a sample of the incubation mixture.
been identified, including the promotion
of wound healing, stimulation of epithelial
cell migration and protection of the intest-
inal epithelial barrier. It is thus believed
that the preparation of ITF can be used in
the prevention and treatment of these dis-
ease conditions.
In general, the expression of a peptide
of less than 100 amino acids proves to be
difficult. When hITF is directly expressed
in rice grain using the Gt1 promoter/sig-
nal peptide, the expression level is about
1 lg per grain, or 0.005% grain weight. To
obtain a higher expression level, a modifi-
cation to the expression system was made.
The relative low expression level of ITF
was not due to lower transcription based
on Northern blot analysis. This suggested
that it could be post-translation modifica-
tion and protein trafficking. Using a fu-
sion partner will generally increase the ex-
pression of a peptide. To apply this specifi-
Fig. 8.7 A fusion strategy to express small peptide
in rice endosperm. (A) Schematic diagram of the
fusion strategy to express small peptide. GOI =
gene of interest. The enterokinase recognition site
was used as a linker between the fusion partner
and GOI. (B, C) ITF fusion protein in Coomassie
blue-stained gel and in Western blot using anti-
ITF antibody. Lanes 1 and 2 represent individual
transgenic lines; lane 3 indicates non-transgenic
TP309. M indicates molecular mass marker.
cally to our rice seed expression system,
proteins such as globulin are selected as
fusion partners. These proteins are se-
lected because they are relatively small,
they have a high expression level in rice
grain, they are targeted to the protein
body, and they are water- and/or salt-solu-
ble. These characteristics are important for
increasing the expression of a peptide as
well as the extraction and purification of
the peptide (Fig. 8.7A).
The human ITF DNA sequence based
on the GenBank accession number L08044
[28] was codon-optimized with rice genetic
codon preference. The codon-optimized
gene was then linked to a rice globulin
gene. Between the codon-optimized gene
and the rice globulin gene was a segment
of DNA encoding for a five amino-acid
peptide, which is an enterokinase recogni-
tion site (Fig. 8.7A).
Transgenic rice grains carrying the ITF
gene were analyzed by Coomassie blue-
stained gel and Western blot analysis (Fig.
8.7B and C). A strong band was observed
which is absent in the non-transgenic
plant TP309 (Fig. 8.7B); this band was
confirmed to be ITF fusion by Western
analysis using anti-ITF antibody (Fig.
8.7C) and anti-GLB antibody. This band is
the strongest and stronger than the native
globulin band, indicating high-level expres-
sion of the fusion protein. Using a refer-
ence marker, it is estimated that the ex-
pression level of the fusion protein was
about 60 lg per grain. Since ITF is about
one-quarter of the fusion protein, about
15 lg ITF per grain (or 0.075% flour
weight) was achieved.
8.3.5
Expression of Lignans in Rice Grain
via Metabolic Engineering
Plant lignans are secondary metabolites
which are most commonly found in woody
stems, roots, seeds, oils, and leaves, and
exist in low levels in cereal endosperm
[30]. Matairesinol and secoisolariciresinol
are two typical plant lignans which are es-
sentially not detectable in rice endosperm.
Plant lignans, once consumed, are then
converted to mammalian lignans by fer-
mentation in the large intestine, where
matairesinol is converted to enterodiol and
secoisolariciresinol to enterolactone.
Enterolactone and enterodiol are the major
lignans found in humans, and are present
in the serum, urine, bile, and seminal
fluid [31]. Studies have shown that lignans
can prevent the development of cancer,
and human populations that consume
high quantities of lignans have a lower in-
cidence of hormonally dependent cancers
than do other populations consuming
high-fat diets [30, 32, 33].
Plant lignans are derived from a process
called the shikamate-chorismate pathway
in phenylpropanoid metabolism, which
leads to the production of coniferyl alcohol
and other metabolites [34]. Specific oxidative
coupling of coniferyl alcohol under the ac-
tion of laccase and dirigent protein gener-
ates pinoresinol (Fig. 8.8A). Pinoresinol,
through the intermediate lariciresinol, is
then converted to secoisolariciresinol, which
is finally modified to form matairesinol [35].
Rice endosperm contains essentially no ma-
tairesinol (Fig. 8.8B); this may be due to a
lack of one or more of the four genes in-
volved in the lignan biosynthetic pathway
(Fig. 8.8A), or that the expression of the
genes in rice endosperm is deregulated.
In order to elevate lignan concentrations
in rice endosperm, Ventria Bioscience and
F
i
g
.
8
.
8
E
x
p
r
e
s
s
i
o
n
o
f
l
i
g
n
a
n
i
n
t
r
a
n
s
g
e
n
i
c
r
i
c
e
e
n
d
o
s
p
e
r
m
.
(
A
)
L
i
g
n
a
n
b
i
o
s
y
n
t
h
e
s
i
s
p
a
t
h
w
a
y
d
e
r
i
v
e
d
f
r
o
m
F
o
r
s
y
t
h
i
a
i
n
t
e
r
m
e
d
i
a
[
3
5
]
.
(
B
)
M
a
t
a
i
r
e
s
i
n
o
l
l
e
v
e
l
i
n
t
r
a
n
s
g
e
n
i
c
g
r
a
i
n
(
4
P
E
-
2
5
6
)
a
n
d
n
o
n
-
t
r
a
n
s
g
e
n
i
c
w
i
l
d
-
t
y
p
e
(
T
P
3
0
9
)
.
G
C
/
M
S
s
c
a
n
s
h
o
w
s
a
c
c
u
m
u
l
a
t
i
o
n
o
f
l
i
g
n
a
n
i
n
t
r
a
n
s
g
e
n
i
c
r
i
c
e
e
n
d
o
s
p
e
r
m
.
(
C
)
M
a
t
a
i
r
e
s
i
n
o
l
l
e
v
e
l
i
n
t
r
a
n
s
g
e
n
i
c
r
i
c
e
g
r
a
i
n
s
o
v
e
r
t
h
r
e
e
g
e
n
e
r
a
t
i
o
n
s
a
s
m
e
a
s
u
r
e
d
b
y
G
C
/
M
S
s
c
a
n
.
Washington State University at Pullman
collaborated to engineer a lignan synthesis
pathway using ExpressTec. The four genes
for laccase, dirigent protein, pinoresinol/
lariciresinol reductase and secoisolariciresi-
nol dehydrogenase involved in lignan syn-
thesis pathway were fused to the Gt1 pro-
moter. Native signal peptides from the
genes of the lignan synthetic pathway
were used. The four constructs, along with
a plasmid containing a plant-selectable
marker, were delivered into rice cells via
particle bombardment. Over 400 trans-
genic plants were generated. Transgenic
seeds harvested from fertile plants were
analyzed using gas chromatography/mass
spectrometry (GC/MS) for elevated lignan,
matairesinol, in the endosperm. The high-
est of these (4PE-256-1) had a matairesinol
level approximately 15-fold that in the cor-
responding wild-type rice. Other families
tested (4PE-103-1, 4PE-115-1, 4PE-131-1)
had matairesinol levels which were up to
3- to 5-fold those in the wild-type. A typical
GC/MS profile is shown in Fig. 8.8B. In
order to determine the generational stabili-
ty of lignan expression in rice endosperm,
a test was carried out using four lines with
R
1
, R
2
, and R
3
seeds tested simultaneously.
This test showed levels in the R
1
seeds to
be similar to the lower levels seen in the
R
2
and R
3
seeds, and the levels of mataire-
sinol to remain fairly consistent across all
three generations. The 4PE-256-1 line was
the best of all the transgenic lines tested,
and consistently gave results well above
3 ng per 100 mg for matairesinol. This
line also showed consistency when the R
1
,
R
2
, and R
3
seeds were tested simulta-
neously, as all three generations showed
elevated levels of matairesinol. An elevated
lignan level in transgenic rice endosperm
not only proved that ExpressTec could be
used to engineer metabolic pathways to
produce secondary metabolites, but also
provided a line expressing a high level of
lignan which could be used to provide a
source of lignans in food for the benefit of
human health.
8.3.6
Protein Expression using ExpressTec
in Barley and Wheat Grains
In addition to expressing polypeptides and
metabolites in rice grains, ExpressTec has
been used to obtain high expression levels
of proteins in different hosts such as bar-
ley or wheat grain. Recombinant hLZ is
expressed at about 0.5% flour weight in
barley and 0.5% flour weight in wheat en-
dosperm. Similarly, rhLF is expressed at
0.7% flour weight in barley endosperm
(see Fig. 8.6A, lane 3).
8.4
Impact of Expression Level on the Cost
of Goods
Expression level has a profound impact on
cost of goods in biopharmaceutical produc-
tion. One way to study this issue is to per-
form computer simulations based on data
from plot production in order to obtain an
early projection of production costs as a
function of the expression level (Nandi et
al., unpublished results). By utilizing the
crop production and process data from the
bench-scale (2 kg per batch), which was
subsequently verified on a pilot scale at
180 kg per batch, an rhLF and rhLZ recov-
ery and purification process is simulated.
With an annual production of 600 kg and
expression of 0.005% flour weight, rhLF
can be produced at $ 382 g
1
(Fig. 8.9);
however, when the expression level in-
creases to 0.5% of flour weight (100), as
was achieved using ExpressTec, the cost
was only $ 5.90 g
1
. Hence, an approxi-
mately 7-fold reduction in direct costs
could be achieved for each 10-fold incre-
mental increase in the expression level. It
is generally believed that the cost of phar-
maceutical production in mammalian cell
culture is about $ 200 g
1
. PMPs would
lose competitiveness if this cost were close
to $ 200 g
1
, particularly if the market size
were to be small. The present analysis in-
dicates that this would require a protein
expression level of at least 0.05% cell
mass.
8.5
Perspectives of Expressing
Biopharmaceuticals in High Plants
Historically, the cost of recombinant bio-
pharmaceutical production has not been a
limiting factor, as drug manufacturers can
easily pass the cost on to the consumers.
Whilst this is still true in some cases, an in-
creasing demand for large volume and nec-
essary low cost for some applications has
forced the drug manufacturers to reduce
their production costs and to make pharma-
ceuticals more affordable. It is believed that
PMPs will be able to meet this demand.
Although efforts made during the past de-
cade have not produced satisfactory results,
one reason for this is that expression levels
in most plant systems have been very low.
However, with the development of Expres-
sTec, a solution has been found to the prob-
lem of low expression level.
Despite obtaining high expression levels
for certain proteins using ExpressTec, not
all proteins are expressed to the same ex-
tent, and this suggests that expression lev-
el is in fact protein-dependent. Both, the
biochemical and biophysical properties of
the biopharmaceutical impact upon indi-
vidual protein expression in cereal endo-
sperm, these factors including protein
folding, conformation, trafficking, deposi-
Fig. 8.9 Impact of expression level on the cost of
goods. Based on the procedures of purifying rhLZ
and rhLF from rice grain, a computer simulation
was conducted assuming an expression level from
0.005 to 1% with the assumption of annual pro-
duction capacity of 600 kg of pharmaceutical
grade lactoferrin and lysozyme.
tion, and accumulation, whilst in some
cases the recombinant proteins showed
changes in solubility when expressed in
rice grain. While the challenge remains
that all soluble recombinant proteins and
multiple polypeptides should be correctly
assembled, these factors will undoubtedly
be the main topics of research in this area
in the future.
References
1 Hwang, Y.-S., Yalda, D., McCullar, C., Wu, L.,
Chen, L., Pham, P., Nandi, S. and Huang, N.
Plant Cell Rep. 2002, 20, 842847.
2 Hwang, Y.-S., McCullar, C. and Huang, N.
Plant Sci. 2001,161, 11071116.
3 Yang, D., Wu, L., Hwang, Y. S., Chen, L. and
Huang, N. Proc Natl Acad Sci USA 2001, 98,
1143811443.
4 Hwang, Y.-S., Ciceri, P., Parsons, R., Moose,
S. P., Schmidt, R. J. and Huang, N. Plant Cell
Physiol. 2004, 45, 15091518.
5 Huang, J., Nandi, S., Wu, L., Yalda, D., Bart-
ley, G., Rodriguez, R. L., Lonnerdal, B. and
Huang, N. Molec. Breed. 2002, 10, 8394.
6 Nandi, S. et al. Plant Sci. 2002, 163, 713722.
7 Akashi, H. Curr. Opin. Genet. Dev. 2001, 11,
660666.
8 Davis, B. K. Prog. Biophys. Molec. Biol. 1999,
72, 157243.
9 Rouwendal, G. J. A., Mendes, O., Wolbert,
E. J. H. and Boer, A. D. d. Plant Molec. Biol.
1997, 33, 989999.
10 Huang, J., Sutliff, T. D., Wu, L., Nandi, S.,
Benge, K., Terashima, M., Ralston, A. H., Dro-
han, W., Huang, N., Rodriguez, R. L. Biotech-
nol Prog. 2001, 17, 126133.
11 Terashima, M., Murai, Y., Kawamura, M., Na-
kanishi, S., Stoltz, T., Chen, L., Drohan, W.,
Rodriguez, R. L., Katoh, S. Appl. Microbiol.
Biotechnol. 1999, 52, 516523.
12 Neuhaus, J. M. and Rogers, J. C. Plant Molec.
Biol. 1998, 38, 127144.
13 Marty, F. Plant Cell 1999, 11, 587600.
14 Muntz, K. Plant Molec. Biol. 1998, 38, 7799.
15 Vitale, A. and Raikhel, N. V. Trends Plant Sci.
1999, 4, 149155.
16 Giddings, G., Allison, G., Brooks, D. and Car-
ter, A. Nature Biotechnol. 2000, 18, 1151
1155.
17 Larrick, J. W., Yu, L., Naftzger, C., Jaiswal, S.
and Wycoff, K. Biomol. Eng. 2001, 18, 8794.
18 Schillberg, S., Fischer, R. and Emans, N. Cell.
Mol. Life Sci. 2003, 60, 433445.
19 Choi, S. B., Wang, C., Muench, D. G., Ozawa,
K., Franceschi, V. R., Wu, Y. and Okita, T. W.
Nature 2000, 407, 765767.
20 Li, X. X., Franceschi, V. R. and Okita, T. W. Cell
1993, 72, 869879.
21 Okita, T. W. and Choi, S. B. Curr. Opin. Plant
Biol. 2002, 5, 553559.
22 Vitale, A. and Galili, G. Plant Physiol. 2001,
125, 115118.
23 Horvath, H., Huang, J., Wong, O., Kohl, E.,
Okita, T., Kannangara, C. G. and von Wett-
stein, D. Proc. Natl. Acad. Sci. USA 2000, 97,
19141919.
24 Yang, D., Guo, F., Liu, B., Huang, N. and Wat-
kins, S. C. Planta 2003, 216, 597603.
25 Sava, G. in: Lysozyme: model enzyme in bio-
chemistry and biology, Jolles, P. (Ed.) 1996.
26 Brock, J. H. Biochem. Cell. Biol. 2002, 80, 16.
27 Hauser, F., Poulsom, R., Chinery, R., Rogers,
L. A., Hanby, A. M., Wright, N. A. and Hoff-
mann, W. Proc. Natl. Acad. Sci. USA 1993,
90, 69616965.
28 Podolsky, D. K., Lynch-Devaney, K., Stow, J. L.,
Oates, P., Murgue, B., DeBeaumont, M.,
Sands, B. E. and Mahida, Y. R. J. Biol. Chem.
1993, 268, 66946702.
29 Chinery, R., Bates, P. A., De, A. and Freemont,
P. S. FEBS Lett. 1995, 357, 5054.
30 Nilsson, M., Harkonen, H., Hallmans, G.,
Knudsen, K. E. B., Mazur, W. and Adlercreutz,
H. J. Sci. Food Agric. 1997, 73, 143148.
31 Borriello, S. P., Setchell, K. D., Axelson, M.
and Lawson, A. M. J. Appl. Bacteriol. 1985, 58,
3743.
32 Thompson, L. U., Seidl, M. M., Rickard, S. E.,
Orcheson, L. J. and Fong, H. H. Nutr. Cancer
1996, 26, 159165.
33 Adlercreutz, H. and Mazur, W. Ann. Med.
1997, 29, 95120.
34 Lewis, N. G. and Sarkanen, S. Lignin and lig-
nan biosynthesis. Oxford University Press,
Washington, DC, USA, 1998.
35 Dinkova-Kostova, A. T., Gang, D. R., Davin,
L. B., Bedgar, D. L., Chu, A. and Lewis, N. G. J.
Biol. Chem. 1996, 271, 2947329482.
Abstract
The use of plants for the production of
recombinant proteins has received a great
deal of recent attention, but production
systems that utilize whole plants lack sev-
eral of the intrinsic benefits of cultured
cells, including the precise control over
growth conditions, batch-to-batch product
consistency, the high level of containment
and the ability to produce recombinant
proteins in compliance with current good
manufacturing practice (cGMP). Plant cell
cultures combine the merits of plant-based
systems with those of microbial and ani-
mal cell cultures, particularly in terms of
downstream processing, a section of the
production pipeline which is rarely given
appropriate prominence when different
production systems are compared. In this
chapter, we discuss the benefits of plant
cell cultures compared to other systems,
the technological requirements for produc-
ing biopharmaceutical proteins in plant
cells, and the unique aspects of down-
stream processing which are applied to
this expression platform.
Abbreviations
Asn asparagine
ATPS aqueous two-phase systems
BY-2 Nicotiana tabacum cv. Bright
Yellow 2
CaMV cauliflower mosaic virus
practice
DMSO dimethylsulfoxide
EBA expanded bed adsorption
FW fresh weight
GM-CSF granulocyte-macrophage colony-
stimulating factor
hGM-CSF human granulocyte-macrophage
colony-stimulating factor
Ni-NTA nickel nitriltriacetic acid
NT-1 Nicotiana tabacum 1
PVP polyvinylpyrrolidone
scFv single chain fragment variable
Ubi1 ubiquitin-1
949
9
Biopharmaceutical Production in Cultured Plant Cells
ISBN: 3-527-31184-X
9.1
Introduction
Plant cells combine some of the more ben-
eficial features of other production systems
and therefore occupy a unique position
among the emerging platforms for com-
mercial recombinant protein production
[1, 2]. Like microbial cells, plant cells are
inexpensive to maintain. They have simple
nutrient requirements, and can be grown
under controlled and defined conditions in
accordance with current good manufactur-
ing practice (cGMP). However, unlike mi-
crobes, plant cells are higher eukaryotic
systems and have the ability to produce
complex, multimeric proteins and glyco-
proteins (see Part IV, Chapter 7). The N-
glycans synthesized in plants are not ex-
actly the same as those synthesized in
mammals, so human glycoproteins pro-
duced in plants do not contain native gly-
can profiles [3]. However, plant-derived re-
combinant proteins are more similar to
their mammalian counterparts than pro-
teins synthesized in bacteria (which do not
glycosylate proteins at all) or in yeast and
filamentous fungi (which produce very dif-
ferent glycans). Other advantages of plant
cells include the lack of endotoxins that
are often present in bacteria, and the ab-
sence of human pathogens such as viruses
or prions (which may be present in mam-
malian cell lines) (see Part IV, Chapter 1).
This high level of safety makes plant cells
suitable for the production of biopharma-
ceuticals [4].
Many different plant-based expression
systems are now available for the produc-
tion of recombinant proteins. Those utiliz-
ing whole plants have been extensively re-
viewed, and will not be discussed in detail
here (see Refs. [58]). Other systems,
based on cultured plant cells or organs, in-
clude hairy roots [9], shooty teratomas [10],
immobilized cells [11] and suspension cell
cultures [12]. With the exception of sus-
pension cells, these systems are heteroge-
neous especially when scaled up, and
hence are difficult to maintain under
cGMP conditions. Therefore, suspension
cells have attracted the most attention
since they are amenable to cGMP and they
can be cultivated in large-scale bioreactors
[13, 14] (see Part IV, Chapter 7).
Suspension cell cultures have been de-
rived from a number of different plant
species, including the widely-used labora-
tory model Arabidopsis thaliana [15], plants
such as Catharanthus roseus and Taxus cus-
pidata which are used to produce valuable
secondary metabolites [16, 17], and impor-
tant domestic crops such as tobacco, rice,
alfalfa, tomato, and soybean [1822]. Be-
cause cell lines from domestic crop species
are well-characterized, they have been the
most frequently used for recombinant pro-
tein production. The most popular cell
lines include those derived from the tobac-
co cultivars Bright Yellow 2 (BY-2) (Fig.
9.1) and Nicotiana tabacum 1 (NT-1) [2].
Plant cell suspensions are typically de-
rived from undifferentiated callus tissue
which has been induced from tissue ex-
plants growing on solid medium. Friable
callus pieces are transferred into liquid
medium and then agitated on rotary sha-
kers or in fermenters to break the callus
into small aggregates and single cells. The
correct balance of plant hormones is pres-
ent in the medium to maintain the undif-
ferentiated state and promote rapid
growth. Transgenic cell suspensions can
be generated by agitating callus derived
from transgenic plant tissue, or transfor-
mation can take place after the cell sus-
pension has been prepared. In the latter
case, transformation is usually achieved
either by cocultivation with Agrobacterium
tumefaciens [23, 24] or particle bombard-
ment [25]. Plant cell suspensions can be
cultivated using conventional fermenter
equipment and the same running modes
as applied to microbial cultures, for exam-
ple, batch, fed-batch, perfusion, and con-
tinuous fermentation [26, 27].
9.2
Recombinant Proteins Produced
in Plant Cell Suspension Cultures
The first recombinant protein expressed in
cultured plant cells was human serum al-
bumin, produced in tobacco suspension
cells derived from transgenic plants [28].
Since then, many different proteins have
been produced in suspension cells from a
variety of plant species, with a focus on
pharmaceutical proteins such as antibod-
ies, cytokines, growth factors, hormones,
and enzymes. A selection of these proteins
is listed in Table 9.1, which also provides
details of the expression construct used in
each case and the final product yield. An
overview of plant-based biopharmaceuti-
cals, their indication and status in clinical
trials is provided by Knblein [52]. Tobacco
cell lines have been used the most widely
because tobacco is the most popular
Fig. 9.1 Tobacco BY-2 suspension cells (A; original magnification
400) are cultivated in shake flasks (B) and bioreactors (C)
(2-L stirred reactor) under sterile and controlled conditions,
allowing the production of recombinant proteins according
to current pharmaceutical production standards.
Table 9.1 Recombinant proteins expressed in cultured plant cells
Expression host Expressed protein Promoter Localization, yield Reference(s)
Tobacco suspension
cells initiated from
transgenic plants
Human serum albu-
min (HSA)
Modified
CaMV 35S
Secretion/apoplast tar-
geting, 0.25 lg mg
1
pro-
tein in supernatant
28
Tobacco suspension
transgenic plants
ScFv antibody frag-
ment
CaMV 35S Secretion, up to
0.5 lg L
1
, up to 0.5% of
TSP
29
Tobacco cv BY-2
suspension cells
Human erythropoietin CaMV 35S Secreted, 1 pg g
1
FW 30, 31
Tobacco cv NT-1
suspension cells
Mouse monoclonal
Heavy chain
CaMV 35S Native heavy-chain secre-
tion signal, ca. 10 lg L
1
extracellular without
PVP, 350 lg L
1
with
PVP
32
Tobacco cv NT-1
suspension cells
Heavy chain mAb CaMV 35S Secreted up to 10 lg L
1
,
with stabilization up to
350 lg L
1
33
Tobacco cv NT-1
suspension cells
Bryodin 1 CaMV 35S Secreted up to 30 mg L
1
34
Tobacco cv NT-1
suspension cells
Human interleukin-2
and interleukin-4
(hIL-2 and hIL-4)
CaMV 35S Secreted (native signal
peptides), 8180 lg L
1
of culture broth
35
Tobacco suspension
cells
Recombinant ricin CaMV 35S 2537.5 lg L
1
36
Rice cv Bengal callus
cells
ScFv antibody frag-
ment
Maize ubi-
quitin-1
Apoplast targeting (opti-
mized Ig leader pep-
tides) and ER-retention,
up to 3.8 lg g
1
callus
FW
37
Tobacco cv Petite Ha-
vana SR1 suspension
transgenic plants
Mouse IgG-2b Enhanced
CaMV 35S
0.3% of TSP or 15 lg g
1
wet weight
38
Rice cv Taipei 309
suspension cells
Human a
1
-antitrypsin
(hAAT)
RAmy3D Secreted, 85 mg L
1
in
shake flask, 25 mg L
1
in
bioreactor
39
Tobacco cv BY-2
suspension cells
BiscFv antibody
fragment
Enhanced
CaMV 35S
Cytosolic (at detection
limit), apoplast-targeted
(up to 0.0064% of TSP),
ER-retained (up to
0.064% of TSP)
40
Tobacco cv NT-1
suspension cells
Human granulocyte-
macrophage colony-
stimulating factor
(hGM-CSF)
CaMV 35S Secreted/targeted to the
apoplast, ca. 250 lg L
1
extracellular,
ca. 150 lg L
1
intra-
cellular (based on
culture volume)
41
Expression host Expressed protein Promoter Localization, yield Reference(s)
Tobacco suspension
transgenic plants
ScFv antibody
fragment
CaMV 35S Apoplast targeting (spor-
amin secretion signal)
1 mg L
1
extracellular,
5 mg L
1
intracellular
42
Rice suspension cells Human a
1
-antitrypsin
(hAAT)
RAmy3D Up to 200 mg L
1
(calli
suspended to 40% (v/v)
cell density in induction
medium)
43
Soybean cv Williams
82 and tobacco cv
NT-1 suspension cells
Hepatitis B surface
antigen (HBsAg)
(ocs)
3
mas Intracellular up to
22 mg L
1
in soybean ca.
2 mg L
1
in tobacco
44
Tobacco suspension
cells
Human granulocyte-
macrophage colony-
stimulating factor
(hGM-CSF)
CaMV 35S 1.66.6 lg mL
1
upon
homogenizing the entire
culture broth
45
Rice cv Taipei 309
suspension cells
Human lysozyme RAmy3D Intracellular (although
RAmy3D signal peptide
was used), up to 34%
of TSP
46
vana SR1 suspension
cells
IL-12 Enhanced
CaMV 35S
Secreted, up to 800lgL
1
of supernatant
47
Tomato cv Seokwang
suspension cells
Human granulocyte-
macrophage colony-
stimulating factor
(hGM-CSF)
Enhanced
CaMV 35S
Secreted, up to 45 lg L
1
of supernatant
21
Tobacco cv NT-1
suspension cells
Hepatitis B surface
antigen (HBsAg)
Arabidopsis
ubq3
Secreted, up to 10 lg L
1
of particulate HBsAg
48
Tobacco cv BY-2
suspension cells
MAb against HBsAg CaMV 35S Secreted, ca. 50/50 be-
tween supernatant and
cells, total max ca.
15 mg L
1
49
Tobacco cv BY-2
suspension cells
Desmodus rotundus
Salivary plasminogen
activator a1 (DSPAa1)
Enhanced
CaMV 35S
Intracellular, up to
1.5 lg g
1
FW and de-
graded when secreted to
the supernatant (3 differ-
ent signal peptides were
used)
50
Tobacco cv BY-2
suspension cells
Thrombomodulin
derivate Solulin
TM
Enhanced
CaMV 35S
Intracellular, up to
27 lg g
1
FW and se-
creted, up to 2.1 mg L
1
of supernatant (3 differ-
ent signal peptides were
used)
51
TSP: total soluble protein, FW: fresh weight.
whole-plant system for recombinant pro-
tein production, and robust expression cas-
settes are available. However, rice cell lines
are now becoming popular due to the
availability of the inducible a-amylase pro-
moter, which in direct comparisons per-
forms better than any tobacco promoter
[53]. The a-amylase system has been used
for the production of GM-CSF [53], alpha-
1-antitrypsin [39, 43, 54, 55] and lysozyme
[46]. Other proteins have been produced in
soybean and tomato suspension cells
[21, 44].
9.3
Challenges and Solutions for the Production
Despite their great promise, several chal-
lenges remain to be addressed before plant
cell cultures can become commercially vi-
able as a production system. As shown in
Table 9.1, many of the proteins that have
been produced in cultured plant cells have
shown relatively poor yields (<10 mg L
1
).
In many cases, proteins have been pro-
duced efficiently for extended periods (e.g.,
[32]), but sometimes the yields decrease
over time, probably as a result of somaclo-
nal variation and other forms of genetic
instability (e.g., [44]). It may be possible to
circumvent such limitations by the careful
choice of stable, high-producing callus
lines (e.g., [46]). Technical obstacles in-
clude poor growth rates (at least compared
to microbial cultures), problems with cul-
ture morphology (e.g., aggregation, ten-
dency for cells to grow on the vessel walls)
and shear sensitivity [5658]. Some of
these issues can be minimized through
improved fermenter design and culture
conditions [59], or though modification of
the nutrient supply [6066]. However, it
may also be necessary to consider expres-
sion strategies in a broader sense, includ-
ing the design of the expression construct,
the medium composition, and the extrac-
tion and purification strategy. We consider
these factors in more detail below.
9.3.1
Design of the Expression Construct
In any platform for the expression of re-
combinant proteins, it is necessary to opti-
mize gene expression to promote accumu-
lation of the desired product. In plant
cells, this means fine-tuning all stages of
gene and protein expression, starting with
transcription and finishing with protein
targeting and post-translational modifica-
tion (see Part IV, Chapter 5). Typically, a
strong and constitutive promoter is pre-
ferred to maximize the rate of transcrip-
tion. The cauliflower mosaic virus (CaMV)
35S promoter has been used in most cell
lines from dicotyledonous plants [2],
although a hybrid ocs-mas promoter from
A. tumefaciens has been used to express
hepatitis B virus surface antigen in tobacco
and soybean cells [44]. For monocotyledo-
nous plants, in which the CaMV promoter
has a lower activity, the constitutive maize
ubiquitin-1 (Ubi1) promoter can be used
[37]. However, the most popular monocot
promoter by far is the rice inducible a-
amylase promoter. The native a-amylase
protein is secreted abundantly from cul-
tured rice cells in response to sucrose star-
vation, so the promoter represents an ideal
inducible system for use with rice suspen-
sion cells (inducible promoters are dis-
cussed in a recent review; see Ref. [67]).
Proteins that have been expressed using
this promoter system are listed in Table
9.1.
While transcriptional activity is impor-
tant for high product yields, the destina-
tion of the recombinant protein probably
has a more significant and direct impact
(see Part IV, Chapter 6). The proteins des-
tination is governed by targeting informa-
tion in the expression construct, specifical-
ly the presence or absence of a signal pep-
tide (which directs the protein into the se-
cretory pathway), plus other signals that
determine whether the protein is actually
secreted or instead accumulates within the
cell. Various different signal peptides have
been used in plant cell culture systems,
and those from plants and animals appear
to work with equal efficiency. Therefore,
when mammalian proteins that are nor-
mally secreted from their native cells are
expressed in plant cells, the native signal
peptide is often retained and used to target
the recombinant protein for secretion in
its new environment [10, 31, 37, 45].
Where there is no native signal peptide,
perhaps because the protein is not se-
creted from its source tissue, a heterolo-
gous plant leader may be included up-
stream of the transgene [29]. Where the
rice a-amylase promoter is used, the rice
a-amylase signal sequence is generally also
included [39, 46, 53, 54].
Proteins targeted to the secretory path-
way in cultured plant cells are generally
secreted into the culture medium, a strat-
egy which simplifies protein recovery and
purification. In the absence of other sort-
ing information, proteins directed to the
secretory pathway in cultured cells will
eventually reach the apoplast, from where
they will either diffuse through the cell
wall and into the culture medium or lodge
in the cell wall matrix. The fate of the pro-
tein depends mostly on its size, since the
pores in the plant cell wall allow the pas-
sage of globular proteins less than 20 kDa
in size. However, some larger pores are
present, and larger proteins can therefore
diffuse through the cell wall, albeit at a
slow rate [68]. Recombinant proteins with
molecular weights significantly in excess
of 20 kDa have been recovered efficiently
from the culture medium [10, 47, 51, 54],
while in other cases even fairly small pro-
teins have remained trapped in the cell
wall and have needed to be released by
disruption or mild enzymatic digestion
[31, 69]. Enzymatic digestion is preferable
to cell disruption because the latter re-
leases phenolic substances and proteases
that can reduce protein yield, and hence
introduces additional processing steps.
9.3.2
Post-translational Modification
of Recombinant Proteins in Plants
The destination of a recombinant protein
within the cell not only affects the yield of
product but also how it is modified. In par-
ticular, it is possible to influence how a pro-
tein is glycosylated by targeting it for retrie-
val to the endoplasmic reticulum (ER) (an
early part of the secretory pathway), since
universal-type high-mannose glycosylation
takes place here, whereas all plant-specific
modifications take place downstream in
the Golgi apparatus [70]. Glycans serve a
number of structural and functional roles
in a glycoprotein, and often affect protein
stability, solubility, folding, biological activ-
ity, longevity, interactions with cellular re-
ceptors, and immunogenicity (see Part IV,
Chapters 2 and 3 and Part VI, Chapter 2).
Therefore, where the precise glycan struc-
tures are important for a particular glyco-
protein, it may be necessary to sacrifice
yield and recover the recombinant protein
from the intracellular environment by cell
disruption rather than secrete an inappro-
priately modified protein into the medium.
Mammalian glycoproteins produced in
plants are glycosylated on the same aspara-
gine (Asn) residues as they are in mam-
malian cells. The first stage of N-glycan
synthesis begins in the ER, involving the
co-translational addition of an oligosac-
charide precursor (Glc
3
Man
9
GlcNAc
2
) to
specific Asn residues within the consensus
sequence Asn-X-Ser/Thr. The N-glycans
then undergo several maturation reactions,
including the removal of certain residues
in the ER and the addition of further resi-
dues in the ER and Golgi apparatus. The
steps occurring in the ER are conserved in
mammals and plants, but they diverge in
the late Golgi apparatus so that core
a(1,6)-linked fucose and terminal sialic
acid residues are added in mammals,
whereas bisecting b(1,2)-xylose and core
a(1,3)-fucose residues are added in plants.
Most of the evidence for differential glyco-
sylation has been obtained through the
analysis of antibodies produced in cultured
mammalian cells and transgenic plants,
but there remains some controversy over
the exact capabilities of plants in terms of
glycan synthesis (see Part IV, Chapter 5).
For example, initial reports suggested that
antibodies produced in tobacco were glyco-
sylated in an heterogeneous manner, com-
prising a mixture of high-mannose-type
and complex-type N-glycans, whereas
those produced in alfalfa were more
homogeneous and contained mostly com-
plex-type glycans [71]. However, more re-
cent studies suggest that this observation
may reflect properties of the antibody itself
rather than the plant (E. Stoger, unpub-
lished data). Recently, it has been reported
that sialylated glycoconjugates can be pro-
duced in Arabidopsis thaliana suspension
cells [72], but this has been disputed and
the exact capability of plant cells regarding
the synthesis of sialic acid remains unclear
[73, 74].
9.3.3
Factors that Affect Protein Stability
Maximizing protein expression and accu-
mulation is only half of the story when
considering how to improve the yield of
recombinant proteins in plant cell cul-
tures. It is also necessary to minimize pro-
tein turnover, which can occur due to pro-
tein degradation within the cell and in the
extracellular environment [10]. Intracellu-
lar degradation probably results from the
release of proteases during homogeniza-
tion, although incomplete folding or as-
sembly may also occur where multimeric
or particularly complex proteins are ex-
pressed [10]. Extracellular degradation, par-
ticularly in the culture medium when the
product is secreted, is likely to reflect a
collection of underlying factors. Significant
causes may include the presence of pro-
teases in the medium [10, 39, 45, 47, 50,
53], the tendency for some proteins to pre-
cipitate or aggregate under non-physiologi-
cal conditions [10], and the adsorption of
certain charged proteins to the walls of the
reactor vessel [32]. The activity of proteases
in the culture medium can be reduced
using various strategies such as adding
protease-inhibitors or gelatin to act as an
alternative substrate [47]. A very successful
strategy is to use rice cell suspensions
with the transgene under the control of
the inducible a-amylase promoter. Rice
cells are thought to secrete less protease
into the culture medium than tobacco
cells, and the promoter can be used to re-
strict protein expression to a defined pro-
duction phase which is separate from the
growth phase, when most proteases are
produced [53].
The most commonly used media for
plant cell cultures are MS medium [75],
Gamborgs B5 medium [76], and Whites
medium [77]. The composition of these
media has been modified in order to opti-
mize the growth of plant cell cultures for
the production of secondary metabolites
[78, 79] or recombinant proteins [41]. The
recovery of recombinant proteins from
plant cell cultures can be improved by add-
ing agents to the medium that interfere
with protein degradation and other unde-
sirable processes, while enhancing protein
synthesis and secretion (Table 9.2). Be-
cause the exact circumstances and require-
ments differ according to the cell line, ex-
pression strategy and product, it is difficult
to predict the effect each reagent may have
under any particular set of conditions. Me-
dium additives must therefore be tested
empirically. Substances that have been
used include simple inorganic compounds
[81], amino acids [38], dimethylsulfoxide
(DMSO) [82], polyethylene glycol [45, 47],
heamin [69], polyvinylpyrrolidone (PVP)
[10, 31, 32, 45, 47, 81], plant hormones
such as gibberellic acid [69], and proteins
such as gelatin [45, 47, 81] and bovine se-
rum albumin (BSA) [41]. Some of these
products enhance cell growth or survival,
Table 9.2 The effect of medium additives on product yields in cultured plant cells
Expression host Expressed protein Medium
additive
Effect Reference(s)
Tobacco cv NT-1
suspension cells
Heavy-chain monoclo-
nal antibody
PVP Addition of 0.75% PVP
360000 increased se-
creted product accumu-
lation 35-fold
32, 33
Tobacco cv BY-2
suspension cells
HGM-CSF BSA, NaCl Enhanced secreted prod-
uct accumulation 100%
(BSA), or 50% (NaCl)
41
Tobacco suspension
transgenic plants
Mouse IgG
1
Brefeldin A Inhibited secretory path-
way and prevented de-
gradation of secreted
protein. Increased mAb
accumulation 2.7-fold
10
Tobacco suspension
transgenic plants
Mouse IgG
1
Reducing
manganese
In manganese-depleted
medium the stability
and accumulation of se-
creted IgG
1
was in-
creased ca. 1.7fold
69
Tobacco cv Petite
Havana SR1
suspension cells
Human granulocyte-
macrophage colony-
stimulating factor
(hGM-CSF)
Pluronic
antifoam,
PEG
Pluronic antifoam in-
creased the growth rate
almost 2-fold, PEG-8000
increased hGM CSF ac-
cumulation 4-fold
80
vana SR1 suspension
cells
HGM-CSF Gelatin,
PVP, PEG
2% gelatin resulted in
4.6-fold improvement in
yield, PVP and PEG
showed no effect
45
vana SR1 suspension
cells
IL-12 heterodimer Gelatin,
PVP, PEG
2% gelatin increased IL-
2 accumulation 7-fold
47
some enhance protein synthesis, some in-
terfere either positively or negatively with
secretion, some act to stabilize extracellu-
lar proteins, and some either inhibit pro-
tease activity or provide alternative sub-
strates for proteases. Care must be exer-
cised because the presence of certain re-
agents may be counterproductive. For
example, BSA and gelatin inhibit proteoly-
sis but can interfere with downstream pro-
cessing. Additionally, these are derived
from animal sources and thus possess the
risk of contamination with human patho-
gens (see Part I, Chapter 6 and Part VII,
Chapter 1).
9.4
Process Engineering
Improvements in product yields can be
gained not only by construct design and
modification of the media, but also by var-
iations in the growth and harvest condi-
tions. Culture process development was
initially directed towards dedicated bioreac-
tor designs suitable for very large working
volumes or providing adequate agitation at
high cell densities with low shear stress
[83, 84]. As the focus shifted towards the
production of recombinant proteins, differ-
ent fermentation strategies (batch, fed-
batch, repeated batch or draw/fill, continu-
ous culture and perfusion culture) were
applied to see if a standardized approach
would be useful, particularly vis--vis regu-
latory approval (see Part VII, Chapter 4)
[27, 85, 86]. It was soon discovered that ge-
netic stability was difficult to maintain in
large-scale cultures [2], so continuous and
perfusion fermentation strategies, which
are long-term processes, were generally
unsuitable. Therefore, batch and fed-batch
processes are the most likely to be favored
when plant suspension cultures are used
for the production of recombinant proteins
on a commercial scale. Another issue aris-
ing from this consideration is the absolute
necessity to establish cryopreservation
techniques and master and working cell
banks for commercial processes. There-
fore, process development for plant cell
cultures has focused on the optimization
of media, medium supplements and physi-
cal parameters.
Another way to improve yields is contin-
uous product removal by the addition of
an inert solid absorber to the fermentation
broth, or by using an absorber in a bypass.
In the case of recombinant antibodies, this
can be achieved simply by cycling the me-
dium through an in-line Protein G column
during the fermentation, and eluting the
bound antibody at the end of the produc-
tion cycle (see Part IV, Chapter 16 and Part
V, Chapter 1). Using this procedure, the
yields of recombinant antibody can be in-
creased up to eight-fold, depending on
variables such as the flow rate, culture per-
iod and column pH [87]. Similar principles
can be applied to any protein expressed
with a suitable affinity tag for example,
an epitope tag, a specific affinity partner
such as glutathione-S-transferase, or a
His
6
tag. The latter has been used to puri-
fy granulocyte-macrophage colony-stimu-
lating factor (GM-CSF) produced in tobac-
co suspension cells by continuous cycling
of the culture medium through a column
packed with metal affinity resin, resulting
in a two-fold increase in yields compared
to cultures that were harvested by conven-
tional methods [41, 87].
One aspect of plant cell culture that
plays a critical role in determining product
yields is the rate of oxygen transfer. Inade-
quate aeration has a limiting effect on cell
growth, and has been shown to limit prod-
uct yields in, for example, an antibody-pro-
ducing tobacco cell line grown in shake
flasks [88]. However, while increased aera-
tion in a stirred 5-L bioreactor helped to
increase the protein yields, excessive aera-
tion caused foaming and had a counter-
productive effect on both cell growth and
product yields [88].
9.5
Downstream Processing
Downstream processing refers to those
post-culture steps which result in the isola-
tion and purification of the product. Re-
gardless of the production system, down-
stream processing is expensive, accounting
for up to 80% of overall production costs
(although this depends on the level of pur-
ity required and is highest for clinical-
grade materials). Downstream processing
involves multiple steps, beginning with
stages that are host-specific but product-
generic and ending with the opposite situ-
ation processes that are host-generic but
product-specific (see Part IV, Chapter 16).
That is, the early stages of downstream
processing are generally similar regardless
of the product being purified, but must
take into account differences between al-
ternative expression platforms, whereas
the last stages are defined very much by
the product itself and how it can be sepa-
rated from unwanted contaminants (see
Part I, Chapter 6 and Part VII, Chapter 1).
In many cases, it is necessary to develop
specific processing steps for each product,
although certain classes of product can be
isolated using a standardized approach
(e.g., the use of affinity chromatography to
isolate recombinant antibodies or recombi-
nant proteins expressed with integral epi-
tope tags). Several aspects of early down-
stream processing must be customized
specifically for plant systems, including
steps for the removal of fibers, oils and
other by-products, and process optimiza-
tion for the treatment of different plant
species and tissues. However, these issues
are much less important in plant cell cul-
tures, which tend to lack the phenolics
and other metabolites produced by whole
plant tissues, and do not accumulate oils
and fibers since these tend to be the prod-
ucts of specialized storage tissues.
In plant cell cultures, the main process
decision is whether to extract the protein
directly from the cell biomass, or to re-
cover it from the medium. As stated
above, secreted proteins are advantageous
from a processing perspective because they
circumvent several unit operations during
the purification process (e.g., cell disrup-
tion, which results in the release of pro-
teases, and the removal of debris, which
can clog filters and chromatography media
therefore interfering with downstream op-
erations) (see Part IV, Chapters 6 and 7).
Furthermore, the culture supernatant is
much simpler than the intracellular mili-
eu: it has fewer contaminating (compet-
ing) proteins and metabolites. On the
other hand, the target protein is highly di-
luted in this environment, so larger vol-
umes of raw material must be processed
to extract the product.
Where the product is secreted, it should
form the major proteinaceous component
of the harvest broth. The best time to har-
vest will depend on the dynamics of pro-
tein expression, secretion and stability, but
in general it is beneficial to seek an appro-
priate harvesting window in the produc-
tion cycle where product expression and
accumulation are maximized but degrada-
tion is limited. For example, in batch or
fed-batch processes, peak production of
the recombinant protein often occurs in
the exponential phase, and begins to de-
cline at the point where the cell mass and
total protein content of the culture reach
their maximum levels [10]. The next opera-
tions include clarification of the medium,
concentration of the product, and its isola-
tion. Large-scale clarification is generally
carried out by a combination of dead-end
and/or cross-flow filtration, sometimes
preceded by bulk cell mass removal using
a decanter, plate separator or (semi)contin-
uous centrifuge. Of these methods, cross-
flow filtration generally provides the best
clarified feed for packed-bed chromatogra-
phy, but it is also requires the greatest
amount of fine-tuning. Hollow-fiber sys-
tems may be more suitable for clarification
if the cell line produces substantial
amounts of extracellular polysaccharides,
as is the case for some transformed BY-2
cell lines. Polysaccharides in the medium
can coat membrane surfaces and form a
gel, which blocks the filters and traps the
desired protein within the gel matrix. An-
other alternative is expanded bed adsorp-
tion (EBA), which may help to capture the
target protein from particulate feed materi-
al. This technique has been used to cap-
ture recombinant proteins from several in-
dustrial-scale microbial and animal cell
cultures (see Part IV, Chapters 4, 12, and
13), and has recently been adapted for use
with plant cells [89]. With further design
improvements, EBA could facilitate the si-
multaneous clarification, concentration,
and initial purification of proteins from
plant cell fermentation broth or cell ex-
tracts [90]. Another alternative is the use
of aqueous two-phase systems (ATPS) for
cultivation and, eventually, in-situ extrac-
tion from plant cell cultures [91].
Where the product must be captured
from the intracellular environment, the
first downstream processing steps are cell
disruption and the removal of debris. Sev-
eral methods have been used for disrup-
tion, including sonication, pressure homo-
genization, enzymatic treatment, and wet
milling. These are all efficient techniques,
but the resulting feedstream is much den-
ser and more particulate than decanted
medium, and the cell debris and fines
generated during homogenization may be
more difficult to remove than intact cells.
The advantages of intracellular expression
for initial downstream processing lie in
the smaller volume of the starting material
and the generally higher concentration of
the target protein, while the major disad-
vantage is the more complex composition
of the feedstream and the liberation of
proteolytic and oxidizing substances.
The final operations inevitably involve
liquid chromatography steps that selec-
tively elute the desired protein. For some
products, such as antibodies, highly selec-
tive capture can be achieved immediately
using Protein A or Protein G affinity col-
umns. Similarly, proteins with affinity tags
can be captured using columns charged
their respective ligands (e.g., His
6
and Ni-
NTA resin). For other proteins, which may
need to be purified using non-affinity-
based methods, several rounds of ion-ex-
change and/or reversed-phase chromato-
graphy may be required. As stated above,
the first chromatographic steps need the
most development with regard to the spe-
cific production system, whereas the final
steps need tuning to the specific product.
For large-scale production, robust and in-
expensive chromatography media are used
in the initial steps, and some loss of selec-
tivity and resolution is anticipated [92].
9.6
Regulatory Considerations
For the production of clinical-grade pro-
teins, plant cell cultures and all down-
stream processing operations need to meet
the standards that have been set for other
biopharmaceutical production systems [93]
(see Part VII, Chapter 4). Regulatory guid-
ance for biopharmaceutical production in
plants currently exists only as draft legisla-
tion, and this does not explicitly discuss
plant cell cultures since the legislation was
drafted to deal with the growing use of ter-
restrial plants [9496]. Furthermore, the
use of plant cell cultures also falls under
the general GMP guidelines for biotechno-
logical production (e.g., Annex 18 of the
EU Guide to Good Manufacturing Prac-
tice) (see Part I, Chapter 6 and Part VII,
Chapter 1). One of the most important re-
quirements is a thorough description of
the genetic background and genetic stabili-
ty of the host cell, as well as the precise
documentation of all events associated
with the introduction of the transgene into
the plant cell (see Part IV, Chapter 6).
However, most of the plant cell lines cur-
rently used for recombinant protein ex-
pression including BY-2 and NT-1 have
a long history in the public domain and
have not been characterized sufficiently to
fulfill GMP requirements, and are not de-
posited in dedicated cell banking facilities.
Since cell banking is a prerequisite for the
reliable supply of well-defined starting ma-
terial, it will be necessary to identify suit-
able wild-type and transformed cell lines
for banking. Also, routine procedures for
the cryopreservation of plant cells will
have to be developed further [97].
9.7
Conclusions
The use of plant suspension cells for the
production of biopharmaceutical proteins
has many advantages, including safety, de-
fined growth conditions, containment, and
the ability to use continuous-culture strate-
gies. However, some challenges remain to
be met before the commercial success of
plant cells is assured. Such challenges in-
clude improving the current low yields, op-
timization of downstream processing, and
demonstration of biological equivalence. If
these issues can be addressed, then plant
cells offer a real potential to compete not
only with transgenic plants, but also with
other fermenter systems for the commer-
cial production of biopharmaceuticals.
References
1 Doran PM (2000) Foreign protein production
in plant tissue cultures. Curr Opin Biotechnol
11: 199204.
2 Hellwig S, Drossard J, Twyman RM, Fischer
R (2004) Plant cell cultures for the production
of recombinant diagnostic and therapeutic
proteins. Nature Biotechnol 22: 14151422.
3 Gomord V, Faye L (2004) Posttranslational
modification of therapeutic proteins in plants.
Curr Opin Plant Biol 7: 171181.
4 Knblein J (2004) Biopharmaceuticals ex-
pressed in plants? A new era in the new mil-
lennium. In: Mller R, Kayser O (Eds.), Appli-
cations in Pharmaceutical Biotechnology, Wiley-
VCH, pp. 3556.
5 Twyman RM, Stoger E, Schillberg S, Christou
P, Fischer R (2003) Molecular farming in
plants: host systems and expression technol-
ogy. Trends Biotechnol 21: 570578.
6 Ma JKC, Drake PMW, Christou P (2003) The
production of recombinant pharmaceutical
proteins in plants. Nature Rev Genet 4: 794
805.
7 Fischer R, Stoger E, Schillberg S, Christou P,
Twyman RM (2004) Plant-based production of
biopharmaceuticals. Curr Opin Plant Biol 7:
152158.
8 Twyman RM, Schillberg S, Fischer R (2005)
The transgenic plant market in the pharma-
ceutical industry. Expert Opin Emerg Drugs
10: 185218.
9 Hilton MG, Rhodes MJC (1990) Growth and
hyoscyamine production of hairy root cultures
of Datura stramonium in a modified stirred
tank reactor. Appl Microbiol Biotechnol 33:
132138.
References 961
10 Sharp JM, Doran PM (2001) Strategies for en-
hancing monoclonal antibody accumulation in
plant cell and organ cultures. Biotechnol Prog
17: 979992.
11 Archambault J (1991) Large-scale (20-L) cul-
ture of surface-immobilized Catharanthus ro-
seus cells. Enzyme Microbial Technol 13: 882
892.
12 Kieran PM, MacLoughlin PF, Malone DM
(1997) Plant cell suspension cultures: some
engineering considerations. J Biotechnol 59:
3952.
13 Schlatmann JE, ten Hoopen HJG, Heijnen JJ
(1996) Large-scale production of secondary
metabolites by plant cell cultures. In: DiCos-
mo F, Misawa M (Eds.), Plant cell culture sec-
ondary metabolism: Toward industrial applica-
tion. CRC Press, Boca Raton, FL, pp. 1152.
Appl Biochem Biotechnol 50: 189216.
14 Wen WS (1995) Bioprocessing technology for
plant cell suspension cultures.
15 Desikan R, Hancock JT, Neill SJ, Coffey MJ,
Jones OT (1996) Elicitor-induced generation of
active oxygen in suspension cultures of Arabi-
dopsis thaliana. Biochem Soc Trans 24: 199S.
16 Seki M, Ohzora C, Takeda M, Furusaki S
(1997) Taxol (Paclitaxel) production using free
and immobilized cells of Taxus cuspidata. Bio-
technol Bioeng 53: 214219.
17 Van Der Heijden R, Verpoorte R, ten Hoopen
HJG (1989) Cell and tissue cultures of Cathar-
anthus roseus (L) Don G. A literature survey.
Plant Cell Tiss Org Cult 18: 231280.
18 Chen MH, Liu LF, Chen YR, Wu HK, Yu SM
(1994) Expression of a-amylases, carbohydrate-
metabolism, and autophagy in cultured rice
cells is coordinately regulated by sugar nutri-
ent. Plant J 6: 625636.
19 Daniell T, Edwards R (1995) Changes in pro-
tein methylation associated with the elicitation
response in cell cultures of alfalfa (Medicago
sativa L.). FEBS Lett 360: 5761.
20 Hoehl U, Upmeier B, Barz W (1988) Growth
and nicotinate biotransformation in batch cul-
tured and airlift fermenter grown soybean cell
suspension cultures. Appl Microbiol Biotech-
nol 28: 319323.
21 Kwon TH, Kim YS, Lee JH, Yang MS (2003)
Production and secretion of biologically active
human granulocyte-macrophage colony stimu-
lating factor in transgenic tomato suspension
cultures. Biotechnol Lett 25: 15711574.
22 Nagata T, Nemoto Y, Hasezawa S (1992) To-
bacco BY-2 cell line as the HeLa cell in the
cell biology of higher plants. Int Rev Cytol
132: 130.
23 Horsch RB, Fry JE, Hoffmann NL, Eichholtz
D, Rogers SG, Fraley RT (1985) A simple and
general method for transferring genes into
plants. Science 227: 12291231.
24 Koncz C, Schell J (1986) The promoter of TL-
DNA gene 5 controls the tissue-specific ex-
pression of chimeric genes carried by a novel
type of Agrobacterium binary vector. Mol Gen
Genet 204: 383396.
25 Christou P (1993) Particle gun-mediated trans-
formation. Curr Opin Biotechnol 4: 135141.
26 Hooker BS, Lee JM, An GH (1990) Cultivation
of plant-cells in a stirred vessel effect of im-
peller design. Biotechnol Bioeng 35: 296304.
27 Ten Hoopen HJG, van Gulik WM, Heijnen JJ
(1992) Continuous culture of suspended plant
cells. In Vitro Cell Dev Biol Plant 28: 115
120.
28 Sijmons PC, Dekker BMM, Schrammeijer B,
Verwoerd TC, van den Elzen PJM, Hoekema
A (1990) Production of correctly processed hu-
man serum albumin in transgenic plants.
Bio/Technology 8: 217221.
29 Firek S, Draper J, Owen MRL, Gandecha A,
Cockburn B, Whitelam GC (1993) Secretion
of a functional single-chain Fv protein in
transgenic tobacco plants and cell suspension
cultures. Plant Mol Biol 23: 861870.
30 Matsumoto S, Ishii A, Ikura K, Ueda M, Sasa-
ki R (1993) Expression of human erythropoie-
tin in cultured tobacco cells. Biosci Biotechnol
Biochem 57: 12491252.
31 Matsumoto S, Ikura K, Ueda M, Sasaki R
(1995) Characterization of a human glycopro-
tein (erythropoietin) produced in cultured to-
bacco cells. Plant Mol Biol 27: 11631172.
32 Magnuson NS, Linzmaier PM, Gao JW,
Reeves R, An GH, Lee JM (1996) Enhanced
recovery of a secreted mammalian protein
from suspension culture of genetically modi-
fied tobacco cells. Protein Express Purif 7:
220228.
33 LaCount W, An GH, Lee JM (1997) The effect
of polyvinylpyrrolidone (PVP) on the heavy
chain monoclonal antibody production from
plant suspension cultures. Biotechnol Lett 19:
9396.
34 Francisco JA, Gawlak SL, Miller M, Bathe J,
Russell D, Chace D, Mixan B, Zhao L, Fell
HP, Siegall CB (1997) Expression and charac-
terization of bryodin 1 and a bryodin 1-based
single-chain immunotoxin from tobacco cell
culture. Bioconjug Chem 8: 708713.
35 Magnuson NS, Linzmaier PM, Reeves R, An
GH, HayGlass K, Lee JM (1998) Secretion of
biologically active human interleukin-2 and in-
terleukin-4from genetically modified tobacco
cells in suspension culture. Protein Expr Purif
13: 4552.
36 Sehnke PC, Ferl RJ (1999) Processing of pre-
proricin in transgenic tobacco. Protein Expr
Purif 15: 188195.
37 Torres E, Vaquero C, Nicholson L, Sack M,
Stoger E, Drossard J, Christou P, Fischer R,
Perrin Y (1999) Rice cell culture as an alterna-
tive production system for functional diagnos-
tic and therapeutic antibodies. Transgenic Res
8: 441449.
38 Fischer R, Liao YC, Drossard J (1999) Affinity-
purification of a TMV-specific recombinant
full-size antibody from a transgenic tobacco
suspension culture. J Immunol Methods 226:
110.
39 Terashima M, Ejiri Y, Hashikawa N, Yoshida
H (1999) Effect of osmotic pressure on hu-
man alpha(1)-antitrypsin production by plant
cell culture. Biochem Eng J 4: 3136.
40 Fischer R, Schumann D, Zimmermann S,
Drossard J, Sack M, Schillberg S (1999) Ex-
pression and characterization of bispecific sin-
gle-chain Fv fragments produced in transgenic
plants. Eur J Biochem 262: 810816.
41 James EA, Wang CL, Wang ZP, Reeves R,
Shin JH, Magnuson NS, Lee JM (2000) Pro-
duction and characterization of biologically ac-
tive human GM-CSF secreted by genetically
modified plant cells. Protein Express Purif 19:
131138.
42 Ramirez N, Lorenzo D, Palenzuela D, Herrera
L, Ayala M, Fuentes A, Perez M, Gavilondo J,
Oramas P (2000) Single-chain antibody frag-
ments specific to the hepatitis B surface anti-
gen, produced in recombinant tobacco cell
cultures. Biotechnol Lett 22: 12331236.
43 Huang JM, Sutliff TD, Wu LY, Nandi S, Benge
K, Terashima M, Ralston AH, Drohan W,
Huang N, Rodriguez RL (2001) Expression
and purification of functional human alpha-1-
antitrypsin from cultured plant cells. Biotech-
nol Prog 17: 126133.
44 Smith ML, Mason HS, Shuler ML (2002) He-
patitis B surface antigen (HBsAg) expression
in plant cell culture: Kinetics of antigen accu-
mulation in batch culture and its intracellular
form. Biotechnol Bioeng 80: 812822.
45 Lee JH, Kim NS, Kwon TH, Jang YS, Yang
MS (2002) Increased production of human
granulocyte-macrophage colony stimulating
factor (hGM-CSF) by the addition of stabiliz-
ing polymer in plant suspension cultures.
J Biotechnol 96: 205211.
46 Huang JM, Wu LY, Yalda D, Adkins Y, Kelle-
her SL, Crane M, Lonnerdal B, Rodriguez RL,
Huang N (2002) Expression of functional re-
combinant human lysozyme in transgenic rice
cell culture. Transgenic Res 11: 229239.
47 Kwon TH, Seo JE, Kim J, Lee JH, Jang YS,
Yang MS (2003) Expression and secretion of
the heterodimeric protein interleukin-12 in
plant cell suspension culture. Biotechnol
Bioeng 81: 870875.
48 Sunil Kuma GB, Ganapathi TR, Revathi CJ,
Prasad KSN, Bapat VA (2003) Expression of
hepatitis B surface antigen in tobacco cell sus-
pension cultures. Protein Expr Purif 32: 10
17.
49 Yano A, Maeda F, Takekoshi M (2004) Trans-
genic tobacco cells producing the human
monoclonal antibody to hepatitis B virus sur-
face antigen. J Med Virol 73: 208215.
50 Schiermeyer A, Schinkel H, Apel S, Fischer
R, Schillberg S (2005) Production of Desmodus
rotundus salivary plasminogen activator a1
(DSPAa1) in tobacco is hampered by proteoly-
sis. Biotechnol Bioeng 89: 848858.
51 Schinkel H, Schiermeyer A, Soeur R, Fischer
R, Schillberg S (2005) Production of an active
recombinant thrombomodulin derivative in
transgenic tobacco plants and suspension
cells. Transgenic Res (in press).
52 Knblein J (2003). Biotech: A New Era In The
New Millennium? Fermentation and Expres-
sion of Biopharmaceuticals in Plants.
SCREENING Trends in Drug Discovery 4:
1416.
53 Shin YJ, Hong SY, Kwon TH, Jang YS, Yang
MS (2003) High level of expression of recom-
binant human granulocyte-macrophage colony
stimulating factor in transgenic rice cell sus-
pension culture. Biotechnol Bioeng 82: 778
783.
54 Terashima M, Murai Y, Kawamura M, Naka-
nishi S, Stoltz T, Chen L, Drohan W, Rodri-
guez RL, Katoh S (1999) Production of func-
tional human alpha(1)-antitrypsin by plant cell
References 963
culture. Appl Microbiol Biotechnol 52: 516
523.
55 Trexler MM, McDonald KA, Jackman AP
(2002) Bioreactor production of human al-
pha(1)-antitrypsin using metabolically regu-
lated plant cell cultures. Biotechnol Prog 18:
501508.
56 Meijer JJ, ten Hoopen HJG, Vangameren YM,
Luyben KCAM, Libbenga KR (1994) Effects of
hydrodynamic stress on the growth of plant-
cells in batch and continuous-culture. Enzyme
Microbial Technol 16: 467477.
57 Offringa R, Degroot MJA, Haagsman HJ,
Does MP, Vandenelzen PJM, Hooykaas PJJ
(1990) Extrachromosomal homologous recom-
bination and gene targeting in plant cells after
Agrobacterium-mediated transformation.
EMBO J 9: 30773084.
58 Yu SX, Kwok KH, Doran PM (1996) Effect of
sucrose, exogenous product concentration,
and other culture conditions on growth and
steroidal alkaloid production by Solanum avi-
culare hairy roots. Enzyme Microbial Technol
18: 238243.
59 Bohme C, Schroder MB, Jung-Heiliger H,
Lehmann J (1997) Plant cell suspension cul-
ture in a bench-scale fermenter with a newly
designed membrane stirrer for bubble-free
aeration. Appl Microbial Biotechnol 48: 149
154.
60 Sakamoto K, Iida K, Sawamura K, Hajiro K,
Asada Y, Yoshikawa T, Furuya T (1993) Effects
of nutrients on anthocyanin production in cul-
tured cells of Aralia cordata. Phytochemistry
33: 357360.
61 Sato K, Nakayama M, Shigeta J (1996) Cultur-
ing conditions affecting the production of an-
thocyanin in suspended cell cultures of straw-
berry. Plant Sci 113: 9198.
62 Ishikawa A, Yoshihara T, Nakamura K (1994)
Jasmonate-inducible expression of a potato ca-
thepsin-D inhibitor-GUS gene fusion in tobac-
co cells. Plant Mol Biol 26: 403414.
63 Ketchum REB, Gibson DM, Croteau RB, Shu-
ler L (1999) The kinetics of taxoid accumula-
tion in cell suspension cultures of Taxus fol-
lowing elicitation with methyl jasmonate. Bio-
technol Bioeng 62: 91105.
64 McCormack BA, Gregory ACE, Kerry ME,
Smith C, Bolwell GP (1997) Purification of an
elicitor-induced glucan synthase (callose
synthase) from suspension cultures of French
bean (Phaseolus vulgaris L.): purification and
immunolocation of a probable M
r
-65000 sub-
unit of the enzyme. Planta 203: 196203.
65 Mirjalili N, Linden JC (1996) Methyl jasmo-
nate induced production of taxol in suspen-
sion cultures of Taxus cuspidata: Ethylene in-
teraction and induction models. Biotechnol
Prog 12: 110118.
66 Yukimune Y, Tabata H, Higashi Y, Hara Y
(1996) Methyl jasmonate-induced overproduc-
tion of paclitaxel and baccatin III in Taxus cell
suspension cultures. Nature Biotechnol 14:
11291132.
67 Padidam M (2003) Chemically regulated gene
expression in plants. Curr Opin Plant Biol 6:
169177.
68 Carpita N, Sabularse D, Montezinos D, Del-
mer DP (1979) Determination of the pore size
of cell walls of living plant cells. Science 205:
11441147.
69 Tsoi BMY, Doran PM (2002) Effect of medium
properties and additives on antibody stability
and accumulation in suspended plant cell cul-
tures. Biotechnol Appl Biochem 35: 171180.
70 Gomord V, Fitchette AC, Lerouge P, Faye L
(2004) Glycosylation of plant-made pharma-
ceuticals. In: Fischer R, Schillberg S (Eds.),
Molecular Farming: Plant-made pharmaceuticals
and technical proteins. Wiley-VCH Verlag
GmbH & Co. KGaA, pp 233250.
71 Bardor M, Loutelier-Bourhis C, Paccalet T, Co-
sette P, Fitchette AC, Vezina LP, Trepanier S,
Dargis M, Lemieux R, Lange C, Faye L, Ler-
ouge P (2003) Monoclonal C5-1 antibody pro-
duced in transgenic alfalfa plants exhibits a
Nglycosylation that is homogenous and suit-
able for glyco-engineering into human-compa-
tible structures. Plant Biotechnol J 1: 451462.
72 Shah MM, Fujiyama K, Flynn CR, Joshi L
(2003) Sialylated endogenous glycoconjugates
in plant cells. Nature Biotechnol 21: 14701471.
73 Seveno M, Bardor M, Paccalet T, Gomord V,
Lerouge P, Faye L (2004) Glycoprotein sialyla-
tion in plants? Nature Biotechnol 22: 1351
1352.
74 Shah MM, Fujiyama K, Flynn CR, Joshi L
(2004) Glycoprotein sialylation in plants? Re-
ply. Nature Biotechnol 22: 13521353.
75 Murashige T, Skoog F (1962) A revised medi-
um for rapid growth and bio assay for tobacco
tissue cultures. Physiol. Plant. 15: 473497.
76 Gamborg OL, Miller RA, Ojima K (1968) Nu-
trient requirements of suspension culture of
soyabean root cells. Exp Cell Res 50: 151158.
77 White PR (1963) The cultivation of animal and
plant cells. Ronald Press, NY
78 Gao WY, Fan L, Paek KY (2000) Yellow and
red pigment production by cell cultures of
Carthamus tinctorius in a bioreactor. Plant Cell
Tiss Org Cult 60: 95100.
79 Raval KN, Hellwig S, Prakash G, Ramos-Plas-
encia A, Srivastava A, Buchs J (2003) Neces-
sity of a two-stage process for the production
of azadirachtin-related limonoids in suspen-
sion cultures of Azadirachta indica. J Biosci
Bioeng 96: 1622.
80 Lee SY, Kim DI (2002) Stimulation of murine
granulocyte macrophage-colony stimulating
factor production by pluronic F-68 and poly-
ethylene glycol in transgenic Nicotiana tabacum
cell culture. Biotechnol Lett 24: 17791783.
81 Wongsamuth R, Doran PM (1997) Production
of monoclonal antibodies by tobacco hairy
roots. Biotechnol Bioeng 54: 401415.
82 Wahl MF, An GH, Lee JM (1995) Effects of di-
methylsulfoxide on heavy-chain monoclonal
antibody production from plant cell culture.
Biotechnol Lett 17: 463468.
83 Singh G, Curtis WR (1995) Reactor design for
plant cell suspension cultures. In: Shargool
PD, Ngo TT (Eds.), Biotechnological applica-
tions of plant cultures. CRC Press, Boca Raton,
FL, pp 151183.
84 Tanaka H, Nishijima F, Suwa M, Iwamoto T
(1983) Rotating drum fermenter for plant-cell
suspension-cultures. Biotechnol Bioeng 25:
23592370.
85 Chang, HN, Sim SJ (1995) Extractive plant
cell culture. Curr Opin Biotechnol 6: 209212.
86 van Gulik WM, ten Hoopen HJG, Heijnen JJ
(2001) The application of continuous culture
for plant cell suspensions. Enzyme Microbial
Technol 28: 796805.
87 James E, Mills DR, Lee JM (2002) Increased
production and recovery of secreted foreign
proteins from plant cell cultures using an af-
finity chromatography bioreactor. Biochem
Eng J 12: 205213.
88 Liu F, Lee JM (1999) Effect of culture condi-
tions on monoclonal antibody production
from genetically modified tobacco suspension
cultures. Biotechnol Bioprocess Eng 4: 259
263.
89 Valdes R, Gomez L, Padilla S, Brito J, Reyes
B, Alvarez T, Mendoza O, Herrera O, Ferro W,
Pujol M, Leal V, Linares M, Hevia Y, Garcia C,
Mila L, Garcia O, Sanchez R, Acosta A, Geada
D, Paez R, Vega JL, Borroto C (2003) Large-
scale purification of an antibody directed
against hepatitis B surface antigen from trans-
genic tobacco plants. Biochem Biophys Res
Commun 308: 94100.
90 Bai Y, Glatz CE (2003) Capture of a recombi-
nant protein from unclarified canola extract
using streamline expanded bed anion ex-
change. Biotechnol Bioeng 81: 855864.
91 Choi JW, Cho GH, Byun SY, Kim DI (2001)
Integrated bioprocessing for plant cell cul-
tures. Adv Biochem Eng Biotechnol 72: 63
102.
92 Bai Y, Glatz CE (2003) Bioprocess considera-
tions for expanded-bed chromatography of
crude canola extract: sample preparation and
adsorbent reuse. Biotechnol Bioeng 81: 775
782.
93 Fahrner RL, Knudsen HL, Basey CD, Galan
W, Feuerhelm D, Vanderlaan M, Blank GS
(2001) Industrial purification of pharmaceuti-
cal antibodies: development, operation, and
validation of chromatography processes. Bio-
technol Genet Eng Rev 18: 301327.
94 FDA (2002) Guidance for industry. Drugs, bio-
logics, and medical devices derived from
bioengineered plants for use in humans and
animals. Food and Drug Administration.
95 CPMP (2002) Points to consider on quality as-
pects of medicinal products containing active
substances produced by stable transgene ex-
pression in higher plants (CPMP/BWP/764/
02), The European Agency for the Evaluation
of Medicinal Products (EMEA).
96 Knblein J, McCaman M (2003) Modern Bio-
pharmaceuticals Recombinant Protein Ex-
pression in Transgenic Plants. SCREENING
Trends in Drug Discovery 6: 3335.
97 Knblein J (2005) Plant-based Expression of
Biopharmaceuticals. In: Meyers RA (Ed.), En-
cyclopedia of Molecular Cell Biology and Molecu-
lar Medicine, 2nd ed., Vol. 10, Wiley & Sons,
pp 489410.
References 965
Abstract
Transgenic plant production technology,
though relatively a new arena in modern
plant science, has an immense potential to
change the shape of agriculture and offer
new opportunities in the field of modern
agrobiotechnology. The use of transgenic
plants in expressing recombinant foreign
proteins in crop plants demonstrates the
feasibility of using this approach for the
production of biopharmaceutical products.
However, several environmental conditions
such as salinity, drought, and extreme
temperatures are major limiting factors for
plant growth and crop productivity. High-
temperature stress is highly unfavorable in
optimal growth of plants, but nearly 25%
of total arable land is affected by heat and
drought stress. It is estimated that more
than one-third of all of the irrigated land
in the world is presently affected by sali-
nity, and this is exclusive of those regions
already classified as arid and desert lands
(which comprise 25% of the total land of
our planet). The problem of drought stress
is even more severe and economically da-
maging. Drought and salinity is predicted
to cause serious salinization of more than
50% of all arable land by the year 2050.
Moreover, the problem is that conventional
plant-breeding tools have been of only par-
tial help so far in alleviating these abiotic
stress problems. Recent microarray studies
have been employed to examine expres-
sion profiles of the whole genomes of
some plants in response to saline and
drought stress. The use of microarray tech-
niques has significantly accelerated efforts
in assigning the functional role of genes
involved in plant responses to saline and
drought stresses and shed light on possi-
ble involvement of regulatory pathways in
stress tolerance. This information is used
for planning new strategies to produce
abiotic, stress-tolerant, transgenic plants
and eventually to produce biopharmaceuti-
cals in areas which are today not farmable
at all.
Abbreviations
ABA abscisic acid
ABRE ABA responsive element
APX ascorbate peroxidase
BADH betaine aldehyde dehydrogenase
CaMV cauliflower mosaic virus
CAT catalases
CDH choline dehydrogenase
CE coupling element
CMO choline monooxygenase
COD choline oxidase
COX cytochrome oxidase
967
10
Producing Biopharmaceuticals in the Desert: Building
an Abiotic Stress Tolerance in Plants for Salt, Heat, and Drought
ISBN: 3-527-31184-X
DRE dehydration response element
GPX glutathione peroxidase
GSH glutathione sulfhydryl
GST glutathione S-transferase
HMW high-molecular weight
HSE heat shock element
HSF heat shock factor
Hsp heat shock protein
HSR heat shock response
HVA LEA family protein
KIRC potassium inward rectifying
channel
KORC potassium outward rectifying
channel
LEA late embryogenesis abundant
LMW low-molecular weight
MAS marker-assisted selection
NAD nicotinamide adenine dinucleo-
tide
P5CR pyrroline-5-carboxylate reductase
P5CS pyrroline-5-carboxylate synthase
PEAMT phosphoethanolamine N-
methyltransferase
ProDH proline dehydrogenase
PS II photosystem II
QTL quantitative trait loci
ROS reactive oxygen species
SOD superoxide dismutase
TCA transgene combining ability
VIC Voltage-independent cation
channels
10.1
General Comments on Abiotic Stresses
Plants have an inherent ability to adjust to
circadian and seasonal fluctuations in envi-
ronmental variables. These variables are, in
fact, advantageous to plants, and plants
have selected these as decisive factors in
controlling their various physiological attri-
butes such as timing of seed germination,
length of the period of vegetative growth,
onset of reproductive cycle, flowering inten-
sity, fruit set or whole plant senescence.
However, apart from the regular circadian
and seasonal variations, there may be cer-
tain irregular, rapid and unpredicted distur-
bances in the environment, and these may
result in stressful conditions. For example,
a paucity of water for long periods due to
lack of irrigation, infrequent rains or lower-
ing of the water table causes drought stress.
On the other hand, excess water through
rain, cyclones or frequent irrigation results
in flooding (also called water-logging, sub-
mergence or anaerobic stress). The cultiva-
tion of plants on saline soils or frequent ir-
rigation with ground water leads to a build-
up in excess salt levels, causing salinity
stress. Sudden atmospheric heating or cool-
ing due to transient changes in wind pat-
terns, cloud formation or excessive sunlight
results in temperature stress.
Most crop plants have been selected by
breeders for yield and yield-related param-
eters, and not for meeting exigencies
caused by different abiotic stress factors.
Therefore, crops have a limited ability to
adapt to extreme abiotic stresses.
However, abiotic stress such as drought,
salinity and extreme temperatures pose ser-
ious threats to agriculture and to natural
vegetation. The loss of farmable land due
to abiotic stresses is directly in conflict with
the needs of the worlds population, which
is projected to increase by 1.5 billion during
the next 20 years, and the challenge of
maintaining the world food supplies.
Although famine in the world nowadays
is originated by complex problems and not
only by an insufficient production of food,
there is no doubt that the gains in food pro-
duction provided by the Green Revolution
have reached their ceiling, while the world
population continues to rise. Therefore, in-
creasing the yield of crop plants in normal
soils and in less-productive lands includ-
10 Producing Biopharmaceuticals in the Desert 968
ing salinized and semi-arid lands and in re-
gions of extreme climates is an absolute
requirement for feeding the world.
10.2
Drought and Salt Tolerance
Agricultural productivity is severely af-
fected by water availability, and the dama-
ging effects of water stress have influenced
ancient and modern civilizations.
Water stress in its broadest context refers
to both, drought and salt stress, which affect
virtually almost every aspect of plant phy-
siology and metabolism. Drought due to
lack of rainfall or irrigation reduces the
amounts of water available for plant growth,
whereas the presence of high salt concen-
trations in saline habitats makes it more dif-
ficult for the plant roots to absorb water
from the environment. Because plant re-
sponses to salt and drought stress are close-
ly related, and the mechanisms for drought
and salt tolerance are partially overlapped,
both aspects will be reviewed together.
It is estimated that more than one-third
of all of the irrigated land in the world is
presently affected by salinity. This is exclu-
sive of the regions classified as arid and de-
sert lands (which comprise 25% of the total
land of our planet). The problem of drought
stress is even more severe and economically
damaging. Drought and salinity is predicted
Stress also plays an important role in deter-
mining how soil and climate limit the distri-
bution of plant species. Thus, an under-
standing of the physiological and biochem-
ical processes that underlie stress injury,
adaptation, and acclimation mechanisms
of plants to environmental stresses is of im-
mense importance for both the agriculture
and the environment.
The need to produce stress-tolerant
crops was evident even in ancient periods
[1]. However, efforts to improve crop per-
formance under environmental stresses
have not yet been very fruitful, mainly be-
cause the fundamental mechanisms of
stress tolerance in plants remain to be
completely understood. A genetic approach
to the development of specific stress-toler-
ant crop varieties requires as a pre-requi-
site the identification of key genetic deter-
minants of stress tolerance-related genes
or quantitative trait loci (QTL). The exis-
tence of salt-tolerant plants (halophytes)
and differences in salt tolerance between
genotypes within salt-sensitive plant (gly-
cophytes) species clearly indicates that
there is a genetic basis to salt response.
Two basic genetic approaches are cur-
rently being utilized to improve stress tol-
erance:
The exploitation of natural genetic varia-
tions, either through direct selection in
stressful environments or through the
mapping of QTLs and subsequent mar-
ker-assisted selection.
The generation of transgenic plants to
introduce novel genes or alter expression
levels of the existing genes to affect the
degree of stress tolerance.
Here, we discuss these approaches and fo-
cus on the recent experimentation with
transgenic plants that has led to increased
salinity and drought tolerance.
10.2.1
Genetics of Salt and Drought Tolerance
Tomato has been a valuable species to ana-
lyze the genetic basis of salinity and
drought tolerance, because making suc-
cessful crosses between wild and cultivated
tomato plants are relatively simple. Lyon
[2] made one of the first attempts to evalu-
ate the inheritance of salt tolerance. An in-
terspecific cross of Lycopersicon esculentum
and L. pimpinellifolium showed that the
fruit yield of the hybrid was more sensitive
to increasing salt (Na
2
SO
4
) than either par-
ent. Other crosses of wild and cultivated
tomato also suggested a complex genetics.
Stem elongation was a dominant trait in
hybrids with L. pennellii, but not with L.
cheesmanii as the parent. Total dry matter
production of another F
1
hybrid between
L. esculentum and L. pennellii showed hy-
brid vigor under saline conditions [3].
Analysis of other plant species has also
suggested that the genetics of salt toler-
ance is complex. In rice, sterility an im-
portant factor in yield under saline condi-
tions is determined by at least three
genes [4, 5]. In a genetic analysis, the ef-
fects of salinity on the seedling stage and
on sterility suggested both additive and
dominance effects, some with high herit-
ability [6, 7]. There is also evidence of
dominance in the salt tolerance of sor-
ghum. Diallele analysis, based on assess-
ing root tolerance to NaCl in salt-treated
as compared with control plants, showed
that there were both additive and domi-
nance effects of NaCl [8]. The above exam-
ples clearly indicate that salinity tolerance,
as in case of growth and yield, appear to
be a complex trait.
Early attempts to evaluate the genetic
basis of stress tolerance in plants were re-
stricted to simple genetic models. How-
ever, with the development of molecular
markers, evaluating the inheritance of sali-
nity tolerance became a more tractable
problem since specific QTLs could be
identified. Through the development of
molecular markers, it has become possible
not only to determine the genetic basis of
the trait, but also to map specific chromo-
somal segments or QTL and to determine
the relative contribution of each QTL to
the variance observed for the trait. Geno-
mic maps have been constructed in var-
ious crops to exploit genetic diversity [9],
tag qualitative and quantitative traits [10],
and analyze the stability of detected QTL
across different environments [11]. Stable
and consistent QTLs provide an excellent
opportunity to improve the efficiency of se-
lection, especially for traits controlled by
multiple genes and highly influenced by
the environment, as is the case for salinity
[12]. The efficiency of marker-assisted se-
lection (MAS) is dependent on a number
of factors such as the distance of observed
QTL from marker loci [12] and the propor-
tion of the total additive variance explained
by the QTL [13]. There is considerable evi-
dence to support the view that salt toler-
ance and its sub-traits are determined by
multiple QTLs. In an inter-generic cross of
tomato, several QTLs were found asso-
ciated with fruit yield in plants growing
under saline conditions [14], although
some of the QTL identified were later
shown to be dependent on the parentage
of the cross [15]. Some QTL associated
with specific aspects of fruit yield were
found regardless of whether the plants
were grown with or without salt; others
were detected only under saline or under
non-saline conditions [16]. Other crosses
have also identified both stress- (salt and
cold) specific and stress-non-specific QTLs.
The stress-non-specific QTL generally ex-
hibited larger individual effects and ac-
counted for a greater portion of the total
phenotypic variation under each condition
than the stress-specific QTL [17]. Recently,
Gur and Zamir evaluated the progress in
breeding for increased tomato (Solanum ly-
copersicum) yield under drought conditions
using genotypes carrying a pyramid of
three independent yield-promoting geno-
mic regions introduced from the drought-
tolerant, green-fruited wild species Sola-
num pennellii. Yield of hybrids was more
than 50% higher than that of a control
market leader variety under both wet and
dry field conditions that received 10% of
the irrigation water [18].
10.2.2
Physiological and Biochemical Strategies
used by Plants to Maintain Water Balance
Adaptation and acclimation to environ-
mental stresses result from integrated
events occurring at all levels of organiza-
tion, from the molecular and cellular level
to the physiological anatomical and mor-
phological level. At the biochemical level,
plants alter metabolism in various ways to
accommodate environmental stresses, in-
cluding producing osmoregulatory com-
pounds and altering ion concentrations in
the different cellular compartments.
10.2.2.1 Osmotic Adjustment
Osmotic adjustment is based on solute ac-
cumulation. Under drought conditions,
when the soil dries, plants can extract
water only as long as their water potential
is lower than that of the soil water. Osmo-
tic adjustment is one of the strategies that
plants use to maintain water balance un-
der drought or salt stress. Osmotic adjust-
ment is the process by which accumula-
tion of solutes by cells causes increase in
osmotic pressure and, as a result, this pre-
vents water loss and the accompanying re-
duction in cellular turgor. Most of the ad-
justments can be attributed to increase in
concentrations of various regular solutes,
including inorganic ions (especially K
+
)
and organic molecules such as, sugars, or-
ganic acids, and amino acids. Since cytoso-
lic enzymes can be severely inhibited by
high concentrations of ions, most of the
ions accumulated in the vacuoles during
osmotic regulation and are kept away from
cytosolic and other subcellular organelles.
However, in order to maintain water po-
tential equilibrium within the cell, the syn-
thesis and accumulation of specific organ-
ic solutes occurs. These solutes desig-
nated as compatible solutes are organic
compounds that do not interfere with en-
zyme functions.
Similarly to drought stress, salinity im-
poses a water-deficit that results from the
relatively high solute concentrations in the
soil, but in addition it may cause ion-spe-
cific stresses resulting from altered K
+
/Na
+
ratios, and also may lead to a build-up in
Na
+
and Cl
concentrations that are detri-

mental to plants. Plants respond to salinity
using two different types of responses.
Salt-sensitive plants restrict the uptake of
salt and adjust their osmotic potential by
the synthesis of compatible solutes (pro-
line, glycinebetaine, sugars, etc.) [19].
Salt-tolerant plants sequester and accu-
mulate salt into the cell vacuoles, control-
ling the salt concentrations in the cytosol
and maintaining a high cytosolic K
+
/Na
+
ratio in their cells [20]. Clearly, ion-exclu-
sion mechanisms could provide a degree
of tolerance to relatively low NaCl concen-
trations, but would not function at high
salt concentrations; this would result in
the inhibition of key metabolic processes,
with concomitant growth inhibition. Here,
we discuss three key processes that con-
tribute to salt tolerance at the cellular lev-
el: 1) the establishment of cellular ion
homeostasis; 2) the synthesis of compati-
ble solutes for osmotic adjustment; and 3)
the increased ability of the cells to neutra-
lize reactive oxygen species generated dur-
ing the stress response (Fig. 10.1).
Ion homeostasis This is a strategy for
adaptation to salinity stress and for the
creation of transgenic plants. Although
Na
+
is required in some plants (particular-
ly halophytes [20]), a high NaCl concentra-
tion is a toxic factor for plant growth. The
alteration of ion ratios in plants is due to
the influx of Na
+
through pathways that
function in the acquisition of K
+
[21]. The
sensitivity to salt of cytosolic enzymes is
similar in both glycophytes and halo-
phytes, indicating that the maintenance of
a high cytosolic K
+
/Na
+
concentration ratio
is a key requirement for plant growth in
high salt [20]. Strategies that plants could
use in order to maintain a high K
+
/Na
+
ra-
tio in the cytosol include: 1) extrusion of
Na
+
ions out of the cell; and 2) vacuolar
compartmentation of Na
+
ions. Under typ-
ical physiological conditions, plants main-
tain a high cytosolic K
+
/Na
+
ratio. Given
the negative membrane potential differ-
ence at the plasma membrane (140 mV)
[22], a rise in extracellular Na
+
concentra-
tion will establish a large electrochemical
Fig. 10.1 Overview of plant responses to salinity
and drought stresses. Stress signals activate sev-
eral molecular responses to re-establish cellular
homeostasis. The stress-responsive mechanisms
include, selective ion transport across membranes
of different cellular compartments, gene activation
by specific transcription factors, synthesis of com-
patible compounds (betaine, proline, etc.) and ac-
tivation of detoxification mechanisms against re-
active oxygen species (ROS). An important aspect
of salinity stress in plants is the stress-induced
production of ROS, including superoxide radicals
(O
2
), hydrogen peroxide (H
2
O
2
) and hydroxyl
radicals (OH).
gradient, favoring the passive transport of
Na
+
into the cells.
Three classes of low-affinity K
+
channels
have been identified. Inward rectifying
channels (KIRC), such as AKT1 [23], acti-
vate K
+
influx upon plasma-membrane hy-
perpolarization and they display a high
K
+
/Na
+
selectivity ratio. A knockout mu-
tant of AKT1 in Arabidopsis (akt1-1) dis-
played similar sensitivity to salt as the
wild-type, suggesting that this channel
does not play a role in Na
+
uptake [24]. K
+
outward rectifying channels (KORCs)
could play a role in mediating the influx
of Na
+
into plant cells. KORC channels
showed a high selectivity for K
+
over Na
+
in barley roots [25], and a somewhat lower
K
+
/Na
+
selectivity ratio in Arabidopsis root
cells [26]. These channels, which open dur-
ing the depolarization of the plasma mem-
brane (i.e., upon a shift in the electrical
potential difference to more positive val-
ues), could mediate the efflux of K
+
and
the influx of Na
+
ions [27]. Voltage-inde-
pendent cation channels (VIC) in plant
plasma membranes have been reported
[28, 29]. These channels have a relatively
high Na
+
/K
+
selectivity, are not gated by
voltage, and provide a pathway for the en-
try of Na
+
into plant cells [27].
Sodium ions can enter the cell through
a number of low- and high-affinity K
+
car-
riers, amongst which is AtHKT1from Ara-
bidopsis. This was shown to function as a
selective Na
+
transporter and, to a lesser
extent, to mediate K
+
transport [30].
AtHKT1 was identified as a regulator of
Na
+
influx in plant roots. This conclusion
was based on the capacity of hkt1 mutants
to suppress Na
+
accumulation and sodium
hypersensitivity in a sos3 (salt-overly-sensi-
tive) mutant background [31], suggesting
that AtHKT1 is a salt-tolerance determi-
nant that controls the entry of Na
+
into
the roots.
Na
+
extrusion from plant cells is powered
by the operation of the plasma membrane
H
+
-ATPase generating an electrochemical
H
+
gradient that allows plasma membrane
Na
+
/H
+
antiporters to couple the passive
movement of H
+
inside the cells, along its
electrochemical potential, to the active ex-
trusion of Na
+
[21]. Recently, AtSOS1 from
Arabidopsis thaliana has been shown to en-
code a plasma membrane Na
+
/H
+
antiport
with significant sequence similarity to plas-
ma membrane Na
+
/H
+
antiporters from
bacteria and fungi [32]. The overexpression
of SOS1 improved the salt tolerance of Ara-
bidopsis, demonstrating that improved salt
tolerance can be attained by limiting Na
+
ac-
cumulation in plant cells [33] (Table 10.1).
The compartmentation of Na
+
ions into va-
cuoles also provides an efficient mechanism
to avert the toxic effects of Na
+
in the cytosol.
The transport of Na
+
into the vacuoles is
mediated by a Na
+
/H
+
antiporter that is dri-
ven by the electrochemical gradient of pro-
tons generated by the vacuolar H
+
-translo-
cating enzymes, the H
+
-ATPase and the
H
+
-PPiase [44]. The overexpression of a
AtNHX1, a vacuolar Na
+
/H
+
antiporter from
Arabidopsis, in Arabidopsis resulted in trans-
genic plants that were able to grow in high
salt concentrations [34]. The paramount role
of Na
+
compartmentation in plant salt toler-
ance has been further demonstrated in
transgenic tomato plants overexpressing
AtNHX1 [35]. The transgenic tomato plants
grown in the presence of 200 mM NaCl
were able to grow, flower, and set fruit.
Although the leaves accumulated high so-
dium concentrations, the tomato fruits dis-
played very low amounts of sodium [35].
Similar results were obtained with trans-
genic Brassica napus (Canola) overexpres-
sing AtNHX1 [36]. Leaves of transgenic
plants grown in the presence of 200 mM
NaCl accumulated sodium to up to 6% of
their dry weight, but the seed yields and
oil quality were not affected, demonstrating
the potential use of this technology for agri-
cultural use in saline soils. Similar results
have been reported in other species. The in-
troduction of a vacuolar Na
+
/H
+
antiporter
from the halophyte Atriplex gmelini con-
ferred salt tolerance in rice [39]. Most re-
cently, the overexpression of GhNHX1 from
cotton into tobacco plants [38] and the over-
expression of AtNHX1 in maize [37] re-
sulted in enhanced salt tolerance.
Additional evidence supporting the role
of vacuolar transport in salt tolerance has
been provided by A. thaliana plants overex-
pressing a vacuolar H
+
-PPiase [40]. Trans-
genic plants overexpressing AVP1, coding
for the vacuolar H
+
-pyrophosphatase, dis-
played enhanced salt tolerance that was
correlated with the increased ion content
of the plants. These results suggest that
the enhanced vacuolar H
+
-pumping in the
transgenic plants provided additional driv-
ing force for vacuolar sodium accumula-
tion via the vacuolar Na
+
/H
+
antiporter.
Synthesis of compatible solutes and produc-
tion of stress-tolerant plants The cellular
response of salt- and drought-tolerant or-
ganisms to both long- and short-term sali-
nity stresses includes the synthesis and ac-
cumulation of a class of osmoprotective
compounds known as compatible solutes.
These relatively small, organic osmolytes
include amino acids and derivatives, poly-
ols and sugars, and methylamines. The os-
molytes stabilize proteins and cellular
structures, and can also increase the osmo-
tic pressure of the cell [45]. This response
Table 10.1 Salt tolerance in transgenic plants expressing genes involved in ion transporters
Gene Gene product Source Cellular role(s) Target plant Parameter
studied
Reference
AtNHX1 Vacuolar Na
+
/
H
+
antiporter
A. thaliana Na
+
vacuolar
sequestration
Arabidopsis Biomass 34
AtNHX1 Vacuolar Na
+
/
H
+
antiporter
A. thaliana Na
+
vacuolar
sequestration
Tomato Biomass,
and fruit
production
35
AtNHX1 Vacuolar Na
+
/
H
+
antiporter
A. thaliana Na
+
vacuolar
sequestration
B. napus Biomass, and
oil production
36
AtNHX1 Vacuolar Na
+
/
H
+
antiporter
A. thaliana Na
+
vacuolar
sequestration
Maize Biomass 37
GhNHX1 Vacuolar Na
+
/
H
+
antiporter
Gossypium
hirsutum
Na
+
vacuolar
sequestration
Tobacco Biomass 38
AgNHX1 Vacuolar Na
+
/
H
+
antiporter
Atriplex
gmelini
Na
+
vacuolar
sequestration
Rice Biomass 39
AtSOS1 Plasma mem-
brane Na
+
/H
+
antiporter
A. thaliana Na
+
extrusion Arabidopsis Biomass 33
AVP1 Vacuolar H
+
-
pyrophosphatase
A. thaliana Vacuolar
acidification
Arabidopsis Biomass 40
HAL1 K
+
/Na
+
trans-
port regulation
S. cerevisiae K
+
/Na
+
homeostasis
Tomato melon
Arabidopsis
Ion content,
plant growth
4143
is homeostatic for cell water status and
protein integrity, which is perturbed in the
face of soil solutions containing higher
amounts of NaCl and the consequent loss
of water from the cell. The accumulation
of osmotically active compounds in the cy-
tosol increases the osmotic pressure to
provide a balance between the apoplastic
(extracellular) solution, which itself be-
comes more concentrated with Na
+
and
Cl
ions, and the vacuolar lumen, which

in halophytes can accumulate up to 1 M
Na
+
(and Cl
). For a short-term stress, this

may provide the cells with the ability to
prevent water loss. However, for continued
growth under salinity stress, an osmotic
gradient (towards the cytosol) must be
kept in order to maintain turgor, water up-
take and facilitate cell expansion.
The enhancement of proline and glycine-
betaine synthesis in target plants has re-
ceived much attention [46]. Two themes
have emerged from the results of these ef-
forts:
there are metabolic limitations on the
absolute levels of the target osmolyte
that can be accumulated; and
the degree to which the transformed
plants are able to tolerate salinity stress
is not necessarily correlative with the
amounts of the osmoprotectants attained.
The metabolic limitations on increasing
the concentration of a given osmoprotec-
tant are well illustrated with both proline
and glycinebetaine. Initial strategies aimed
at engineering higher concentrations of
proline began with the overexpression of
genes encoding the biosynthetic enzymes
pyrroline-5-carboxylate (P5C) synthase
(P5CS) and P5C reductase (P5CR) that cat-
alyze the two steps between the substrate
(glutamic acid) and the product (proline).
P5CS overexpression in tobacco dramati-
cally elevated free proline in transgenic to-
bacco [47] (Table 10.2). However, the regu-
lation of free proline does not appear to be
straightforward. Proline catabolism, via
proline dehydrogenase (ProDH), is up-
regulated by free proline, and there exists
a strong evidence that free proline inhibits
P5CS [64]. Further, a two-fold increase in
free proline was achieved in tobacco plants
transformed with a P5CS modified by site-
directed mutagenesis [65]. This modifica-
tion alleviated the feedback inhibition by
proline on the P5CS activity and resulted
in an improved germination and growth
of seedlings under salt stress. Free cellular
proline levels are also transcriptionally and
translationally controlled. P5CR promoter
analysis revealed that P5CR transcripts
have reduced translational initiation. A 92-
bp segment of the 5' untranslated region
(UTR) of P5CR was sufficient to provide
increased mRNA stability and translational
inhibition under salt stress to the GUS re-
porter gene that was ligated at the 3' end
to this small region [66]. These results
highlighted the complex regulation of
P5CR during stress, and emphasized the
importance of stability and translation of
P5CR mRNA during salt stress. An alter-
native approach to attain significant free
proline levels, where antisense cDNA
transformation was used to decrease
ProDH expression, was utilized [55]. Levels
of proline in the transgenic Arabidopsis
were twice (100 lg g
1
fresh weight) that
of control plants grown in the absence of
stress, and three times higher (600 lg g
1
fresh weight) than in control plants grown
under stress. The high levels of proline
were correlated with an improvement in
tolerance to salinity, albeit for a short dura-
tion exposure to 600 mM NaCl.
Considerably more experimentation has
been directed at the engineering of glyci-
nebetaine synthesis than for any other
compatible solute. Unlike proline, glycine-
betaine degradation is not significant in
plants [67], but the problems of metabolic
fluxes, compounded with the compartmen-
tation of the substrate and product pools,
has made the engineering of appreciable
levels of glycinebetaine problematic. In
plants that are naturally glycinebetaine ac-
cumulators (e.g., spinach and sugarbeet),
synthesis of this compound occurs in the
chloroplast, with two oxidation reactions
from choline to glycinebetaine. The first
oxidation to betaine aldehyde is catalyzed
by choline monooxygenase (CMO), an
iron-sulfur enzyme. Betaine aldehyde oxi-
dation to glycinebetaine is catalyzed by be-
taine aldehyde dehydrogenase (BADH), a
Table 10.2 Salt and drought tolerance in transgenic plants expressing genes involved in osmolyte biosynthesis
Gene Gene product Source Cellular
role(s)
Target plant Parameter
studied
Reference
betA Choline
dehydrogenase
E. coli Glycine-
betaine
Tobacco Dry weight 48, 49
BADH Betaine
dehydrogenase
Atriplex
hortensis
Glycine-
betaine
Tomato Root growth 50
EctA,
EctB,
EctC
L-2,4-diamino-
butyric acid
acetyltransferase,
L-2,4-diamino-
butyric acid trans-
aminase, L-ectoine
synthase
Halomonas
elongata
Ectoyne Tobacco Salinity
tolerance
51
OtsA
OtsB
Trehalose-6-P
synthase
Trehalose-6-P
phosphatase
E. coli Trehalose Tobacco
Rice
Increased
biomass,
morphogenesis,
growth
52, 53
TPS1 Trehalose-6-P
synthase
S. cerevisiae Trehalose Tobacco Improved
drought
tolerance
54
P5CS Dl
1
-Pyrroline-5-
carbo xylate synthase
V. aconitifolia Proline Tobacco Increased
proline; plant
growth
47
ProDH Proline dehydro-
genase
Arabidopsis
thaliana
Proline Arabidopsis Inflorescence
lodging in
response to
NaCl stress
55
IMT1 Myo-inositol-O-
methyl transferase
M. chrystalli-
num
D-Ononitol Tobacco Seed germi-
nation
56, 57
COD1;
COX
Choline oxidase A. globiformis;
A. panescens
Glyc-
inebetaine
Arabidopsis
Rice
Brassica
Seed germi-
nation; plant
growth
5862
HAL3 FMN-binding
protein
S. cerevisiae K
+
/Na
+
homeostasis
Arabidopsis Seedlings 63
non-specific soluble aldehyde dehydrogen-
ase [68]. In E. coli, these reactions are cyto-
solic; in this species, the first reaction is
catalyzed by the protein encoded by the
beta locus choline dehydrogenase (CDH),
which is an NAD
+
-dependent enzyme, and
BADH in E. coli, is encoded by the betB lo-
cus. In Arthrobacter globiformis, the two oxi-
dation steps are catalyzed by one enzyme
choline oxidase (COD), which is encoded
by the codA locus [69]. The codA gene of
A. globiformis offers an attractive alternative
to the engineering of glycinebetaine syn-
thesis as it necessitates only a single gene
transformation event. This strategy was
employed for engineering glycinebetaine
synthesis in Arabidopsis [70]. The 35S pro-
moter driven construct for transformation
included the transit peptide for the small
subunit of Rubisco so that the COD pro-
tein would be targeted to the chloroplast.
Improved salinity tolerance was obtained
in transgenic Arabidopsis that accumulated,
as a result of the transformation,
1 lmol g
1
fresh weight glycinebetaine.
The same construct was used by for trans-
formation of Brassica juncea [62] and toler-
ance to salinity during germination and
seedling establishment was improved
markedly in the transgenic lines. COX
from Arthrobacter panescens, which is ho-
mologous to the A. globiformis COD, was
used to transform Arabidopsis, Brassica na-
pus and tobacco [61]. This set of experi-
ments differs from those above in that the
COX protein was directed to the cytoplasm
and not to the chloroplast. Improvements
in tolerance to salinity and drought and
freezing were observed in some trans-
genics from all three species, but the toler-
ance was variable. The levels of glycinebe-
taine in the transgenic plants were not sig-
nificantly higher than those of wild-type
plants, but increased significantly with the
exogenous supply of choline to plants, sug-
gesting that the supply of choline is a sig-
nificant constraint on the synthesis of gly-
cinebetaine [61].
Two important issues emerge from the
results of the above discussion. The first is
that the concentrations of glycinebetaine
in the transgenic plants were much lower
than the concentrations noted in natural ac-
cumulators. Despite the fact that these lev-
els are not high enough to be osmotically
significant, a moderate (and significant) in-
crease in tolerance to salinity and other
stresses was conferred. This raises the pos-
sibility that the protection offered by glyci-
nebetaine is not only osmotic, which is a
point raised by several of the above groups.
[71]. Compatible solutes (including manni-
tol) may also function as scavengers of oxy-
gen radicals, which may be supported by
the results of Alia et al. [72], where the pro-
tection to photosystem II in plants express-
ing codA was observed. An alternative pos-
sibility, not necessarily exclusive of the first,
is that the increased level of peroxide gener-
ated by the COD/COX oxidation of choline
causes an up-regulation of ascorbate perox-
idase and catalase [49] which may also im-
prove the tolerance to salinity stress [46].
The second issue is that the level of glycine-
betaine production in the transgenics is
limited by choline. Because betaine synthe-
sis takes place in the chloroplast, the free
choline pool may not reflect its availability
to the chloroplast, which may be limited
in this compartment by the activity and/or
abundance of choline transporters. How-
ever, a dramatic increase in glycinebetaine
levels (to 580 lmol g
1
dry weight in Arabi-
dopsis) was shown in the transgenic plants
when they were supplemented with choline
in the growth medium [61]. This limitation
was not explored in the transgenic tobacco
expressing E. coli enzymes CDH and BADH
in the cytoplasm [49]. Although these trans-
genic plants demonstrated an improved tol-
erance to salinity, glycinebetaine levels were
on the order of those mentioned above. Sa-
kamoto and Murata [69] also asserted that
despite the similarities in tolerance exhib-
ited by transgenic plants engineered to
synthesize betaine in either the chloroplast
or cytoplasm, the site of synthesis of betaine
may play a role in the degree of tolerance
shown. Indeed, if the betaine present in
these plants is localized primarily in the
chloroplast, it may be present at significant
concentrations (50 mM) [58]. However, Sa-
kamoto and Murata [69] downplayed the
limitation of the metabolic pool of choline
on the levels of glycinebetaine obtained in
the engineered plants, by suggesting that
the choline-oxidizing activity may be the
limiting factor. This argument seems to be
supported by Huang et al. [61], who found
that the levels of glycinebetaine correlated
with the levels of COX activity measured
in each plant. The increase in glycinebe-
taine with exogenous choline argues against
this notion. Stronger evidence for the lim-
itations of choline metabolism have been
presented by McNeil et al. [73]. By over-ex-
pressing spinach phosphoethanolamine N-
methyltransferase (PEAMT), which cata-
lyzes the three methylation reactions re-
quired for the conversion of phosphoetha-
nolamine to phosphocholine, up to a 50-
fold increase in free choline was obtained.
This led to an increase in glycinebetaine lev-
els (+60%) in plants that were expressing
spinach CMO and BADH in the chloro-
plast. Further, the addition of ethanolamine
to the plant growth medium caused an in-
creased choline and glycinebetaine levels,
showing that the metabolic flux through
this pathway is also limited by the supply
of ethanolamine. As PEAMT is itself inhib-
ited by phosphocholine, further engineer-
ing efforts need to include:
the modification of PEAMT to remove
this inhibition [73];
increasing the supply of ethanolamine
by overexpression of serine decarboxy-
lase; and
resolving the compartmentation problem
of choline supply and choline oxidation,
either by use of choline oxidation in the
cytoplasm or by finding the appropriate
transporters to improve choline supply
to the chloroplast [46].
Finally, as the compatible solutes are non-
toxic, the interchangeability of these com-
pounds between species has held much in-
terest (see Table 10.1). The recent exam-
ples include the engineering of: 1) ectoine
synthesis with enzymes from the halophi-
lic bacterium Halomonas elongata in plants
[51,74]; and 2) trehalose synthesis which
occurs in bacteria, yeast, and in extremely
desiccation-tolerant plants [75] into potato
[76] and rice [53]. An intriguing report on
the improved tolerance to salinity in tobac-
co expressing yeast invertase in the apo-
plast highlights the potential of manipulat-
ing sucrose metabolism [77]. The overex-
pression of polyols, such as mannitol [78]
and d-ononitol [57], have been shown to
contribute to enhanced drought and salt
tolerance in transgenic tobacco plants.
10.2.2.2 Antioxidant Protection Improves
Salt Tolerance
An important aspect of salinity stress in
plants is the stress-induced production of
reactive oxygen species (ROS), including
superoxide radicals (O
2
), hydrogen perox-
ide (H
2
O
2
), and hydroxyl radicals (OH).
ROS are a product of altered chloroplast
and mitochondrial metabolism during
stress. These ROS cause oxidative damage
to different cellular components including
membrane lipids, proteins, and nucleic
acids [79]. The alleviation of this oxidative
damage could provide enhanced plant re-
sistance to salt stress. Plants use low mo-
lecular mass antioxidants such as ascorbic
acid and reduced glutathione (GSH) and
employ a diverse array of enzymes such as
superoxide dismutases (SOD), catalases
(CAT), ascorbate peroxidases (APX), GSH-
S-transferases (GST), and GSH peroxi-
dases (GPX) to scavenge ROS. Transgenic
rice overexpressing yeast mitochondrial
Mn-dependent SOD displayed enhanced
salt tolerance [80] (Table 10.3). The overex-
pression of a cell wall peroxidase in tobac-
co plants improved germination under os-
motic stress [82]. Transgenic tobacco plants
overexpressing both GST and GPX dis-
played improved seed germination and
seedling growth under stress [84]. Subse-
quent studies [83] demonstrated that, in
addition to increased GST/GPX activities,
the transgenic seedlings contained higher
levels of GSH and ascorbate than wild-type
seedlings, showed higher levels of mono-
dehydroascorbate reductase activity, and
the GSH pools were more oxidized. These
results would indicate that the increased
GSH-dependent peroxidase scavenging ac-
tivity and the associated changes in GSH
and ascorbate metabolism led to reduced
oxidative damage in the transgenic plants
and contributed to their increased salt tol-
erance. During salt stress, plants display
an increase in the generation of H
2
O
2
and
other ROS [83,85]. The major substrate for
the reductive detoxification of H
2
O
2
is as-
corbate, which must be continuously re-
generated from its oxidized form. A major
function of GSH in protection against oxi-
dative stress is the reduction of ascorbate
via the ascorbate-GSH cycle, where GSH
acts as a recycled intermediate in the re-
duction of H
2
O
2
[86]. Ruiz and Blumwald
[87] investigated the enzymatic pathways
leading to GSH synthesis during the re-
sponse to salt stress of wild-type and salt-
tolerant Brassica napus L. (Canola) plants
overexpressing a vacuolar Na
+
/H
+
antipor-
ter [36]. Wild-type plants showed a marked
increased in the activity of enzymes asso-
ciated with cysteine synthesis (the crucial
step for assimilation of reduced sulfur into
Table 10.3 Salt tolerance in transgenic plants expressing genes in-
volved in redox reactions
Gene Gene product Source Cellular role(s) Target plant Parameter
studied
Reference
MnSOD Superoxide
dismutase
S. cerevisiae Reduction of
O
2
content
Rice Photosynthetic
electron trans-
port
80
GlyI Glyoxylase B. juncea S-D-Lactoyl-
glutathione
Tobacco Chlorophyll
content of
detached leaves
81
TPX2 Peroxidase N. tabaccum Change cell
wall properties
Tobacco Germination;
water retention
in seed walls
82
GST
GPX
Glutathione
S-transferase
Glutathione
peroxidase
N. tabaccum
N. tabaccum
ROS scaven-
ging
Tobacco Germination
and growth
83
organic compounds such as GSH) result-
ing in a significant increase in GSH con-
tent. On the other hand, these activities
were unchanged in the transgenic salt-tol-
erant plants, and their GSH content did
not change with salt stress. These results
clearly showed that salt stress induced an
increase in the assimilation of sulfur, and
the biosynthesis of cysteine and GSH
aimed to mitigate the salt-induced oxida-
tive stress. The small changes seen in the
transgenic plants overexpressing the vac-
uolar Na
+
/H
+
[35] suggested that the accu-
mulation of excess Na
+
in the vacuoles
(and the maintenance of a high cytosolic
K
+
/Na
+
ratio) greatly diminished the salt-
induced oxidative stress, highlighting the
important role of Na
+
homeostasis during
salt stress.
10.2.2.3 Expression of Stress-responsive
Genes may Confer Improved
Stress Tolerance
The developments of stress-resistant trans-
genic plants are largely predicted on the
capability of the products of overexpressed
stress-responsive genes to protect other-
wise sensitive plants. However, it must be
noted that not all the identified stress-re-
sponsive genes have proven successful in
conferring stress resistance.
A large group of genes, the expression
of which is regulated by osmotic stress,
was first identified in seeds during ma-
turation and desiccation, and are known to
encode for so-called LEA proteins (named
for late embryogenesis abundant). Some of
these proteins are known also to increase
in vegetative tissues of plants exposed to
stresses related to water-deficit phenom-
ena. Although the exact defense mecha-
nism and their function is still unknown,
the overexpression of HVA1 an LEA III
family protein in rice and wheat has
been shown to increase tolerance to water-
deficit stresses [88, 89]. The LEA proteins
are hydrophilic, and their protective role
might be associated with their abilities to
retain water and to prevent the crystalliza-
tion of major proteins during desiccation
processes.
Recent investigations have focused on
the large-scale identification of genes asso-
ciated with root growth in corn, and re-
sults have demonstrated the spatial expres-
sion pattern of specific set of genes under
water-deficit conditions. These genes are
predicted to be involved in processes re-
lated to stress tolerance, and also in basic
cellular processes associated with cell ex-
pansion [90]. The products of the candi-
date genes related to stress can be classi-
fied into two groups: 1) those that can di-
rectly protect against stress; and 2) those
that are associated with signal transduc-
tion in the stress response.
Several of the stress-responsive genes re-
ported in the literature are regulated by ab-
scisic acid (ABA), a plant hormone known
to play a crucial role in plant responses to
water stress, and its levels increase under
these conditions. Studies of the promoters
of several stress-induced genes have led to
the identification of specific regulatory se-
quences for genes associated with water-
deficient stresses. Some of the promoters
contain a six-nucleotide sequence element
referred as the ABA responsive element
(ABRE) which binds transcriptional factors
involved in ABA-dependent gene activation
[91]. Additional promoter sequences, called
coupling elements (CEs), function with
the ABRE to control the expression of
ABA-regulated genes. CEs may provide ad-
ditional regulatory element involved in tis-
sue-specific and temporally regulated pat-
terns.
Not all of these water-stressed responsive
genes are induced by ABA. The cis-acting
promoter element, the dehydration re-
sponse element (DRE), plays an important
role in regulating gene expression in re-
sponse to drought, salt and freezing stres-
ses, but not by ABA. It has been shown
that the transcription factor DREB1A spe-
cifically interacts with the DRE and in-
duces expression of stress-tolerance genes.
Moreover, overexpression of the cDNA en-
coding DREB1A in transgenic Arabidopsis
and tobacco plants activated the expression
of many of these stress-tolerance genes
and resulted in improved tolerance to
drought, salt loading, and freezing [92, 93].
However, use of the strong constitutive
35S cauliflower mosaic virus (CaMV) pro-
moter for driving the expression of DRE-
B1A resulted also in a problem of severe
growth retardation under normal growing
conditions. In contrast, expression of DRE-
B1A under the control of the stress-induci-
ble rd29A promoter caused minimal inhi-
bitory effects on plant growth while pro-
viding an even greater tolerance to stress
conditions than did expression of the gene
from the CaMV promoter. These results
suggest an optimal combination of a gene
and promoter for the production of stress-
tolerant transgenic plants. Use of the
stress-inducible rd29A promoter rather
than the constitutive promoter for overex-
pressing the DREB1A is an excellent ex-
ample for improving abiotic stress toler-
ance of agriculturally important crops by
gene transfer [93].
Recent microarray studies have been used
to examine the expression profiles of whole
genomes of some plants in response to sa-
line and drought stress [9496]. The use of
microarray techniques has significantly ac-
celerated efforts in assigning the functional
role of genes involved in plant responses to
saline and drought stresses, and will shed
light on the possible involvement of regula-
tory pathways in stress tolerance. This infor-
mation will be used for planning new strat-
egies to produce abiotic stress-tolerant
transgenic plants.
10.3
High-temperature Stress
High-temperature stress is very unfavor-
able for the optimal growth of plants. Al-
most 25% of the total arable land world-
wide is affected by heat and drought
stress. The annual mean air temperature
of 23% of the Earths land surface is above
408C [97], and with the increasing concen-
tration of greenhouse gases the Earths sur-
face temperature is expected to increase
further by 1.25.88C by 2100 ad (http:/
www.epa.gov). This rise in ambient tem-
perature would clearly create a warmer cli-
mate in most parts of the world.
In the main, all stages of crop growth
are vulnerable to high-temperature stress.
Processes such as seed germination, seed-
ling emergence and vigor and survival of
the seedling/plants are adversely affected
by high-temperature stress during the veg-
etative phase. High-temperature stress
counteracts the developmental process by
causing damage to leaf photosynthesis,
protein metabolism (including enzyme ac-
tivities), membrane stability, respiration,
and transpiration. The reproductive phase
in crops is highly susceptible to supra-opti-
mal temperatures, and results in a dimin-
ished crop yield. Pollen viability is one of
the principal factors affected negatively by
high-temperature stress.
In general, cool-season crops are more
sensitive to heat stress than are warm-sea-
son annuals. In barley, a high temperature
leads to a reduction in tiller number, a
shortened duration of the tillering stage,
abortion of spikelets, and a decline in su-
crose synthase activity and starch deposi-
tion. The abortion of flowers is a conspicu-
ous effect of heat stress in brassicas. In
common bean, high-temperature stress
causes an increase in flower abscission, a
reduction in pollen viability, a shriveled ap-
pearance of pollen grains, decrease in num-
ber of pollen tubes penetrating stigma, and
a decline in fertility, poor pod-setting and
seed-setting and a high number of parthe-
nocarpic pods. In rice, the major processes
affected by heat stress include reduction in
pollen viability and floret fertility, a short-
ened duration of grain filling, a reduction
in starch content of grains, and increase
packing of starch within the endosperm
causing a chalky appearance of kernels.
10.3.1
Heat Shock Response
and Heat Shock Proteins
High-temperature stress is, in general,
simulated in laboratory experiments by
subjecting biological systems to heat shock
(HS) treatment. Plants mount resistance
to HS stress by eliciting specific metabolic
adjustments which, as a whole, are re-
ferred to as a heat shock response (HSR)
(Fig. 10.2). The temperature required to in-
duce the HSR varies amongst different
plant species, but an increase of 5108C
over and above the ambient temperature is
sufficient to elicit HSR. The molecular ba-
sis of HS response was revealed for the
first time when Ritossa [98] reported that
temperature elevation brings about altered
puffing pattern of polytene chromosomes
in Drosophila. Tissieres et al. [99] for the
first time showed that the HS condition
results in an altered protein profile in Dro-
sophila cells. Further research established
that almost all organisms ranging from
bacteria to human respond to HS by
synthesizing a new set of proteins called
heat shock proteins (Hsps). The first re-
ports on HS-induced alterations in the
protein profile of plants appeared during
the 1980s [100].
Hsps are broadly classified on the basis
of their molecular weights as high- molec-
ular weight (HMW, 70100 kDa) and low-
molecular weight (LMW, 1520 kDa) Hsp.
The level of accumulation of different
Hsps is often variable; for example, under
stress conditions the LMW-Hsps may com-
prise up to 1% of the cellular proteins.
During the past 25 years of research, Hsp
have been extensively analyzed for their
physiological, biochemical, cellular, and
molecular properties. Consequently, it has
been shown that Hsp are highly conserved
proteins that is, plant Hsps are similar
to Hsps reported from animal and micro-
bial cells. Detailed characterization of Hsp
in terms of their molecular weights, differ-
ent inducers that trigger Hsp synthesis,
cellular locations, expression characteris-
tics, synthesis under field conditions, cel-
lular levels, and conservation of amino
acid sequence have been discussed at
length elsewhere [101, 102].
Besides HS, selective plant Hsps are in-
duced in response to different abiotic
stresses such as heavy-metal stress, water
stress, wounding stress, salt stress, cold
shock, and anoxia stress. Plants, in gener-
al, survive lethal temperature stress more
efficiently after prior exposure to a mild
stress as against a direct response to lethal
stress. This phenomenon is termed ac-
quired thermotolerance [103]. Hsp are be-
lieved to be important for the protection of
cells against heat injury both in basal ther-
motolerance (i.e., thermotolerance shown
without prior heat shock) as well as in ac-
quired thermotolerance responses.
Over the years, a large number of genes
that encode Hsp (referred to as heat shock
genes or hsp) have been isolated, se-
quenced, and cloned. This has been
achieved from a large number of plant
species representing diverse taxonomic
classes. The availability of the complete
genomic sequence data of Arabidopsis has
provided vast information on different
families of the hsp [104,105]. As the geno-
mic sequence of rice is almost complete,
and programs to sequence the genomes of
several other plant species are in progress,
enormous amounts of information on
plant hsp can be expected in the future.
The induction of hsp genes is noted to be
mediated by specific cis-acting sequences
present in the promoters of heat shock
genes. These cis-acting sequences are spe-
cifically termed as heat shock elements
(HSEs). There is clear evidence to show
that HSEs interact with positively-acting
regulatory proteins called heat shock fac-
tors (HSFs) to bring about increased tran-
scription of heat shock genes.
Fig. 10.2 Synthesis and regulation of cytoplasmic
heat shock proteins. Plants mount resistance to
heat shock (HS) stress by eliciting specific meta-
bolic adjustments, together referred to as heat
shock response (HSR). The temperature for the
induction of the HSR varies amongst different
plant species, but an increase of 5108C over and
above the ambient temperature is sufficient to eli-
cit HSR. Heat shock proteins (Hsp) are synthe-
sized in the cytoplasm using hsp mRNA tran-
scribed in the nucleus. Most Hsp act as chaper-
ones and protect cellular proteins from the heat
shock (HS)-based denaturation process. Synthesis
of hsp mRNA from hsp genes is governed by heat
shock factor (HSF). HSF is present in the cyto-
plasm as inactive molecule (monomer state) that
is unable to bind to DNA. Under high-temperature
stress, HSF is activated (trimer state). Accumula-
tion of denatured proteins also leads to activation
of HSF. Active HSF binds to the promoter regions
of hsp genes and mediate hsp gene transcription.
The activity of HSF is down-regulated by Hsp.
10.3.2
Production of High-temperature-tolerant
Transgenic Plants
A detailed understanding of the molecular
mechanisms that underlie HSRs in plants
(including heat shock genes/proteins, pro-
moters, trans-acting factors and signaling
components) has played a vital role in the
production of high-temperature-tolerant
transgenic plants, as discussed and sum-
marized (Table 10.4) in the following sec-
tions.
10.3.2.1 Transgenics Made
for Altered Levels of Hsps
Cellular proteins reportedly lose their bio-
logical activity upon HS due to aggrega-
tion and/or misfolding. There is evidence
that the stress-induced accumulation of ag-
gregated and misfolded proteins proves de-
leterious to the cells, and that the abnor-
mal state of proteins triggers the HSR in
living organisms [106]. HS is also known
to enhance the synthesis of certain pro-
teases involved in the degradation of ab-
normal proteins. Hsp are reported to func-
tion as molecular chaperones that coop-
erate as a functional-network in protecting
cells against heat damage. Hsp16.9,
Hsp17.1, Hsp17.3, and Hsp18.1 are shown
to prevent the aggregation or denaturation
of proteins during heat shock [107110].
Hsp100 is shown to rescue the heat-in-
duced protein aggregates by their resolubi-
lization during the recovery phase [111].
Certain other Hsps such as Hsp40, Hsp60,
Hsp70, and Hsp90 (alone or in coopera-
tion) are known to stabilize the heat-dena-
tured proteins [111113].
The level of expression of different Hsps
has been genetically manipulated, the aim
being to achieve an enhanced thermotoler-
ance in plants. Malik et al. [114] produced
transgenic carrot lines and plants in which
carrot small Hsp17.7 was overexpressed
under the control of CaMV35S promoter
(high-strength, constitutive promoter
mainly used for expressing alien genes
mainly in dicotyledonous plants). Thermo-
tolerance was assessed in terms of cell via-
bility and growth, as well as electrolyte
leakage in plants after severe stress. The
modified expression of Hsp17.7 in this ex-
periment resulted in an enhanced survival
of transgenic cell lines and plants at high
temperature. Park and Hong [115] raised
transgenic tobacco plants overexpressing
tobacco small hsp. In these studies, seed-
lings transformed with sense construct
(that leads to synthesis of sense transcript)
showed a higher cotyledon opening rate
compared to seedlings transformed with
antisense construct (that leads to synthesis
of antisense transcript). Transgenic rice
plants overexpressing Oshsp17.7 gene have
shown an increased thermotolerance as
well as increased resistance to UV-B irra-
diation [116]. Tomato Lehsp (M) gene over-
expressed in tobacco showed that trans-
genics transformed with sense construct
were more thermotolerant at 488C than
those transformed with antisense con-
struct [117]. Sun et al. [118] overexpressed
Arabidopsis hsp17.6A gene in Arabidopsis.
As a matter of contrast, this study showed
that transformants are more tolerant to os-
motic stress, but not heat stress.
Apart from LMW-Hsp, there are selec-
tive instances where levels of HMW-Hsp
have also been manipulated. Queitsch et
al. [119] reported the production of trans-
genic Arabidopsis plants by modifying level
of Hsp100 protein. In this study, 14-day-
old transgenic plants were tested for their
performance at high temperature. Trans-
genic plants survived up to 458C tempera-
ture stress for 1 h as they showed vigorous
growth after the removal of stress. The
T
a
b
l
e
1
0
.
4
S
e
l
e
c
t
i
v
e
r
e
p
o
r
t
s
o
n
p
r
o
d
u
c
t
i
o
n
o
f
h
i
g
h
-
t
e
m
p
e
r
a
t
u
r
e
,
s
t
r
e
s
s
-
t
o
l
e
r
a
n
t
t
r
a
n
s
g
e
n
i
c
c
r
o
p
s
G
e
n
e
P
r
o
t
e
i
n
S
o
u
r
c
e
C
e
l
l
u
l
a
r
r
o
l
e
(
s
)
A
.
R
e
g
u
l
a
t
o
r
y
g
e
n
e
s
C
o
m
m
e
n
t
s
R
e
f
e
r
e
n
c
e
T
r
a
n
s
-
h
o
s
t
P
r
o
m
o
t
e
r
u
s
e
d
T
r
a
n
s
c
r
i
p
t
i
o
n
f
a
c
t
o
r
g
e
n
e
s
h
s
f
1
H
e
a
t
s
h
o
c
k
t
r
a
n
s
c
r
i
p
t
i
o
n
a
l
f
a
c
t
o
r
1
A
.
t
h
a
l
i
a
n
a
T
r
a
n
s
c
r
i
p
t
i
o
n
f
a
c
t
o
r
A
.
t
h
a
l
i
a
n
a
C
a
M
V
3
5
S
T
r
a
n
s
f
o
r
m
a
n
t
s
e
x
h
i
b
i
t
e
d
t
h
e
r
m
o
t
o
l
e
r
a
n
c
e
a
n
d
t
h
e
r
e
w
a
s
c
o
n
s
t
i
t
u
t
i
v
e
e
x
p
r
e
s
s
i
o
n
o
f
t
h
e
h
s
p
g
e
n
e
s
a
t
n
o
r
m
a
l
t
e
m
p
e
r
a
t
u
r
e
1
0
7
h
s
f
3
-
d
o
-
A
.
t
h
a
l
i
a
n
a
T
r
a
n
s
c
r
i
p
t
i
o
n
f
a
c
t
o
r
A
.
t
h
a
l
i
a
n
a
C
a
M
V
3
5
S
I
n
c
r
e
a
s
e
d
t
h
e
r
m
o
t
o
l
e
r
a
n
c
e
1
2
1
h
s
f
1
-
d
o
-
L
.
e
s
c
u
l
e
n
t
u
m
T
r
a
n
s
c
r
i
p
t
i
o
n
f
a
c
t
o
r
L
.
e
s
c
u
l
e
n
t
u
m
C
a
M
V
3
5
S
O
v
e
r
e
x
p
r
e
s
s
i
n
g
l
i
n
e
s
h
o
w
e
d
i
n
c
r
e
a
s
e
d
t
h
e
r
m
o
t
o
l
e
r
a
n
c
e
w
h
i
l
e
p
o
s
t
-
t
r
a
n
s
c
r
i
p
t
i
o
n
a
l
l
y
s
i
l
e
n
c
e
d
l
i
n
e
p
l
a
n
t
s
a
n
d
f
r
u
i
t
s
w
e
r
e
t
h
e
r
m
o
s
e
n
s
i
t
i
v
e
1
2
3
h
s
f
3
-
d
o
-
A
.
t
h
a
l
i
a
n
a
T
r
a
n
s
c
r
i
p
t
i
o
n
f
a
c
t
o
r
A
.
t
h
a
l
i
a
n
a
C
a
M
V
3
5
S
I
n
c
r
e
a
s
e
d
t
h
e
r
m
o
t
o
l
e
r
a
n
c
e
1
2
2
h
s
f
1
b
-
d
o
-
A
.
t
h
a
l
i
a
n
a
T
r
a
n
s
c
r
i
p
t
i
o
n
f
a
c
t
o
r
L
y
c
o
p
e
r
s
i
c
u
m
(
T
o
m
a
t
o
i
n
b
r
e
d
l
i
n
e
L
4
7
8
3
)
C
a
M
V
3
5
S
T
r
a
n
s
g
e
n
i
c
s
s
h
o
w
e
d
e
n
h
a
n
c
e
d
t
h
e
r
m
o
t
o
l
e
r
a
n
c
e
a
n
d
c
h
i
l
l
i
n
g
t
o
l
e
r
a
n
c
e
1
2
4
S
i
g
n
a
l
t
r
a
n
s
d
u
c
t
i
o
n
c
o
m
p
o
n
e
n
t
g
e
n
e
s
d
b
f
2
C
e
l
l
c
y
c
l
e
r
e
g
u
l
a
t
e
d
p
h
o
s
p
h
o
-
p
r
o
t
e
i
n
A
.
t
h
a
l
i
a
n
a
P
r
o
t
e
i
n
k
i
n
a
s
e
A
.
t
h
a
l
i
a
n
a
C
a
M
V
3
5
S
T
r
a
n
s
f
o
r
m
a
n
t
s
s
h
o
w
e
d
s
t
r
i
k
i
n
g
t
o
l
e
r
a
n
c
e
t
o
h
e
a
t
,
s
a
l
t
,
c
o
l
d
a
n
d
o
s
m
o
t
i
c
s
t
r
e
s
s
u
p
o
n
o
v
e
r
e
x
p
r
e
s
-
s
i
o
n
1
2
5
S
t
r
u
c
t
u
r
a
l
g
e
n
e
s
D
e
t
o
x
i
f
i
c
a
t
i
o
n
c
o
m
p
o
n
e
n
t
g
e
n
e
s
a
p
x
1
A
s
c
o
r
b
a
t
e
p
e
r
o
x
i
d
a
s
e
H
.
v
u
l
g
a
r
i
s
P
e
r
o
x
i
s
o
m
a
l
a
s
c
o
r
b
a
t
e
p
e
r
o
x
i
d
a
s
e
A
.
t
h
a
l
i
a
n
a
C
a
M
V
3
5
S
T
r
a
n
s
f
o
r
m
a
n
t
s
w
e
r
e
s
i
g
n
i
f
i
c
a
n
t
l
y
m
o
r
e
t
o
l
e
r
a
n
t
t
o
h
e
a
t
s
t
r
e
s
s
a
s
c
o
m
p
a
r
e
d
t
o
w
i
l
d
-
t
y
p
e
1
3
1
T
a
b
l
e
1
0
.
4
(
c
o
n
t
i
n
u
e
d
)
G
e
n
e
P
r
o
t
e
i
n
S
o
u
r
c
e
C
e
l
l
u
l
a
r
r
o
l
e
(
s
)
A
.
R
e
g
u
l
a
t
o
r
y
g
e
n
e
s
C
o
m
m
e
n
t
s
R
e
f
e
r
e
n
c
e
T
r
a
n
s
-
h
o
s
t
P
r
o
m
o
t
e
r
u
s
e
d
P
H
G
P
x
G
l
u
t
a
t
h
i
o
n
e
p
e
r
o
x
i
d
a
s
e
L
.
e
s
c
u
l
e
n
t
u
m
P
h
o
s
p
h
o
l
i
p
i
d
h
y
d
r
o
-
p
e
r
o
x
i
d
e
g
l
u
t
a
t
h
i
o
n
e
p
e
r
o
x
i
d
a
s
e
N
.
t
a
b
a
c
u
m
(
t
r
a
n
s
i
e
n
t
e
x
p
r
e
s
s
i
o
n
)
D
u
p
l
i
c
a
t
e
d
P
V
X
c
o
a
t
p
r
o
t
e
i
n
p
r
o
-
m
o
t
e
r
T
r
a
n
s
i
e
n
t
e
x
p
r
e
s
s
i
o
n
o
f
L
e
P
H
G
P
x
p
r
o
t
e
c
t
e
d
t
o
b
a
c
c
o
l
e
a
v
e
s
f
r
o
m
s
a
l
t
a
n
d
h
e
a
t
s
t
r
e
s
s
M
o
r
e
o
v
e
r
,
s
t
a
b
l
e
e
x
p
r
e
s
s
i
o
n
o
f
L
e
P
H
G
P
x
i
n
t
o
b
a
c
c
o
c
o
n
f
e
r
r
e
d
p
r
o
t
e
c
t
i
o
n
a
g
a
i
n
s
t
t
h
e
f
u
n
g
a
l
p
h
y
t
o
p
a
t
h
o
g
e
n
B
o
t
r
y
t
i
s
c
i
n
e
r
e
a
1
3
2
F
a
t
t
y
a
c
i
d
m
e
t
a
b
o
l
i
s
m
g
e
n
e
s
f
a
d
7
O
m
e
g
a
-
3
f
a
t
t
y
a
c
i
d
d
e
s
a
t
u
r
a
s
e
A
.
t
h
a
l
i
a
n
a
C
a
u
s
e
s
r
e
d
u
c
t
i
o
n
o
f
t
r
i
e
n
o
i
c
f
a
t
t
y
a
c
i
d
s
a
n
d
h
e
x
a
d
e
c
a
t
r
i
e
n
o
i
c
a
c
i
d
N
.
t
a
b
a
c
u
m
C
a
M
V
3
5
S
T
r
a
n
s
f
o
r
m
a
n
t
s
s
h
o
w
i
n
g
s
i
l
e
n
-
c
i
n
g
o
f
t
h
e
g
e
n
e
w
e
r
e
b
e
t
t
e
r
a
b
l
e
t
o
a
c
c
l
i
m
a
t
e
t
o
h
i
g
h
e
r
t
e
m
p
e
r
a
t
u
r
e
1
3
0
H
e
a
t
s
h
o
c
k
g
e
n
e
s
h
s
p
1
7
.
6
A
H
e
a
t
s
h
o
c
k
P
r
o
t
e
i
n
1
7
.
6
A
A
.
t
h
a
l
i
a
n
a
M
o
l
e
c
u
l
a
r
c
h
a
p
e
r
o
n
e
(
i
n
v
i
t
r
o
)
A
.
t
h
a
l
i
a
n
a
C
a
M
V
3
5
S
T
r
a
n
s
f
o
r
m
a
n
t
s
w
e
r
e
t
o
l
e
r
a
n
t
t
o
o
s
m
o
t
i
c
s
t
r
e
s
s
b
u
t
n
o
t
h
e
a
t
s
t
r
e
s
s
1
1
8
h
s
p
1
7
.
7
H
e
a
t
s
h
o
c
k
p
r
o
t
e
i
n
1
7
.
7
D
.
c
a
r
o
t
a
H
e
a
t
s
h
o
c
k
p
r
o
t
e
i
n
D
.
c
a
r
o
t
a
C
a
M
V
3
5
S
T
r
a
n
s
f
o
r
m
a
n
t
s
e
x
p
r
e
s
s
e
d
t
h
e
h
s
p
1
7
.
7
g
e
n
e
i
n
t
h
e
a
b
s
e
n
c
e
o
f
h
e
a
t
s
h
o
c
k
a
n
d
s
h
o
w
e
d
i
n
-
c
r
e
a
s
e
d
t
h
e
r
m
o
t
o
l
e
r
a
n
c
e
1
1
4
h
s
p
1
0
1
H
e
a
t
s
h
o
c
k
p
r
o
t
e
i
n
1
0
1
A
.
t
h
a
l
i
a
n
a
-
d
o
-
A
.
t
h
a
l
i
a
n
a
C
a
M
V
3
5
S
T
r
a
n
s
f
o
r
m
a
n
t
s
c
o
n
s
t
i
t
u
t
i
v
e
l
y
e
x
p
r
e
s
s
i
n
g
h
s
p
1
0
1
t
o
l
e
r
a
t
e
d
s
u
d
d
e
n
s
h
i
f
t
s
t
o
e
x
t
r
e
m
e
t
e
m
-
p
e
r
a
t
u
r
e
b
e
t
t
e
r
t
h
a
n
t
h
e
c
o
n
t
r
o
l
s
1
1
9
-
d
o
-
-
d
o
-
-
d
o
-
-
d
o
-
O
.
s
a
t
i
v
a
U
b
i
1
T
r
a
n
s
f
o
r
m
a
n
t
s
e
x
p
r
e
s
s
i
n
g
h
s
p
1
0
1
s
h
o
w
e
d
e
n
h
a
n
c
e
d
t
o
l
e
r
a
n
c
e
t
o
h
i
g
h
t
e
m
p
e
r
a
t
u
r
e
1
2
0
T
a
b
l
e
1
0
.
4
(
c
o
n
t
i
n
u
e
d
)
G
e
n
e
P
r
o
t
e
i
n
S
o
u
r
c
e
C
e
l
l
u
l
a
r
r
o
l
e
(
s
)
A
.
R
e
g
u
l
a
t
o
r
y
g
e
n
e
s
C
o
m
m
e
n
t
s
R
e
f
e
r
e
n
c
e
T
r
a
n
s
-
h
o
s
t
P
r
o
m
o
t
e
r
u
s
e
d
C
y
t
o
s
o
l
i
c
s
m
a
l
l
h
s
p
S
e
n
s
e
a
n
d
a
n
t
i
s
e
n
s
e
H
e
a
t
s
h
o
c
k
p
r
o
t
e
i
n
N
.
t
a
b
a
c
u
m
H
e
a
t
s
h
o
c
k
p
r
o
t
e
i
n
N
.
t
a
b
a
c
u
m
C
a
M
V
3
5
S
S
e
n
s
e
s
e
e
d
l
i
n
g
s
s
h
o
w
e
d
h
i
g
h
e
r
c
o
t
y
l
e
d
o
n
o
p
e
n
i
n
g
r
a
t
e
c
o
m
p
a
-
r
e
d
t
o
t
h
e
a
n
t
i
s
e
n
s
e
s
e
e
d
l
i
n
g
s
1
1
5
O
s
h
s
p
1
7
.
7
-
d
o
-
O
.
s
a
t
i
v
a
H
e
a
t
s
h
o
c
k
p
r
o
t
e
i
n
O
.
s
a
t
i
v
a
C
a
M
V
3
5
S
I
n
c
r
e
a
s
e
d
t
h
e
r
m
o
t
o
l
e
r
a
n
c
e
,
i
n
c
r
e
a
s
e
d
r
e
s
i
s
t
a
n
c
e
t
o
U
V
-
B
1
1
6
L
e
h
s
p
(
M
)
(
b
o
t
h
s
e
n
s
e
a
n
d
a
n
t
i
s
e
n
s
e
)
-
d
o
-
L
.
e
s
c
u
l
e
n
t
u
m
H
e
a
t
s
h
o
c
k
p
r
o
t
e
i
n
N
.
t
a
b
a
c
u
m
C
a
M
V
3
5
S
S
e
n
s
e
t
r
a
n
s
g
e
n
i
c
s
w
e
r
e
t
h
e
r
m
o
-
t
o
l
e
r
a
n
t
a
t
4
8
8
C
w
h
i
l
e
A
n
t
i
s
e
n
s
e
p
l
a
n
t
s
w
e
r
e
s
e
v
e
r
e
l
y
d
a
m
a
g
e
d
1
1
7
vector-transformed control plants could
not regain growth during the post-stress
recovery period. The critical role of
Hsp100 in providing thermotolerance was
thus shown in this experiment. Katiyar-
Agarwal et al. [120] successfully raised
transgenic basmati rice overexpressing
hsp100 cDNA. After exposure to different
levels of high-temperature stress, the sur-
vival rate of transgenic rice plants was
compared with that of untransformed con-
trol plants. Hsp100 overexpression was
seen to provide a distinct growth advan-
tage to transgenic rice during the post-
stress recovery period.
10.3.2.2 Transgenics Made for Altered
Levels of Regulatory Proteins
As mentioned above, the transcriptional
regulation of hsp is mediated by HSEs lo-
cated in the promoter region. HSFs com-
prise a complex and highly-conserved fami-
ly of proteins that bind to HSEs and coordi-
nate binding of the RNA polymerase and
other related factors so as to actively tran-
scribe hsp transcripts. The hsp gene induc-
tion system has emerged as a powerful tar-
get for manipulating levels of Hsps in trans-
genic experiments. Lee et al. [107] altered
the expression of hsf1 gene in Arabidopsis
thaliana. Transgenic Arabidopsis plants pro-
duced in this study were shown to express
Hsp even at normal temperatures. It was
further noted that transgenic plants could
tolerate temperatures up to 508C, while
wild-type plants were killed above 468C
stress. Prandl et al. [121] overexpressed Ara-
bidopsis hsf3 in Arabidopsis using a CaMV35
promoter. Transgenic plants in this experi-
ment showed a clearly enhanced thermoto-
lerance. In an independent study, hsf3 gene
overexpressed in Arabidopsis also resulted in
increased thermotolerance [122]. Mishra et
al. [123] overexpressed tomato hsfA1 gene
in tomato plants. In these studies, overex-
pressing lines showed an increased thermo-
tolerance, while transgenic lines in which
trans-gene was silenced were thermosensi-
tive. By contrast, Li et al. [124] overexpressed
Arabidopsis hsf1B in tomato and demon-
strated an enhanced thermotolerance and
chilling tolerance in transgenic plants.
Apart from the transcription factor genes,
attempts have also been made to alter signal
transduction component genes to manipu-
late levels of the Hsp. In one such study,
Lee et al. [125] raised transgenic Arabidopsis
plants that overexpressed Atdbf2 gene en-
coding for cell cycle-regulated phosphopro-
tein (that has a cellular role as protein kin-
ase). The transformants in these studies
demonstrated a striking tolerance to heat,
salt, cold, and osmotic stress.
Levels of Osmolytes
Manipulating levels of Hsps through hsp
genes or through regulatory genes that
change the expression levels of Hsps (as
mentioned in the above two categories) are
not the only means by which thermotoler-
ance appears to be governed. Indeed, sev-
eral other target sites have been identified
which, upon manipulation in transgenic
experiments, are important in governing
the thermotolerance property. As men-
tioned for salinity and for drought toler-
ance, there are certain low molecular-
weight compounds such as amino acids
(e.g., proline), polyamines (e.g., putres-
cine), quaternary ammonium compounds
(e.g., glycinebetaines), sugars (e.g., manni-
tol) and sugar alcohols (e.g., polyols) that
help plants to acclimatize against osmotic
stresses. Glycinebetaine protects the photo-
synthetic machinery by stabilizing the oxy-
gen-evolving photosystem II (PS II) com-
plex [126]. As mentioned earlier, metabolic
engineering for the biosynthesis of glyci-
nebetaine in Arabidopsis has been achieved
by introducing the bacterial codA gene
(that encodes choline oxidase protein).
When the effect of high-temperature stress
on these transgenics was examined at the
imbibition and germination stage of seeds,
the seeds of transgenic plants were seen to
be more tolerant to heat stress than the
wild-type [127]. The overproduction of gly-
cinebetaine also provided a significant ad-
vantage to the growth of young transgenic
seedlings at supra-optimal temperatures.
Levels of Membrane Lipids
Living cells adapt to changes in extracellu-
lar temperature by altering the composi-
tion of their membrane lipids. Many years
ago, it was established that membrane
fluidity increases due to increased unsa-
turation of membrane lipids in response
to low temperature [128]. Conversely, the
saturation of membrane lipids is noted to
increase when cells are subjected to supra-
optimal temperatures, thereby increasing
membrane rigidity [129]. Murakami et al.
[130] produced transgenic tobacco plants
in which the gene encoding the chloro-
plast-localized fatty acid desaturase was si-
lenced. Significant reduction in the
amount of trienoic fatty acids in homozy-
gous transgenic lines in comparison to
wild-type plants was observed. Thermoto-
lerance assays revealed that transgenic to-
bacco plants were resistant to high-tem-
perature stress (418C for 2 h), whereas
wild-type plants could not survive such ex-
treme temperature treatment. This report
showed that lowering the content of unsa-
turated fatty acid reduces the sensitivity of
plants to heat stress.
Levels of Cell Detoxification
Proteins
Detoxification pathways play a major role
when plants are subjected to different
abiotic stresses. The components of cell
detoxification mechanisms have been used
in specific experiments to alter thermoto-
lerance responses in transgenic plants.
Overexpression of barley hvapx1 gene (en-
coding for peroxisomal ascorbate peroxi-
dase) in Arabidopsis caused an increased
thermotolerance in transgenic plants as
compared to wild-type plants [131]. In a re-
cent study, tomato gene encoding for glu-
tathione peroxidase was overexpressed in
tobacco; subsequent transient expression
of the transgene protected transgenic
leaves from salt and heat stress [132].
10.4
Conclusions and Perspectives
The use of a variety of transgenic plants
for the production of biopharmaceuticals
is feasible. However, several environmental
conditions such as salinity, drought, and
extreme temperatures are major limiting
factors for plant growth and crop produc-
tivity. High-temperature stress is highly
unfavorable for the optimal growth of
plants, yet almost 25% of the total arable
land worldwide is affected by heat and
drought stress. Moreover, it is estimated
that more than one-third of all irrigated
land worldwide is presently affected by
salinity. The problem of drought stress is
even more severe and economically dama-
ging, the main difficulty being that
drought and salinity together is predicted
Conventional breeding programs for
raising abiotic stress-tolerant genotypes
have met with limited success. In evaluat-
ing the possibility of improving stress tol-
erance in plants, a number of considera-
tions should be considered. First, whilst it
has been recognized by many researchers
that dramatic changes in gene expression
are associated with all types of stresses,
the promoters that are most commonly
used for transgene introductions are pri-
marily constitutively expressed, including
the CaMV35S promoter, ubiquitin, and ac-
tin promoters [133]. Recent studies have
noted that the overexpression of specific
stress-induced genes under the control of
stress-induced or tissue-specific promoters
often display a better phenotype [134, 135].
Second, whilst there have been a number
of successes in the production of abiotic
stress-tolerant plants using tobacco or Ara-
bidopsis, there is a clear need to begin to
introduce these tolerance genes into crop
plants. Third, it is likely that the effective-
ness of a specific transgene will be based
on the specific genetic background into
which it is transformed. One component
of this is the well-known phenomenon of
position effect, although in addition the
ability of a transgene to function may well
be determined by the overall genetic back-
ground, independently of the chromosom-
al location of the transgene, referred to as
Transgene Combining Ability (TCA).
Finally, we also need to establish better
comparative systems. At the same time,
we need to look at rational concepts for
combining genes, just as the researchers
of disease resistance are currently doing
with gene stacking. For example, the over-
expression of AtSOS1 in meristems (non-
vacuolated cells) and AtNHX1 (for vac-
uolar Na
+
accumulation), together with the
overproduction of compatible solutes,
would not only provide the ability of using
NaCl as an osmoticum during vegetative
growth but also would provide the seed-
lings with the ability to reduce Na
+
toxicity
during early growth and seedling estab-
lishment. Wherever applicable, genes for
protection against oxidative stress must be
combined, particularly in actively photo-
synthesizing cells that are prone to more
chloroplast damage due to ROS.
While progress in improving stress toler-
ance has been slow in the past, there are a
number of opportunities and reasons for
optimism. Over the past ten years, there
has been the development of a number of
the functional tools that can allow us to
dissect many of the fundamental questions
associated with stress tolerance. These in-
clude: 1) the development of molecular
markers for gene mapping and the con-
struction of associated maps; 2) the devel-
opment of EST libraries; 3) the complete
sequencing of plant genomes including
Arabidopsis, rice, and maize; 4) the produc-
tion of T-DNA or transposon-tagged muta-
genic populations of Arabidopsis; and 5)
the development of a number of forward
genetics tools that can be used in gene
function analysis such as TILLING [136].
And indeed, of late, transgenic plants have
been raised that in fact show increased re-
sistance to abiotic stresses such as heat,
drought, and salinity. For example, trans-
genic Arabidopsis plants expressing Hsp
could tolerate temperatures up to 508C,
while wild-type plants were killed above
468C stress, and transgenic, drought-toler-
ant, tomatoes showed 50% higher yields
under dry field conditions with only 10%
of the irrigation water. Also, transgenic
plants overexpressing a Na
+
/H
+
antiporter
could be obtained with remarkable salt tol-
erance. Transgenic tomato plants for exam-
ple grown in the presence of 200 mM
NaCl were able to grow, flower, and set
fruit. Similar results could be obtained
with transgenic canola leaves which grew
in the presence of 200 mM NaCl and accu-
mulated sodium up to 6% (!) of their dry
weight. However, the seed yields and oil
quality were not affected an impressive
demonstration of the application of such
technology for agricultural use in saline
soils.
Clearly, we need to counteract against a
changing environment and climate: in-
creases in ambient temperature, drought
and salinity. Therefore, we must focus on
examining the comparative effects and in-
teraction of specific transgenes within a
defined genetic background, combine the
improved genetic elements, and determine
the efficacy of these approaches in the
field.
In the future, we will be able to produce
even high-value traits for example, bio-
pharmaceuticals in areas of the world
which, today, are not farmable at all. More-
over, as recommended by Knblein, this is
what we should focus on because then, at
the dawn of this new millennium, we would
for the first time be capable of producing
sufficient amounts of biopharmaceuticals
to treat everybody on our planet [137].
References
1 T. Jacobsen, R. M. Adams, Science 1958, 128,
12511258.
2 C. Lyon, Bot. Gazette 1941, 103, 107122.
3 Y. Saranga, D. Zamir, A. Marani, J. Rudich,
J. Am. Soc. Hort. Sci. 1991, 116, 10671071.
4 M. Akbar, T. Yabuno, S. Nakao, Jap. J. Breed-
ing 1972, 22, 277284.
5 M. Akbar, T. Yabuno, Jap. J. Breeding 1977, 27,
237240.
6 S. Moeljopawiro, H. Ikehashi, Euphytica 1981,
30, 291300.
7 M. Akbar, G. S. Khush, R. Hille, D. Lambers,
IRRI 1986, 399409.
8 F. M. Azhar, T. McNeilly, Plant Breeding 1988,
101, 114121.
9 C. E. Thormann, M. E. Ferreira, L. E. A. Camar-
go, J. G. Tivang, T. C. Osborn, Theor. Appl.
Genet. 1994, 88:973980.
10 D. V. Butruille, R. P. Guries, T. C. Osborn, Ge-
netics 1999, 153, 949964.
11 S. Hittalmani, H. E. Shahidhar, P. G. Bagali,
N. Huang, J. S. Sidhu, V. P. Singh, G. S.
Khush, Euphytica 2002, 125, 207214.
12 J. W. Dudley, Crop Sci. 1993, 33, 660668.
13 R. Lande, R. Thompson, Genetics 1990, 124,
743756.
14 M. P. Breto, M. J. Asins, E. A. Carbonell, Theor.
Appl. Genet. 1994, 88, 395401.
15 J. Monforte, M. J. Asins, E. A. Carbonell, Theor.
Appl. Genet. 1997, 95, 284293.
16 J. Monforte, M. J. Asins, E. A. Carbonell, Theor.
Appl. Genet. 1997, 95, 706713.
17 M. R. Foolad, G. Y. Lin, F. Q. Chen, Plant
Breeding 1999, 118, 167173.
18 A. Gur, D. Zamir, PLoS Biol. 2004, 2(10):
e245, 00010006.
19 H. Greenway, R. Munns, Annu. Rev. Plant
Physiol. 1980, 31, 149190.
20 E. P. Glenn, J. J. Brown, E. Blumwald, Crit.
Rev. Plant Sci. 1999, 18, 227255.
21 E. Blumwald, G. S. Aharon, M. P. Apse, Bio-
chim. Biophys. Acta 2000, 1465, 140151.
22 N. Higinbotham, Annu. Rev. Plant Physiol.
1973, 24, 2546.
23 H. Sentenac, N. Bonneaud, M. Minet, F. La-
croute, J. M. Salmon, F. Gaymard,
C. Grignon, Science 1992, 256, 663665.
24 E. P. Spalding, R. E. Hirsch, D. R. Lewis,
Q. Zhi, M. R. Sussman, B. D. Lewis, J. Gen.
Physiol. 1999, 113, 909918.
25 L. H. Wegner, K. Raschke, Plant Physiol. 1994,
105, 799813.
26 F. J. M. Maathuis, D. Sanders, Planta 1995,
197, 456464.
27 F. J. M. Maathuis, A. Amtmann, Ann. Bot.
1999, 84, 123133.
28 S. K. Roberts, M. Tester, J. Exp. Bot. 1997, 48,
839846.
29 H. de Boer, L. H. Wegner, J. Exp. Bot. 1997,
48, 441449.
30 N. Uozumi, E. J. Kim, F. Rubio, T. Yamaguchi,
D. Muto, A. Tsuboi, E. P. Bakker, T. Naka-
mura, J. I. Scroeder, Plant Physiol. 2000, 122,
12491259.
31 A. Rus, S. Yokoi, A. Sharkhuu, M. Reddy,
B. H. Lee, T. K. Matsumoto, H. Koiwa, J. K.
Zhu, R. A. Bressan, P. M. Hasegawa, Proc.
Natl. Acad. Sci. USA 2001, 98, 1415014155.
32 H. Shi, M. Ishitani, C. Kim, J. K. Zhu, Proc.
Natl. Acad. Sci. USA 2000, 97, 68966901.
References 991
33 H. Shi, B. H. Lee, S. J. Wu, J. K. Zhu, Nature
Biotechnol. 2003, 21, 8185.
34 M. P. Apse, G. S. Aharon, W. A. Snedden, E.
Blumwald, Science 1999, 285, 12561258.
35 H. X. Zhang, E. Blumwald, Nature Biotechnol.
2001, 19, 765768.
36 H. X. Zhang, J. N. Hodson, J. P. Williams, E.
Blumwald, Proc. Natl. Acad. Sci. USA 2001,
98, 68966901.
37 Z. Y. Yin, A. F. Yang, K. W. Zhang, J. R. Zhang,
Acta Bot. Sin. 2004, 46, 854861.
38 A. Wu, G. D. Yang, Q. W. Meng, C. C. Zheng,
Plant Cell Physiol. 2004, 45, 600607.
39 M. Ohta, Y. Hayashi, A. Nakashima, A. Hama-
da, A. Tanaka, T. Nakamura, T. Hayakawa,
FEBS Lett. 2002, 532, 279282.
40 R. A. Gaxiola, J. Li, S. Unurraga, L. M. Dang,
G. J. Allen, S. L. Alper, G. R. Fink, Proc. Natl.
Acad. Sci. USA 2001, 98, 1144411449.
41 M. Bordas, C. Montesinos, M. Dabauza, A.
Salvador, L. A. Roig, R. Serrano, V. Moreno,
Trans. Res. 1997, 6, 4150.
42 C. Gisbert, A. M. Rus, M. C. Bolarin, J. M. Lo-
pez-Coronado, I. Arrillaga, C. Montesinos, M.
Caro, R. Serrano, V. Moreno, Plant Physiol.
2000, 123, 393402.
43 S. X. Yang, Y. X. Zhao, Q. Zhang, Y. K. He,
H. Zhang, H. Luo, Cell Res. 2001, 11, 142148.
44 E. Blumwald, Physiol. Plant. 1987, 69, 731
734.
45 P. H. Yancey, M. E. Clark, S. C. Hand, R. D.
Bowlus, G. N. Somero, Science 1982, 217,
12141222.
46 D. Rontein, G. Basset, A. D. Hanson, Metabolic
Eng. 2002, 4, 4956.
47 P. B. K. Kishor, Z. Hong, G. H. Miao, C. A. A.
Hu, D. P. S. Verma, Plant Physiol. 1995, 108,
13871394.
48 G. Lilius, N. Holmberg, L. Bulow, Biotechnol-
ogy 1996, 14, 177180.
49 K. O. Holmstrom, S. Somersalo, A. Mandal, T.
E. Palva, B. Welin, J. Exp. Bot. 2000, 51(343),
177185.
50 X. Jia, Z. Q. Zhu, F. Q. Chang, Y. X. Li, Plant
Cell Rep. 2002, 21, 141146.
51 H. Nakayama, K. Yoshida, H. Ono, Y. Muroo-
ka, A. Shinmyo, Plant Physiol. 2000, 122,
12391247.
52 E. A. H. Pilon-Smits, N. Terry, T. Sears, H.
Kim, A. Zayed, S. B. Hwang, K. Van Dun, E.
Voogd, T. C. Verwoerd, R. W. H. Krutwagen,
O. J. M. Goddijn, J. Plant Physiol. 1998, 152,
525532.
53 K. Garg, J. K. Kim, T. G. Owens, A. P. Ranwala,
Y. Do Choi, L. V. Kochian, R. J. Wu, Proc. Natl.
Acad. Sci. USA 2002, 99, 1589815903.
54 C. Romero, J. M. Belles, J. L. Vaya, R. Serrano,
F. A. Culianez-Macia, Planta 1997, 201, 293
297.
55 T. Nanjo, M. Kobayashi, Y. Yoshiba, Y. Kaku-
bari. K. Yamaguchi-Shinozaki, K. Shinozaki,
FEBS Lett. 1999, 461, 205210.
56 D. M. Vernon, M. C. Tarczynski, R. G. Jensen,
H. J. Bohnert, Plant J. 1993, 4, 199205.
57 E. Sheveleva, W. Chmara, H. J. Bohnert, R. G.
Jensen, Plant Physiol. 1997, 115(3), 12111219.
58 H. Hayashi, L. Alia Mustardy, P. Deshnium,
M. Ida, N. Murata, Plant J. 1997, 12(1), 133
142.
59 Alia, H. Hayashi, A. Sakamoto, N. Murata,
Plant J. 1998, 16(2), 155161.
60 A. Sakamoto, Alia, N. Murata, Plant Mol. Biol.
1998, 38(6), 10111019.
61 J. Huang, R. Hirji, L. Adam, K. L. Rozwadows-
ki, J. K. Hammerlindl, W. A. Keller, G. Selvar-
aj, Plant Physiol. 2000, 122(3), 747756.
62 V. S. K. Prasad, P. Sharmila, P. A. Kumar, P. P.
Saradhi, Mol. Breeding 2000, 6, 489499.
63 A. Espinosa-Ruiz, J. M. Belles, R. Serrano,
F. A. Culiaez-Macia, Plant J. 1999, 20, 529
539.
64 N. H. Roosens, R. Willem, Y. Li, I. I. Verbrug-
gen, M. Biesemans, M. Jacobs, Plant Physiol.
1999, 121, 12811290.
65 Z. L. Hong, K. Lakkineni, Z. M. Zhang, D. P.
S. Verma, Plant Physiol. 2000, 122, 11291136.
66 X. J. Hua, B. Van De Cotte, M. Van Montagu,
N. Verbruggen, Plant J. 2001, 26, 157169.
67 L. Nuccio, S. D. McNeil, M. J. Ziemak, A. D.
Hanson, R. K. Jain, G. Selvaraj, Metabol. Eng.
2000, 2, 300311.
68 B. Rathinasabapathi, Ann. Bot. 2000, 86, 709
716.
69 A. Sakamoto, N. Murata, J. Exp. Bot. 2000, 51,
8188.
70 H. Hayashi, Alia, L. Mustardy, P. Deshnium,
M. Ida, N. Murata, Plant J. 1997, 12(1), 133
142.
71 H. J. Bohnert, B. Shen, Sci. Hort. 1999, 78,
237260.
72 Alia, Y. Kondo, A. Sakamoto, H. Nonaka,
H. Hayashi, P. P. Saradhi, T. H. H. Chen, N.
Murata, Plant Mol. Biol. 1999, 40, 279288.
73 S. D. McNeil, M. L. Nuccio, M. J. Ziemak, A. D.
Hanson. Proc. Natl. Acad. Sci. USA 2001, 98,
1000110005.
74 H. Ono, K. Sawads, N. Khunajakr, T. Tao,
M. Yamamoto, M. Hiramoto, A. Shinmyo, M.
Takano, Y. Murooka, J. Bacteriol. 1999, 181,
9199.
75 O. J. M. Goddijn, K. Van Dun, Trends Plant Sci.
1999, 4, 315319.
76 E. T. Yeo, H. B. Kwon, S. E. Han, J. T. Lee, J. C.
Ryu, M. O. Byu, Molecules and Cells 2000, 10,
263268.
77 E. Fukushima, Y. Arata, T. Endo, U. Sonne-
wald, F. Sato, Plant Cell Physiol. 2001, 42,
245249.
78 M. Tarczynski, R. Jensen, H. Bohnert, Science
1993, 259, 508510.
79 B. Halliwell, J. M. C. Gutteridge, Arch. Bio-
chem. Biophys. 1986, 246, 501514.
80 Y. Tanaka, T. Hibin, Y. Hayashi, A. Tanaka, S.
Kishitani, T. Takabe, S. Yokota, T. Takabe,
Plant Sci. 1999, 148, 131138.
81 J. Veena, V. S. Reddy, S. K. Sopory, Plant J.
1999, 17, 385395.
82 I. Amaya, M. A. Botella, M. De La Calle, M. I.
Medina, A. Heredia, R. A. Bressan, P. M. Ha-
segawa, M. A. Quesada, V. Valpuesta, FEBS
Lett. 1999, 457, 8084.
83 V. P. Roxas, S. A. Lodhi, D. K. Garrett, J. R. Ma-
han, R. D. Allen, Plant Cell Rep. 2000, 41,
12291234.
84 V. P. Roxas, R. K. Smith, E. R. Allen, R. D. Al-
len, Nature Biotechnol. 1997, 15, 988991.
85 Y. Gueta-Dahan, Z. Yaniv, B. A. Zilinskas, G.
Ben-Hayyim, Citrus. Planta 1997, 203, 460469.
86 C. Foyer, B. Halliwell, Planta 1976, 133, 2125.
87 M. Ruiz, E. Blumwald, Planta 2002, 214, 965
969.
88 D. Xu, X. Duan, B. Wang, B. Hong, T. Ho, R.
Wu, Plant Physiol. 1996, 110(1), 249257.
89 E. Sivamani, A. Bahieldin, J. M. Wraithc, T.
Al-Niemia, W. E. Dyera, T. H. D. Hod, R. Qu,
Plant Sci. 2000, 155, 19.
90 M. Bassani, P. M. Neumann, S. Gepstein,
Plant Mol. Biol. 2004, 56(3), 367380.
91 M. J. Guiltinan, W. R. Marcotte, R. S. Quatrano,
Science 1990, 250, 267271.
92 M. Kasuga, Q. Liu, S. Miura, K. Yamaguchi-
Shinozaki, K. Shinozaki, Nature Biotechnol.
1999, 17, 287291.
93 M. Kasuga, S. Miura, K. Shinozaki, K. Yama-
guchi-Shinozaki, Plant Cell Physiol. 2004,
45(3), 346350.
94 S. Kawasaki, C. Borchert, M. Deyholos,
H. Wang, S. Brazille, K. Kawai, D. Galbraith,
H. J. Bohnert, Plant Cell 2001, 13, 889905.
95 M. Seki, J. Ishida, M. Narusaka, M. Fujita,
T. Nanjo, T. Umezawa, A. Kamiya, M. Naka-
jima, A. Enju, T. Sakurai, M. Satou, K.
Akiyama, K. Yamaguchi-Shinozaki, P. Car-
ninci,
J. Kawai, Y. Hayashizaki, K. Shinozaki,
Funct. Integr. Genomics 2002, 2(6), 282291.
96 B. Sottosanto, A. Gelli, E. Blumwald, Plant
J. 2004, 40(5), 75271.
97 Y. Klueva, E. Maestri, N. Marmiroli, H. T.
Nguyen, In: A. S. Basra (Ed.), Crop Responses
and Adaptations to Temperature Stress. Food
Products Press, Binghamton, NY, 2001,
pp. 177217.
98 M. Ritossa, Experientia 1962, 18, 571573.
99 A. Tissieres, H. K. Mitchell, U. M. Tracey, J.
Mol. Biol. 1974, 84, 389398.
100 T. M. Barnett, C. Altohuler, N. McDaniel,
J. P. Mascarenhas, Dev. Genet. 1980, 1, 331
340.
101 M. Agarwal, N. Sarkar, A. Grover. J. Plant
Biol. 2003, 30(2), 141149.
102 W. Wang, B. Vinocur, O. Shoseyov, A. Alt-
man, Trends Plant Sci. 2004, 9(5), 244252.
103 S. L. Singla, A. Pareek, A. Grover, In:
M. N. V. Prasad (Ed.), Plant Ecophysiology.
John Wiley and Sons, Inc., New York, 1997,
pp 101127.
104 M. Agarwal, S. Katiyar-Agarwal, C. Sahi, D.
R. Gallie, A. Grover, Cell Stress Chaperone
2001, 6(3), 219224.
105 A. Grover, Cell Stress Chaperone 2002, 7(1),
15.
106 J. Ananthan, A. L. Goldberg, R. Voellmy,
Science 1986, 232, 522524.
107 J. H. Lee, A. Hubel, F. Schoffl. Plant J.
1995, 8(4), 603612.
108 C. H. Yeh, P. F. L. Chang, K. W. Yeh, W. C.
Lin, Y. M. Chen, C. Y. Lin, Proc. Natl. Acad.
Sci. USA 1997, 94, 1096710972.
109 L. S. Young, C. H. Yeh, Y. M. Chen, C. Y. Lin,
Biochem. J. 1999, 344, 3138.
110 D. Low, K. Brandle, L. Nover, C. Forreiter,
Planta 2000, 211, 575582.
111 J. R. Glover, S. Lindquist, Cell 1998, 94, 73
82.
112 F. U. Hartl, R. Hlodan, T. Langer, Trends Bio-
chem. Sci. 1994, 19, 2025.
113 J. Buchner, Trends Biochem. Sci. 1999, 24,
136141.
114 K. Malik, J. P. Slovin, C. H. Hwang, J. L.
Zimmerman. Plant J. 1999, 20, 8999.
References 993
115 S. M. Park, C. B. Hong, J. Plant Physiol.
2002, 159, 2530.
116 T. Murakami, S. Matsuba, H. Funatsuki,
K. Kawaguchi, H. Saruyama, M. Tanida,
Y. Sato, Molecular Breeding 2004, 13 (2), 165
175.
117 K. Sanmiya, K. Suzuki, Y. Egawa , M. Sho-
no, FEBS Lett. 2004, 557, 265268.
118 W. Sun, C. Bernard, B. van de Cotte, M. Van
Montagu, N. Verbruggen, Plant J. 2001, 27
(5), 407415.
119 C. Queitsch, S. W. Hong, E. Vierling,
S. Lindquist, Plant Cell 2000, 12(4), 479492.
120 S. Katiyar-Agarwal, M. Agarwal, A. Grover,
Plant Mol. Biol. 2003, 51(5), 677686.
121 R. Prndl, K. Hinderhofer, G. Eggers-Schu-
macher, F. Schffl, Mol. Gen. Genet. 1998,
258, 269278.
122 I. Panchuk, R. A. Volkov, F. Schoffl, Plant
Physiol. 2002, 129(2), 838853.
123 S. K. Mishra, J. Tripp, S. Winkelhaus, B.
Tschiersch, K. Theres, L. Nover, K. D. Scharf,
Genes Dev. 2002, 16, 15551567.
124 Y. Li, C. S. Chang, L. S. Lu, C. A. Liu, M. T.
Chan, Y. Y. Chang, Bot. Bull. Acad. Sin.
2003, 44, 129140.
125 H. Lee, M. V. Montagu, N. Verbruggen, Proc.
Natl. Acad. Sci. USA, 1999, 96, 58735877.
126 G. C. Papageorgiou, N. Murata, Photosynth
Res. 1995, 44, 243252.
127 Alia, H. Hayashi, A. Sakamoto, N. Murata,
Plant J. 1998, 16, 155161.
128 N. Murata, Plant Cell Physiol. 1983, 24, 81
86.
129 G. Thomas, P. J. Dominy, L. Vigh, A. R.
Mansourian, P. J. Quinn, W. P. Williams,
Biochim. Biophys. Acta 1986, 849, 131140.
130 Y. Murakami, M. Tsuyama, Y. Kobayashi, H.
Kodama, K. Iba, Science 2000, 287(5452),
4769.
131 W. M. Shi, Y. Muramoto, A. Ueda, T. Takabe,
Gene, 2001, 273, 2327.
132 S. Chen, Z. Vaghchhipawala, W. Li, H.
Asard, M. B. Dickman, Plant Physiol. 2004,
135(3), 16301641.
133 A. Grover, P. K. Aggarwal, A. Kapoor, S. Ka-
tiyar-Agarwal, M. Agarwal, A. Chandramou-
li, Curr. Sci. 2003, 84, 355367.
134 B. Zhu, J. Su, M. Chang, D. P. S. Verma,
Y.-L. Fan, R. Wu, Plant Sci. 1998, 139, 4148.
135 M. Kasuga, Q. Liu, S. Miura, K. Yamaguchi-
Shinozaki, K. Shinozaki, Nature Biotechnol.
1999, 17(3), 287291.
136 T. Colbert, B. J. Till, R. Tompa, S. Reynolds,
M. N. Steine, A. T. Yeung, C. M. McCallum,
L. Comai, S. Henikoff, Plant Physiol. 2001,
126, 480484.
137 J. Knblein, Biopharmaceuticals expressed
in plants a new era in the new Millen-
nium. In: R. Mller, O. Kayser (Eds.), Appli-
cations in Pharmaceutical Biotechnology.
Wiley-VCH, 2004.
Abstract
Antithrombin (AT) concentrates derived
from pooled human plasma have been
used for the management of hereditary
and acquired AT deficiencies since the
early 1980s. The development of a recom-
binant version of AT would alleviate supply
and safety concerns associated with the
use of the plasma-derived biotherapeutic.
However, the complex structure of the AT
molecule, and the large doses usually re-
quired in supplementation treatments,
have precluded the use of traditional bacte-
rial and cell culture bioreactors for com-
mercial production. GTC Biotherapeutics
has applied the transgenic animal expres-
sion system to the production of recombi-
nant AT (tradename ATryn
). A herd of
transgenic dairy goats expressing high lev-
els of human AT in their milk was charac-
terized and expanded, providing a homo-
geneous, well-defined and abundant sup-
ply of AT. This chapter describes the clini-
cal development of ATryn
(including
eight clinical studies), and the production
of this modern biopharmaceutical in trans-
genic goats.
Abbreviations
ACT activated clotting time
ARDS adult respiratory distress syn-
drome
AT antithrombin
BSE bovine spongiform encephalo-
pathy
CD circular dichroism
CJD CreutzfeldJakob disease
CPB cardiopulmonary bypass
DIC disseminated intravascular coag-
ulation
DVT deep-vein thrombosis
ELISA enzyme linked immunosorbent
assay
FFP fresh-frozen plasma
FISH fluorescence in-situ hybridiza-
tion
HD hereditary-deficient
HSPG heparan sulfate proteoglycans
HUVEC human umbilical vein endothe-
lial cells
IL interleukin
LC/MS liquid chromatography/mass
spectrophotometry
nvCJD new-variant Creutzfeld-Jakob
disease
995
11
The First Biopharmaceutical from Transgenic Animals: ATryn

ISBN: 3-527-31184-X
rhAT recombinant human anti-
thrombin
TAT thrombinantihrombin (com-
plex)
TSE transmissible spongiform ence-
phalopathy
11.1
Introduction
In the fall of 1992, a goat kid was born at
Tufts University School of Veterinary Medi-
cine: a normal healthy male goat (Fig. 11.1)
with added genetic material in his chromo-
somes. It was a DNA construct that would
imbue his progeny with a unique ability to
produce a recombinant human protein, an-
tithrombin, in their milk. The offspring of
this genetically engineered goat are living
bioreactors that produce a recombinant pro-
tein much more efficiently than traditional
cell culture bioreactors. This is the story of
how the small biotech company that engi-
neered that animal proceeded down the
road to testing the product in the clinic. It
is the story of this transgenic products de-
velopment path towards market.
The serpin antithrombin (AT), also
known as antithrombin III or AT-III, is
the main regulator of many serine pro-
teases generated during activation of the
clotting cascade (see Part II, Chapters 1
and 3). AT is a 58-kDa single-chain glyco-
protein composed of 432 amino acids with
four oligosaccharide chains, and with a
plasma concentration of approximately
150 lg mL
1
(23 lM). It is a key factor in
the prevention of clotting. In the presence
of heparin, it is not only a strong inhibitor
996
Fig. 11.1 (A) Aerial view of the GTC Biotherapeutics farm in
Charlton (MA. USA) with rhAT-producing transgenic goats.
(B) 155-92, the transgenic founder. (C) rhAT transgenic does,
descendants of 155-92.
of thrombin and Factor Xa, but it is an
equally effective inhibitor of Factor IX, as
well as, to a lesser extent, Factors XIa, XIIa
(and its fragment), trypsin, plasmin, and
kallikrein [16] (see Part III, Chapter 6).
AT also weakly neutralizes Factor VIIa [7,
8]. AT inhibits thrombin by forming a 1: 1
stoichiometric complex between the two
components via a reactive site (arginine)-
active center (serine) interaction. Heparin,
the first-line anticoagulant in cardiovascu-
lar medicine, works as an anticoagulant
because its binding to AT induces an AT
configuration change, making it more
than 1000-fold more active towards throm-
bin [9]. Once the equimolar thrombinan-
tithrombin (TAT) complex is formed, both
molecules are incapacitated. The TAT com-
plex is then removed by hepatic receptor
identified as the LDL receptor-related pro-
tein [10]. The half-life of native AT is much
longer (55 h) than that of the TAT complex
(<5 min) [11].
Hereditary deficiency of AT is associated
with an increased risk of venous throm-
boembolism, often beginning in adoles-
cence and frequently accompanied by pul-
monary emboli [1215]. It is a rare disease,
with a prevalence of 1: 2000 to 1: 5000 [13]
and is characterized by a decreased AT ac-
tivity of 2560% of normal. The pattern of
inheritance is autosomal dominant. The
null mutation, in mice, is a fetal lethal [16]
(see Part III, Chapter 4). The majority of
patients with a hereditary deficiency disor-
der experience at least one thrombotic epi-
sode between the ages of 10 and 35 years,
usually triggered in high-risk situations
such as surgery, birth delivery, pregnancy,
trauma, or severe infection [17]. Currently,
plasma-derived human AT concentrates are
being used prophylactically for the preven-
tion of thromboembolic complications in
these high-risk situations in patients with
hereditary AT deficiency. The development
of rhAT will provide a stable supply of an-
tithrombin and reduce the perceived viral
risks associated with the current plasma-
derived products.
Acquired deficiencies of AT are much
more frequently encountered, either due
to increased consumption [sepsis, dissemi-
nated intravascular coagulation (DIC), sur-
gery, extensive thrombosis], increased loss
(burns, trauma, nephrotic syndrome), drug
treatment (heparin treatment, estrogen
use, L-asparaginase treatment), or de-
creased production (prematurity, liver dis-
eases) [1820]. In several of these acquired
conditions, the decrease in AT levels is as-
sociated with a poor outcome, especially
for septic patients [18, 2025]. The value of
AT supplementation with concentrates in
the case of acquired deficiencies is contro-
versial. Numerous small trials have shown
a favorable impact of AT replacement ther-
apy on the severity of sepsis-induced DIC
[2630]. However, the small size of these
trials did not permit a determination of
whether AT treatment had a significant
impact on mortality. The only large clinical
trial that examined the effect of AT in the
severe sepsis patient was the KyberSept
trial that used plasma-derived product.
This was an international, multi-center,
placebo-controlled, double-blind, random-
ized clinical trial involving over 2300 pa-
tients [31, 32]. The KyberSept trial did not
meet its primary end-point of a significant
reduction in 28-day, all-cause mortality, the
mortality rate being equivalent in the
treated and placebo groups. These results
were particularly disappointing, in that
treatment with antithrombin had been
convincingly shown to be of survival ad-
vantage in several Phase II trials [28, 29],
and small non-randomized clinical studies
[33, 34]. Further review of the KyberSept
trial data revealed a strong treatment inter-
action between antithrombin and heparin.
When the subgroup of patients treated
with antithrombin without heparin was
compared with placebo patients that did
not receive heparin, the 90-day mortality
rate was significantly reduced (p <0.05) in
the antithrombin group [31, 35]. Further
large trials will be required to demonstrate
the therapeutic value of AT supplementa-
tion in sepsis as well as for other acquired
deficiency indications.
11.2
Recombinant Production of AT
11.2.1
The Milk Expression Technology
This protein expression technology is
based on producing recombinant proteins
in the milk of transgenic animals. It be-
gins with linkage of the promoter region
of a milk-specific gene to a gene of inter-
est. This DNA construct is then introduced
into the chromosomes of a mammalian
embryo to be transferred to surrogate
mothers. Animals carrying this transgenic
are then capable of producing the gene
product in their milk (Fig. 11.2). During
lactation, the recombinant protein is
synthesized in the mammary gland and
secreted into the milk. The product can
then be purified from the milk to pharma-
ceutical-grade purity [3645].
One advantage of the milk-expression
system is that the constructs generated are
expressed both in the mouse model sys-
tem and in the large dairy animal produc-
tion system at similar levels. This allows
optimization of the expression transgene
in the mouse system, before embarking
on the large-animal portion of the project.
The predominant protein in goat milk is
beta casein (1020 mg mL
1
); this is
thought to comprise 2550% of the total
protein (~30 mg mL
1
). The beta casein
998
Fig. 11.2 Representation of the production pro-
cess for rhAT. The hAT cDNA was linked to capr-
ine mammary gland-specific regulatory elements.
This transgene was microinjected into goat em-
bryos that were then transferred to the oviducts of
recipients and carried to term. Offspring were
tested for the presence of the transgene. Female
and male transgenic founders were induced to
lactate to evaluate hAT expression in milk. The se-
lected founder, 155-92, was mated to non-trans-
genic females to generate the rhAT production
herd.
gene is regulated in both a tissue-specific
and temporal fashion, with maximal mam-
mary gland expression occurring post par-
turition. The goat beta casein gene was
cloned as an 18.5 kilobase pairs (kb) frag-
ment in a lambda EMBL3 vector from a
Saanen goat genomic library and charac-
terized in transgenic mice [46]. The coding
sequence of the goat beta casein was re-
placed with an XhoI cloning site immedi-
ately upstream of the initiating ATG co-
don. The 3' downstream region begins at
the end of exon 7 and continues through
the 3' untranslated region through 6.2 kb
of downstream genomic sequence. This
type of construct has now been used to ex-
press nearly 100 different proteins in the
milk of mice. It also has been now used
successfully to secrete high levels (g L
1
) of
14 of these proteins in the milk of trans-
genic goats. It has a unique ability to ex-
press cDNA constructs at gram levels,
which has been a challenge for other milk
promoters.
11.2.2
Generation of the AT Founder
11.2.2.1 The AT Transgene
The human AT cDNA was obtained from
Dr. G. Zettlemeissl (Behringwerke A.G.,
Marburg, Germany) in plasmid pbAT6.
The sequence of the cDNA is the same as
that published by Bock et al., 1982 [47],
with the exception of the silent nucleotide
changes at bp 1050 (T?C), 1317 (C?T)
and bp 1371 (A?G). The cDNA was engi-
neered with an XhoI site at the 5 end by
site-directed mutagenesis to allow excision
of the cDNA as a 1.45 kb XhoI to Sa1I
fragment. The transgene (pBC6; Fig. 11.3)
was then assembled by linking the AT
cDNA to the XhoI site of the goat beta
casein vector. The resulting 14.8 kb trans-
gene was excised from bacterial sequences
by digesting pBC6 with NotI and SalI, re-
striction sites which flank the eukaryotic
sequences. The DNA fragment was puri-
fied by agarose gel electrophoresis. The re-
sulting fragment was then used to gener-
ate transgenic animals.
Transgenic mice were used to test ex-
pression of the AT transgene. The genera-
tion of transgenic mice has become rou-
Fig. 11.3 Schematic representation of the BC6
transgene for recombinant expression of rhAT in
the milk of transgenic animals. The cDNA encod-
ing hAT (striped box) replaces the coding region
of the caprine beta-casein gene. This cDNA is
flanked by the promoter and the by 3' untrans-
lated caprine beta-casein sequences. Black boxes
indicate the non-coding exons of the casein gene.
R, restriction enzyme EcoRI; N, restriction enzyme
NotI; S, restriction enzyme SalI.
tine in academic and commercial laborato-
ries (see Part III, Chapter 4). However, in
the early 1990s it could still be a challenge
to generate transgenic mice. Microinjec-
tion of pronuclear preimplantation em-
bryos was used to introduce the heterolo-
gous DNA (BC6 transgene) into the ge-
nome. Of more than 200 founder animals
born from microinjected embryos, only
one transgenic mouse was identified. Fe-
males belonging to the line derived from
this founder consistently produced nearly
1 mg mL
1
of recombinant human AT
(rhAT) in their milk, as measured by Wes-
tern analysis. With this success, it was
decided to generate transgenic goats with
the same construct.
11.2.2.2 Pronuclear Microinjection
Genzyme Corporation (Genzyme Trans-
genics Corporation, the progenitor of GTC
Biotherapeutics was spun off by Genzyme
in 1993) had established a research colla-
boration with Dr. Karl Ebert at Tufts Uni-
versity School of Veterinary Medicine. Dr.
Ebert had carried out the microinjections
of the mice to test many of the early milk
vector constructs. Dr. Ebert also led the
team that adapted the pronuclear microin-
jection technique to development of trans-
genic goats [48]. The AT microinjections
were started early in 1991. In a process
that is very similar to that used to generate
transgenic mice, Dr. Ebert microinjected
goat embryos. Female goats were supero-
vulated and then mated. One-cell, pronuc-
lear stage embryos were collected surgi-
cally from these does and submitted to the
microinjection procedure (see Part I,
Chapter 11). (This technique is impres-
sively shown on the supplement CD-ROM
in a microscopic video kindly provided by
Professor Hwang.) A few picoliters of a
DNA solution containing the purified
transgene fragment was injected into one
or both pronuclei of the embryo using mi-
cromanipulators and a micropipette. The
microinjected embryos were then trans-
ferred into the oviduct of suitably prepared
surrogate female recipients that carried
the progeny to term.
As the microinjected DNA randomly in-
tegrates into the embryonic genome at a
low rate, through an unknown mecha-
nism, it is believed that the integration
event can take place at anytime as the em-
bryo divides. This may give rise to mosa-
icism in which not all of the cells of the
resulting transgenic animal will carry the
transgene. In addition, the transgene may
integrate in several locations in the genome.
The transgenic male that was identified as
the founder for the AT program had these
characteristics; he was mosaic with multiple
chromosomal integration sites.
The microinjection process was carried
out on 139 caprine embryos, which were
transferred into 57 surrogate mothers. Of
the 70 progeny born, five were identified
as being transgenic for the beta casein
hAT construct (Table 11.1), BC6. Two foun-
der females (223-92, 225-92) and three
males (155-92, 222-92, and 224-92) were
identified. A later series of injections with
the genomic version of hAT gave rise to
one founder female (227-92). The proce-
dure to identify transgenic animals re-
quired both ear tissue and blood cells to
be analyzed. DNA was isolated from each
source and a polymerase chain reaction
(PCR) specific for the hAT gene was used
to determine which animals carried the
BC6 construct. In addition to PCR, South-
ern blot analyses were also performed to
confirm the integrity of the coding se-
quence and the region of the beta casein
vector flanking the hAT gene.
Hormones were used to induce peri- or
pre-pubertal goats to lactate. At the age of
1000
5 months, does could produce up to a few
liters of milk following induction. Analysis
of the milk for expression of rhAT allowed
prediction of whether the transgene would
be expressed during natural lactation,
which would not occur for another 5
months following successful breeding. The
223-92 doe was induced to lactate and pro-
duced 0.5 mg mL
1
of rhAT in its milk,
whereas the other female, 225-92, pro-
duced nothing. It is also possible to induce
approximately 50% of males to lactate by
using the same hormonal regime. Only
young male 155-92 produced a few millili-
ters of milk, and this contained 1
2 mg mL
1
of rhAT, as determined by Wes-
tern blot analysis. This male served as the
genetic founder for the hAT transgenic
production herd. Milk obtained from fe-
male transgenic goats, the offspring de-
rived from 155-92, is the source material
for purification of rhAT.
11.2.3
Establishment of the hAT Transgenic Herd
One of the challenges to breeding a trans-
genic production herd is that the expres-
sion level of the founder needs to be as-
sured before committing significant effort
to herd expansion. It is assumed that if
the founder can produce the recombinant
protein, then the progeny that inherit the
transgene will also express the protein.
During the 14-year period that we have
carried out these types of program, this
has been the case.
11.2.3.1 Germline Transmission
In order to establish the hAT production
herd, it was necessary to expand the num-
ber of animals by breeding. The founder
male, 155-92, was bred to 14 does during
the first year. Of the progeny generated,
only five kids carried the transgene less
than 50% inheritance of the transgene
marker. This indicated that the founder
155-92 was mosaic. Mosaicism can be an
issue in transgenic herd expansion since it
increases the number of animals that
must be mated to yield a sufficient num-
ber of transgenic females. Even with 50%
transmission (non-mosaic), only 25% of
the offspring will be the desired transgenic
females. Several of the transgenic male
offspring are kept for herd expansion,
since they are not mosaic and would pass
the transgene onto 50% of their progeny.
In the case of the hAT transgenic herd, it
was necessary to identify a male carrying a
single expressively active integration site.
Southern blot analysis of the 155-92 foun-
der showed four to six copies of the trans-
gene. However, when the progeny of 155-
92 were analyzed, four of the five carried
Table 11.1 Summary of BC6 founder goat information
Animal ID Date of birth Sex Germline
transmission
Induced lactation
rhAT level [g L
1
]
Natural lactation
rhAT level [g L
1
]
155-92 02/05/92 M 4/23 1.3 ND
222-92 09/30/92 M ND ND ND
223-92 10/07/92 F 1/5 0.03 <0.15
224-92 10/07/92 M ND ND ND
225-92 10/07/92 F 0/2 Undetectable Undetectable
ND, Not determined.
four to six copies, but one had only one
copy. This was initially disturbing, since it
suggested that there was instability of the
transgene, similar to what had been shown
with the alpha-1 antitrypsin transgenic
founder sheep [49]. To establish the genetic
pattern of these animals, fluorescence in-
situ hybridization (FISH) was used to visua-
lize and follow the chromosomal integra-
tion sites throughout the breeding program
[50]. Semi-quantitative FISH showed that
the founder male, 155-92, carried at least
four integration sites, each with a different
extent of mosaicism (Fig. 11.4). The most
highly represented site, located on chromo-
some 5 (C5), carried four copies of the
transgene. The other sites on chromosomes
1, 12, and 23 (C1, C12, C23) each appeared
1002
Fig. 11.4 FISH analysis of BC6 transgene integra-
tion sites carried by offspring of the 155-92 foun-
der. Integrations located on chromosome 5 (1),
chromosome 12 (2), chromosome 1 (3), and
chromosome 23 (4) are respectively represented.
For each panel, (a) is the respective chromosome
ideogram, while (b) shows a representative pic-
ture of each of the founders corresponding chro-
mosome, with the arrow indicating the transgene-
specific signal.
to carry only a single copy of the transgene.
During breeding, these chromosomes with
their transgenes can segregate indepen-
dently. The progeny of 155-92 have different
combinations of these sites. Analysis of the
milk from those progeny in which these
sites showed full or partial segregation con-
firmed that it was only the four- to six-copy
C5 transgene integration site that was ex-
pressed in the 155 line. Transgenic off-
spring that carried any combination of the
C1, C12, or C23 sites in the absence of the
C5 integration never expressed rhAT, indi-
cating that these integration sites are non-
functional.
11.3
Characterization of rhAT
11.3.1
Purification of rhAT
A process (summarized in Fig. 11.5) was
designed for purification of rhAT from
goat milk with a cumulative yield greater
than 50% [51]. The source material for
purification is milk produced by trans-
genic goats expressing the rhAT protein at
approximately 2 g L
1
. Goats typically lac-
tate for 300 days each year, producing 2
4 L of milk daily. The process normally
produces 300 g of purified rhAT per batch
from no more than 375 L of milk contain-
ing approximately 600 g of rhAT. Milk con-
taining rhAT is diluted with EDTA buffer
and is then clarified by tangential flow fil-
tration with a nominal 500-kDa pore-size
Hollow Fiber membrane filter (Step 1).
The filter permeate is cycled through a
closed loop linking the filtration system to
the Heparin-Hyper D column (Step 2) un-
til >90% of the rhAT is captured (about
eight volume cycles). Heparin affinity
chromatography is also used routinely in
the production of all commercially avail-
able hpAT in the European Union [52].
The Heparin-Hyper D column is washed
and then eluted with a 2.5 M sodium chlo-
ride buffer. Once the Heparin-Hyper D el-
Fig. 11.5 Schematic representation of the rhAT purification from the milk of transgenic goats.
uate is obtained, it is transferred into a
downstream processing area.
The eluate is then filtered through a Pall
DV-20 viral removal filter (Step 3), concen-
trated, and diafiltered by membrane filtra-
tion to adjust the ionic strength for the ap-
plication of the rhAT onto the ANX-Se-
pharose column. After loading, the ANX-
Sepharose column is washed and the rhAT
is eluted with 0.32 M buffer (Step 4). The
ANX-Sepharose eluate is conditioned with
sodium citrate and applied to the Methyl
HyperD column (Step 5). The Methyl Hy-
perD column is washed and the rhAT
eluted from the resin with a lower concen-
tration sodium citrate buffer.
The final formulation is achieved by
concentration and diafiltration into a ci-
trate, glycine, sodium chloride buffer with
the proper ionic strength and dilution to
the final protein concentration of approxi-
mately 25 mg mL
1
. The product is filled
into vials (10 mL, containing ~250 mg pro-
tein), lyophilized, and then treated in a va-
lidated terminal viral inactivation step
(Step 6).
11.3.2
Structural Characterization
The recombinant human AT purified from
transgenic goat milk is indistinguishable
from plasma-derived AT (hpAT) with the
exception of the carbohydrates (see Part
IV, Chapters 2 and 7). Recombinant hu-
man AT made in the milk of transgenic
goats contains the same 432 amino acids
as hpAT preparations as determined by
amino-terminal sequence analysis, peptide
mapping, and liquid chromatography/
mass spectrophotometry analysis (LC/MS)
[51]. N-terminal sequence analysis con-
firmed that the rhAT had the correct N-ter-
minal sequence. The reduced and pyridyl-
ethylated peptide map of rhAT was essen-
tially identical to that of Thrombate
III
(a plasma-derived preparation of AT com-
mercialized in the United States by Bayer
Corporation). The only differences noted
were in the regions of the glycopeptides
due to the glycosylation heterogeneity in
the rhAT. The primary sequence of rhAT
was confirmed by on-line LC/MS analysis
of an endoproteinase Lys-C digest. Both
rhAT and hpAT contain the same three dis-
ulfide bonds (Cys 8128, Cys 2195, Cys
247430) as determined by peptide map-
ping under non-reducing conditions [51].
AT contains four methionine residues,
which may be prone to oxidation under
forced conditions in vitro. Normally, there
is low oxidation of AT. In a comparative
study of rhAT, Thrombate
III and Kyber-

nin
(a commercial plasma-derived AT
preparation), all three AT preparations
were found to have similar low levels of
methionine oxidation [53]. It was also
shown that methionine oxidation had little
impact on the inhibitory activity of rh or
hpAT.
The conformation of rhAT was analyzed
further by circular dichroism (CD) spec-
troscopy [51]. The far-UV CD spectrum
was similar for both rhAT and hpAT pro-
teins, and was characterized by two nega-
tive bands and a positive maximum indica-
tive of the presence of both a-helix and b-
sheet, which was consistent with the crys-
tal structure data of hpAT [54]. In the far-
UV CD spectrum, the addition of heparin
produced little change, which suggested
that secondary structures were not altered.
The near-UV spectra of both proteins were
also similar [51], and in excellent agree-
ment with previously published spectra of
AT derived from human and bovine plas-
ma [55]. The near-UV spectrum showed a
dramatic increase in band intensity across
the whole region when heparin was added
to both proteins. This increase was attrib-
1004
utable to the conformational change of
buried and exposed tryptophan residues
upon heparin binding.
11.3.3
Glycosylation of rhAT
Using on-line LC/MS analysis of an endo-
proteinase Lys-C digest, the only post-
translational modifications detected were
at the known N-glycosylation sites on
either rhAT or hpAT [51]. Both rhAT and
hpAT contain the same four N-linked gly-
cosylation sites (Asn96, 135, 155, 192), as
determined by peptide mapping and by
LC/MS. No evidence of O-linked glycosyla-
tion was observed during LC/MS analysis
of both proteins. Human plasma AT lacks
glycosylation at the Asn 135 (the b-form)
in 515% of the total AT found in plasma
[56, 57]. LC/MS data indicated that the
rhAT had glycosylation at Asn135 greater
than 80% of the time.
As inferred previously from peptide
maps, the monosaccharide composition of
rhAT was different from that of Thromba-
te
III [51]. hpAT has predominantly iden-

tical oligosaccharides on the four N-linked
glycosylation sites [58, 59], although 15
30% of the chains may lack terminal sialic
acid [60, 61]. The main glycosylation differ-
ences observed for the rhAT were the pres-
ence of fucose and GalNAc, a higher level
of mannose, and a lower level of galactose
and sialic acid. There is also substitution
of 4050% of the N-acetylneuraminic acid
with N-glycolylneuraminic acid. As ex-
pected from the monosaccharide composi-
tional analysis, the LC/MS analysis was
more complex for all the rhAT glycopep-
tides than for hpAT. The terminal sialic acid
in the rhAT contained the same 2,6 linkage
found in hpAT [61, 62]. Several laboratories
have determined that differences in glycosy-
lation of AT do not affect the intrinsic rate
constant of the uncatalyzed or heparin-cata-
lyzed inhibition of thrombin, indicating
that the carbohydrate chains solely affect
heparin binding and not heparin activation
or proteinase binding functions [6365].
Thus, glycosylation differences do not im-
pact the major biological activity of AT
which is thrombin inhibition, but did ex-
plain the differences in affinity for heparin
and in pharmacokinetics.
11.3.4
Inhibitor Activity and Heparin Binding
Affinity of rhAT
Antithrombin is a serine protease inhibitor
that inhibits thrombin and Factors IXa
and Xa and, to a lesser extent, Factors
VIIa, XIa, XIIa, as well as proteases such
as trypsin, plasmin, and kallikrein [16].
The addition of heparin increases its inhi-
bitory activity 300- (Factor Xa) to 1000-
(thrombin) fold by inducing an allosteri-
cally transmitted conformational change in
the reactive center loop of the AT molecule
[66]. Thrombin and AT interact to produce
a tightly bound TAT complex that is essen-
tially irreversible and is cleared quickly
from the circulation [10, 67]. Heparin
binding to the AT molecule plays a cataly-
tic role in increasing the inhibitory activity
of AT toward thrombin and Factor Xa.
There are two forms of AT in human plas-
ma; these have different heparin affinities,
but the same inhibitory activity toward
thrombin [56, 57]. Some 8590% of circu-
lating hpAT has glycosylation on four Asn
residues; this fully glycosylated form is re-
ferred to as the a-form. Between 5 and
15% of circulating hpAT (referred to as
the b-form) lacks glycosylation at Asn135,
and has a 3- to 10-fold higher heparin af-
finity than the a-form [56].
The specific activity of rhAT was identi-
cal to that of hpAT (Thrombate III) in an
in vitro thrombin inhibition assay in the
presence of excess heparin (~6 IU mg
1
)
[51], and very similar to that reported for
Atenativ
(Pharmacia) [52, 68]. Equivalent

inhibition for rhAT and Thrombate III was
also seen in an in vitro Factor Xa inhibi-
tion assay in the presence of excess hepa-
rin [51]. Heparin cofactor activation of
rhAT versus hpAT was determined by vary-
ing the amount of heparin used in either
inhibition assay. A lower concentration of
heparin was required for rhAT than hpAT
for inhibition of both enzymes, similar to
the b-form of hpAT. Thus, rhAT closely re-
sembles hpAT with respect to its activity
for both thrombin and Factor Xa in the
presence of saturating levels of heparin
[51]. By using a tryptophan fluorescence
assay, a four-fold higher affinity for hepa-
rin was observed with rhAT when com-
pared with Thrombate III, but similar to
that reported for the b-form of hpAT. The
fluorescence values at saturating heparin
were indistinguishable for rhAT and
Thrombate III [51]. RhAT has a three- to
four-fold higher overall heparin affinity
than the a-form of hpAT, due to the glyco-
sylation differences between these mole-
cules. However, in the presence of excess
heparin moieties, as found on the surface
of vascular endothelial cells or with exoge-
nous heparin supplementation, the alpha,
beta, and recombinant forms of AT have
identical inhibitory activities against throm-
bin [56] because the heparin is not itself in-
volved in the inhibition.
in vitro studies were also conducted to
assess the behavior of rhAT in assays used
routinely to monitor hpAT in patient thera-
py. Cooper et al. [69] compared rhAT to di-
lutions of normal pooled plasma in three
assays: 1) a thrombin-based assay using
60, 180 and 300 s incubation of thrombin
(Dade Behring kit); 2) a Factor Xa-based
assay (Chromogenix kit); and 3) an hAT
ELISA (Dako antibodies). Antithrombin
level in the rhAT concentrate was assayed
against the 8th British Blood Coagulation
Factors Standard to compare the stated
dose with the assayed dose per vial. Hepa-
rin binding was also assessed by two-di-
mensional electrophoresis with or without
heparin in the first dimension, and by
heparin-Sepharose gel filtration. The con-
centration of rhAT in the vial was 89% of
the stated value by this thrombin-based as-
say, 85% by Xa-based assay, and 119% by
the antigen assay.
11.3.5
Viral and Prion Safety
Since it is obtained from the milk of closed
herd of transgenic goats, rhAT is inherently
unlikely to transmit human blood-borne
viruses and other human infectious agents.
Moreover, no human-derived protein is
added during the production, isolation, or
formulation of rhAT. The production goat
herd is closed and highly controlled. A high
level of donor control and testing is a key pa-
rameter in the viral safety strategy for rhAT.
In addition, the rhAT-containing milk is as-
sessed in vitro on three cell cultures (human
MRC-5, monkey Vero, and goat turbinate) to
screen for adventitious viruses, including
any emerging unknown virus, that may be
present in the milk pool used as the starting
material for production of rhAT. To date, all
milk pools from the GTC Farm have tested
negative in the in vitro cell line screening. In
addition, a nanofiltration step has been in-
corporated into the current rhAT process
for viral removal between the heparin affin-
ity column and the ion-exchange column.
The validated viral removal capacity of the
rhAT purification process is summarized
in Table 11.2. The viral validation studies
demonstrated that a significant virus reduc-
tion of 8.5 to 25.3 log
10
was accom-
1006
plished across the distinctly different modes
of the rhAT process.
Transmissible spongiform encephalopa-
thies (TSE), such as new-variant Creutz-
feld-Jakob disease (nvCJD) in humans, bo-
vine spongiform encephalopathy (BSE) in
cattle and scrapie in sheep and goats, also
must be considered in assuring the safety
of products made from human or rumi-
nant sources. Human donors are moni-
tored for CJD and nvCJD, and potentially
contaminated blood, plasma pools and
products made from them have been re-
called or traced when a contributing donor
has been diagnosed with CJD. All GTC
goats are certified free of scrapie in the 5-
year United States Department of Agricul-
ture (USDA) Voluntary Scrapie Flock Cer-
tification Program, and various risk-mini-
mization measures have been instituted to
reduce any potential risk from this TSE in
this highly controlled, closed donor goat
population. In addition, the rhAT purifica-
tion process has been validated for its abil-
ity to remove 11.3 log
10
scrapie.
In aggregate, these data strongly support
the conclusion that the rhAT manufactur-
ing and production process is safe for hu-
man use with respect to potential viral and
prion adventitious agent contamination.
11.4
Preclinical Studies
Antithrombin is a complex protein with
multiple biologically important activities. It
is the most critical modulator of coagula-
tion, and has potent anti-inflammatory
properties independent of its effects on co-
agulation [35, 7073]. Several in vitro and
in vivo animal studies, some of which used
hpAT as a direct comparator, were underta-
ken to evaluate the potency of rhAT.
11.4.1
In vitro Efficacy Studies
Antithrombin has been shown to promote
the release of prostacyclin from endothelial
cells [74] and to attenuate ischemia-in-
duced leukocyte extravasation [75]. In addi-
tion, chemotactic effects of AT on neutro-
phils and signaling via interaction with
cell-surface heparan sulfate proteoglycans
Table 11.2 Validation of log
10
viral removal or inactivation by the RhAT process
Process step Pseudora-
bies virus
Xenotropic
murine
retrovirus
Human
adenovirus
Porcine
parvovirus
Poliovirus
a
Mouse
adenovirus
a
VirA/Gard 500-kDa 5.1 3.7 5.3 2.4 4.1 3.5
Heparin Hyper-D 1.8 1.2 <1.0 1.4 4.0 2.3
Pall DV-20 Filter 4.8 3.8 6.3 3.7 ND ND
ANX-Sepharose 3.6 1.0 7.1 1.1 2.4 <1.0
Methyl HyperD 5.6 3.5 4.8 5.7 5.2 2.7
Heat treatment 2.8 5.0 1.8 2.4 1.9 ND
Total reduction 23.7 18.2 25.3 16.7 17.6 8.5
a) Poliovirus and mouse adenovirus were only used once in pre-
liminary validation runs. All the others were run in duplicate
and, in some cases, three times.
ND, Not determined.
(HSPG), possibly via syndecan-4, have
been reported [76].
A study was undertaken to characterize
the mechanism by which AT regulates the
migration of neutrophils, which are in-
volved in a variety of conditions including
inflammatory diseases. (A dramatic video
animation of neutrophils is available on
the supplement CD-ROM.) Human neu-
trophils were obtained from healthy volun-
teers, and migration was measured in
modified Boyden chambers. Either Kyber-
nin P (a commercial hpAT preparation) or
rhAT was used as an attractant. rhAT was
at least as effective in deactivating neutro-
phil chemotaxis as Kybernin P [77]. To in-
vestigate the role of intact HSPG on the
neutrophil surface for AT-induced cell mi-
gration, neutrophils were pretreated with
heparinase or chondroitinase. Chemotactic
effects of hpAT (1 U mL
1
) or rhAT
(1 U mL
1
) were completely abolished by
pretreatment with both agents. Antibodies
to syndecan-4 also inhibited rhAT-induced
migration of neutrophils. Collectively,
these data suggest that AT regulates neu-
trophil migration via effects on its hepa-
rin-binding site on cell surface syndecan-4.
Aspects of these investigations with rhAT
confirm the findings of previous studies of
the binding of hpAT to syndecans on the
cell surface and AT inhibition of neutro-
phil chemotaxis. Similar effects were ob-
served with rhAT and hpAT on human eo-
sinophils [78].
In another study [79], the same group of
investigators tested whether rhAT might
influence lipopolysaccharide (LPS)-induced
enhancement of adhesion of neutrophils
to human endothelial cells and the in-
volvement of glycosaminoglycans. Using
human umbilical vein endothelial cells
(HUVEC), experiments were performed
aiming to quantitate fluorometrically the
adhesion of neutrophils. Treatment with
LPS and interleukin-1 (IL-1) increased
neutrophil adhesion to HUVEC, which
was inhibited by rhAT. Concomitant incu-
bation of rhAT and an inhibitory pentasac-
charide reversed rhATs effects on adhe-
sion. Treatment of endothelial cells with
heparinase and chondroitinase to release
HSPG which normally bind AT led to
higher neutrophil adhesion. Treatment of
endothelial cells with antibodies to synde-
can-4, which is a receptor for AT, en-
hanced the adhesion of neutrophils. Wes-
tern blotting studies showed that LPS-in-
duced signaling was diminished by rhAT,
and that the effect was reversible by chon-
droitinase or heparinase. From these re-
sults, it was concluded that LPS-induced
adhesion of leukocytes to endothelium is
reversed by ligation of rhAT with synde-
can-4. Complementary experiments show-
ed that rhAT attenuated the expression the
beta-2 integrin CD11/CD18 on activated
neutrophils and monocytes [80], providing
a potential direct molecular mechanism
for the effect of AT on adhesion.
11.4.2
In vivo Efficacy Studies
Animal disease models for acquired AT de-
ficiency, such as Escherichia coli-induced
sepsis, were used to assess the biological
consistency of the rhAT compared to hpAT.
Although in these models the exact contri-
butions of the anti-coagulant and anti-in-
flammatory properties of AT are still some-
what controversial [81], they do allow for
some comparison of the properties of
rhAT with hpAT. Human plasma-derived
AT has been shown to prevent the lethal
effects of experimentally induced sepsis in
several animal models [8289], and to
block cytokine production in vitro. To date,
several dose response studies have been
performed with rhAT in rats (GTC Biother-
1008
apeutics, unpublished results [90]) and ba-
boons [91] that have been lethally chal-
lenged with E. coli, K. pneumoniae or LPS.
Both rhAT and hpAT preparations were
equally effective in preventing sepsis and
septic shock in these models. Additional
studies have been performed demonstrat-
ing protective effects of rhAT in a smoke
inhalation sepsis model in sheep, and in a
xenotransplantation model in primates.
Recombinant human AT was initially
tested by Taylor et al. in a lethal E. coli ba-
boon sepsis model. Infusion of rhAT was
found significantly to improve the survival
of treated baboons [91]. This was also ob-
served previously, by the same group, with
hpAT [82]. Five adult animals were used,
and the dosing regimen involved adminis-
tration of 500 U kg
1
rhAT as 0.5-h infu-
sions at t =1 h and at t =+3 h and
250 U kg
1
rhAT as a bolus at the time of
E. coli challenge (t =0 h). Administration of
rhAT protected three of the five baboons
from the challenge. One of the two ba-
boons that died was not administered the
third dose of rhAT; this animal died due to
sepsis. The cause of death of the other ba-
boon was attributed to capillary leakage in
the lungs consistent with adult respiratory
distress syndrome (ARDS); there was no
evidence of DIC. These results indicated
that, when given in the appropriate doses,
rhAT protects against DIC and that, in
three of four cases, rhAT protects against
death from a lethal dose of E. coli. The
protective effect of rhAT was due to a com-
bination of anticoagulation and anti-in-
flammatory effects. As previously noted
for the hpAT study, rhAT attenuated the co-
agulation response and fibrinolysis. Signif-
icantly, treatment with rhAT also markedly
attenuated the release of the pro-inflam-
matory cytokines IL-6 and IL-8.
Further evidence of the anti-inflamma-
tory activity of rhAT was found in an endo-
toxemic rat model [90]. In this model,
rhAT treatment reduced endotoxin-
mediated mesenteric venule leukocyte ad-
hesion and small intestine mucosal barrier
injury. Rolling and firm adhesion of leuko-
cytes in mesenteric venules of endotoxe-
mic rats was measured using intravital mi-
croscopy. Endotoxemia was induced by in-
travenous administration of 10 mg kg
1
of
endotoxin, after which rats were treated
either with saline or rhAT (250 and
500 U kg
1
). Following anesthesia, the dis-
tal ileum was exteriorized and the mesen-
tery inserted in an intravital microscopy
chamber fitted with a video camera sys-
tem; the mesenteric circulation was then
examined. Flux of rolling leukocytes was
measured as the number of white blood
cells that could be seen rolling past a fixed
perpendicular line in the venule during a
1-minute interval. Quantification of venu-
lar endothelium leukocyte adherence was
performed off-line by playing back video-
taped images and counting the number of
leukocytes that stuck and remained sta-
tionary for a period of >30 seconds. rhAT
was shown to attenuate both endotoxin-in-
duced venular leukocyte rolling and adhe-
sion in a dose-dependent manner. Pretreat-
ment with indomethacin, a prostaglandin
synthesis inhibitor, completely abolished
the effect on leukocytes rolling and adhe-
sion, suggesting that the effect of AT could
be mediated by an effect on prostacyclin
production.
This effect on leukocyte adhesion ob-
tained with rhAT is similar to the activity
of hpAT observed in related models. For
example, in the skinfold of endotoxemic
Syrian hamster, multiple injections of
250 U kg
1
of hpAT attenuated LPS-in-
duced arteriolar and venular leukocyte ad-
hesion [92, 93]. Here again, this effect was
completely abolished by pretreatment with
indomethacin. Previously, in a feline me-
sentery ischemia/reperfusion model using
intravital microscopy to monitor leukocyte
rolling and adhesion, pretreatment with
hpAT (250 U kg
1
) reduced neutrophil roll-
ing and adhesion to pre-ischemic levels
during reperfusion [75].
The previously described baboon study
showed survival after E. coli challenge when
rhAT was used in a pretreatment regime.
Murakami et al. [94] investigated the effect
of post-treatment with rhAT on sepsis after
smoke inhalation in sheep. Acute lung in-
jury frequently arises after smoke inhala-
tion complicated by pneumonia induced
by Pseudomonas aeruginosa. Previously,
these investigators had shown that high-
dose hpAT administration attenuated endo-
toxin-induced acute lung injury in rats. In
addition, they found that hpAT administra-
tion promoted prostacyclin production,
which inhibited leukocyte activation.
In the present study, Pseudomonas aeru-
ginosa was instilled into the lungs of an-
esthetized sheep after insufflations of cool
cotton smoke. One group of sheep also re-
ceived rhAT by continuous infusion, start-
ing 1 hour after injury and continuing for
the next 24 hours at 1000 U kg
1
per 44-
hour period. Plasma AT levels fell signifi-
cantly in the control animals, but were
maintained at baseline levels in rhAT-
treated animals, as was previously demon-
strated in other sepsis models. In addition,
rhAT attenuated septic shock in these ani-
mals and acute lung injury was improved
histologically, with a reduction of cast for-
mation in the airways. All animals receiv-
ing rhAT were negative for fibrin degrada-
tion products, in contrast to untreated ani-
mals. Platelet levels fell in control animals;
however, platelet counts in the rhAT-
treated group at 24 hours were not differ-
ent from baseline values, which suggested
that rhAT administration attenuated the
coagulation abnormalities observed with
sepsis. The investigators concluded that
post-treatment with rhAT was effective in
sepsis after smoke inhalation in sheep.
In a subsequent study [95], nebulized
rhAT was used in the same model and
was even more effective than intravenously
administered rhAT at half the dose. In ad-
dition, pulmonary gas exchange, shunt
fraction and lung wet:dry weight ratio
were significantly attenuated by AT nebuli-
zation, thereby underscoring the protective
effect of rhAT in this sepsis model.
The hypothesis that treatment with rhAT
would prevent or at least delay the onset
of rejection and coagulopathy was tested
using a life-supporting pig-to-baboon renal
xenotransplantation model [96]. Non-im-
munosuppressed baboons were trans-
planted with transgenic pig kidneys ex-
pressing the human complement regula-
tors CD55 and CD59. The baboons were
treated with rhAT by intravenous infusion
every 8 hours, with or without heparin. No
bleeding complications were observed.
RhAT-treated baboons had preservation of
normal renal function for 45 days, which
was twice as long as untreated animals,
and developed neither thrombocytopenia
nor significant coagulopathy during this
period. Thrombin clotting times were rela-
tively normal in the rhAT-treated baboons
for 45 days, and platelets and clotting fac-
tors were not consumed faster than they
could be replaced. The relative importance
of the anticoagulant and anti-inflammatory
properties of AT in the xenografts setting
remains to be determined.
11.4.3
Toxicology, Pharmacokinetic,
and Mutagenicity Studies
A series of single-dose and repeat-dose tox-
icological studies were conducted to exam-
ine the safety of rhAT administered intrave-
1010
nously to rats, dogs and non-human pri-
mates at doses up to 10 times the highest
anticipated dose in man. In addition, the
safety of rhATand heparin administered in-
travenously to rats was evaluated in a single-
dose toxicological study. In all of these stud-
ies, rhAT was well tolerated. Most clinical
observations and/or adverse reactions were
related to the pharmacological anticoagu-
lant properties of the test article, and were
seen only at the highest doses tested.
Pharmacokinetic studies were performed
to determine if gender, dose and/or re-
peated administration affect the kinetic
disposition of rhAT. The single-dose phar-
macokinetic studies were performed in
Sprague-Dawley rats, beagle dogs, cyno-
molgus monkeys, and baboons. The re-
peat-dose pharmacokinetic studies were
performed in Sprague-Dawley rats and cy-
nomolgus monkeys. The results indicate
that the kinetic disposition of rhAT was
non-linear in all species examined. Clear-
ance decreased as a function of increasing
dose concurrent with an increase in half-
life. Single or repeated administration of
rhAT did not alter the kinetics of rhAT, nor
were there any gender-related effects asso-
ciated with this compound.
Three studies were performed to evalu-
ate the potential mutagenicity or genotoxi-
city of rhAT. The studies performed in-
cluded an Ames assay that evaluated muta-
genicity in five different Salmonella typhi-
murium strains and two E. coli strains, a
mouse in vivo micronucleus assay that
evaluated genotoxicity, and a CHO cell in
vitro assay that evaluated chromosomal
aberration. The results obtained did not
show any potential mutagenicity or geno-
toxicity associated with rhAT.
11.5
Clinical Trials with rhAT
At this point, eight clinical studies have
been undertaken with ATryn
(Table 11.3).
Two clinical indications were pursued:
Heparin resistance in patients under-
going cardiac surgery involving cardio-
pulmonary bypass (CPB).
Prevention of deep-vein thrombosis
(DVT) in patients who have a hereditary
Table 11.3 Summary of clinical studies
Phase Indication Type of study Total
patients [n]
I Not applicable Open-label, PK in normal healthy volunteers 17
I Not applicable Open-label, cross-over PK trial in normal healthy
volunteers
26
I/II Hereditary deficiency Open-label PK in patients with hereditary deficiency 15
II Heparin resistance Dose-ranging trial, nine dose groups, CABG 36
III Heparin resistance Double-blind, placebo-controlled in heparin-resistant
patients/CPB
54
III Heparin resistance Double-blind, placebo-controlled in heparin-resistant
patients/CPB
52
III Heparin resistance Active-controlled trial in heparin-resistant patients 47
III Hereditary deficiency Open-label in hereditary deficient patients 14
CABG, Coronary artery bypass grafting; CPB, Cardiopulmonary
bypass; PK, Pharmacokinetics.
deficiency of AT and who are in high-
risk situations such as birth delivery or
surgery.
In addition, rhAT was used in a compas-
sionate-use program for United States AT
hereditary deficient patients undergoing a
high-risk event and that could not access
hpAT due to lack of availability. In all the
human studies completed to date, rhAT
has proved safe and met the primary end-
points of that study. All studies showed
that rhAT was well-tolerated and safe in
these patient populations.
11.5.1
Single-dose Pharmacokinetic Study
in Healthy Volunteers
A single-center, randomized, parallel group
Phase I study was conducted in 17 male
volunteers aged between 20 and 28 years.
The subjects were divided into four groups
each of three volunteers, with individual
subjects receiving either 50, 100, 150, or
200 U AT per kg body weight. An addi-
tional five volunteers were enrolled to do-
nate plasma for the determination of en-
dogenous baseline AT concentrations.
There were no serious adverse events. Re-
view of all hematology and biochemistry
profiles, urinalysis, vital signs, electrocar-
diogram, activated clotting time (ACT) and
coagulation parameters revealed no clini-
cally significant changes in any group, or
any differences between those subjects re-
ceiving active drug and those that did not
receive AT. There were four mild adverse
events. One subject in the 150 U kg
1
dose
group experienced a non-serious headache
which was judged as possibly being related
to the drug. Patients were tested before
and after treatment for the presence of an-
tibodies to rhAT. No antibodies could be
detected either before or after treatment.
Lu et al. [97] investigated the pharmaco-
kinetics of rhAT in these study subjects.
The concentrations of AT, after initial
doses given over 30 minutes, were best de-
scribed by a weight-normalized, two-com-
partment model. The fast compartment
volume was 41.1 ml kg
1
and the volume
of distribution was 115.4 ml kg
1
. Inter-
compartmental clearance was 0.0763 ml
kg
1
min
1
, and elimination clearance was
0.0383 ml kg
1
min
1
. These variables are
equivalent to a distribution half-life of
196 minutes and an elimination half-life of
2568 minutes. Approximately 75% of the
supplemental dose was removed from
plasma by the initial distribution process.
In conclusion, rhAT when given as a
30-minute infusion at doses up to
171 U kg
1
was shown to be safe and
well-tolerated in healthy male volunteers.
11.5.2
Heparin Resistance
11.5.2.1 Heparin Resistance in Coronary
Artery Bypass Graft Patients:
Dose-finding Study
Acquired AT deficiency may render hepa-
rin less effective during cardiac surgery
and CPB. This Phase II study was de-
signed to examine the pharmacodynamics
and optimal dose of rhAT needed to main-
tain normal AT activity during CPB, to op-
timize the anticoagulant response to hepa-
rin, and to attenuate excessive activation of
the hemostatic system in patients under-
going coronary artery bypass grafting
(CABG). During CPB, AT activity fre-
quently decreases to as little as 3050% of
normal. Low AT concentrations during car-
diac surgery are likely to develop because
of the preoperative use of heparin, the ef-
fect of hemodilution on the pump, and
CPB- associated excessive hemostatic sys-
tem activation [98]. Anticoagulation is used
1012
during cardiac surgery to prevent thrombo-
sis of the extracorporeal circuit and to
minimize CPB-related activation of the he-
mostatic system. In some cases, when
heparin alone is not effective, either fresh-
frozen plasma (FFP) [99] or hpAT concen-
trates [100102] have been used in patients
that show an appreciable heparin resis-
tance prior to initiation of CPB. However,
AT concentrate has not been approved for
this indication in the US.
A single-center, open-label, single-dose,
dose-escalation study was conducted in 36
patients, aged between 18 and 80 years,
admitted for primary cardiac surgery re-
quiring CPB [103]. All patients underwent
elective primary CABG and had been re-
ceiving heparin therapy for at least
12 hours prior to surgery. Thirty patients
received rhAT, and six received placebo.
Patients receiving active drug were divided
into groups of three, and assigned to one
of nine dosing cohorts. The individual
treatment dosing cohorts were 10, 25, 50,
75, 100, 125, 150, 175, and 200 U kg
1
rhAT.
A tenth, placebo, cohort was added which
included an additional three patients.
None of the patients that had post-drug
samples taken developed circulating anti-
bodies to AT following treatment. Supple-
mentation of rhAT significantly (p<0.0001)
improved heparin responsiveness, as mea-
sured by an increase in the ACT
(844+191 s) as compared to heparin ad-
ministration alone (531+180 s). Further-
more, AT supplementation resulted in sig-
nificantly (p=0.001) better inhibition of
thrombin (as measured by a decrease in fi-
brin monomer) and fibrinolysis (as mea-
sured by a decrease in D-dimer) at doses
up to 125 U kg
1
. There was also a re-
duced impairment of platelet function
after CPB, which is thought to be the most
important hemostatic defect after CPB. Re-
sults suggest that single rhAT doses of
75 U kg
1
and higher will maintain the AT
activity level at greater than 100% through-
out the course of CPB.
Based on the results of this Phase II
trial, two Phase III studies were conducted
to test the potential benefit of supplement-
ing rhAT levels in patients with an ac-
quired deficiency state who demonstrate
heparin resistance.
11.5.2.2 Heparin Resistance
in CABG Patients
Two identical Phase III placebo-controlled,
double-blind, multicenter studies were
conducted in heparin-resistant patients
scheduled for cardiac surgery requiring
CPB. The study objective was to establish
whether heparin-resistant patients who re-
ceive rhAT are less likely to require FFP as
a source of AT to achieve an ACT
>480 seconds, as compared to patients re-
ceiving placebo [104, 105]. The secondary
objectives were to compare the effect of
rhAT and FFP on laboratory measures of
plasma levels of AT, thrombin activity, and
fibrinolysis. Specifically, the trials ad-
dressed whether rhAT, at a dose of
75 U kg
1
, increases plasma AT levels and
inhibits thrombin activity and fibrinolysis
more effectively, than 2 units of FFP
alone. One study (AT97-0502) enrolled 54
patients (rhAT: placebo=27: 27). The sec-
ond study (AT97-0504) enrolled 52 patients
(rhAT:placebo=28: 24).
One dosing cohort received 75 U kg
1
rhAT, and the other cohort received a nor-
mal saline placebo (both as single bolus
intravenous injection). Heparin resistance
was defined as failure to achieve an ACT
of >480 seconds after receiving a total dose
of 400 U kg
1
heparin intravenously after
anesthesia induction and surgical incision,
but just prior to CPB. The proportion of
rhAT patients who required administration
of 2 U FFP to achieve an ACT of >480 sec-
onds was significantly less (p <0.001) than
that of placebo patients (19% versus 81%
in the AT97-0502 study; 21% versus 91%
in the AT97-0504 study).
In summary:
Administration of rhAT precluded the
need for FFP for 44/55 patients versus
only 7/50 patients treated with placebo.
The mean AT activity level and change
in AT activity level was significantly
greater after rhAT administration than
placebo throughout the study period.
RhAT replacement therapy maintained
AT activity within or above normal range
throughout the study period.
AT activity continued to decline in the
placebo group due to continued con-
sumption and hemodilution, despite 2
units of FFP not providing adequate AT
replacement therapy.
ACTs, the standard measure of coagula-
tion status, were significantly increased
in the rhAT-treated group compared to
the placebo group.
Compared to placebo patients, patients
who received rhAT showed significant
inhibition of the generation of two
markers of thrombin activation pro-
thrombin fragment 1.2 and thrombin
antithrombin complex.
Some trends were observed for a de-
crease in the production of D-dimer
after rhAT treatment compared to place-
bo treatment.
The safety profiles of the rhAT and pla-
cebo groups were comparable.
No evidence for an immune response
(measured by patient immune response
assays) was observed.
11.5.3
Hereditary Deficiency
11.5.3.1 Compassionate Use of rhAT
for Hereditary AT-deficient Patients
in High-risk Situations
Hereditary AT deficiency is associated with
a significant risk of venous thrombosis in
high-risk situations such as birth delivery
1014
Table 11.4 Hereditary AT-deficient patients treated with rhAT in a compassionate-use basis
Subject
no.
Age
[years]
Sex Baseline
AT level
[%]
Surgery type Total rhAT
received
[U]
Vascular
duplex
ultra-
sound
Clinical evi-
dence of
thrombosis
Anti
rhAT
antibody
1a 22 M 40 Laparoscopic
splenectomy
74480 Negative
(d5)
Negative Negative
1b 22 M 40 Bilateral hip
replacement
294000 Not
tested
Negative Negative
2 47 F 49 Hysterectomy 101921 Negative
(d7)
Negative Negative
3 36 F 58 C-section 65000 Negative
(w6)
Negative Negative
4 72 M 52 Coronary artery
bypass
39200 Not
tested
Negative Not
tested
5 71 M 53 Knee replace-
ment
73480 Negative Negative Negative
and surgery. Plasma-derived AT concen-
trate (Bayer Corp.; Thrombate
III) has
been approved in the United States for re-
placement when anticoagulation is inter-
rupted in these patients. However, Throm-
bate III supplies have been limited and
there have been periods when it was not
available.
Five patients with hereditary AT defi-
ciency and a prior history of thromboem-
bolism were treated with rhAT on a com-
passionate use basis for six surgical proce-
dures [106]. One patient had two surgical
procedures 6 weeks apart, and received
rhAT on each occasion. Patients were
treated perioperatively, receiving multiple
doses of rhAT for 216 days. Dosing was
determined individually by the investiga-
tors, with the goal of maintaining an AT
activity of 80150% of normal. Patients
were followed for clinical evidence of
thrombosis, bleeding, adverse events, and
development of antibodies to the rhAT.
All six surgical events were successfully
treated with rhAT, as shown by the ab-
sence of clinical evidence of thrombosis
for these patients that all had an history of
thromboembolism. In two patients, where
initial pre-and post-treatment levels were
available, there was a 1.69 and 1.66% per
U per kg increase, which is similar to the
1.39 and 2.05% per U per kg reported for
hpAT. There was no clinical evidence of
thrombosis or bleeding, and no adverse
events related to the drug were reported.
Four of the six surgical events were fol-
lowed-up by vascular duplex ultrasound of
the lower extremities, with no clinical evi-
dence of acute thrombosis (Table 11.4).
Four of the five patients, who receive mul-
tiple doses of rhAT, were also screened for
antibody formation against rhAT several
weeks postoperatively. None of the patients
developed detectable antibodies to the
rhAT.
11.5.3.2 rhAT Use for Hereditary
AT-deficient Patients
(009 and 02001)
Two trials studying the use of rhAT in he-
reditary-deficient (HD) patients were re-
cently completed [107, 108; see also von
Depka et al., in preparation].
Pharmacokinetic study The Phase I/II
(009) trial was a pharmacokinetic study to
obtain clearance data from asymptomatic
HD patients after single-dose administra-
tion of rhAT, and to use these data to sim-
ulate steady-state concentration profiles
and develop dosing regimens. Fifteen HD
patients were administered a 50 or
100 IU kg
1
dose of rhAT as a bolus dose.
Plasma samples were collected over a 72-
hour period and analyzed for AT activity.
In this study, the non-compartmental phar-
macokinetics (mean SD) were: C
max
(100 U kg
1
) 212.6725.51%; C
max
(50 U
kg
1
) 147.3325.83%; T
max
0.0770.083 h;
K
el
0.1840.266 L h
1
; T
1/2
10.497.19 h;
Vd area 167.65122.25 mL kg
1
; MRT
15.66 10.02 h;, Cl 15.7812.44 mL h
1
kg
1
, and incremental recovery 1.530.34
(IU mL
1
)/(IU kg
1
administered).
The data were also modeled using a
two-compartment model. Median values
for the modeled baseline, terminal half-
life, Vd area, K
10
, K
12
and K
21
were
66.87%, 2.4 h, 66.33 mL kg
1
, 0.30 h
1
,
0.31 h
1
, and 7.87 h
1
, respectively. These
median model parameters were used to
simulate plasma concentrations when AT
is given once, twice, three, and four times
per day and as a continuous infusion (with
a loading dose) at various doses and pa-
tient baseline values. The results indicated
that a combination of a short infusion
loading dose and a continuous infusion
maintenance dose appeared to be the opti-
mal dosing regimen to maintain AT levels
between 80 and 120%.
Efficacy study The 02001 Phase III effi-
cacy study was open-label, blinded evalua-
tor trial of rhAT in at least 12 hereditary
AT-deficient patients being treated prior to,
during, and following high-risk events.
The trial assessed the incidence of DVT
following prophylactic intravenous rhAT
administration to hereditary AT-deficient
patients during situations associated with
a high-risk event. Dosing of rhAT was
achieved by a loading infusion of rhAT fol-
lowed by continuous infusion for at least 3
days, with the objective of maintaining AT
plasma activity between 80% and 120% of
normal.
The primary study end-point was inci-
dence of DVT and other thromboembolic
events assessed clinically and by both lo-
cally and centrally assessed ultrasonogra-
phy. The duplex-ultrasounds of the lower
extremities were used to confirm or ex-
clude the occurrence of DVT. These proce-
dures were performed and interpreted by
qualified specialists within the same hospi-
tal/institution on a real-time basis for the
timely and appropriate clinical care of the
patient. Furthermore, duplex ultrasound
studies were videotaped for a subsequent
standardized blinded interpretation by a
qualified, independent laboratory that pro-
vided an unbiased evaluation of the inci-
dence of DVT. Thromboembolic events
other than DVT that occurred during the
study period were assessed, and the inves-
tigator established the clinical relationship
of the event to treatment with rhAT. Sec-
ondary end-points were safety, adverse
events and immunogenicity.
The study was initiated in December
2003, and completed in the fourth quarter
of 2003, enrolling 14 patients comprising
nine birth deliveries and five surgeries.
None of the patients showed clinical signs
of DVT or thromboembolism, nor devel-
oped antibodies against rhAT. Upon local
review of duplex-ultrasounds, one patient
was evaluated to have an acute DVT which
was resolved by day 7 after treatment.
Upon centralized review, an additional pa-
tient was evaluated to have a DVT which
was resolved by day 30 after treatment.
Neither patient exhibited clinical signs of
thrombosis. The patient in whom DVT
was detected only with central review was
a birth delivery patient who was evaluated
locally to be exhibiting chronic changes,
and no special treatment was initiated. For
the other patient in whom DVT was de-
tected both centrally and locally, the find-
ings developed following a hip replace-
ment (a highly thrombogenic procedure,
even in non-AT-deficient patients), treat-
ment with rhAT in combination with ther-
apeutic low molecular-weight heparin and
vitamin K antagonists was continued. The
patient remained without symptoms, and
the DVT resolved.
11.6
Conclusions
Following completion of the 02001 efficacy
trial, a European regulatory filing was sub-
mitted in January 2004, for the use of
rhAT in the prophylaxis of DVT in heredi-
tary AT-deficient patients in a high-risk sit-
uation. If this application is approved,
this will constitute the first approval of a
transgenically produced biopharmaceutical.
Indeed, this will constitute the first ap-
proval of a biologic manufactured in a
new recombinant production system since
approval of the first product manufactured
in cell culture in the early 1990s. The de-
velopment of a recombinant option for an-
tithrombin will provide a safe and reliable
supply of this important factor, and will fa-
cilitate the resumption of clinical trials
aimed at acquired deficiencies of anti-
1016
thrombin such as cardiovascular surgery,
severe burns, and severe sepsis. Other
transgenically expressed recombinant
products which are currently in develop-
ment include human albumin, C1-esterase
inhibitor and alpha-1 antitrypsin, as well
as several monoclonal antibodies. In sum-
mary, the emergence of this transgenic
manufacturing platform will provide an at-
tractive option for the recombinant pro-
duction of complex biopharmaceuticals
that are needed in large amounts.
Acknowledgments
The authors are deeply indebted to their
colleagues at GTC Biotherapeutics and
Genzyme. The development of a new drug
is truly an enormous endeavor, and entire
scientific teams have dedicated years of
their professional lives to the filing of
ATryn. In addition, the authors extend spe-
cial thanks to Suzanne Groet, Tom New-
berry, and Dick Scotland for their critical
reading and constructive suggestions, as
well as to Jennifer Williams and Merry
Harvey for their generous help with the il-
lustrations.
References
1 U. Abildgaard, Scand. J. Clin. Lab. Invest.
1969, 24, 2327.
2 E. T. Yin, S. Wessler, P. J. Stoll, J. Biol. Chem.
1971, 246, 37123719.
3 R. D. Rosenberg, P. S. Damus, J. Biol. Chem.
1973, 248, 64906505.
4 J. S. Rosenberg, P. McKenna, R. D. Rosenberg,
J. Biol. Chem. 1975, 250, 88838888.
5 K. Kurachi, K. Fujikawa, G. Schmer, E. W. Da-
vie, Biochemistry 1976, 15, 373377.
6 N. Stead, A. P. Kaplan, R. D. Rosenberg, J.
Biol. Chem. 1976, 251, 64816488.
7 S. Kondo, W. Kisiel, Thromb. Res. 1987, 46,
325335.
8 M. A. Blajchman, R. Austin, F. Fernandez-Ra-
chubinski, W. P. Sheffield, Blood 1992, 80,
21592171.
9 L.-O. Andersson, L. Engman, E. Henningson,
J. Immunol. Methods 1977, 14, 271281.
10 M. Z. Kounnas, F. C. Church, W. S. Argraves,
D. K. Strickland, 1996, 271, 65236529.
11 S. V. Pizzo, Am. J. Med. 1989, 87 (Suppl. 3B),
10S.
12 O. Egeberg, Thromb. Diath. Hemorrh. 1965,
14, 473489.
13 E. Thaler, K. Lechner, Clin. Haematol. 1981,
10, 369390.
14 K. Lechner, P. A. Kyrle, Thromb. Haemost.
1995, 73, 340348.
15 H. H. van Boven, D. A. Lane, Semin. Hematol.
1997, 34, 188204.
16 K. Ishiguro, T. Kojima, K. Kadomatsu, Y. Na-
kayama, A. Takagi, M. Suzuki, N. Takeda, M.
Ito, K. Yamamoto, T. Matsushita, K. Kusuga-
mi, T. Muramatsu, H. Saito, J. Clin. Invest.
2000, 106, 873878.
17 D. Menache, J. P. OMalley, J. B. Schorr, B.
Wagner, C. Williams & The Cooperative Study
Group, Blood 1990, 75, 3339.
18 H. R. Buller, J. W. ten Cate, Am. J. Med. 1989,
87(3B), 44S48S.
19 E. F. Mammen, Semin. Thromb. Haemost.
1995, 32, 26.
20 W. F. Jr. Baker, R. L. Bick, Semin. Thromb. Hae-
most. 1999, 25, 387406.
21 N. Smith-Erichsen, A. O. Aasen, M. J. Galli-
more, E. Amundsen, Circ. Shock. 1982, 9,
491497.
22 R. F. Wilson, A. Farag, E. F. Mammen, Y. Fujii,
Am. Surg. 1989, 55, 450456.
23 F. Fourrier, C. Chopin, J. Goudemand, S.
Hendrycx, C. Caron, A. Rime, A. Marey, P.
Lestavel, Chest 1992, 101, 816823.
24 R. F. Wilson, E. F. Mammen, J. G. Tyburski,
K. M. Warsow, S. M. Kubinec, J. Trauma 1996,
40, 384387.
25 J. T. Owings, M. Bagley, R. Gosselin, D. Ro-
mac, E. Disbrow, J. Trauma 1996, 41, 396405.
26 B. Blauhut, S. Necek, H. Vinazzer, H. Berg-
mann, Thromb. Res. 1982, 27, 271278.
27 F. Fourrier, C. Chopin, J. J. Huart, I. Runge,
C. Caron, J. Goudemand, Chest 1993, 104,
882888.
28 B. Eisele, M. Lamy, L. G. Thijs, H. O. Kei-
necke, H. P. Schuster, F. R. Matthias, F. Four-
rier, H. Heinrichs, U. Delvos, Intensive Care
Med. 1998, 24, 663672.
References 1017
29 F. Baudo, T. M. Caimi, F. de Cataldo, A. Raviz-
za, S. Arlati, G. Casella, D. Carugo, G. Palare-
ti, C. Legnani, L. Ridolfi, R. Rossi, A. DAnge-
lo, L. Crippa, D. Giudici, G. Gallioli, A. Wol-
fler, G. Calori, Intensive Care Med. 1998, 24,
336342.
30 H. Vinazzer, Semin. Thromb. Haemost. 1999,
25, 257263.
31 B. L. Warren, A. Eid, P. Singer, S. S. Pillay, P.
Carl, I. Novak, P. Chalupa, A. Atherstone, I.
Penzes, A. Kubler, S. Knaub, H. O. Keinecke,
H. Heinrichs, F. Schindel, M. Juers, R. C.
Bone, S. M. Opal, JAMA 2001, 286, 1869
1878.
32 D. Rublee, S. M. Opal, W. Schramm, H. O.
Keinecke, S. Knaub, Crit. Care 2002, 6, 349356.
33 S. M. Opal, Crit. Care Med. 2000, 28 (9 Suppl.),
S34S37.
34 F. Fourrier, M. Jourdain, A. Tournoys, Crit.
Care Med. 2000, 28 (9 Suppl.), S38S43.
35 S. M. Opal, C. M. Kessler, J. Rmisch, S.
Knaub. Crit. Care Med. 2002, 30 (Suppl.),
S325S331.
36 J. P. Simons, M. McClenaghan, A. J. Clark, Na-
ture 1987, 328, 530532.
37 K. Gordon, E. Lee, J. A. Vitale, A. E. Smith, H.
Westphal, L. Hennighausen, Biotechnology
(NY) 1987, 5, 11831187.
38 H. Meade, L. Gates, E. Lacy, N. Lonberg, Bio-
technology (NY) 1990, 8, 443446.
39 A. J. Clark, J. Mammary Gland Biol. Neopl.
1998, 3, 337350.
40 H. M. Meade, Y. Echelard, C. A. Ziomek, M. W.
Young, M. Harvey, E. S. Cole, S. Groet, T. E.
Smith, J. M. Curling, Gene Expression Systems:
Using Nature For The Art of Expression, Aca-
demic Press, USA, 1998.
41 D. P. Pollock, J. P. Kutzko, E. Birck-Wilson,
J. L. Williams, Y. Echelard, H. M. Meade, J. Im-
munol. Methods 1999, 137, 147157.
42 L. M. Houdebine, Transgenic Res. 2000, 9, 305
320.
43 E. A. Maga, J. D. Murray, Biotechnology (NY)
1995, 13, 14521457.
44 A. L. Archibald, M. McClenaghan, V. Hornsey,
J. P. Simons, A. J. Clark, Proc. Natl. Acad. Sci.
USA 1990, 87, 51785182.
45 J. Denman, M. Hayes, C. ODay, T. Edmunds,
C. Bartlett, S. Hirani, K. M. Ebert, K. Gordon,
J. M. McPherson, Biotechnology (NY) 1991, 9,
839843.
46 B. Roberts, P. DiTullio, J. Vitale, K. Hehir, K.
Gordon, Gene 1992, 121, 255262.
47 S. C. Bock, K. L. Wion, G. A. Vehar, R. M.
Lawn, Nucleic Acids Res. 1982, 10, 81138125.
48 K. M. Ebert, J. P. Selgrath, P. DiTullio, J. Den-
man, T. E. Smith, M. A. Memon, J. E. Schind-
ler, G. M. Monastersky, J. A. Vitale, K. Gordon,
Biotechnology (NY) 1991, 9, 835838.
49 A. S. Carver, M. A. Dalrymple, G. Wright, D. S.
Cottom, D. B. Reeves, Y. H. Gibson, J. L. Keen-
an, J. D. Barrass, A. R. Scott, A. Colman, et al.,
Biotechnology (NY) 1993, 11, 12631270.
50 J. Williams, F. A. Ponce de Leon, P. Midura,
M. Harrington, H. Meade, Y. Echelard, Therio-
genology 1998, 49, 398.
51 T. Edmunds, S. M. Van Patten, J. Pollock, E.
Hanson, R. Bernasconi, E. Higgins, P. Mana-
valan, C. Ziomek, H. Meade, J. M. McPherson,
E. S. Cole Blood 1998, 91, 45614571.
52 P. Hellstern, U. Mober, M. Ekblad, C. U. An-
ders, B. Faller, S. Muller, Haemostasis 1995,
25, 193201.
53 S. M. Van Patten, E. Hanson, R. Bernasconi,
K. Zhang, P. Manavalan, E. S. Cole, J. M.
McPherson, T. Edmunds, J. Biol. Chem. 1999,
274, 1026810276.
54 H. A. Schreuder, B. de-Boer, R. Dijkema, J.
Mulders, H. J. Theunissen, P. D. Grootenhuis,
W. G. Hol, Nat. Struct. Biol. 1994, 1, 4854.
55 B. Nordenman, C. Nystrom, I. Bjork, Eur. J.
Biochem. 1977, 78, 195203.
56 B. Turk, I. Brieditis, S. C. Bock, S. T. Olson, I.
Bjork, Biochemistry 1997, 36, 66826691.
57 J. Swedenborg, Blood Coagul. Fibrinolysis 1998,
9 (Suppl. 3), S7S10.
58 L.-E. Franzen, S. Svensson, J. Biol. Chem.
1980, 255, 50905093.
59 T. Mizouchi, J. Fujii, K. Kurachi, A. Kobata,
Arch. Biochem. Biophys. 1980, 203, 458465.
60 G. Zettlmeissl, H. S. Conrad, M. Nimtz, H. E.
Karges, J. Biol. Chem. 1989, 264, 2115321159.
61 B. Fan, B. C. Crews, I. V. Turko, J. Choay, G.
Zettlmeissl, P. Gettins, J. Biol. Chem. 1993,
268, 1758817596.
62 E. Munzert, J. Muthing, H. Buntemeyer, J.
Lehmann, Biotechnol. Prog. 1996, 12, 559563.
63 E. Ersdal-Badju, A. Lu, X. Peng, V. Picard, P.
Zendehrouh, B. Turk, I. Bjork, S. T. Olson,
S. C. Bock, Biochem. J. 1995, 310, 323330.
64 S. T. Olson, A. M. Frances-Chmura, R. Swan-
son, I. Bjork, G. Zettlmeissl, Arch. Biochem.
Biophys. 1997, 341, 212221.
65 H. Biescas, M. Gensana, J. Fernandez, P. Ris-
tol, M. Massot, E. Watson, F. Vericat, Haema-
tologica 1998, 83, 305311.
1018
66 J. L. Meagher, J. M. Beechem, S. T. Olson,
P. G. W. Gettins, J. Biol. Chem. 1998, 273,
2328323289.
67 D. H. Perlmutter, G. I. Glover, M. Rivetna,
C. S. Schasteen, R. J. Fallon, Proc. Natl. Acad.
Sci. USA 1990, 87, 37533757.
68 T. W. Barrowcliffe, C. A. Eggleton, M. Mah-
moud, Br. J. Haematol. 1983 55, 3746.
69 P. C. Cooper, S. Cooper, S. Kitchen, M. Mak-
ris, J. Thromb. Haemost. 2003, 1 (Suppl. 1),
1016.
70 K. Okajima, Immunol. Rev. 2001, 184, 258274.
71 V. De Palo, C. Kessler, S. M. Opal, Adv. Sepsis
2001, 1, 114124.
72 J. Rmisch, E. Gray, J. N. Hoffmann, C. J. Wie-
dermann, Blood Coagul. Fibrinolysis 2002, 13,
657670.
73 C. J. Wiedermann, J. Rmisch, Acta Med. Aus-
triaca 2002, 29, 8992.
74 M. Uchiba, K. Okajima, K. Murakami, H.
Okabe, K. Takatsuki, Thromb. Res. 1995, 80,
201208.
75 L. Ostrovsky, R. C. Woodman, D. Payne, D.
Teoh, P. Kubes, Circulation 1997, 96, 2302
2310.
76 S. Dunzendorfer, N. C. Kaneider, A. Rabenstei-
ner, C. Meierhofer, C. Reinisch, J. Rmisch,
C. J. Wiedermann, Blood 2001, 97, 10791085.
77 N. C. Kaneider, P. Egger, S. Dunzendorfer,
C. J. Wiedermann, Biochem. Biophys. Res. Com-
mun. 2001, 287, 4246.
78 C. Feistritzer, N. C. Kaneider, D. H. Sturn, C. J.
Wiedermann, Clin. Exp. Allergy 2004, 34, 696
703.
79 N. C. Kaneider, E. Frster, B. Mosheimer,
D. H. Sturn, C. J. Wiedermann, Thromb. Hae-
most. 2003, 90, 11501157.
80 D. Gritti, A. Malinverno, C. Gasparetto, C. J.
Wiedermann, G. Ricevuti, Int. J. Immuno-
pathol. Pharmacol. 2004, 17, 2732.
81 B. Risberg, Blood Coagul. Fibrinolysis 1998, 9
(Suppl. 3), S3S6.
82 F. B. Taylor, T. E. Emerson, R. Jordan, A. K.
Chang, K. E. Blick. Circ. Shock. 1988, 26, 227
235.
83 G. Dickneite, E. P. Paques, Thromb. Haemost.
1993, 69, 98102.
84 T. E. Emerson, Blood Coagul. Fibrinolysis 1994,
5 (Suppl. 1), S37S45.
85 C. M, Kessler, Z. Tang, H. M. Jacobs, L. M. Szy-
manski, Blood 1997, 89, 43934401.
86 G. Dickneite, Semin. Thromb. Hemost. 1998,
24, 6169.
87 M. Uchiba, K. Okajima, K. Murakami,
Thromb. Res. 1998, 89, 233241.
88 G. Dickneite, B. Leithaser, Arterioscler.
Thromb. Vasc. Biol. 1999, 19, 15661572.
89 G. Dickneite, M. Kroez, Blood Coagul. Fibri-
nolysis 2001, 12, 19.
90 R. Nevire, A. Tournoys, S. Mordon, X.
Marchal, F. Song, M. Jourdain, F. Fourrier,
Shock 2001, 15, 220225.
91 M. C. Minnema, A. C. Chang, P. M. Jansen,
Y. T. Lubbers, B. M. Pratt, B. G. Whittaker,
F. B. Taylor, C. E. Hack, B. Friedman, Blood
2000, 95, 11171123.
92 J. N. Hoffmann, B. Vollmar, D. Inthorn,
F. W. Schildberg, M. D. Menger, Am. J. Phys-
iol. Cell Physiol. 2000, 279, C98C107.
93 J. N. Hoffmann, B. Vollmar, J. Rmisch, D.
Inthorn, F. W. Schildberg, M. D. Menger,
Crit. Care Med. 2002, 30, 218225.
94 K. Murakami, R. McGuire, R. A. Cox, J. M.
Jodoin, F. C. Schmalstieg, L. D. Traber, H. K.
Hawkins, D. N. Herndon, D. L. Traber, Crit.
Care Med. 2003, 31, 577583.
95 K. Murakami, P. Enkhbaatar, R. A. Cox,
H. K. Hawkins, L. D. Traber, D. L. Traber,
Am. J. Physiol. (in press)
96 P. J. Cowan, A. Aminian, H. Barlow, A. A.
Brown, K. Dwyer, R. J. A. Filshie, N. Fisicaro,
D. M. A. Francis, H. Gock, D. J. Goodman, J.
Katsoulis, S. C. Robson, E. Salvaris, T. A.
Shinkel, A. B. Stewart, A. J. F. dApice, Am. J.
Transplant. 2002, 2, 520525.
97 W. Lu, T. G. K. Mant, J. H. Levy, J. M. Bailey,
Anesth. Analg. 2000, 90, 531534.
98 W. Dietrich, M. Spannagl, W. Schramm, W.
Vogt, A. Barankay, J. A. Richter, J. Thorac.
Cardiovasc. Surg. 1991, 102, 505514.
99 A. H. Sabbagh, G. K. T. Chung, P. Shuttle-
worth, B. J. Applegate, W. Gabrhel. Ann.
Thorac. Surg. 1984, 37, 466468.
100 W. Dietrich, A. Schroll, A. Barankay, J. A.
Richter, Anesthesiology 1985, 63, 3A.
101 K. Hashimoto, M. Yamagishi, T. Sasaki, M.
Nakano, H. Kurosawa, Ann. Thorac. Surg.
1994, 58, 799804.
102 M. R. Williams, A. B. DAmbra, J. R. Beck,
T. B. Spanier, D. L. Morales, D. N. Helman,
M. C. Oz, Ann. Thorac. Surg. 2000, 70, 873
877.
103 J. H. Levy, G. J. Despotis, F. Szlam, P. Olson,
D. Meeker, A. Weisinger, Anesthesiology 2002,
96, 10951102.
References 1019
104 H. van Aken, G. Hartlage, E. Martin, J.
Motsch, D. E. Birnbaum, C. Schmidt, E. Ott,
R. O. Feneck, R. D. Latimer, A. Vuylsteke,
J. B. Streisand, J. H. Levy, G. J. Despotis, Eu-
ropean Association of Cardiothoracic Anesthe-
siologists Annual Meeting, Aarhus, Denmark
June 2000. Meeting abstract.
105 M. S. Avidan, J. H. Levy, H. van Aken, R. O.
Feneck, R. D. Latimer, E. Ott, E. Martin,
D. E. Birnbaum, J. Bonfiglio, G. J. Despotis.
Anesthesiology 2005, 102, 276284.
106 B. A. Konkle, K. A. Bauer, R. Weinstein, A.
Greist, H. E. Holmes, J. Bonfiglio, Transfu-
sion 2003, 43, 390394.
107 R. C. Tait, M. Morfini, I. D. Walker, M. Mak-
ris, G. Dolan, R. Menon, P. K. Noonan, J.
Frieling, J. Bonfiglio, Blood 2002, 100, Ab-
stract 3970.
108 R. C. Tait, B. A. Konkle, K. A. Bauer, J. Bonfi-
glio, G. Dolan, J. Frieling, A. Greist, H. E.
Holmes, M. Makris, M. Morfini, B. Noonan,
J. B. Streisand, R. Weinstein, I. D. Walker, Y.
Echelard, Ann. Hematol. 2003, 82 (Suppl. 1),
S84.
1020
Abstract
Today, recombinant protein production in-
volves many options. In addition to E. coli,
several yeast systems (see Part IV, Chapter
13), insect cells (see Part IV, Chapter 14),
different mammalian expression systems
(CHO, BHK, NS0, HKB11, PER.C6) (see
Part II, Chapter 3 and Part IV, Chapters 1
and 3) other alternative expression systems
are currently under development for the
production of biopharmaceuticals. These
include transgenic animals or plants, and
will be discussed in Part IV, Sub-Part 2 of
this book. This chapter will focus on E.
coli, a still-modern secretory Saccharomyces
cerevisiae system, and the recently devel-
oped mammalian HKB11 expression sys-
tem. An E. coli host/vector system is de-
scribed that was originally developed for
the efficient production of an interleukin-4
variant. Later, it transpired that this system
is ideally suited to the expression of other
proteins and Fab fragments. The secretory
S. cerevisiae system has long been known
in biotechnology, but remains highly at-
tractive; the expression of a protease inhib-
itor will be presented as an example. Re-
combinant human glycoprotein therapeu-
tics should at best have structural identity
with the natural product. A hybrid clone,
designated HKB11 (hybrid of human kid-
ney and B cells) is a favorable cell host for
the production of human biopharmaceuti-
cal proteins. The host/vector system sup-
ports the production of gram quantities of
proteins in a large-scale transient transfec-
tion format, as well as the development of
stable cell lines. These systems, together
with the baculovirus insect system, are
used routinely within BayerHealthCare
Pharma Biotechnology to produce biophar-
maceutical proteins for research purposes
[ultra high-throughput screening (uHTS)],
for POC (proof-of-concept) studies, and for
therapeutic applications.
1021
ISBN: 3-527-31184-X
12
Producing Modern Biopharmaceuticals:
The Bayer HealthCare Pharma Experience with a Range
of Expression Systems
Heiner Apeler
Alea Non Iacta Est Improving Established Expression Systems
Abbreviations
BPTI bovine pancreatic trypsin inhibitor
CAI codon adaptation index
cDNA complementary DNA
Dhfr dihydrofolate reductase
EBV Epstein-Barr virus
Fab fragment antigen binding
HAT hypoxanthine-aminopterin-thymi-
dine
HKB11 hybrid of human kidney and B
cells
IL interleukin
IL-2 SA interleukin 2-selective agonist
IPTG isopropyl-b-d-thiogalactopyrano-
side
MBP maltose-binding protein
MFa mating factor a
MTX methotrexate
PEG polyethyleneglycol
POC proof-of-concept
Rop repressor of primer
RSCU relative synonymous codon usage
uHTS ultra high-throughput screening
UV ultra violet
12.1
The Escherichia coli Expression Platform
Mature human interleukin-4 (IL-4) is com-
posed of 129 amino acids. IL-4 variants
that are able to block both IL-4 and IL-13
activities have been described [1]. These
antagonistic properties are regarded as
useful for the treatment of diseases which
involve T
H2
development and/or IgE pro-
duction [2]. For the set-up of a viable com-
mercial production system for an IL-4 vari-
ant (IL-4 v), a broad screening of host or-
ganisms and expression systems on a
small scale was conducted (data not
shown).
After this screening, it became clear that
intracellular expression of IL-4 v forming
inclusion bodies in E. coli was by far the
most suitable and productive system. With
many of the evaluated systems it was pos-
sible to express IL-4 v, but the yield was
generally rather low as compared to the E.
coli inclusion body expression system.
Therefore, it was clear that this system
should be optimized for the intracellular
expression of IL-4 v inclusion bodies in E.
coli.
The main criteria for an efficient and
safe expression system are:
a high product yield,
regulatable stable expression, and
stability of the expression vector.
The initial production process for Bay IL-4 v
relied upon a thermoinducible system in-
volving the kP
R
-promoter and the heat-
sensitive cI857 repressor [1]. In comparison
to a heat-induced expression system, an iso-
propyl-b-d-thiogalactopyranoside (IPTG) or
lactose-induced system has several advan-
tages (no heat shock proteins are induced,
modulation of induction is possible, and
E. coli can be grown at its temperature
optimum). Promoters that are inducible
through IPTG or lactose include the trc,
T5, and T7 promoters. All of these promo-
ters in different vector backgrounds were
tested for the expression of IL-4 v. In addi-
tion to the promoter, several other features
of an expression plasmid are important for
the criteria listed above. These are:
ribosomal binding site (rbs),
codon usage of the corresponding gene,
transcriptional terminator,
resistance gene,
regulation of expression, and
origin of replication (ori).
Expression plasmids for IL-4 v with modi-
fications in all of these elements were gen-
12 Producing Modern Biopharmaceuticals 1022
erated during the course of the develop-
ment process. The quality and suitability
of the corresponding expression system
was ranked mainly according to five crite-
ria shown in Table 12.1.
The finally selected expression vector
pRO2.1.O, in combination with the host
cell E. coli W3110, ideally fulfills all crite-
ria. The expression vector pRO2.1.O con-
tains the following elements.
12.1.1
T5 Promoter
The E. coli phage T5 promoter, together
with two lac operator, sequences is derived
from the pQE30 plasmid belonging to the
pDS family of plasmids [3, 4].
12.1.2
T7 g10 Ribosomal Binding Site
The ribosomal binding site (rbs) is derived
from the region upstream from gene 10 of
the phage T7 (T7 g10 leader). Gene 10 of
phage T7 codes for the coat protein, which
is the major protein expressed after T7 in-
fection. The T7 g10 rbs was obtained from
the vector pET-9a [5]. It is one of the
strongest rbs known. The T7 g10 leader
spans a region of about 100 bp. In the fi-
nal construct pRO2.1.O the region up-
stream of the XbaI site is deleted [6]. Olins
et al. described that the section of mRNA
between the XbaI site within the T7 g10
leader and the initiator methionine codon
is sufficient for its function. The T7 g10
leader sequence in pRO2.1.O now spans
42 bp and harbors one base exchange from
G to A in position 16 (starting from the
XbaI site).
12.1.3
Codon Usage of the IL-4 v Gene
As an effective measure of synonymous
codon usage bias, the codon adaptation in-
dex (CAI), can be useful for predicting the
level of expression of a given gene [7, 8].
The CAI is calculated as the geometric
mean of relative synonymous codon usage
(RSCU) values corresponding to each of
the codons used in a gene, divided by the
maximum possible CAI for a gene of the
same amino acid composition. RSCU val-
ues for each codon are calculated from
very highly expressed genes of a particular
organism (e.g., E. coli), and represent the
observed frequency of a codon divided by
the frequency expected under the assump-
tion of equal usage of the synonymous co-
dons for an amino acid. Highly expressed
genes (e.g., genes encoding ribosomal pro-
teins) have generally high CAI values
Table 12.1 Criteria for evaluation of an E. coli inclusion body expression system
Criterion Experimental evaluation
Yield of IL-4 v SDS-PAGE and capillary gel electrophoresis
Plasmid stability Replica plating over 78 generations without antibiotic selection
Maintenance of induction
capability
Small-scale expression studies (SDS-PAGE analysis) over 78 gen-
erations without antibiotic selection
Performance of the fermentation
process
Fermentation at the 10-L scale
Performance of the renaturation
and purification process
Routine renaturation and purification of material derived from
10 L fermentation
0.46. Poorly expressed genes (e.g., lacI
and trpR in E. coli) have low CAI values
0.3 [7].
The calculated E. coli CAI value for the
natural IL-4 v gene is 0.733. This means
that the natural gene should be and in
fact is well-suited for high-level expres-
sion in E. coli. Nevertheless it was felt that
a synthetic gene with optimal E. coli codon
usage (CAI value=1) has the potential to
further increase the expression level.
Often, restriction sites are introduced
into the synthetic gene sequence to allow
for the assembly and cloning of small sec-
tions of the gene. In this case, however, in-
ternal restriction sites were omitted in or-
der to make no compromises regarding
the optimal codon usage. An NdeI and a
BamHI site were added to the 5' and 3'
ends of the synthetic gene, respectively.
The NdeI site includes the ATG start co-
don, and the BamHI site immediately fol-
lows the TGA stop codon.
12.1.4
Transcriptional Terminator
A T7 DNA fragment containing the tran-
scription terminator T} is derived from
the vector pET-9a [5]. Transcriptional termi-
nators determine the points where the
mRNA-RNA polymeraseDNA complex dis-
sociates, thereby ending transcription. The
presence of a transcriptional terminator at
the end of a highly expressed gene has sev-
eral advantages: they minimize sequester-
ing of RNA polymerase that might be en-
gaged in unnecessary transcription; they re-
strict the mRNA length to the minimal,
thus limiting energy expense; as strong
transcription may interfere with the origin
of replication, a transcriptional terminator
increases plasmid stability due to copy
number maintenance [9].
12.1.5
Resistance Gene
The kan resistance gene is derived from
the vector pET-9a [5]. Originally, this is the
kan gene of Tn903from the vector pUC4
KISS [10]. In the final vector pRO2.1.O the
kan gene and the IL-4 v gene have oppo-
site orientations, so there should be no in-
crease in the kan gene product after induc-
tion due to read-through transcription
from the T5 promoter. Kanamycin was
chosen as selective marker because it is
the preferred antibiotic for GMP-purposes.
In addition, kan gene-based vectors are
more stable than ampicillin-resistant (bla)
plasmids. Ampicillin selection tends to be
lost in cultures as the drug is degraded by
the secreted b-lactamase enzyme. In con-
trast to that, the mode of bacterial resis-
tance to kanamycin relies upon an amino-
glycoside phosphotransferase that inacti-
vates the antibiotic.
12.1.6
Regulation of Expression
Controlled gene expression is absolutely
necessary for the set-up of a stable plas-
mid system, particularly if the protein of
interest is deleterious to the host cell. The
expression vector pRO2.1.O uses a lac-
based inducible system consisting of a lac
repressor gene (lacI) and two synthetic lac
operator sequences fused downstream to
the E. coli phage T5 promoter. The lacI
q
promoter and the lacI structural gene were
isolated from the vector pTrc99A [11]. I
q
is
a promoter mutation which leads to over-
production of the lacI repressor. The wild-
type lac repressor is a tetrameric molecule
comprising four identical subunits of 360
amino acids each. The lac repressor tetra-
mer is a dimer of two functional dimers.
The four subunits are held together by a
four-helix bundle formed from residues
340 to 360. Due to the isolation of the lacI
gene from the vector pTrc99A by a NarI
restriction enzyme cut, the residues be-
yond amino acid 331 are deleted and 10
amino acids not normally encoded in the
lacI gene are added. It is known that mu-
tations or deletions that occur in the C-ter-
minal part of lacI, beyond amino acid 329,
result in functional dimers that appear
phenotypically similar to the wild-type re-
pressor [12].
12.1.7
Origin of Replication (ori)
The origin of replication (ori) of the ex-
pression plasmid pRO2.1.O is derived
from the vector pET-9a, the ori of which
originates from pBR322. pRO2.1.O there-
fore carries the pMB1 (ColE1) replicon.
Plasmids with this replicon are multicopy
plasmids that replicate in a relaxed fash-
ion. A minimum of 1520 copies of plas-
mid are maintained in each bacterial cell
under normal growth conditions. The ac-
tual number for pRO2.1.O has been deter-
mined to be 1425 copies per cell. Replica-
tion of pRO2.1O from its ColE1-type ori is
initiated by a 555-nucleotide RNA tran-
script, RNA II, which forms a persistent
hybrid with its template DNA near the ori.
The RNA IIDNA hybrid is then cleaved
by RNase H at the ori to yield a free 3
OH that serves as a primer for DNA poly-
merase I. This priming of DNA synthesis
is negatively regulated by RNA I, a 108-nu-
cleotide RNA molecule complementary to
the 5 end of RNA II. Interaction of the
antisense RNA I with RNA II causes a
conformational change in RNA II that in-
hibits binding of RNA II to the template
DNA and consequently prevents the initia-
tion of plasmid DNA synthesis. The bind-
ing between RNAs I and II is enhanced by
a small protein of 63 amino acids (the Rop
protein, Repressor of primer), which is en-
coded by a gene located 400 nucleotides
downstream from the origin of replication
[13, 14]. Deletion of the rop gene leads to
an increase in copy number and, due to a
gene dosage effect, to enhanced expression
levels of the plasmid-encoded heterologous
gene. This observation was also made for
the IL-4 v expression vectors tested. How-
ever, it transpired that rop
-plasmids are
instable and lost very rapidly during fer-
mentation under non-selective conditions.
Therefore, the replicon of pRO2.1.O con-
tains the rop gene to ensure high plasmid
stability.
The vector pRO2.1.O lacks the mob gene
that is required for mobilization and is
therefore incapable of directing its own
conjugal transfer from one bacterium to
another. The nic/bom site is present in
pRO2.1.O.
In the process of cloning of pRO2.1.O, all
elements not necessary for plasmid replica-
tion, resistance and regulatable expression
were deleted. Therefore, pRO2.1.O can be
termed a cleaned-up vector. This is impor-
tant for regulatory purposes.
With this expression system (W3110/
pRO2.1.O) very high IL-4 v yields of
10.2 g L
1
and a specific IL-4 v content of
248 mg g
1
cell dry weight are obtained
under high cell density conditions. With
other IL-4 v expression vectors these yields
were never achieved (see Table 12.2). Table
12.2 also summarizes the results from the
plasmid stability tests. The vector pRO2.1.O
remains fully stable over 78 generations
without antibiotic selection (Fig. 12.1). In
contrast, the expression vector pRO21.1.O,
which is based on the pET-30a plasmid, is
lost very rapidly. Only 16% of the colonies
contain the plasmid after 78 generations
without kanamycin selection. In addition,
it is very important that the capability of
Table 12.2 Important features of selected IL-4 v expression plasmids
Vector Remarks Strain Plasmid
stability
a)
Expression
level
[mg IL-4 g
1
dry weight]
Fermentation
yield
[g L
1
]
b)
Expression
before
induction
pW4A2 kP
R
-promoter JM103 recA
-
++++ 86 0.2 low cell
density
complex
medium
No
pAD49 T7-promoter
rop
+
pET-9a
W3110 (DE3) ++++ 85 4.4 No
pRO17.4.M T7-promoter
rop
cleaned up
c)
W3110 (DE3) +++ 230 6.3 Yes
pRO17.12.M T7-promoter
lacIq rop
+
cleaned up
W3110 (DE3) +++ 140 3.7 Yes
pRO21.1.O T7-promoter
lacIq rop
+
W3110 (DE3) ++ 209 7.2 No
pRO2.1.O T5-promoter
lacIq rop
+
cleaned up
W3110 ++++ 248 10.2 No
a) ++++ 100% stability after 80 generations; +++ 80100% stability
after 80 generations; ++ 6080% stability after 80 generations.
b) Fed-batch fermentation, mineral medium, 10-L scale.
c) All elements not necessary for plasmid replication and IL-4 v
expression removed.
Fig. 12.1 Plasmid stability of the IL-4 v expression vectors
pRO21.1.O (T7-promoter based) and pRO2.1.O (T5 promoter-
based). For details, see Table 12.2.
regulatable induction and high yield expres-
sion is maintained under non-selective con-
ditions. During development of the IL-4 v
expression vector, it became clear that some
of the T7-promoter based plasmids are very
stable (see for example the vector pAD49 in
Table 12.2), but that the ability of expression
of the desired protein is lost very rapidly.
This is due to mutations in the T7 RNA
polymerase gene within the DE3 prophage
(H. Wehlmann, unpublished data). With
the vector pRO2.1.O, expression of IL-4 v
as well as regulatable induction with the
inductor IPTG, are very stable under non-
selective conditions. After 78 generations
without kanamycin selection the expression
system is as effective as a fresh culture
(Fig. 12.2).
In conclusion, with the development of
the W3110/pRO2.1.O expression system, a
viable production system for an IL-4 variant
has been established. In addition, the newly
developed host/vector system is ideally sui-
ted for the expression of other cytokines
and cytokine muteins. The vector has been
adapted to harbor N- and C-terminal affinity
tags as well as solubility tags such as mal-
tose-binding protein (MBP), NusA and
thioredoxin for the high-throughput pro-
duction of proteins derived from genomics
and proteomics approaches. Data have also
been acquired which show that the system
can be used for the periplasmic expression
of Fab-fragments. In our hands, this system
is the one of choice for the expression of
many different proteins.
12.2
The Saccharomyces cerevisiae Expression
Platform
Aprotinin which is also known as bovine
pancreatic trypsin inhibitor (BPTI) be-
longs to the family of Kunitz-type inhibi-
tors, and inhibits serine proteases such as
trypsin, chymotrypsin, plasmin, and plas-
ma kallikrein [15]. Aprotinin consists of
58 amino acids. The aprotinin variant
(DesPro(2)-Ser(10)-Arg(15)-Ala(17)-Asp(24)-
Thr(26)-Glu(31)-Asn(41)-Glu(53)-aprotinin)
was designed by means of rational muta-
genesis, and differs from aprotinin by two
amino acids in the active site and by seven
amino acids in the backbone. The changes
in the active site of the aprotinin variant
increase the potency towards inhibition of
plasma kallikrein, whereas the inhibition
of plasmin is only marginally reduced.
Fig. 12.2 Maintenance of induction and expres-
sion capability of the IL-4 v expression vector.
SDS-PAGE (15%) analysis of total extracts from
the E. coli/pRO2.1.O strain grown for different
numbers of generations without antibiotic selec-
tion after 5 h of induction with 0.5 mM IPTG.
The gel was run under reducing conditions and
stained with Coomassie brilliant blue. Lane 1:
Molecular weight marker (Low Range, Life Tech-
nologies), lane 2: IL-4 v (5 lg); lane 3: E. coli/
pRO2.1.O (eight generations) before induction;
lane 4: eight generations after induction; lane 5:
18 generations after induction; lane 6: 28 genera-
tions after induction; lane 7: 38 generations after
induction; lane 8: 48 generations after induction;
lane 9: 58 generations after induction; lane 10:
68 generations after induction; lane 11: 78 genera-
tions before induction; lane 12: 78 generations
after induction; lane 13: molecular weight marker
(Low Range, Life Technologies); lane 14: molecu-
lar weight marker (High Range, Life Technolo-
gies).
Expression of recombinant aprotinins
has been achieved in E. coli K12 as a fu-
sion with parts of the MS-2 polymerase
gene [16]. In this case the fusion protein is
deposited as inclusion bodies. Functionally
active aprotinin can be obtained after solu-
bilization and purification of the fusion
protein, CNBr (cyanobromide) cleavage
and renaturation of the aprotinin moiety.
Low-level periplasmic expression of native,
properly folded aprotinin has been shown
in E. coli employing the E. coli alkaline
phosphatase signal sequence [17]. With re-
spect to expression level and ease of purifi-
cation, it transpired that secretory expres-
sion in the yeast Saccharomyces cerevisiae is
by far the most attractive system for the
production of this aprotinin variant [18].
In addition, due to the absence of an N-
glycosylation site there are no problems
with non-human glycosylation of the pro-
tein in yeast.
The backbone of the yeast expression
vector is derived from the commercially
available vector pYES2 (Invitrogen). The
GAL1 promoter and f1 ori region were re-
moved from this vector and replaced by
the S. cerevisiae mating factor a1 (MFa1)
promoter and part of the MFa1 precursor
sequence (the MFa1 pre-pro-sequence).
The DesPro2-Ser10-Arg15-Ala17-Asp24-
Thr26-Glu31-Asn41-Glu53 aprotinin vari-
ant cDNA was fused to the S. cerevisiae a-
mating factor pre-pro-signal sequence. The
yeast expression vector for the aprotinin
variant and a detailed description of the
MFa1 leader processing is shown in
Fig. 12.3. Processing of the pre-pro-a-factor
aprotinin variant fusion construct requires
two different proteolytic activities. A signal
peptidase, localized in the endoplasmic re-
ticulum, first cleaves between the pre and
the pro-sequence. The glycosylated pro-a-
factor aprotinin variant is subsequently
cleaved by an endoproteinase at the car-
boxyl side of the dibasic sequence Lys-Arg,
thereby releasing the aprotinin variant into
the supernatant. This protease is the prod-
uct of the KexII gene [19].
Yeast cells transformed with the corre-
sponding vector are capable of secreting
large amounts of correctly processed apro-
Fig. 12.3 Yeast expression vector for the DesPro2-Ser10-Arg15-Ala17-Asp24-Thr26-
Glu31-Asn41-Glu53 aprotinin variant and processing of the MFa1 leader peptide.
tinin variant into the supernatant. In shak-
ing-flask cultures, yields between 20 and
40 mg L
1
can be obtained by employing
the MFa1 promoter. Fed-batch fermenta-
tions at the 10-L scale resulted in product
concentrations in the supernatant of 150
230 mg L
1
.
Aprotinin variants with a deletion of the
amino acid proline in position 2 (DesPro2)
can be effectively produced using the de-
scribed system.
Expression of recombinant aprotinin
variants with the natural N-terminal se-
quence Arg-Pro-Asp is not possible be-
cause the KexII protease is unable to
cleave off the MFa1 pre-pro signal peptide
in the corresponding context.
Aprotinin variants with the natural N-
terminal sequence Arg-Pro-Asp can be
secreted from yeast by completely remov-
ing the a-factor-pro sequence and fusing
the aprotinin variant gene directly to the a-
factor pre sequence. In this construct, the
processing of the a-factor aprotinin fusion
only depends on the signal peptidase and
avoids the classical KexII protease activ-
ity. Large amounts of homogeneously pro-
cessed aprotinin variants can be obtained
using this approach.
In conclusion, the described yeast sys-
tem is well suited for the secretory expres-
sion of small disulfide-bonded, non-glyco-
sylated proteins. It should be emphasized
that the performance of yeast cells for ex-
pression of a particular protein can be
further optimized by conventional strain
development programs using UV-irradia-
tion or treatment with mutagenic com-
pounds.
12.3
The HKB11 Expression Platform
HKB11 is a human/human hybrid cell
line that was generated by PEG-fusion of
human embryonic kidney cells (293S) and
Burkitts lymphoma cells (2B8). 2B8 is a
G418-resistant, HAT (hypoxanthine-ami-
nopterin-thymidine)-sensitive clone derived
from HH514-16. HH514-16 harbors a
non-transforming and heterogeneous
Chet/DNA-free EBV (Epstein-Barr virus) ge-
nome that does not produce EBV particles.
Generation of the cell line has been de-
scribed in detail [20]. The initial purpose
of establishing a hybrid cell line of 293S
and a human suspension B-cell line (2B8)
was to resolve the aggregation problem of
293S cells, which tend to clump when
grown in suspension culture. The HKB11
cell line was selected for non-aggregating
properties. Adaptation of HKB11 cells to se-
rum-free suspension conditions is readily
possible. Compared to 293S cells, HKB11
cells form smaller and looser aggregates un-
der serum-free suspension conditions.
Although the transformation or hybridiza-
tion events, required in most cases to pro-
duce stable cell lines, may result in altered
glycosylation profiles, it was found that
HKB11 cells possess typical human glycosy-
lation enzymes such as alpha(2,3) and al-
pha(2,6) sialyl transferases. Moreover, pro-
teins produced from transfected HKB11
cells were found to be capped with sialic
acid of alpha(2,3) and alpha(2,6) linkages.
The characteristics of the HKB11 cell line
are summarized in Table 12.3.
Due to the expression of EBNA1,
HKB11 cells are an excellent cell host for
oriP expression vectors. Expression of an
IL-2 selective agonist (IL-2 SA) using an
oriP episomal vector (pCEP4 from Invitro-
gen) has been described [21]. The pCEP4
vector harbors (besides oriP and EBNA1) a
functional hygromycin resistance gene and
a CMV enhancer/promoter. Expression
was further optimized by incorporating
the HIV transactivator Tat and its response
element TAR into this vector [21]. A syner-
gistic effect was observed on the expres-
sion of IL-2 SA when Tat/TAR and oriP/
EBNA1 elements were present in a single
vector employing a transient transfection
approach.
Tat and TAR sequences were subcloned
into different TAR-oriP and Tat-oriP vec-
tors due to the large plasmid size of the
Tat/TAR-oriP vector (12.1 kbp) and to
further optimize the cloning procedure by
using smaller-sized plasmids [22]. Methods
have been developed to use these vectors
for the production of milligram and gram
quantities of proteins and recombinant
antibody candidates in a large-scale transi-
ent transfection scheme. Usually, transfec-
tion is performed with 500 lg of plasmid
DNA containing the gene of interest and
DMRIE-C transfection reagent (Invitrogen)
in a 500-mL suspension culture of HKB11
cells. At 3 days post transfection, the cul-
ture from the shake-flask is transferred to
a Cellbag (Wave Biotech LLC) and the cul-
ture volume sequentially increased to 10 L.
At 10 days post transfection, the cells or
supernatant are harvested for purification
of proteins [22]. Under hygromycin B drug
selection this process can be extended, and
gram quantities of proteins and recombi-
nant antibodies can be obtained in a mat-
ter of weeks.
Although HKB11 is not a dhfr (dihydro-
folate reductase)-negative cell line, stable
clones can be generated in a selection me-
dium containing 100 nM MTX (methotrex-
ate). Amplification of the gene cassette
with increasing concentrations of MTX is
possible comparable to the CHO dhfr
-
sys-
tem.
In conclusion, the described HKB11
host/vector system supports the rapid pro-
duction of milligram and gram quantities
of proteins and recombinant antibodies in
a large-scale transient transfection format
to generate material for proof of concept
Table 12.3 Comparison of HKB11 with parental cell lines
Characteristics 293S 2B8 HKB11
Chromosome number 64 47 90110
EBV genome Negative Positive Positive
EBNA Positive Positive
VCA ND Negative
VCA after transfection with BZLF1 ND Positive (0.2% of cell population)
EBV transformation Negative ND
het DNA Negative ND
Surface Ig-l Negative Positive Negative
Aggregates Large (tight) Small Small (loose)
Sialyltransferase a(2,3), a(2,6) ND a(2,3), a(2,6)
MSX sensitivity (lM) ND ND 150
MTX sensitivity (nM) ND ND 50
ND = not determined; EBNA = Epstein-Barr virus nuclear anti-
gen; VCA = EBV capsid antigen; BZLF1 = EBV latency-interrupt-
ing gene involved in the viral lytic cycle; MSX = methionine sul-
foximine; MTX = methotrexate.
(POC) studies. In addition, the HKB11 ex-
pression system can be used to develop
cell lines for stable production and clinical
manufacturing of biopharmaceutical prod-
ucts.
12.4
Outlook and Conclusion
Although, during the past few years, mam-
malian cells have surpassed microbial sys-
tems for the production of biopharmaceu-
ticals, it is extremely important to have es-
tablished different expression platforms.
In addition to the systems described in
this chapter (E. coli inclusion body, secre-
tory S. cerevisiae and HKB11), we routinely
use E. coli for periplasmic expression of
Fab fragments, insect cells (only for re-
search purposes) and CHO cells in batch
and perfusion cultures. Based on experi-
ence with different classes of proteins, a
broad screening of all available expression
systems is in most cases not warranted.
Yeast (S. cerevisiae) is an option with rather
small secreted and non-glycosylated pro-
teins. The engineering of yeast systems
(Pichia, Hansenula) for human glycosyla-
tion or defined glycosylation (e.g., for tis-
sue targeting) will clearly boost the utiliza-
tion of this system for therapeutic applica-
tions (see Part IV, Chapter 7). E. coli (and
other emerging bacterial expression sys-
tems) can also in the future be the system
of choice, for example in the secretory ex-
pression of Fab fragments or single chain
antibodies. It may even be worthwhile con-
sidering inclusion body-based processes
for this type of biopharmaceutical.
Whereas small peptides are in most cases
produced by chemical synthesis, E. coli of-
fers the possibility for the production of
large amounts of peptides, and in the fu-
ture even peptides containing non-protei-
nogenic amino acids may be accessible.
Higher cell densities in bioreactors, engi-
neered host cells and the application of
high-throughput screening equipment for
the screening and selection of high-produ-
cer clones will be the driving force for
mammalian cells. Scaled-up approaches
for transiently transfected cells such as
the HKB11 system described here are
able to produce grams of recombinant pro-
tein and antibody in a relatively short time
frame. These systems may offer advan-
tages compared to the conventional proce-
dures, and may therefore become more ap-
plicable on a broader basis, perhaps with
respect to personalized medicine or the de-
velopment of other modern biopharmaceu-
ticals.
References
1 N. Kruse, H. P. Tony, W. Sebald, EMBO J.
1992, 11, 32373244.
2 J. J. Ryan, J. Allergy Clin. Immunol. 1997, 99,
15.
3 H. Bujard, R. Gentz, M. Lanzer, D. Stber, M.
Mller, I. Ibrahimi, M. T. Huptle, B. Dobber-
stein, Methods Enzymol. 1987, 155, 416433.
4 D. Stber, H. Matile, G. Garotta, in: I. Lefko-
vits and B. Pernis (Eds), Immunological Meth-
ods. Academic Press, Inc., 1990, Vol. IV,
pp. 121152.
5 F. W. Studier, A.H. Rosenberg, J. J. Dunn, J. W.
Dubendorff, Methods Enzymol. 1990, 185,
6089.
6 P. O. Olins, C. S. Devine, S. H. Rangwala, K. S.
Kavka, Gene 1988, 73, 227235.
7 P. M. Sharp, W.-H. Li, Nucleic Acids Res.
1987, 15, 12811295.
8 H. Apeler, U. Gottschalk, D. Guntermann, J.
Hansen, J. Mssen, E. Schmidt, K.-H. Schnei-
der, M. Schneidereit, H. Rbsamen-Waig-
mann, Eur. J. Biochem. 1997, 247, 890 895.
9 P. Balbas, F. Bolivar, Methods Enzymol. 1990,
185, 1437.
10 F. Barany, Gene 1985, 37, 111123.
11 E. Amann, B. Ochs, K.-J. Abel, Gene 1988, 69,
301315.
References 1031
12 H. C. Pace, M. A. Kercher, P. Lu, P. Markie-
wicz, J. H. Miller, G. Chang, M. Lewis, Trends
Biochem. Sci. 1997, 22, 334339.
13 J. Sambrook, E. F. Fritsch, T. Maniatis, Molec-
ular Cloning: A Laboratory Manual, 2nd edn.,
Cold Spring Harbor Laboratory, Cold Spring
Harbor, NY 1989, Vol. 1.
14 R. Lahijani, G. Hulley, G. Soriano, N. A. Horn,
M. Marquet, Hum. Gene Ther. 1996, 7, 1971
1980.
15 W. Gebhard, H. Tschesche, H. Fritz, in: A. J.
Barrett and G. S. Salvesen (Eds), Proteinase
Inhibitors. Elsevier Science Publ. BV 1986,
pp. 375387.
16 E.-A. Auerswald, D. Hrlein, G. Reinhardt,
W. Schrder, E. Schnabel, Biol. Chem. Hoppe
Seyler 1988, 369, Suppl., 2735.
17 C. B. Marks, M. Vasser, P. Ng, W. Henzel,
S. Anderson, J. Biol. Chem. 1986, 261 (16),
71157118.
18 H. Apeler, J. Peters, W. Schrder, K.-H.
Schneider, G. Lemm, V. Hinz, G.J. Rossouw,
K. Dembowsky, ArzneimForsch DrugRes
2004, 54 (8), 483497.
19 A. J. Brake, Methods Enzymol. 1990, 185,
408421.
20 M. S. Cho, H. Yee, S. Chan, J. Biomed.
Science 2002, 9, 631638.
21 M. S. Cho, H. Yee, C. Brown, K.-T. Jeang,
S. Chan, Cytotechnology 2001, 37, 2330.
22 M. S. Cho, H. Yee, C. Brown, B. Mei, C. Mi-
renda, S. Chan, Biotechnol. Prog. 2003, 19,
229232.
Abstract
Expression of insulin precursors in the
yeast Saccharomyces cerevisiae has in recent
years formed the basis for the manufac-
ture of the majority of the human recom-
binant insulin and insulin analogues sold
by Novo Nordisk A/S. In this chapter, we
describe the composition of the yeast ex-
pression system and review a strategy for
modification of the insulin precursors in
order to improve the expression system.
This strategy involves adaptation of the in-
sulin precursor for efficient secretion and
processing, and has highlighted the impor-
tance of addressing every event leading to
secretion of the insulin precursor from the
initial translocation and folding in the en-
doplasmic reticulum via transport through
the Golgi apparatus, to maturation by kex-
in and ultimately secretion via correct lo-
calization to secretory vesicles. The modifi-
cations of the insulin precursors described
in this chapter have led to significant im-
provements in the expression system, re-
sulting in robust scalable expression sys-
tems amenable to industrial fermentation
and downstream processing of the insulin
precursor.
Abbreviations
FLP family of lambda phage (FLIP)
POT plasmid encoded selection
SV secretory vesicles
YPS1 yapsin 1 aspartyl endoprotease
13.1
Introduction
13.1.1
Insulin and Diabetes
Insulin is a naturally occurring peptide
hormone produced by the b-cells in the
Langerhans islets of the pancreas in re-
sponse to hyperglycemia. Insulin facilitates
the entry of glucose into target tissues
such as muscle, adipose tissue and liver
by binding to and activating specific mem-
brane receptors on these cells. Diabetes
mellitus is a group of metabolic diseases
characterized by high blood sugar (glu-
cose) levels, which result from defects in
insulin secretion, or action, or both. In
Type 1 diabetes this may be due to b-cell
destruction, and in Type 2 diabetes to a
combination of b-cell failure and resistance
of target tissues to insulin action (insulin
resistance). The latter disease can, in its
1033
13
Advanced Expression of Biopharmaceuticals in Yeast
at Industrial Scale: The Insulin Success Story
ISBN: 3-527-31184-X
early stages, be helped by low calorie non-
sugar diet and/or treatment with oral anti-
diabetic drugs, while the later stages and
Type 1 diabetes require insulin treatment
(see also Part VI, Chapter 4). During the
first 60 years after the discovery of insulin
by Frederick Banting and Charles Best in
19211922, and their successful treatment
of diabetics, insulin extracted from bovine
or porcine pancreases was the drug avail-
able to treat Type 1 diabetics.
13.1.2
Human Insulin
With the rapid increase in the incidence of
diabetes among the population, it is no
longer possible to satisfy the pharmaceuti-
cal requirement (estimated to be 15
20 tonnes per year in 2005) from animal
sources. Furthermore, the animal-extracted
insulins are slightly different from human
insulin, which might cause formation of
insulin-binding antibodies and allergic re-
actions. Porcine insulin, which differs
from human insulin only by a single ami-
no acid in position B30, can be converted
to human insulin in a transpeptidation re-
action, in which an alanine is replaced
with a threonine [1].
13.1.3
Biosynthetic Human Insulin
Recent developments in molecular biology
and biotechnology have opened up new pos-
sibilities, among which is the biosynthesis
of human insulin. Insulin is composed of
two disulfide-linked peptide chains referred
to as the A chain and B chain. The first re-
combinant approach used Escherichia coli as
host and expression of A- and B-chains as
fusion-proteins in separate strains [2]. In a
later approach in E. coli, proinsulin (B-
chain-Connecting-peptide-A-chain) was ex-
pressed also as a fusion protein [3]. In both
of these systems the fusion-proteins were
isolated as inclusion bodies, and several
chemical steps were needed for dissolution,
cleavage, folding, and formation of disulfide
bridges. Later, Thim et al. [4] reported the
successful expression of single chain insu-
lin precursors with a mini C-peptide pro-
duced and secreted in the yeast Saccharo-
myces cerevisiae, and which containing the
correct disulfide bridges. Subsequently,
Markussen et al. [5] reported the successful
expression and conversion of other mini C-
peptide insulin precursors to human insu-
lin, with minimal post-fermentation chem-
istry, and purification of the product. The
S. cerevisiae insulin expression system [5]
could be scaled up for the stable large-scale
production of insulin [6].
13.1.4
Novel Administration Methods
and Insulin Analogues
With the introduction of novel routes for
administration of insulin, such as the in-
trapulmonary route (see Part VI, Chapter
4), where bioavailability is markedly lower
than with subcutaneous administration,
the demand for insulin will increase. The
demand for a more optimal treatment of
the patient has called for the design and
development of new fast- and slow-acting
insulin analogues [710] (see Introduction)
and has required alterations of the yeast
process. The development and optimiza-
tion of the S. cerevisiae expression system
for biosynthesis of human insulin has re-
cently been extensively reviewed [11, 12].
13.1.5
S. cerevisiae Host Expression System
The yeast S. cerevisiae (known as bakers
yeast) has played a major role in food pro-
13 Advanced Expression of Biopharmaceuticals in Yeast at Industrial Scale: The Insulin Success Story 1034
duction for several thousand years by its
use in wine, beer, and bread making. In
later years, S. cerevisiae has also been de-
veloped as an efficient eukaryotic expres-
sion system for biotechnology (for a re-
view, see [13]). S. cerevisiae secretes very
few endogenous proteins, which makes
the purification of secreted heterologous
proteins easier. In addition, it is robust
and well adapted to the physical stress in
large-scale cultivation.
Here, we review the possibilities for op-
timization of the production process for
human insulin and its analogues in a spe-
cific S. cerevisiae host system described be-
low and used by Novo Nordisk A/S.
The S. cerevisiae host cell (reviewed in
[12]) is a diploid with the genotype (MATa/
MATa pep4-3/pep4-3 leu2-3,112/leu2-3,112
HIS4/his4 tpi1::LEU2/tpi1::LEU2 cir
+
). The
pep4-3 mutation impairs the normal pro-
teolytic activities in the vacuole [14], and is
thought to improve heterologous protein
production. The deletion of TPI1 allows
autoselection of the POT expression S. cer-
evisiae-E. coli 2l-pBR322 shuttle plasmid
(Fig. 13.1) and the use of the strong con-
stitutive TPI1 promoter and terminator as
initiating and termination signals for the
transcription of DNA encoding heterolo-
gous proteins inserted between the two.
The non-transformed host strain grows
poorly, when glucose is the only carbon
source. Thus, transformation with the
POT expression plasmid allows selection
by the ability to grow on glucose [15, 16].
The Schizosaccharomyces pombe POT gene
is homologous to TPI1, but is poorly ex-
pressed in S. cerevisiae, and multiple co-
pies of the POT plasmid are therefore re-
quired to generate sufficient gene product
to allow growth on glucose as the sole car-
bon source [15, 16]. The POT plasmid copy
number is probably influenced by counter-
acting selective forces, including the need
for complementation of the host marker
(tpi1D) by the plasmid-encoded selection
gene (POT) and a selective advantage of
cells with a low copy number enabling
them to limit the burden of fusion protein
expression and secretion [12]. The 2l part
of the plasmid has been slightly modified
to inhibit recombination events. The FLP-
gene has been truncated and the inverted
Fig. 13.1 S. cerevisiae expression vector used for
production of recombinant insulin. The S. cerevi-
siae/E. coli shuttle vector is composed of the fol-
lowing genetic units: 1) The transcription promo-
ter and terminator of the S. cerevisiae TPI1 (triose
phosphate isomerase) gene flanking the DNA en-
coding the leader-insulin precursor. 2) The TPI1
gene from Schizosaccharomyces pombe (TPI1
p
)
used as a selectable marker. 3) Part of pBR322
containing E. coli origin of replication to ensure
replication in E. coli as well as the bacterial ampi-
cillin gene (Amp
R
). 4) Major parts of the 2l plas-
mid of S. cerevisiae (including the origin of replica-
tion) are included to ensure replication, amplifica-
tion and inheritable stability in yeast.
repeats mutated. The pBR322 part of the
shuttle vector contains the bla-gene adding
ampicillin resistance, when grown in E.
coli. Methods to delete this gene, which is
an environmental issue with a high public
concern, have been developed [17].
The host cell itself could be thought of as
a target for further mutations or other ge-
netic modifications that could improve its
secretory, folding and posttranslational
modification abilities. Deletion of specific
genes that enhance protein secretion has
been suggested [18, 19]; additionally, the
overexpression of pse1 [21] and sso2 [22] im-
proved the yield of other heterologous pro-
tein products. However, these observations
were related to the expression of recombi-
nant proteins unrelated to insulin. More
specifically, Kozlov et al. [20] found im-
provement of proinsulin expression in res-
piratory-deficient rho
S. cerevisiae strains.
13.2
Design and Optimization
of the Insulin Precursor Molecule
13.2.1
The Insulin Precursor Molecule: Background
The efficient secretion of heterologous pro-
teins from eukaryotic cells often requires
more than just a signal peptide N-termin-
ally fused to the appropriate protein.
Transport through the yeast secretory path-
way includes translocation through the en-
doplasmic reticulum (ER) membrane into
the ER-lumen, attachment of N-linked car-
bohydrate chains and folding and oxida-
tion of disulfides, transport from the ER to
the Golgi apparatus, post-translational
modifications in the Golgi apparatus,
transport by secretory granules to the cell
membrane, and finally exocytic exit to the
extracellular space. Especially small pep-
tides such as the insulin molecule require
molecular assistance for efficient synthesis
and secretion in S. cerevisiae. Proinsulin,
fused to the a-factor leader (see below),
was not readily expressed and secreted in
S. cerevisiae [4] (Fig. 13.2). However, this a-
leader fusion protein could be modified
for successful expression and secretion in
S. cerevisiae by substitution of insulins
long C-peptide with a dibasic amino acid
sequence, creating insulin precursors that
subsequently could be converted to human
insulin by the action of the enzymes tryp-
sin and carboxypeptidase B [4]. In a more
subtle, optimized form this insulin fusion
protein was changed by deletion of threo-
nine
B30
of the insulin precursor and lysi-
ne
B29
was connected to glycine
A1
either di-
rectly or using a short C-peptide contain-
ing a C-terminal lysine (e.g., AAK or SK)
and fused to the a-factor leader [5]. Such
single-chain insulin precursors, lacking
threonine
B30
, could be converted into hu-
man insulin by transpeptidation using por-
cine trypsin (see Section 13.3.2.2, Trans-
peptidation), but with differences in over-
all yields and conversion rates.
13.2.2
Optimization of the Insulin Precursor
In the following sections, optimization of
the individual segments constituting the
signal peptide-leader-insulin precursor will
be discussed (see Fig. 13.2).
13.2.2.1 Signal Peptide
Eukaryotic signal peptides have universal
properties, and many can be used as substi-
tutes for the a-factor signal peptide in di-
recting heterologous proteins to the secre-
tory pathway in yeast [13]. The yapsin 1 as-
partyl endoprotease (YPS1) signal peptide
has been shown to provide efficient secre-
tion of heterologous proteins (especially in-
sulin) in yeast in combination with a pro-
peptide [23]. The hydrophobicity of the sig-
nal peptide can be decisive for co- or post-
translational translocation of the pre-pro-fu-
sion protein into the ER [24], and a pre-
viously unrecognized diversity in the signal
peptides was found which carries specific
structural information that serves to identi-
fy the translocation route [25].
13.2.2.2 Leader Peptide
S. cerevisiae cells of mating type a secrete a
13-residue peptide pheromone (a-factor)
[26]. The a-factor is the product of the
MFa1 gene which encodes a 165-residue
polypeptide (pre-pro-a-factor) consisting of
a signal peptide, a leader and four repeats
of a pro-a-factor (see Fig. 13.2), which are
maturated during secretion [26, 27] (see
Part IV, Chapter 12). It was early recog-
nized that the S. cerevisiae a-factor signal-
leader, was able to confer secretory compe-
tence on proteins expressed in S. cerevisiae
[28, 29]. Although other leaders have been
found, the a-factor leader has become the
primary leader choice for secretory expres-
sion in S. cerevisiae, and also in alternative
yeast species [13].
The presence of a leader sequence has
been shown to be necessary for insulin pre-
cursor secretion, since a signal peptide
fused directly to the insulin precursor does
not result in secretion of insulin. In order
to explore the possibility of developing lea-
ders conferring greater secretory compe-
tency onto the insulin precursor, a stepwise
approach was initiated starting from a mini-
mal designed leader composed of a signal
peptidase site, an N-glycosylation site, and
a kexin processing site [30]. This minimal
Fig. 13.2 Schematic representation of pre-pro-a-
factor and pre-pro-insulin. The a-factor is ex-
pressed in multiple copies as a pre-pro-polypep-
tide precursor. Each a-factor is released by kexin
cleavage, followed by removal of the N-terminal
spacer peptide (E) by the dipeptidyl aminopepti-
dase Ste13p. Insulin is expressed in the b-cell as
pre-pro-insulin. The C-peptide of pro-insulin is re-
moved by the action of prohormone convertases.
Combining the a-factor pre-pro with pro-insulin
does not lead to secretion of pro-insulin from
yeast; however, the introduction of very short C-
peptides allows insulin precursor secretion. Kexin
cleavage can be optimized by N-terminal exten-
sions on the insulin precursor. S = Signal peptide;
L = leader; E = spacer peptide; B = insulin B-
chain; A = insulin A-chain (together denoted I in
the text); C = connecting peptide. Small arrows
denote the dibasic kexin cleavage sites.
leader allowed secretion of small amounts
of insulin precursor, though after extensive
optimization it was possible to identify a
number of synthetic leaders equivalent to
or better than the a-factor leader at facilitat-
ing export of the insulin precursor to the
culture supernatant [30, 31].
The three N-linked glycosylation sites of
the a-leader have been shown to be impor-
tant but not essential for the ability of
the a-factor leader to secrete a-factor [32,
33]. In addition, mutation of all three-con-
sensus sites for N-linked glycosylation de-
creased the quantity of secreted insulin
precursor to &10% [34]. In contrast to
what was found with the a-leader, elimina-
tion of the two consensus sites for N-
linked glycosylation in one of the newly
developed leaders, actually improved the
ability to facilitate secretion of the insulin
precursor [31].
13.2.2.3 Spacer Peptide
In the pro-a-factor (see Fig. 13.2), each 13-
residue a-factor is preceded by a dibasic
kexin processing site (KR) connected to a
spacer peptide with two to three dipeptidyl
repeats (EA or DA) designed for subse-
quent processing with the Ste13p dipepti-
dyl peptidase [26, 27, 35]. When the a-fac-
tor leader with spacer peptide initially was
employed for recombinant insulin expres-
sion, the efficiency of the dipeptidyl ami-
nopeptidase was shown to be a limiting
step in maturation of insulin [4]; therefore,
the spacer peptide was deleted and the
protein fused directly to the a-leader in
most of the initial studies. However, this
resulted in the secretion of substantial
amounts of uncleaved hyperglycosylated
leader-insulin precursor from the yeast cell
(see Fig. 13.3 for details of the secretion
process). The introduction of a spacer pep-
tide (EA)
3
K, resembling the a-factor spacer
peptide, but with a lysine (K) introduced
C-terminally to allow tryptic digestion at a
later step in the production process, signif-
icantly improved expression of the insulin
precursor [36]. One of the reasons for this
was an improved cleavage in the Golgi of
the fusion-protein by kexin, which allowed
more maturated insulin precursor to be se-
creted. However, the previously described
inefficiency of Ste13p resulted in the secre-
tion of a mixed population of insulin pre-
cursors, complicating recovery and conver-
sion to insulin. The addition of a glutamic
acid residue to the spacer peptide N-termi-
nus (to give E(EA)
3
K) solved this problem
by preventing processing by Ste13p. When
this secreted insulin precursor was ana-
lyzed it was found that the extension had
been cleaved off, resulting in the presence
of maturated insulin precursor. This
pointed to the presence of a specific pro-
tease activity [36], and characterization of
this proteolytic activity indicated that the
aspartyl endoprotease yapsin 1 (YPS1) [37]
was responsible for removing the spacer
peptide from the insulin precursor. A dele-
tion of yps1 is non-essential and solves the
problem [38]; however, an alternative solu-
tion was found in that the spacer peptide
could be further modified by insertion of
different amino acid residues N-terminally
to the lysine residue, thereby preventing
proteolytic cleavage by yapsin 1. A proline
residue was found to be most efficient to
inhibit yapsin 1 activity and at the same
time allow tryptic in vitro digestion C-ter-
minally to the lysine residue [36]. Different
modifications of the spacer peptide were
subsequently generated (e.g., EEGEPK)
which further increased the insulin precur-
sor fermentation yield, without losing the
in vivo stability and at the same time sup-
porting in vitro conversion to insulin [12].
The in vitro conversion of the single-chain
precursor that is required to obtain human
insulin can be an advantage in the recov-
ery and purification process, because this
conversion changes pI, hydrophobicity, hy-
drophilicity and the size of the peptide,
thus providing optimal conditions for the
purification of insulin that is, the re-
moval of impurities.
13.2.2.4 Connecting Peptide (C-peptide)
and Insulin Aspart
The primary optimized connecting mini-
peptides, which support secretion and in
vitro conversion, was briefly described
above. Further optimization in this region,
leading to improved expression, was ob-
tained by fixing K in the N-terminal posi-
tion to G
A1
, and using a C-peptide with 2
16 residues and the most flexible amino
acid glycine immediately N-terminal to K.
Later, it was shown that especially gluta-
mic or aspartic acid in C1 with G
C2
K
C3
results in an optimized version of a mini
C-peptide [40].
A special and interesting case has re-
cently been described [41], dealing with an
insulin analogue substituted in B28 (proline
to aspartic acid). This insulin analogue pre-
cursor can be converted to a fast-acting in-
sulin aspart analogue by transpeptidation
(see Section 13.3.2 and Fig. 13.4; see also
the Introduction). The insulin analogue
precursor is not well expressed with the
AAK or the B29-A1 peptide bond, as mini-
C-peptide and extractive refolding of the
precursor was studied as one solution to in-
crease yield [42]. However, precursor opti-
mization was also successfully pursued. It
Fig. 13.3 Schematic representation of the S. cere-
visiae insulin precursor expression system. The ex-
pression plasmids (2l) are located in the nu-
cleus. mRNA is transcribed and translated into
Signal-leader-insulin precursor which subsequently
enters the endoplasmic reticulum (ER). The signal
peptide (S) is cleaved off by a signal peptidase
and the leader core-glycosylated (yellow filled cir-
cles), insulin folds and disulfides are formed. Lea-
der-insulin precursor is transported through Golgi,
where glycosylations are extended and upon enter-
ing late-Golgi kexin cleaves at the dibasic site
thereby separating insulin precursor from leader.
The insulin precursor enters secretory vesicles
(SV) and is secreted from the yeast cell. Part of
the insulin precursor may misroute to the vacuole.
Inefficient kexin cleavage leads to secretion of hy-
perglycosylated leader-insulin precursor; this can
be remedied by inclusion of an N-terminal exten-
sion on the insulin precursor.
was hypothesized based on knowledge of
the 3-D structure of Asp
B28
insulin [43]
and the location of a phenol binding pocket
in the insulin hexamer that the side chain
of an aromatic amino acid localized in the
mini C-peptide could improve the stability
of the insulin precursor [41] (see below, Sta-
bility). The interaction of this hydrophobic
amino acid side chain with the hydrophobic
core of the insulin precursor molecule
could be energetically favorable and conse-
quently enhance the folding stability of the
precursor. Folding stability is important
for secretory efficiency [12] (see Section
13.2.2.5), and therefore it was further hy-
pothesized that enhancing the folding sta-
bility of the precursor would facilitate trans-
port through the secretory pathway and
thus increase the overall productivity. The
introduction of an aromatic amino acid (X)
into the mini C-peptide (e.g., EXK) did in-
deed increase the expression yield of the in-
sulin aspart precursor. Tryptophan had the
greatest positive influence (5-fold) on the ex-
pression yield, whereas phenylalanine and
tyrosine increased the expression yield ap-
proximately 3.5-fold [41]. This is an example
of a highly specific calculated optimization;
Asp
B28
is disturbing the ordinary C-peptide
turn and the overall folding, which is more
than compensated by introduction of an
aromatic amino acid in C2 in the C-peptide.
13.2.2.5 Structural Mutations and Relation-
ship between Expression Yields
and In vitro Folding Stability
The insulin fold is composed of two A-chain
a-helixes connected by a loop, and one cen-
tral B-chain a-helix flanked by an N-terminal
sequence in extended conformation (in the
structural T-state of insulin) and a b-sheeted
C-terminal. Enhanced folding stability can
for example be engineered by mutations,
leading to the removal of hydrophobic
side-chains at the protein surface as well
as substitutions near the N- and C-terminal
of the a-helixes stabilizing these helixes [44].
Experiments examining the expression of a
series of insulin analogues in S. cerevisiae
found a positive correlation between insulin
analogue in vitro folding stability and the fer-
Fig. 13.4 Conversion of insulin precursor to re-
combinant human insulin using transpeptidation.
The C-peptide is removed from the partially puri-
fied insulin precursor using a serine protease
(e.g., trypsin). By inclusion of a threonine ester
(T') and using appropriate reaction conditions, it
is possible to couple threonine to Lys
B29
using
trypsin as a transpeptidase, thereby obtaining re-
combinant human insulin. The N-terminal exten-
sion (spacer peptide in text) on the B-chain of in-
sulin (X)
n
K is removed in the transpeptidation
step. The insulin A-chain and B-chain are shown
in red and blue, respectively. Disulfide bonds are
depicted in yellow.
mentation yield of the corresponding insu-
lin analogue precursor (expressed fused to
the a-factor leader and with the AAK mini
C-peptide connecting B29 and A1) [12]. Sta-
bilization of folding of individual parts of
the precursor molecule is important for
the overall stability, and this seems to have
a positive impact on the fermentation yield.
This principle can probably be extended to
cover all parts of the signal peptide-leader-
insulin precursor.
13.2.2.6 Polymerization of the Insulin
Precursor Influences Secretion
Yield and Retention in the Vacuole
The polymerization of insulin and proin-
sulin to dimers and hexamers is a very im-
portant process which takes place in the
pancreas and impacts upon the pharma-
ceutical application of insulin, because the
dissociation of the polymers is the rate-
limiting processes in the absorption and
action in the tissues of the biological active
monomeric insulin [45]. The polymeriza-
tion diminishes osmotic pressure and hy-
drophobicity and improves solubility. A
positive correlation between expression
yield and the degree of polymerization was
found [46] that supports yeast in vivo poly-
merization of insulin precursors and ac-
centuates its importance. However, poly-
merization can also account for a draw-
back by retention of product in the vacuole
[47]. Intracellular retention of a substantial
quantity of the synthesized insulin precur-
sor indicated that the insulin precursor fol-
lowed two different intracellular routes in
the late secretory pathway [12, 47]. Consti-
tutive secretion to the culture supernatant
may reflect saturation of a sorting mecha-
nism in the late Golgi due to overexpres-
sion. The kexin cleaves the leader-insulin
precursor peptide in a late Golgi compart-
ment to yield free insulin precursor, and
this change is provoking accumulation in
the vacuole [47]. Zhang et al. [47] applied a
genetic screen to identify genes influenc-
ing insulin precursor accumulation and
identified several vps-mutants important
for accumulation of insulin precursors and
concomitant possible targets for regula-
tion. The uncleaved leader-precursor was
not accumulated, but was secreted to the
medium. Thus, the leader seems to hide
the sorting signal or prevent its formation.
The sorting mechanism was suggested to
be a consequence of endoproteolytic ma-
turation and multimeric assembly of insu-
lin [47].
13.3
Production of Insulin
A S. cerevisiae strain optimized for the pro-
duction of insulin by some of the recombi-
nant strategies outlined previously must
comply with the cultivation conditions pre-
vailing at industrial scale.
13.3.1 Fermentation
A prerequisite for industrial strains to be
used for continuous cultivations is the
ability to remain stable for extended peri-
ods of time when grown at large scale (see
also the Introduction and Part IV, Chap-
ters 1 and 12). This has previously been
shown to be the case for S. cerevisiae
strains based on a POT-selection plasmid
and an insulin precursor transcribed by
the constitutive TPI1-promoter [6]. During
growth of S. cerevisiae on carbohydrates as
the major carbon source, ethanol forma-
tion should be avoided since formation of
ethanol leads to lower biomass yields and
consequently decreasing product titers.
This issue of fermentative metabolism is
often solved by monitoring the oxygen
consumption and carbon dioxide evolution
rates, and using these variables as inputs
for control algorithms adjusting the fer-
mentation process parameters.
13.3.2
Transpeptidation
The fermentation broth leaving the fer-
menter is traditionally clarified for yeast
cells by means of centrifugation or filtra-
tion processes, followed by purification by
chromatography. Conversion of the insulin
precursor lacking Thr
B30
to human insulin
is elegantly achieved through the use of
enzymatic processes based on serine endo-
peptidases, for example porcine trypsin
(EC 3.4.21.4) or lysyl endopeptidase from
Achromobacter lyticus (EC 3.4.21.50) (Fig.
13.4). When used initially as hydrolases in
water-rich solvents, the C-peptide is re-
moved by cleavage at the peptide bond of
Lys
B29
and Lys
A1-1
, the lysine that joins the
C-peptide to the insulin A-chain. Following
this hydrolysis the des(B30)-insulin mole-
cule is coupled to a threonine ester, this
time exploiting the enzymes capability to
function in mixtures of water and organic
solvents [5]. Changing the equilibrium
from hydrolysis to synthesis is accom-
plished by adjusting the solvent composi-
tion and the reaction conditions (pH, sub-
strate concentration and water activity).
Transpeptidation, where the equilibrium
can be displaced towards peptide bond syn-
thesis, can be obtained by a one-step proce-
dure with a pre-programmed adjustment of
the water activity prevailing in the reaction
mixture. Spacer peptides, if applied, are ele-
gantly removed in the same operations that
remove the C-peptide. The intermediate
product with an esterified carboxy-terminus
of Thr
B30
is purified by chromatography,
and finally the human insulin-ester is hy-
drolyzed leading to human insulin.
13.4
Conclusions and Future Aspects
The expression of insulin precursors in
the yeast S. cerevisiae has, in recent years,
formed the basis for the manufacture of
the majority of human recombinant insu-
lin and insulin analogues sold by Novo
Nordisk A/S (see Introduction). Studying
the expression of insulin in S. cerevisiae
has highlighted the importance of addres-
sing every event leading to secretion of the
insulin precursor from the initial translo-
cation and folding in the ER, transport
through the Golgi apparatus, to matura-
tion by kexin and ultimately secretion via
correct localization to secretory vesicles.
Modification of the insulin precursor, as
described in this chapter, has led to signifi-
cant improvements in the expression sys-
tem, resulting in robust scalable expres-
sion systems which are amenable to indus-
trial fermentation and downstream proces-
sing of the insulin precursor. Some
observations indicate that there are limita-
tions in the secretory capacity of S. cerevi-
siae [12, 48] and that these point to alterna-
tive initiatives. Focus on the actual secre-
tion machinery in S. cerevisiae and alterna-
tive hosts systems such as Saccharomyces
kluyveri [49] or Pichia pastoris [50, 51] are
options for further exploitation and the
large-scale production of other biopharma-
ceuticals.
References
1 Markussen, J. US Patent 4343898 (1982).
2 Chance, R. E., Hoffmann, J. A, Kroeff, E. P.,
Johnson, M. G., Schirmer, E. W., Bromer,
W. W., Ross, M. J., and Wetzel, R. In: Peptides.
Synthesis-Structure-Function. Proceedings of the
Seventh American Peptide Symposium (D.H.
Rich and E. Gross, eds). Pierce Chemical
Company, p 721 (1981).
3 Frank, B. H., Pettee, J. M., Zimmerman, R. E.,
and Burck, P. J. In: Peptides. Synthesis-Struc-
ture-Function. Proceedings of the Seventh
American Peptide Symposium (D.H. Rich and
E. Gross, Eds). Pierce Chemical Company,
p 729 (1981).
4 Thim, L., Hansen, M. T., Norris, K., Hoegh,
I., Boel, E., Forstrom, J., Ammerer, G. and
Fiil, N. P. Proc. Natl. Acad. Sci. USA 83, 6766
6770 (1986).
5 Markussen, J., Damgaard, U., Diers, I. Fiil,
N., Hansen, M. T., Larsen, P., Norris, F., Nor-
ris, K., Schou, O., Snel, L., Thim, L., and
Voight, H. O. In: Peptides 1986 (D. Theodoro-
poulos, Ed.), pp 189194. Berlin: Walter de
Gruyter & Co. (1987).
6 Diers, I. V., Rasmussen, E., Larsen, P. H., and
Kjaersig, I.-L. In: Drug Biotechnology Regula-
tion, Scientific Basis and Practices (Y.-Y. H. Chiu
and J. L. Gueriguian, Eds), pp 166176, New
York, NY: Marcel Dekker, Inc. (1991).
7 Brange, J., Ribel, U., Hansen, J. F., Dodson,
G., Hansen, M. T., Havelund, S., Melberg,
S. G., Norris, F., Norris, K., Snel, L., Srensen,
A. R., and Voigt, H. O. Nature 333, 679682
(1988).
8 Markussen, J., Hougaard, P., Ribel, U., Sren-
sen, A. R., and Srensen, E. Protein Eng. 1,
205213 (1987).
9 Markussen, J., Diers, I., Engesgaard, A., Han-
sen, M. T., Hougaard, P., Langkjaer, L., Norris,
K., Ribel, U., Srensen, A. R., and Srensen,
E. Protein Eng., 1, 215223 (1987).
10 Markussen, J., Diers, I., Hougaard, P., Langk-
jaer, L., Norris, K., Snel, L., Srensen, A. R.,
Srensen, E., and Voigt, H.O. Protein Eng. 2,
157166 (1988).
11 Kjeldsen, T. Appl. Microbiol. Biotechnol. 54,
277386 (2000).
12 Kjeldsen, T., Balschmidt, P., Diers, I., Hach,
M., Kaarsholm, N. C., and Ludvigsen, S. Biote-
chol. Genet. Eng. Rev. 18, 89121 (2001).
13 Romanos, M. A., Scorer, C. A., and Clare, J. J.
Yeast 8, 423488 (1992).
14 Ammerer, G., Hunter, C. P., Rothman, J. H.,
Saari, G. C., Valls, L. A., and Stevens, T. H.,
Mol. Cell. Biol. 6, 24902499 (1986).
15 MacKay, V. L., Yip, C., Welch, S., Gilbert, T.,
Seidel, P., Grant, F., and OHara, P. In: Re-
combinant Systems in Protein Expression (K. K.
Alitalo, M.-L. Huhtala, J. Knowles, and A. Va-
heri, Eds), pp. 2536. Amsterdam: Elsevier
Science Publishers BV (1990).
16 Kawasaki, G. H. and Bell, L. US Patent
5,871,957 (1999).
17 Kjeldsen, T. and Vad, K. US Patent 6358705
(2002).
18 Smith R. A., Duncan, M. J., and Moir, D. T.
Science 229, 12191224 (1985).
19 Bartkeviciute, D. and Sasnauskas, K. FEMS
Yeast Res. 4, 833840 (2004).
20 Kozlov, D. G., Prahl, N., Efremov, B. D., Peters,
L., Wambut, R., Karpychev, I. V., Eldarov,
M. A., and Benevolensky, S. V. Yeast 11, 713
724 (1995).
21 Chow, T. Y., Ash, J. J., Dignard, D., and Tho-
mas, D. Y. J. Cell Sci. 101, 709719 (1992).
22 Ruohonen, L., Toikkanen, J., Tieaho, V., Outo-
la, M., Soderlund, H., and Keranen, S. Yeast
13, 337351 (1997).
23 Christiansen, L. and Petersen, J. G. L. (1994)
Patent application WO95/02059.
24 Ng, D. T., Brown, J. D., and Walter, P. J. Cell
Biol. 134, 269278 (1996).
25 Plath, K., Mothes, W., Wilkinson, B. M., Stir-
ling, C.J., and Rapoport, T.A. Cell 94, 795807
(1998).
26 Thorner, J. In: The Molecular Biology of the
Yeast Saccharomyces cerevisiae: Life Cycle and
Inheritance (J. N. Strathern, E. W. Jones, and
J. R. Broach, Eds), pp. 143180. New York NY:
Cold Spring Harbor Laboratory (1981).
27 Kurjan, J. and Herskowitz, I. Cell 30, 933943
(1982).
28 Bitter, G. A., Chen, K. K., Banks, A. R., and
Lai, P.-H. Proc. Natl. Acad. Sci. USA 81, 5330
5334 (1984).
29 Brake, A. J., Merryweather, J. P., Coit, D. G.,
Heberlein, U. A., Masiarz, F. R., Mullenbach,
G. T., Urdea, M. S., Valenzuela, P., and Barr,
P. J. Proc. Natl. Acad. Sci. USA 81, 46424646
(1984).
30 Kjeldsen, T., Pettersson, A. F., Hach, M.,
Diers, I., Havelund, S., Hansen, P. H., and
Andersen, A. S. Protein Expr. Purif. 9, 331336
(1997).
31 Kjeldsen, T., Hach, M., Balschmidt, P., Have-
lund, S., Pettersson A. F., and Markussen, J.
Protein Expr. Purif. 14, 309316 (1998).
32 Julius, D., Schekman, R., and Thorner, Cell
36, 309318 (1984).
33 Caplan, S., Green, R., Rocco, J., and Kurjan, J.
J. Bacteriol. 173, 627635 (1991).
34 Kjeldsen, T., Andersen, A. S., Hach, M., Diers,
I., Nikolajsen J., and Markussen, Biotechnol.
Appl. Biochem. 27, 109115 (1998).
References 1043
35 Julius, D., Blair, L., Brake, A., Sprague, G.,
and Thorner. Cell 32, 839852 (1983).
36 Kjeldsen, T., Brandt, J. Andersen, A. S., Egel-
Mitani, M., Hach, M., Pettersson, A. F., and
Vad, K. Gene 170, 107112 (1996).
37 Egel-Mitani, M., Flygenring, H. P., and Han-
sen, M. T. Yeast 6, 127137 (1990).
38 Egel-Mitani, M., Brandt, J., and Vad, K. US
Patent 6110703 (2000).
39 Kjeldsen, T. and Ludvigsen, S. US Patent
6777207 (2004).
40 Kjeldsen, T. and Ludvigsen, S. WO200279251
(2002).
41 Kjeldsen, T., Ludvigsen, S., Diers, I., Balsch-
midt, P., Srensen, A. R., and Kaarsholm,
N. C. J. Biol. Chem. 277, 1824518248 (2002).
42 Diers, I. Patent application WO 200146453
(2000).
43 Whittingham, J. L., Edwards, D. J., Antson,
A. A., Clarkson, J. M., and Dodson, G. G. Bio-
chemistry 37, 1151611523 (1998).
44 Kaarsholm, N. C., Norris, K., Jrgensen, R. J.,
Mikkelsen, J., Ludvigsen, S., Olsen, H. B., Sr-
ensen, A. R., and Havelund, S. Biochemistry
32, 1077310778 (1993).
45 Dodson, E. J., Dodson, G. G., Hubbard, R. E.,
Moody, P. C. E., Turkenburg, J., Whittingham,
J., Xiao, B., Brange, J., Kaarsholm, N., and
Thogersen, H. Phil. Trans. R. Soc. Lond. A
345, 153164 (1993).
46 Kjeldsen, T. and Pettersson, A. F. Protein Expr.
Purif. 27, 331337 (2003).
47 Zhang B., Chang, A. Kjeldsen, T. B., and Ar-
van, P. J. Cell Biol. 153, 11871198 (2001).
48 Egel-Mitani, M., Hansen, M. T., Norris, K.,
Snel, L., and Fiil, N. P. Gene, 73, 113120
(1988).
49 Srivastava, A., Piskur, J., Nielsen, J., and Egel-
Mitani, M. PCT patent application WO
200014258 (2000).
50 Kjeldsen, T., Pettersson, A. F., and Hach, M.
Biotechnol. Appl. Biochem. 29, 7989 (1999).
51 Wang, Y., Liang, Z.-H., Zhang, Y. S., Yao, S.-Y.,
Xu, Y.-G., Tang, Y. H., Zhu, S. Q., Cui, D. F.,
and Feng, Y.-M. Biotechnol. Bioeng. 73, 7479
(2001).
Abstract
Baculovirus-based protein expression in in-
sect cell culture represents today a ripened
method for the rapid production of proteins
up to pilot scale. Recent advances on all lev-
els of the production process such as opti-
mized expression vectors, chemically de-
fined media, optimized nutritional and ki-
netic parameters as well as the design of
novel cultivation systems contribute to dras-
tically increased protein yields with conco-
mitant reduction of process time and costs.
This chapter provides a comprehensive
overview on recent research and achieve-
ments on the successive steps in a protein
expression process, with a focus on how to
integrate these findings for achieving over-
all optimized performance. A case study
on the production of a secreted protein il-
lustrates the impact of the different optimi-
zation steps on the overall process yields.
Abbreviations
AcNPV Autographa californica nuclear
polyhedrosis virus
BEVS baculovirus expression
vector system
BTK brutons tyrosine kinase
variant
DO dissolved oxygen
MOI multiplicity of infection
OTR oxygen transfer rate
OUR oxygen uptake rate
PCD peak cell density
s-ICAM-1 secreted intercellular adhesion
molecule-1
TOH time of harvest
TOI time point of infection
14.1
Introduction
The vast variety of different drug targets
and yet-to-be-characterized open reading
frames (ORFs) emerging from post-geno-
mic research initiatives has substantiated
the need for rapid and generic pilot-scale
production of desired proteins for bio-
chemical/functional studies as well as for
target validation. The baculovirus expres-
sion vector system (BEVS) developed for
heterologous protein production in insect
cell cultures almost three decades ago is
one of todays preferred pilot-production
technology owing to BEVS superior pro-
tein titers and unmatched from-gene-to-
protein process speed, which still sur-
passes currently available mammalian cell-
based production processes.
1045
14
Baculovirus-based Production of Biopharmaceuticals
using Insect Cell Culture Processes
ISBN: 3-527-31184-X
The baculovirus which prevails in todays
production processes belongs to the Euba-
culoviridae [1], a subfamily of double-
stranded DNA viruses. Following infection
of host arthropods, these viruses produce
proteinaceous structures known as occlu-
sion bodies representing up 50% of the to-
tal cellular protein [1]. The baculoviruss su-
perior protein production capacity resides
in the viral very late polyhedrin and p10 pro-
moters, which are often engineered to ex-
press desired transgenes rather than struc-
tural virus genes forming occlusion bodies
[2, 3]. During the past decade, biotechnolo-
gically relevant BEVS research has focused
on improving process engineering parame-
ters including nutrient/oxygen supply, in-
fection kinetics and heterologous protein
production (for a review, see Ref. [4]). In-
depth understanding of BEVS-related pro-
cess parameters associated with maximum
production levels has resulted in generic
protocols for the rapid implementation of
large-scale bioreactor operation in an as-
sembly line-like production scenario [5].
With product yield at its near maximum,
the BEVS community is currently focusing
on the improvement of product quality by
humanizing insect cell glycosylation pat-
terns. The unique combination of transi-
ent expression implementation, high-yield
protein production capacity and the pro-
spect of human glycoprofiles in insect cul-
tures indicates a bright future for BEVS
technology in the production of biophar-
maceuticals [4, 6, 7].
14.2
Molecular Tools for the Construction
of Transgenic Baculoviruses
The most widely used baculoviral expres-
sion systems are derived from the Autogra-
pha californica nuclear polyhedrosis virus
(AcNPV) [1, 8]. Owing to its extended ge-
nome of 129 kb, baculoviruses are not
amenable to standard cloning procedures
and require multi-step split-genome strate-
gies for the design of transgenic viral vec-
tors. The classic BEVS capitalizes on a small
helper vector encoding the transgene ex-
pression unit flanked by viral sequences,
which recombines onto the baculoviral ge-
nome following co-transfection into insect
cells. BEVS are commercially available in
different variants, including b-galactosi-
dase-encoding helper vectors, which facili-
tate straightforward screening for desired
recombinant baculovirus genomes (e.g.,
BacPAK, BD Biosciences; pPBac, Strata-
gene; Bac-N-Blue, Invitrogen). Recent pro-
gress in transgenic baculovirus design en-
abled helper vectorbaculovirus genome re-
combination in Escherichia coli, from which
transgenic genomes are recovered and pre-
pared for subsequent transfection and virus
production in insect cells (Bac-to-Bac, Invi-
trogen). The latest generation of baculovirus
design includes lambda integrase-mediated
in vitro recombination of the transgene ex-
pression unit onto an engineered baculo-
virus genome. Such in vitro design technol-
ogy has further been refined so that the
transgene expression unit replaces a thymi-
dine kinase expression unit on the baculo-
virus chromosome, thereby enabling gancy-
clovir-mediated selection of desired recom-
binants (BaculoDirect, Invitrogen).
Baculovirus stocks are typically produced
by the aforementioned methods, followed
by optional amplification rounds and titra-
tion. The traditional titration technology is
based on plaque assays where viral stock
dilutions are applied on confluent insect
cell cultures and clonal plaques scored by
visual inspection after several days. More
rapid technologies include: 1) immunode-
tection of viral proteins ([9], BacPAK Rapid
Titer Kit, BD Biosciences); 2) fluorescent
14 Baculovirus-based Production of Biopharmaceuticals using Insect Cell Culture Processes 1046
virus-encoded marker proteins profiled by
microscopy [10]; or 3) quantitative real-
time PCR specific for viral sequences [11].
Despite advanced quantification technolo-
gies, only plaque assays enable direct pro-
filing of infectious particles; all other
methods have relied on empiric correction
factors when considering non-infectious/
defective baculoviral particles.
14.3
Insect Cell Culture
While first-generation baculovirus-based
expression technology included infection
of entire insect larvae [2, 12], advanced
technologies take advantage of insect-de-
rived serum-free suspension cultures for
baculovirus maintenance and protein pro-
duction [4, 13, 14]. The most prominent
insect cell lines include the Spodoptera fru-
giperda-derived ovarian cell lines Sf-21 and
its subclone Sf-9, as well as the Trichoplu-
sia ni egg cell homogenate-derived cell line
BI-TN-5B1-4, known as High Five
TM
cells.
While High Five
TM
cells provide increased
specific and volumetric productivities of
secreted proteins [1517], Sf-9 continues to
be the preferred cell line for expression of
intracellular or membrane proteins, while
Sf-21 prevailed for isolation and propaga-
tion of viral stocks. Cell culture media re-
quirements for insect cells are comparable
to those for mammalian cells, with the ex-
ception of a lower pH (6.26.9) and the
tolerance of higher free amino acid and
glucose levels without switching to over-
flow metabolism [14].
14.4
Insect Cell Glycosylation
and Glycoengineering
Since insect cells fail to synthesize com-
plex N-glycans and produce potentially al-
lergenic structures of the Fuca(1,3)Glc-
NAc-Asn type, BEVS-based production of
protein pharmaceuticals for clinical use re-
mains compromised [7]. Several strategies
for overcoming such limitations are cur-
rently competing for industrial implemen-
tation:
Screening for insect cell lines which pro-
duce more complex glycoforms. The Da-
nau plexippis-derived DpN1 cell line
showed up to 26% of complex glyco-
forms compared to less than 5% for
High Five
TM
cells, yet its overall product
titers were less competitive [7, 17].
Sophisticated feeding strategies based
on N-acetylmannosamine (ManNAc) [18]
supplementation or addition of specific
glycosyltransferase inhibitors [19]: Wata-
nabe and co-workers reported the capaci-
ty of High Five
TM
for interferon-c N-gly-
can sialylation following cultivation of
these insect cells in the presence of an
hexosaminidase inhibitor (2-AND).
Metabolic engineering of insect cells: Ec-
topic expression of epimerases and ki-
nases, such as the bifunctional enzyme
SAS and UDP-GlcNAc 2-epimerase/
ManNAc kinase initiating sialic acid bio-
synthesis in mammalian cells, resulted
in N-acetylneuraminic acid (Neu5Ac)
synthesis in the absence of any media
supplements [18]. A particular Sf-9-de-
rived insect cell line [6], engineered for
simultaneous expression of five different
glycosyltransferases, produced bianten-
nary terminally sialylated N-glycans re-
miniscent of glycoprofiles typically
found in mammalian cells (Mimic
TM
Sf-
9 cells, Invitrogen [6]).
14.5
Nutrient and Kinetic Considerations
for Optimized BEVS-based
Protein Production
In a baculovirus-infected insect cell cul-
ture, nutrients are converted into biomass,
virus progeny and desired product protein,
as well as diverse by-products [20, 21]. The
relative quantities of these culture prod-
ucts can be influenced by modification of
critical kinetic process parameters, includ-
ing the time point of infection (TOI, typi-
cally expressed as cell density at infection)
and the multiplicity of infection (MOI,
number of infective viral particles per cell)
[2022]. Different strategies to modulate
TOI and MOI along with nutrient supply
and timely infection are covered in the fol-
lowing sections.
14.5.1
Nutritional Parameters
Nutrient and oxygen supply appear to be
the only limiting factors for BEVS to reach
high cell densities and recombinant pro-
tein titers [23, 24]; no other limiting effects
such as contact inhibition or accumulation
of toxic metabolites (including alanine or
ammonium) have been observed for Sf-9
cells [23]. Several feeding strategies have
been devised to increase cell density and
product titer. Complete medium exchange
[25, 26] or perfusion systems [27, 28] ini-
tially appeared to be promising for supply-
ing additional nutrients, but these
approaches turned out to require tedious
and time-consuming separation steps or
special reactor configurations, which are
often incompatible with streamline recom-
binant protein production. Fed-batch sys-
tems, which supply specific nutrient con-
centrates to increase cell density, seem to
prevail in current insect cell-based produc-
tion processes. The influence of several
nutrient cocktails on insect cell densities
has been quantified either by scoring spe-
cific nutrient consumption rates [26, 29] or
by fractional (factorial) experimental de-
sign [5, 30, 31] consisting of individual nu-
trient effects being evaluated individually.
Maximum insect cell densities are reached
most efficiently following medium supple-
mentation with yeastolate and lipid con-
centrates [32], amino acids, vitamins and
trace elements [30, 31], as well as glucose
and glutamine [23].
Beyond cell densities of 110
6
cells mL
1
,
oxygen supply is a critical issue [23, 3335],
in particular for baculovirus-infected cul-
tures which require 3040% more oxygen
compared to standard insect cell cultures
[36, 37]. Typically, dissolved oxygen (DO)
levels above 30% air saturation are consid-
ered to be sufficient to prevent oxygen lim-
itation [33, 35, 36]. As well as oxygen supply,
CO
2
removal is a critical issue for insect cul-
ture, since high concentrations of dissolved
hydrogen-carbonate not only modulate the
cultures pH, but also negatively impact on
cell growth [38]. For example, CO
2
accumu-
lation was suggested to be a major limiting
factor for the expression of TGF-b receptor
production on a 150-L scale. Although the
culture medium was sparged with pure oxy-
gen to minimize shear stress, the low volu-
metric gas flow was insufficient to strip CO
2
off the bioreactor, which resulted in de-
creased TGF-b receptor production [39].
By integrating of all the aforementioned
parameters into a sophisticated feeding
scheme for glucose, amino acids, yeasto-
late, lipids, vitamins and trace elements
combined with optimized oxygen supply,
Elias and co-workers [40] reached 5.210
7
Sf-9 cells mL
1
, the current record in maxi-
mum density for this cell line operated in
fed-batch mode.
14.5.2
Infection-related Kinetic Parameters
Recombinant protein production using
baculovirus-infected insect cell cultures
has a well-defined endpoint: lysis of the
host cell. Therefore, maximum product
yield can only be achieved when baculo-
virus infection is precisely timed so that
maximum cell numbers complete protein
production just before the nutrients be-
come limiting [21, 22]. Owing to Poisson-
like infection distribution, infection and
maximum cell density timing are best syn-
chronized when insect cells are infected at
an MOI of at least five [41]. However, in-
fection of high cell density cultures at an
MOI of five requires substantial virus
amounts, resulting in undesired addition
of conditioned medium to the production
culture. In order to halt this vicious circle,
concentrated high-titer viral stocks are typ-
ically produced prior to infection [40]. Al-
ternatively, an in situ virus amplification
step can be initiated by infecting a low-
density cell culture at a low MOI (e.g.,
0.1), so that the few infected cells produce
progeny virus (approx. 500 virus per in-
fected cell; [22]), which subsequently in-
fects the remaining culture. Using in situ
virus amplification, baculovirus infections
can be performed at MOIs as low as
310
5
[42, 43]. Yet, if several in situ am-
plification steps are performed, small er-
rors in virus quantification or fluctuations
in cell culture kinetics are also amplified,
and may lead to premature infection of
the entire culture or to a large proportion
of uninfected cells should nutrients be-
come limiting [42]. When making the
choice between high-MOI infections and
low-MOI infections using in situ virus am-
plification, several protein-specific parame-
ters should be considered (see Table 14.1).
14.5.3
Protease Activities in Infected Insect
Cell Cultures
Proteases released during the infection
process may lead to protein inhomogene-
ities and product degradation [4, 5]. Sev-
eral baculovirus-encoded proteases and
chitinases have been identified which,
upon insect larvae infection, induce lique-
faction of the infected host [44, 45]. The
cysteine protease v-cath identified in A. ca-
lifornica shows functional homologies to
mammalian cathepsin L [46]. Several stud-
ies have focused on the neutralization of
14.5 Nutrient and Kinetic Considerations for Optimized BEVS-based Protein Production 1049
Table 14.1 Decision-making parameters for determination of a high or low multiplicity of infection (MOI) strategy
Low MOI infection with
in situ virus amplification
High MOI infection
Stable protein Unstable protein
Intracellular protein Secreted protein
Non-toxic protein Toxic/growth-retarding protein
Known infection kinetics/cell growth rate Unknown infection kinetics/cell growth rate
Only low-titer/limited virus stocks available Concentrated virus stocks available
Bioreactor operation at high infection cell
densities at or beyond 1210
6
cells mL
1
(fed-batch processes)
Bioreactor operation at medium to low cell
densities (batch processes)
baculovirus-derived proteolytic activities by:
1) genetic engineering of the viral genome
(BacVector-3000, [47]); 2) using earlier viral
promoters exemplified by the basic protein
promoter [48]; or 3) addition of cysteine
protease-specific inhibitors including leu-
peptin or pepstatin A [35, 49]. All of these
strategies resulted in decreased protease
activities, which correlated with increased
protein titers and improved product homo-
geneities, as required for high-end protein
applications as in crystallography or NMR
studies. A comprehensive overview of the
different strategies for improving nutrient
supply and optimizing infection kinetics is
provided in Table 14.2.
14.6
Scaling-up Baculovirus-based
Protein Production
While laboratory-scale BEVS-based protein
production can be achieved in standard cell
culture flasks, manufacturing of milligram
to gram quantities of a desired protein re-
quires more sophisticated production hard-
ware, including roller bottles, classical stir-
red-tank and perfusion bioreactors, or the
recently developed rotating wall vessel and
Wave
TM
bioreactors. Although perfusion re-
actors [27, 28] or rotating-wall vessels [4]
reach higher peak cell densities and show
low shear forces, they remain limited by
complex hardware management and the
lack of generic operation protocols. While
stirred-tank bioreactors remain the pre-
ferred cultivation systems for insect cells
up to several hundred-liter scale [39], the re-
cently developed Wave
TM
bioreactor system
is gathering decisive momentum in this ter-
ritory [4, 5, 50]. In capitalizing on a com-
pletely disposable reactor chamber, the Wa-
ve
TM
system is particularly efficient for
short-batch processes (36 days) as typical
bioreactor maintenance processes including
disassembly, cleaning, assembly and sterili-
zation become obsolete [5]. Wave
TM
bioreac-
tor operation is characterized by low shear
forces and high volumetric oxygen transfer
rates mediated by the continuous renewal
of the medium surface. Although Wave
TM
and classic stirred-tank bioreactors reached
identical titers of a secreted model product
protein when operated at a 10-L scale, the
operating costs for Wave
TM
were 40% lower
compared to the stirred-tank bioreactor [5,
51]. The latest-generation Wave
TM
bioreac-
tors managing culture volumes of up to
500 L are now at eye level with standard stir-
red-tank bioreactors as far as scale is con-
cerned [52] (Wave Biotech, Tagelswangen,
Switzerland).
14.7
Generic Protocol of Optimized
Protein Production
There are three major parameters deter-
mining the overall product yield and quali-
ty of BEVS processes: 1) abundance of nu-
trients; 2) timing of baculovirus infection;
and 3) harvest timing. In considering the
state-of-the-art know-how related to these
three parameters, we provide a genetic
protocol for the rapid determination of op-
timized BEVS-based process parameters.
Optimal parameters for small-scale pro-
cesses are derived by the following two-
step procedure:
1. Optimize nutritional factor to supply
BEVS with maximum substrate quanti-
ties. Peak cell density (PCD) is consid-
ered to be the best lumped value for the
assessment of nutritional parameters as
it integrates all factors modulating
growth and viability [29].
2. With nutritional supply at its optimum,
the kinetic parameters can be adjusted
Table 14.2 Analysis and optimization of nutritional and kinetic pa-
rameters in baculovirus-based protein production
Parameter(s)
studied
Result Reference
MOI, TOI, TOH,
yeastolate
Generic protocol for fast determination of optimum parameters
with subsequent scale-up for production runs.
5
MOI, TOI Modeling of the relationship between MOI and TOI and their
correlation with protein titer.
21
MOI, TOI Experimental validation and extension of the models developed
by Licari and Bailey [21].
20
MOI, pH,
protease activity
Proteolytic activities as a function of MOI. High MOI preferable
when protease activity is critical. pH-dependent activities
of proteases.
56
MOI, TOI Experimental validation of the models developed by Licari
and Bailey [21].
57
MOI, TOI Extremely low MOIs (0.00003) feasible but difficult to control. 42
MOI,
protease activity
Simulation and experimental validation of the relationship
between MOI and proteolytic activities.
53
Low MOI infection Virus-like particle production at low MOI. 43
Protease activity Determination of protease activity at different time
points after infection.
58
Diverse nutrients Multiparameter study to determine optimum nutrient
supply to reach 310
7
Sf-9 cells mL
1
.
24
Diverse nutrients,
MOI
Factorial experimental design for determination of optimal
protein production.
30
MOI, TOI Modeling of infection and protein expression kinetics with
experimental validation. Determination of critical rate constants.
22
Optimization of
feeding regime
Fed-batch on demand resulting in 5.210
7
Sf-9 cells mL
1
.
Production of b-galactosidase at high cell density.
40
Glucose,
amino acids
Scale-up for High Five
TM
cells based on nutrient consumption
analysis and specific nutrient supplementation.
29
Multi-parameter
study
Analysis of optimum conditions using factorial experimental
design and response surface experiments.
31
TOI Effect of infection time point in suspension cultures
and scale-up in bioreactor.
25
Nutrient
consumption rates
Determination of specific nutrient consumption rates
for sugars and amino acids prior to and after infection.
26
Yeastolate Production and investigation of yeastolate for insect cell culture. 32
Medium
development
Review of insect cell culture medium development. 14
Pluronic F-68 Protective effects of Pluronic F-68 on insect cells. 59
MOI, multiplicity of infection; TOI, time of infection; TOH, time of harvest.
to the desired product protein by select-
ing the best MOI, TOI, and time of har-
vest (TOH) [21, 53].
14.7.1
Optimization of Nutritional Parameters
Fed-batch processes are the method of
choice to reach maximum cell densities
and product yield because of their straight-
forward setup and optimal compatibility
with standard bioreactor hardware. Recent
research has highlighted a variety of medi-
um supplements including yeastolate [31,
32], amino acids [24, 30] and lipid mixes
[31,] which, when combined with sophisti-
cated feeding strategies, resulted in higher
cell densities of up to 5.210
7
Sf-9
cells mL
1
[40]. However, for high-through-
put pilot-scale production of different bio-
pharmaceutical proteins a well-balanced
protocol, which considers maximum cell
density and product titer as well as bio-
reactor operation parameters exemplified
by the oxygen transfer rate, is required. A
generic protocol for routine bioreactor-
based fed-batch operation has recently
been outlined for the Wave
TM
bioreactor
[5, 50] and Sf-9 cells, where the SF900II
medium was supplemented once with
4 g L
1
yeastolate at cell densities of 3
410
6
cells mL
1
, thereby resulting in
peak cell densities of up to 8.810
6
cells mL
1
. Oxygen supply was optimally
adjusted by regulating the inflow gas com-
position, as well as the agitation speed by
means of control software integrating a
predictive mass balance-based model of
oxygen transfer and consumption of insect
cells in a Wave
TM
bioreactor.
14.7.2
Optimization of Infection Kinetics
In-vivo and in silico research has high-
lighted the need for the optimal adjust-
ment of infection kinetics [5, 2022, 26].
Infection kinetics are predominantly deter-
mined by the MOI and TOI. Whereas an
optimum MOI is best chosen based on the
criteria in Table 14.1, the correlating opti-
mum TOI must be determined experimen-
tally for each product protein. A protocol
for rapid determination of optimum TOI
in small-scale batch cultures and its subse-
quent translation into large-scale fed-batch
production configurations has recently
been established [5, 50]. This generic pro-
tocol is based on findings by Radford [26,
54], Bdard [24], Elias [40], Chan [31] and
Power [22], whose models suggest that: 1)
peak cell densities in uninfected cultures
as well as maximum product titers in in-
fected cultures are limited by the same nu-
trients; and 2) that the different steps of
viral infection, replication and protein syn-
thesis practically show the same kinetic
behavior at different cell densities, pro-
vided that the insect cell density exceeds
1.510
6
cells mL
1
at infection. Equation
(14.1) [5] describes the correlation between
the peak cell density (PCD) of uninfected
insect cell populations cultivated at various
nutrient levels using different feeding
strategies with the optimum TOI [5].
TOI
lownutrient
PCD
lownutrient
TOI
highnutrient
PCD
highnutrient
14:1
Using Eq. (14.1), the optimum TOI for
small-scale cell culture operation in simple
batch settings can first be determined and
then translated into large-scale fed-batch
bioreactor management using a high nu-
trient supply for the final production run,
provided that the uninfected peak cell den-
sities have been determined for both sys-
tems beforehand.
14.7.3
Determination of the Optimum TOH
Determination of the optimum TOH is
critical to enable sufficient time for maxi-
mum protein production before cell leak-
age and proteolysis occur. When perform-
ing small-scale production experiments for
the assessment of optimum TOI and MOI
(as suggested in Section 14.7.2), different
TOHs should also be evaluated. Transla-
tion of the optimum TOH from small-
scale to large-scale production scenarios is
not straightforward, as cell growth rate as
well as protein production kinetics may
differ between cell culture plates and bio-
reactors [29]. Trypan blue-based staining
was suggested for monitoring the infection
and determination of the TOH [5]. During
the infection process the cell membrane
integrity decreases [22], and this results in
protein leakage [54] and trypan blue trans-
fer into the cells. In contrast to mamma-
lian cells, which are stained exclusively by
trypan blue when dead, trypan blue stain-
ing scores the progress of baculovirus in-
fection in insect cells [36]. Using the frac-
tion of trypan blue-stained cells as a mea-
sure of the cultures infection state, TOH
translation from small- to large-scale pro-
duction processes becomes possible [5].
14.8
Case study: Rapid Optimization
of Expression Conditions and Large-scale
Production of a Brutons Tyrosine Kinase
Variant (BTK)
In this section we will exemplify imple-
mentation of the aforementioned generic
protocols for production optimization of
BTK (720 mg purified protein) within 2
weeks, starting from a virus master stock.
14.8.1
Determination of Optimum MOI, TOI,
and TOH in Small-scale Roller Bottles
Sf-9 cells were cultivated in 450 cm
2
roller
bottles (70 mL Sf900II medium) and in-
fected with different MOIs at different
TOIs (expressed in cell density at the in-
fection time point). Relative protein titer
and protein quality were examined by: 1)
Coomassie blue-stained SDSPAGE; 2)
Western blot analysis at 48 h and 72 h
after infection; and 3) trypan blue staining.
The results at 48 h post infection are
shown in Table 14.3; at 72 h, a lower
protein quality and titers were observed
post infection (data not shown). These
data suggested infection of insect cells at
an MOI of 1 and as the cell density reaches
31% (1.710
6
cells mL
1
) of the uninfected
peak cell density under the same culture
conditions (5.510
6
cells mL
1
), and that
the protein should be harvested when 76%
of the cells are stained by trypan blue.
14.8.2
Scale-up to 10-L Wave
TM
Bioreactors
The BTK production run was performed
in two parallel 10-L Wave
TM
bioreactors
(Fig. 14.1) operated in fed-batch mode with
4 g L
1
yeastolate supplementation at infec-
tion, an MOI of 1, and a TOI of 2.810
6
cells mL
1
(32% of 8.810
6
cells mL
1
, the
peak cell density under fed-batch condi-
tions). During the growth and production
phase, oxygen levels were controlled by ad-
justing the in-gas composition and the agi-
tation speed based on an oxygen control
software simulating oxygen transfer and
consumption in a Wave
TM
bioreactor. The
softwares underlying model is displayed
14.8 Case study: Rapid Optimization of Expression Conditions and Large-scale Production of a (BTK) 1053
in Fig. 14.2, and the corresponding volu-
metric oxygen transfer rates are shown in
Table 14.4. The culture was harvested
when 77% of the cells were stained with
trypan blue, resulting in 720 mg BTK after
purification. The entire protocol was final-
ized within 2 weeks, with optimization
taking place in the first week and produc-
tion in the second. A detailed overview on
the process as well as operation parame-
ters for the Wave
TM
system are shown in
Tables 14.5 and 14.6.
Table 14.3 Analysis of recombinant protein titer and the fraction
of trypan blue-stained cells as a function of MOI and TOI at 48 h
post infection
Run no. TOI
[10
6
cells mL
1
]
MOI Protein titer
[IOD]
Stained cells
[%]
1 1.0 0.1 56.5 82
2 1.7 0.1 74.6 ND
3 2.4 0.1 100.6 80
4 3.5 0.1 87.7 76
5 1.0 1.0 93.8 82
6 1.7 1.0 117.0 76
7 2.4 1.0 110.8 60
8 3.5 1.0 108.6 73
IOD, integrated optical density derived from quantification of Western blot bands.
ND, not detected.
Fig. 14.1 Wave
TM
bioreactor system equipped with inocula-
tion bottle as well as medium supply for fed-batch mode.
14.8.3
Process Cost-based Choice
of the Large-scale Cultivation System
For detailed cost analysis, a 10-L roller bot-
tle culture, a conventional stirred-tank re-
actor as well as a Wave
TM
system were
compared. Since recombinant protein ti-
ters for a secreted model glycoprotein (se-
creted intercellular adhesion molecule 1; s-
ICAM-1) were comparable for all three sys-
tems (93 mg L
1
, 1241 mg L
1
and
113 mg L
1
, respectively), a detailed pro-
cess cost analysis was performed [5, 50].
Fig. 14.2 Oxygen transfer and consumption for in-
sect cell culture in a Wave
TM
bioreactor. The mod-
el can be used to adjust rocking rate, rocking an-
gle and the in-gas composition to achieve opti-
mum oxygen supply in the different growth
phases. l, specific growth rate [h
1
]; lmax, maxi-
mum specific growth rate without oxygen limita-
tions [h
1
]; KS, Dissolved oxygen concentration at
which l=0.5lmax (Monod kinetic) [g mL
1
]; H,
Henrys constant [Pa
1
]; kLa, volumetric oxygen
mass transfer coefficient [h
1
]; c
O2,in
, oxygen con-
centration in in-gas [g mL
1
]; m*
O2,m
, dissolved
oxygen in medium under saturation conditions
[g]; m
O2,h
, oxygen in reactor headspace [g]; m
O2,m
,
dissolved oxygen in medium [g]; OTR, volumetric
oxygen transfer rate [g mL
1
h
1
]; OUR volumetric
oxygen uptake rate [g mL
1
h
1
]; p, atmospheric
pressure [Pa]; q
a
, aeration rate [mL h
1
]; q
s
, specif-
ic oxygen uptake rate [g cell
1
h
1
]; c
c
, cell density
[cell mL
1
]; q
O2
, specific oxygen uptake rate
[g cell
1
h
1
]; t, time [h]; VH, reactor headspace
volume [mL]; VR, reactor liquid volume [mL].
Table 14.4 Volumetric oxygen mass transfer coefficients (k
L
a [h
1
])
in a Wave
TM
bioreactor
a)
at different rocking rates and rocking
angles
Rocking angle Rocking rate [min
1
]
[8]
20 24 28
5 1.15 1.89 5.77
7.5 2.87 5.51 7.33
10 5.25 6.29 8.18
a) The Wave
TM
reactor was equipped with a 20-L-Cellbag
TM
filled
with 10 L of water.
Table 14.5 Protocol for cultivation of Sf-9 cells in 20-L Cellbags
TM
Time Operation
a)
Settings Remark
Monday Per 20-L CellBag
TM
inoculate four roller
bottles (850 cm
2
) with 310
5
Sf-9 cells
mL
1
in 250 mL Sf-900II medium.
Assemble and sterilize sample adapter,
inoculum bottle and medium tank. Fill
medium tank with 9 L Sf-900II medium
plus 0.1% Pluronic F68 and incubate
for 2 days at 288C for sterility control.
Thursday Connect sample adapter, inoculum
bottle and medium reservoir to
CellBag
TM
under laminar flow.
Insert DO-probe
if needed
Place CellBag
TM
on rocking unit,
inflate with air and fill with 9 L
medium.
Rocking rate=
20 min
1
; angle=
7.58; temperature=
28 8C; turn off aera-
tion and close all in-
lets.
Friday Control inoculum cultures and
CellBag
TM
for sterility. Fill inoculum
into inoculum bottle under laminar
flow. Inoculate the culture with final
density of approx. 310
6
cells mL
1
.
Start stop watch.
Rocking rate=
20 min
1
; angle=
7.58; temperature=
288C; set air flow
to 3 L h
1
.
Inoculum
culture should
be at 2.53.010
6
cells mL
1
Sunday Count cells and perform Medium
analysis.
Set air flow to
12 L h
1
to meet
increased oxygen
requirements.
Monday Count cells and perform Medium
analysis.
Adjust rocking rate
and aeration para-
meters to meet
oxygen requirements.
Tuesday Count cells and perform Medium
analysis. Add virus and nutrients
at desired cell density via inoculum
bottle.
Adapt oxygen concen-
tration in inlet gas to
maintain DO concen-
tration between 50
and 100% of satura-
tion.
a) The aeration parameters (air and oxygen flow rates, rocking
rate and rocking angle) can be determined by calculating the
oxygen flux using the model described in Fig. 14.2 and Table
14.4.
Table 14.6 Protocol for BEVS-based recombinant protein produc-
tion. When sufficient virus stock is available the entire procedure
(determination of optimum expression conditions and expression
at large scale) can be performed within 2 weeks
Time Operation Remark
Thursday
evening
Inoculate four roller bottles (850 cm
2
)
with 310
5
cells mL
1
in 250 mL Sf-900II
medium each
Monday
morning
Inoculate four roller bottles (850 cm
2
) with
310
5
cells mL
1
in 250 mL Sf-900II medium each
Inoculum cultures for
Wave
TM
Bioreactor
Assemble and sterilize sample port, inoculum bottle
and medium tank. Fill with 9 L Sf-900II medium
plus 0.1% Pluronic F-68 and incubate for 2 days
at 288C for sterility control.
Preparation for Wave
TM
Bioreactor
Inoculate eight roller bottles (490 cm
2
)
with 0.7; 1.4; 2.1 and 2.810
6
cells mL
1
,
two cultures per cell density, culture volume:
80 mL Sf-900II medium.
Cultures for determination
of optimum expression
conditions (different MOI
and TOI are tested)
Infect the eight roller cultures with two different
MOI per cell density (e.g., 0.1 and 1 MOI).
Inoculate two roller bottles (490 cm
2
) with 1.810
6
cells mL
1
in 80 mL Sf-900II medium
Cultures for determination
of peak cell density
a)
Wednesday Determine fraction of trypan blue-stained cells in
infected roller cultures (mix samples with trypan
blue and wait 3 min before counting)
Infected cells become
slowly colored, counting
directly after mixing
results in underestimation
of stained cells.
Take 1.5-mL samples from each infected culture,
centrifuge, and store protein containing fraction
appropriately.
Filtration is also possible
when secreted proteins
are produced
Count the two cultures for PCD determination
a)
Thursday Same as on Wednesday
Perform electrophoresis with samples from kinetic
studies on two gels, use one for Western blotting,
the other for Coomassie blue staining. Block the
Western blot over night, dry the Coomassie
blue-stained over night.
Test on expression level
and protein integrity
Connect sample port, inoculum bottle and
medium reservoir to CellBag
TM
under laminar flow.
Preparation of CellBag
TM
Friday
morning
Finish Western blot from Thursday. Perform quanti-
fication of bands on the gel. If available, use more
exact analytical methods (ELISA, HPLC).
Determine optimum MOI, TOI and harvest time point
with regard to expression level and protein integrity.
Integration of fixed costs, consumables
and manpower requirements on a typical
industry cost basis showed that the
Wave
TM
system was far the most econom-
ical solution (1 1297 per run), while roller
bottles (1 1789 per run) and the stirred-
tank reactor (1 2224 per run) were up to
70% more cost-intensive (Fig. 14.3, Table
14.7).
14.9
Conclusion
Today, the baculovirus expression vector sys-
tem represents a well-established technol-
ogy for routine pilot production of small
to medium quantities of desired product
proteins, mainly for research purposes.
The characteristics of BEVS, including cul-
ture media design, virus production and
quantification as well as large-scale cell cul-
ture, have been extensively studied during
the past decade, and this has resulted in
Time Operation Remark
Friday
afternoon
Install CellBag
TM
on rocking unit, fill
with 9 L medium and wait until temperature
of 288C is stable. Inoculate then with starter
cultures from Monday.
Install CellBag
TM
Monday
morning
Calculate optimum infection cell density with
Eq. (14.1) based on a PCD of 8.810
6
cells mL
1
(Sf-9 cells with yeastolate addition)
Perform culture analysis in CellBag
TM
according
to above protocol.
Following days Perform culture analysis in CellBag
TM
and adjust
culture parameters according to above protocol.
Perform infection when optimum infection cell
density is reached.
Supply 1 yeastolate or yeastolate ultrafiltrate
at approx. 310
6
cells mL
1 b)
Use yeastolate for intra-
cellular proteins and
yeastolate ultrafiltrate
for secreted and
membrane-bound ones.
Determine fraction of trypan blue-stained cells
twice a day. Harvest product when reaching the
same fraction as on the optimum time point of
harvest in optimization studies.
Harvest time typically
on Thursday or Friday.
a) Determination of peak cell density can be omitted if this value
is known from former experiments. Peak cell density is
slightly dependent on inoculum density, therefore an inocu-
lum cell density is chosen which is typically the average of the
tested TOIs.
b) If infection occurs at cell densities <310
6
cells mL
1
, add
yeastolate or yeastolate ultrafiltrate at the time point when an
uninfected culture would reach this density.
Fig. 14.3 Process costs for
three different cultivation sys-
tems based on a 10-L culture
scale. Fixed costs, consum-
ables and manpower costs
were calculated on a typical
industry cost basis.
Table 14.7 Process cost calculations for the three different cell culture systems operated at 10 L scale
System Roller bottles Stirred-tank reactor Wave
TM
bioreactor
20 Roller bottles with
500 mL culture volume
10 L Stirred-tank reactor Wave Bioreactor with
20-L CellBag
TM
Fix costs
a)
Incubator 8900, 1 Bioreactor and process
control unit 35 600, 1
Wave
TM
reactor and
equipment 12 460, 1
Costs per day 5, 1 Costs per day 20, 1 Costs per day 7, 1
Process time 7 d Process time
b)
9 d Process time 7 d
Total Fix costs 35, 1 180, 1 49, 1
Consumables
process set-up
20 Roller bottles 150, 1 DO-Probe membrane
8, 1
20-L CellBag
TM
220, 1
15 Pipettes 4, 1 O-rings 13, 1
Sterile filter 6, 1 Tubing, fittings 46, 1 Tubing, fittings 25, 1
Sterile filters 25, 1 Sterile filters 12, 1
Sterilization 200, 1 Sterilization 50, 1
Total set-up 160, 1 Total set-up 292, 1 Total set-up 307, 1
Consumables
per day
5 Pipettes 1.5, 1
5 Medium analysis
80, 1
Medium analysis 16, 1 Medium analysis 16, 1
3 L oxygen 0.04, 1 370 L oxygen 5, 1 60 L oxygen 0.8, 1
Total material
per day 82, 1
Total material
per day 21, 1
Total material
per day 17.80, 1
Culture time/process 7 d Culture time/process 7 d Culture time/process 7 d
Total daily 574, 1 Total daily 147, 1 Total daily 125, 1
Cleaning Sterilization
c)
50, 1 Sterilization
d)
100, 1 Sterilization
e)
16, 1
Washing machine 95, 1
Total cleaning 50, 1 Total cleaning 195, 1 Total cleaning 16, 1
Total
Consumables
784, 1 634, 1 448, 1
System Roller bottles Stirred-tank reactor Wave
TM
bioreactor
20 Roller bottles with
500 mL culture volume
10 L Stirred-tank reactor Wave Bioreactor with
20-L CellBag
TM
Manpower
Process set-up
Bioreactor assembly 3.2 h CellBag
TM
assembly 1.0 h
Bioreactor installation 1.0 h CellBag
TM
installation 0.5 h
Probe calibration 0.8 h
Inoculation 1.0 h Inoculation 1.5 h Inoculation 1.0 h
Virus addition 0.4 h Virus addition 0.3 h Virus addition 0.3 h
Feeding 0.4 h Feeding 0.3 h Feeding 0.3 h
Total set-up 1.8 h Total set-up 7.1 h Total set-up 3.1 h
Manpower
per day
Medium analysis and cell
counting (for 5 rollers)
f)
0.8 h
Medium analysis,
cell counting 0.5 h
Medium analysis,
cell counting 0.5 h
Flushing rollers with
oxygen 0.3 h
Total per day 1.1 h Total per day 0.5 h Total per day 0.5 h
Typical process time 7 d Typical process time 7 d Typical process time 7 d
Total 7.7 h Total 3.5 h Total 3.5 h
Manpower
cleaning
Sterilization 0.2 h Sterilization 0.5 h Sterilization 0.5 h
Disassembly 1 h Disassembly 0.4 h
Cleaning 2 h Cleaning 0.5 h
Total 0.2 h Total 3.5 h Total 1.4 h
Total
Man-power
costs
Total manpower 9.7 h 14.1 h 8.0 h
Manpower costs
per hour
g)
100, 1
100, 1 100, 1
970, 1 1410, 1 800, 1
Total Process
costs
1789, 1 2224, 1 1297, 1
a) Depreciation period of 5 years (1780 days).
b) 2 days for washing and assembling included.
c) 25% of autoclave volume required.
d) 50% of autoclave volume required.
e) 8% of autoclave volume required.
f) Per day, five roller bottles out of the 20 are sampled to obtain
a representative value for dissolved oxygen concentration, cell
number and viability.
g) Labor costs include costs for lab technician as well as all other
material, administration and infrastructure costs. Costs which
are common to all systems (consumables and manpower for
medium preparation, harvesting) were not considered.
straightforward implementation protocols
which are applicable to the vast majority
of product proteins. Future research and de-
velopment in the area of BEVS will likely
shift from culture medium design and pro-
cess engineering to metabolic engineering
for more efficient nutrient utilization [55]
and/or production of mammalian-like gly-
cosylation profiles [56] (see Part IV, Chap-
ters 2 and 7). A new era of genetically opti-
mized insect cells and baculoviral vectors
will almost certainly boost the success of
previous process engineering investiga-
tions, and this may eventually result in
BEVS-produced biopharmaceuticals.
Acknowledgments
The authors thank Cornelia Weber for crit-
ical comments on the manuscript. These
studies were supported by Cistronics Cell
Technology GmbH, Einsteinstrasse, P.O.B.
145, CH-8093 Zurich, Switzerland, the
Swiss National Science Foundation (grant
no. 631-065946) and the Swiss Federal Of-
fice for Education and Science (BBW)
within EC Framework 6.
References
1 L. A. King, R. D. Possee: The baculovirus ex-
pression system. Chapman & Hall, London,
1992.
2 V. A. Luckow, M. D. Summers, Bio/Technology
1988, 6, 14061410.
3 R. D. Possee, Curr. Opin. Biotechnol. 1997, 8,
569572.
4 L. Ikonomou, Y. J. Schneider, S. N. Agathos,
Appl. Microbiol. Biotechnol. 2003, 62, 120.
5 W. Weber, E. Weber, S. Geisse, K. Memmert,
Cytotechnology 2002, 38, 7785.
6 J. Hollister, E. Grabenhorst, M. Nimtz, H.
Conradt, D. L. Jarvis, Biochemistry 2002, 41,
1509315104.
7 N. Tomiya, M. J. Betenbaugh, Y. C. Lee, Acc.
Chem. Res. 2003, 36, 613620.
8 J. E. Vialard, M. A. Basil, C. D. Richardson: In-
troduction to the molecular biology of baculo-
viruses. In: Methods in molecular biology, C. D.
Richardson (Ed.). Humana Press; 1995: 203
224. Baculovirus expression protocols, vol. 39.
9 M. S. Kwon, T. Dojima, M. Toriyama, E. Y.
Park, Biotechnol. Prog. 2002, 18, 647651.
10 B. Philipps, M. Forstner, L. M. Mayr, Biotech-
niques 2004, 36, 8083.
11 H. R. Lo, Y. C. Chao, Biotechnol. Prog. 2004,
20, 354360.
12 S. Maeda, Annu. Rev. Entomol. 1989, 34, 351
372.
13 L. Ikonomou, G. Bastin, Y. J. Schneider, S. N.
Agathos, In Vitro Cell. Dev. Biol. Anim. 2001,
37, 549559.
14 E. J. Schlaeger, Cytotechnology 1996, 20, 57
70.
15 T. J. Wickham, T. Davis, R. R. Granados, M. L.
Shuler, H. A. Wood, Biotechnol. Prog. 1992, 8,
391396.
16 T. R. Davis, T. J. Wickham, K. A. McKenna,
R. R. Granados, M. L. Shuler, H. A. Wood, In
Vitro Cell Dev. Biol. Anim. 1993, 29A, 388
390.
17 L. A. Palomares, C. E. Joosten, P. R. Hughes,
R. R. Granados, M. L. Shuler, Biotechnol. Prog.
2003, 19, 185192.
18 K. Viswanathan, S. Lawrence, S. Hinderlich,
K. J. Yarema, Y. C. Lee, M. J. Betenbaugh, Bio-
chemistry 2003, 42, 1521515225.
19 S. Watanabe, T. Kokuho, H. Takahashi, M. Ta-
kahashi, T. Kubota, S. Inumaru, J. Biol.
Chem. 2002, 277, 50905093.
20 K. T. K. Wong, C. H. Peter, P. F. Greenfield, S.
Reid, L. K. Nielsen, Biotechnol. Bioeng. 1996,
49, 659666.
21 P. Licari, J. E. Bailey, Biotechnol. Bioeng.
1992, 39, 432441.
22 J. F. Power, S. Reid, K. M. Radford, P. F.
Greenfield, L. K. Nielsen, Biotechnol. Bioeng.
1994, 44, 710719.
23 R. A. Taticek, M. L. Shuler, Biotechnol. Bioeng.
1997, 54, 142152.
24 E. Bedard, S. Perret, A. Kamen, Biotechnol.
Lett. 1997, 19, 629632.
25 A. W. Caron, J. Archambault, B. Massie, Bio-
technol. Bioeng. 1990, 36, 11331140.
26 K. M. Radford, S. Reid, P. F. Greenfield, Bio-
technol. Bioeng. 1997, 56, 7381.
27 V. Jger, Cytotechnology 1996, 20, 191198.
References 1061
28 E. Chico, V. Jger, Biotechnol. Bioeng. 2000,
70, 574586.
29 J. D. Yang, P. Gecik, A. Collins, S. Czarnecki,
H. H. Hsu, A. Lasdun, R. Sundaram, G.
Muthukumar, M. Silberklang, Biotechnol.
Bioeng. 1996, 52, 696706.
30 C. Bedard, A. Kamen, R. Tom, B. Massie, Cy-
totechnology 1994, 15, 129138.
31 L. C. Chan, P. F. Greenfield, S. Reid, Biotech-
nol. Bioeng. 1998, 59, 178188.
32 B. Maiorella, D. Inlow, A. Shauger, D. Harano,
Bio/Technology 1988, 6, 14061410.
33 P. E. Cruz, A. Cunha, C. C. Peixoto, J. Clem-
ente, J. L. Moreira, M. J. Carrondo, Biotechnol.
Bioeng. 1998, 60, 408418.
34 M. Y. Wang, T. R. Pulliam, M. Valle, V. N. Va-
kharia, W. E. Bentley, J. Biotechnol. 1996, 46,
243254.
35 Y. C. Hu, W. E. Bentley, Biotechnol. Prog.
1999, 15, 10651071.
36 A. Kamen, C. Bedard, R. Tom, S. Perret, B.
Jardin, Biotechnol. Bioeng. 1996, 50, 3648.
37 A. Kamen, R. Tom, A. W. Caron, C. Chavarie,
B. Massie, J. Archambault, Biotechnol.
Bioeng. 1991, 38, 619628.
38 C. Mitchell-Logean, D. W. Murhammer, Bio-
technol. Prog. 1997, 13, 875877.
39 A. Garnier, R. Voyer, R. Tom, S. Perret, B. Jar-
din, A. Kamen, Cytotechnology 1996, 22, 5363.
40 C. B. Elias, A. Zeiser, C. Bedard, A. A. Kamen,
Biotechnol. Bioeng. 2000, 68, 381388.
41 K. U. Dee, M. L. Shuler, Biotechnol. Prog.
1997, 13, 1424.
42 J. M. Liebmann, D. LaScala, W. Wang, P. M.
Steed, BioTechniques 1999, 26, 3642.
43 L. Maranga, T. F. Brazao, M. J. Carrondo, Bio-
technol. Bioeng. 2003, 84, 245253.
44 T. Ohkawa, K. Majima, S. Maeda, J. Virol.
1994, 68, 66196625.
45 R. E. Hawtin, T. Zarkowska, K. Arnold, C. J.
Thomas, G. W. Gooday, L. A. King, J. A. Kuzio,
R. D. Possee, Virology 1997, 238, 243253.
46 J. M. Slack, J. Kuzio, P. Faulkner, J. Gen. Virol.
1995, 76 (Pt 5), 10911098.
47 S. A. Monsma, M. Scott, Innovations 1997,
1619.
48 B. C. Bonning, P. W. Roelvink, J. M. Vlak, R. D.
Possee, B. D. Hammock, J. Gen. Virol. 1994,
75 (Pt 7), 15511556.
49 L. E. Pyle, P. Barton, Y. Fujiwara, A. Mitchell,
N. Fidge, J. Lipid Res. 1995, 36, 23552361.
50 W. Weber, E. Weber, S. Geisse, K. Memmert,
Catching the Wave: the BEVS and the Bio-
wave, E. Lindner-Olsson, N. Chatzissavidou,
E. Lllau (Eds) Dordrecht, Kluwer Academic,
2001.
51 V. Singh, Cytotechnology 1999, 30, 149158.
52 L. N. Pierce, P. W. Shabram, BioProcessing J.
2004, 3, 15.
53 P. Licari, J. E. Bailey, Biotechnol. Bioeng.
1991, 37, 238246.
54 K. M. Radford, C. Cavegn, M. Bertrand, A. R.
Bernard, S. Reid, P. F. Greenfield, Cytotech-
nology 1997, 24, 7381.
55 C. B. Elias, E. Carpentier, Y. Durocher, L. Bis-
son, R. Wagner, A. Kamen, Biotechnol. Prog.
2003, 19, 9097.
56 M. J. Betenbaugh, N. Tomiya, S. Narang, J. T.
Hsu, Y. C. Lee, Curr. Opin. Struct. Biol. 2004,
14, 601606.
Abstract
The widespread use of cell-free systems in
biomedical research laboratories reflects
their usefulness in producing functional
proteins. However, cell-free methods have
typically yielded only nanogram to micro-
gram quantities of proteins, which has lim-
ited their utility to functional studies. Cell-
free systems derived from many cell types
have been described in the scientific litera-
ture. For a small number of these cell types,
significant advances made in recent years
have seen the development of robust, cost-
effective and highly efficient cell-free ex-
pression systems suitable for the prepara-
tion of proteins in milligram quantities. In
this chapter, the advantages of cell-free pro-
tein synthesis methods, with particular em-
phasis on applications to structural biology,
will be discussed. Escherichia coli and wheat
embryo systems are the best-characterized
prokaryotic and eukaryotic high-efficiency
systems, respectively, and are the focus
here. The bulk of the chapter will be de-
voted to a discussion of recent advances in
cell-free synthesis methods that have facili-
tated the production of proteins in high
yield that are soluble, intact, and functional.
Recent advances in cell-free expression sys-
tems that are amenable to automation and
high-throughput screening, and therefore
well-suited for the development of biophar-
maceuticals, will also be discussed.
Abbreviations
AMP adenosine monophosphate
ARS aminoacyl-tRNA synthetase
CAT chloramphenicol acetyl-
transferase
CDP cytosine diphosphate
CTP cytidine triphosphate
dNTP desoxynucleotide triphosphate
DTT dithiothreitol
EF elongation factor
GSH reduced glutathione
GSSG oxidized glutathiones
HEPES 2-[4-(2-hydroxyethyl)-1-pipera-
zinyl] ethanesulfonic acid
IAM iodoacetamide
IF initiation factor
MAD multi-wavelength anomalous
dispersion
MTF methionyl-tRNA transformy-
lase factor
MWCO molecular weight cut-off
NAD nicotinamide adenine
dinucleotide
NMR nuclear magnetic resonance
1063
15
Robust and Cost-effective Cell-free Expression of Biopharmaceuticals:
Escherichia Coli and Wheat Embryo
Luke Anthony Miles
ISBN: 3-527-31184-X
PA plasminogen activator
PDI protein disulfide isomerase
PEP phosphoenol pyruvate
pIVEX plasmid for In Vitro
EXpression
PURE protein synthesis using
recombinant elements
RIL plasmid in E. coli strain BL21
RRF ribosome recycling factor
SDS-PAGE sodium dodecyl sulfate-poly-
acrylamide gel electrophoresis
TCA trichloroacetic acid
WG wheat germ
15.1
Introduction
15.1.1
Background
Cell-free protein synthesis refers to a fami-
ly of techniques in which ribosomes and
translation factors are isolated from cells
to synthesize polypeptides in vitro from a
messenger RNA template. Proteins can
also be expressed in vitro from a DNA
template by exploiting the ability of bacte-
riophage RNA polymerases to synthesize
mRNA transcripts from DNA.
Cell-free systems have long been used to
test the expression of constructs and to as-
sess the properties of the translation prod-
uct such as solubility, stability, and activity.
Instability and low efficiency limited their
usefulness to small-scale synthesis, and in
vivo methods were relied on for up-scaled
protein production to meet the demands
of consuming techniques such as those
used in structural biology.
Since the early investigations of Zubay on
developing Escherichia coli extracts for in vitro
translation[1], mucheffort has beenspent on
increasing the yields fromE. coli cell-free sys-
tems. Many laboratories have made modifi-
cations to extract preparation protocols and
reaction conditions that have seen substan-
tial increases in stability and efficiency of
these systems, and none more so than the
group of Yokoyama (RIKEN, Yokohama, Ja-
pan). This group has developedan E. coli cell-
free system capable of producing proteins at
a rate approaching 500 lg mL
1
reaction
mixture in the first hour of reaction [2, 3].
A significant leap forward was made in
1988 by Spirin et al. [4], who developed a
continuous-flow apparatus for the continu-
ous supplementation of reaction mixtures
with substrates required for protein syn-
thesis and continuous removal of reaction
by-products. In this way, it was shown that
the activity of a cell-free system could be
sustained for many hours, compared with
batch-mode reactions which became inac-
tive after approximately 45 minutes. Since
then, many reports have been made de-
scribing the use of simple dialysis systems
[3, 5] that can maintain the high productiv-
ity of a reaction over many hours, without
the use of a cumbersome apparatus.
A combination of these developments
has seen expression yields as high as
6 mg mL
1
achieved over 20 hours in an E.
coli cell-free system [3], making cell-free
methods a serious alternative to in vivo
methods of large-scale protein production.
A push has been made more recently to
develop productive systems that do not
rely on flow apparatus or dialysis mem-
branes, in order to make the methods
compatible with robot-controlled formats.
This has seen the emergence of fed-batch
mode reactions where reactions are peri-
odically supplemented with aliquots of re-
action substrates [6]. Such a system has
15 Robust and Cost-effective Cell-free Expression of Biopharmaceuticals: Escherichia Coli and Wheat Embryo 1064
been shown to prolong synthesis activity
for up to 3 hours, producing 750 lg of pro-
tein per mL reaction mixture [7].
Endo and co-workers at Ehime Univer-
sity, Matsuyama, Japan, have led the devel-
opment of the most promising eukaryotic
cell-free system to date, based on wheat
embryos. A significant advance made by
this group was the development of pEU
expression vectors that have overcome
many of the difficulties associated with
mRNA synthesis for translation in a eukar-
yotic system [8]. In addition to extensive
optimization of reaction conditions that
have seen improvements in protein syn-
thesis rates, Endo and colleagues have im-
proved wheat extract embryo preparation
protocols to enhance the stability of these
systems to a remarkable extent [9]. When
coupled with the dialysis mode of reaction,
Endo et al. were able to maintain transla-
tional activity in a coupled transcription/
translation wheat embryo reaction for
150 hours, producing 5 mg of enzymati-
cally active protein per mL reaction mix-
ture [10]. This again represents a serious
alternative to in vivo methods of large-scale
protein production.
15.1.2
Some Advantages to Cell-free Synthesis
Cell-free synthesis is often used to express
troublesome proteins. For example, it is
well-suited to the expression of toxic pro-
teins because there is no need to consider
the action of a gene product on the viabili-
ty of a host cell. The open nature of cell-
free methods makes them useful for pre-
paring proteins prone to mis-folding, ag-
gregation into inclusion bodies, and pro-
teolytic digestion. This is because the ex-
pression reaction conditions can be readily
manipulated to promote proper protein
folding and inhibit proteolysis. In addition,
polymerase chain reaction (PCR)-generated
DNA templates can be used in cell-free
systems, obviating the need for time-con-
suming cloning procedures in the initial
steps of DNA template and reaction condi-
tion optimization. This also makes cell-
free synthesis well-suited to high-through-
put and robot-controlled formats. Unpuri-
fied reaction mixtures can often be used
directly in functional assays, greatly expe-
diting the analysis of expression trials.
In the broadest terms, cell-free synthesis
offers advantages over traditional in vivo
methods of recombinant protein expres-
sion, as it enables the experimentalist
strictly to control conditions under which
expression reactions are performed. Bene-
fits of such an open expression system are
exemplified by the fact that cell-free reac-
tion conditions can be sufficiently modi-
fied to support co-translational folding of
active disulfide-bonded proteins.
15.1.3
Cell-free Synthesis in Structural Biology
Many techniques used in structural biology
and biophysics rely on preparing proteins
incorporating costly isotope and heavy-atom
labels. The group of Kainosho prepared pro-
tein samples that were isotopically labeled
with a chemically synthesized amino acid
using in vitro and in vivo E. coli expression
systems [2]. Consumption of the expensive
amino acid was shown to be two orders of
magnitude lower in the cell-free system
than in the in vivo expression system. Low
levels of endogenous amino acids in cell-
free systems also result in high-level incor-
poration of labeled amino acids [2]. In addi-
tion, scrambling effects resulting from the
metabolism of amino acids in vivo from
one type to another is not observed in vitro.
For all these reasons, cell-free synthesis
methods have become indispensable to
many structural biology and biophysics la-
boratories (see Part VII, Chapter 2).
Crystals from selenium-methionine- and/
or selenium-cysteine-labeled proteins can
be studied by multi-wavelength anomalous
dispersion (MAD) phasing techniques that
can facilitate the solution of an X-ray crystal
structure from a single crystal form [11].
However, if in vivo expression systems are
used to prepare selenium-labeled proteins,
amino acid metabolism and the toxicity of
Se-methionine can result in low protein
yields and low incorporation rates.
All but the smallest proteins studied by
nuclear magnetic resonance (NMR) need
to be labeled with isotopes of hydrogen, ni-
trogen, or carbon (
2
H,
15
N,
13
C), or combi-
nations thereof. Cell-free systems can be
used for uniform, selective, or site-specific
labeling of proteins with isotopically en-
riched amino acids, and relatively inexpen-
sive algal amino acids can be used for this
purpose. Unlike the in vivo expression of
labeled proteins in minimal media, expres-
sion conditions and synthesis yields are
the same for expression of natural abun-
dance and isotopically enriched proteins.
Selective labeling with single amino acid
types can be carried out in tandem to help
make rapid and unambiguous resonance
assignments from NMR spectra.
15.2
Transcription
Transcription and translation are coupled
in the cytoplasm of prokaryotes. In eukar-
yotes, transcription of DNA and subse-
quent processsing of the messenger RNA
transcript both occur in the nucleus, while
translation of the mature mRNA template
occurs in the cytoplasm. Cell-free systems
lack mechanisms for mRNA processing,
and therefore require fully matured
mRNA templates for translation. DNA
templates are similarly transcribed by bac-
teriophage RNA polymerases (T7, SP6,
and T4) (see Part III, Chapter 2) for trans-
lation in both E. coli (prokaryotic) and
wheat embryo (eukaryotic) systems. How-
ever, these systems require differently
structured mRNA templates for efficient
translation to occur. Details of the con-
struction and treatment of different DNA
templates for generating translatable
mRNA strands are discussed here.
15.2.1
Plasmid DNA Templates for E. coli Systems
Any plasmid designed to express recombi-
nant proteins in E. coli under the control
of a bacteriophage RNA polymerase pro-
moter can be used in an E. coli cell-free
system. As a rule of thumb, constructs
that express well in vivo tend to express
well in vitro. T7 RNA polymerase is often
considered intrinsically more efficient than
other RNA polymerases (see Part IV, Chap-
ter 12). However, this differential efficiency
can often be attributed to differential sen-
sitivity to salt concentration. Because of
the relative robustness and efficiency of
the T7 polymerase, plasmids under the
control of a T7 promoter are almost exclu-
sively used in E. coli cell-free systems.
pIVEX (plasmid for In Vitro EXpression)
vectors (Roche) have been developed spe-
cifically for the optimal expression in E.
coli extracts, and are available with a range
of affinity tags for detection and purifica-
tion. pIVEX vectors are very high copy
number, which is useful for large-scale
protein synthesis in vitro where large
quantities of highly purified DNA template
are required. However, one limitation of
pIVEX vectors is that they are non-induci-
ble and therefore cannot be used for in-
duced expression trials in bacteria.
15.2.2
Plasmid DNA Templates for Wheat Embryo
Systems
Efficient translation in eukaryotic systems
traditionally required messenger RNA
template with a capped 5' end, optimally
structured 5' and 3' untranslated regions
(UTRs), and a polyadenylated tail in the 3'-
UTR of the transcript. 5'-capped mRNA
has traditionally been synthesized in vitro
by transcribing the DNA template in the
presence of cap analogues such as 7-
methylguanosine in addition to the
dNTPs. The ability of 5' cap analogues to
bind initiation factor eIF-4E means that
free cap analogue can also occupy eIF-4E
and render it unavailable for translation.
In fact, researchers who work to develop
improved cap analogues test the activity of
a candidate molecule for its ability to inhi-
bit in vitro translation in its free form.
Naturally, the use of 5' cap analogues to
produce capped mRNA dictates that tran-
scription and translation reactions must be
performed separately. Their use also neces-
sitates careful optimization of cap analo-
gue concentration in the transcription re-
action or purification of the mRNA tem-
plate before use in a translation reaction.
In an effort to eliminate the require-
ment for 5' cap analogue incorporation,
Endo and Sawasaki explored the genomes
of plant RNA viruses that have positive-
sense RNA lacking 5' caps and 3' polyA
tails [8]. In those studies, the sequence en-
coding the 5'-UTR in a pSP65 vector was
replaced with the omega 71 (X71) se-
quence of tobacco mosaic virus. Uncapped
mRNA from this construct encoding for
luciferase proved to be highly active in a
wheat embryo cell-free assay. Further en-
hancement in activity was achieved by
modifying the 5' end of the X71 sequence
to GAA. With the 5'-UTR comprised of
GAA-X71, Endo and Sawasaki systemati-
cally modified the length and sequence of
the 3'-UTR that works synergistically with
the 5'-UTR to enhance transcript stability
and initiation of translation. It was found
that the activity of transcripts were more
sensitive to the length of the 3'-UTR than
to variations in the 3'-UTR sequence.
mRNA synthesized with a 1626-nucleotide
3'-UTR, and the GAA-X71 5'-UTR transla-
tion enhancer proved to be almost as ac-
tive as the capped and polyadenylated
mRNA transcribed from the unmodified
pSP65 vector.
As a result of those studies, Endo et al.
developed the pEU series of vectors with
various affinity tags for efficient expression
in wheat embryo cell-free systems without
the need for expensive cap analogues. pI-
VEX WG (wheat germ) vectors are now
commercially available (Roche) which sim-
ilarly rely on translation-enhancing un-
translated regions flanking the multi-clon-
ing site.
15.2.3
Linear DNA Templates
Some regions of a plasmid vector such as
the origin of replication and antibiotic re-
sistance genes are not required for gene
expression. Linear DNA constructs are
suitable templates for in vitro expression if
the elements that regulate transcription
(RNA polymerase promoter/terminator)
and translation (ribosome binding site,
start/stop codons) in a vector are appended
to the gene to be expressed. These tem-
plates are generated and amplified by the
PCR and can be used to rapidly screen
protein expression and folding without the
need for cloning into a plasmid vector.
PCR products can generally be used direct-
ly in a cell-free reaction, without the need
for purification of the template.
Kits for the generation of DNA tem-
plates by PCR can be purchased from sev-
eral suppliers, including E. coli and wheat
embryo pIVEX linear DNA template kits
(Roche). Similarly, Endo and Sawasaki [8]
have demonstrated efficient translation in
a wheat embryo system from PCR-gener-
ated linear DNA templates based on the
structure of their pEU plasmid series.
PCR-generated DNA templates are excel-
lent for rapid batch-mode screening of do-
main boundaries, product solubility and
expression levels, and for generating mu-
tants. PCR conditions must be optimized
to ensure that a single product is obtained,
and high-fidelity enzymes should be used
to minimize the introduction of random
mutations. For scale-up to dialysis mode,
linear templates are less useful as they are
prone to degradation by nucleases and
plasmids are less expensive to amplify.
Successful PCR-generated constructs can
be sub-cloned directly into an appropriate
plasmid vector.
15.2.4
Plasmid Purification
DNA template preparations are a major
source of contamination in cell-free syn-
thesis. In vitro translation reactions are
sensitive to contamination by salts,
RNases, detergents, and alcohol, all of
which are typically used in commercial
plasmid purification kits. As such, plas-
mids prepared by resin-based purification
protocols must be further purified by phe-
nol/chloroform and chloroform extraction.
DNA should be precipitated with isopropa-
nol, washed in ethanol, dried, and taken
up in RNase-free water to a concentration
of about 1 mg mL
1
. The DNA should be
stored frozen, in small aliquots.
15.2.5
Transcription Reaction
Bacteriophage RNA polymerases are avail-
able from a range of commercial suppli-
ers, and transcription-only reactions
should be performed according to the
manufacturers instructions. The concen-
tration of mRNA added to a subsequent
translation reaction must be established
empirically. Conditions for coupled tran-
scription/translation are described in Sec-
tion 15.6.
15.3
Translational
15.3.1
The PURE System
An important contribution to fundamental
cell-free science has come from the labora-
tory of Ueda et al. at the University of To-
kyo, Japan [12]. This group reconstituted a
cell-free system with purified recombinant
translation factors used by E. coli, produc-
ing 160 lg mL
1
h
1
protein in a simple
batch-mode reaction. That system is re-
ferred to as the PURE (Protein synthesis
Using Recombinant Elements) system.
Ribosomes needed for translation in the
PURE system are isolated from E. coli
using sucrose-density gradient centrifuga-
tion. The protein factors necessary for
translation in E. coli are recombinantly ex-
pressed as His-tagged fusions, and puri-
fied to homogeneity. These include the fac-
tors for initiation (IF1, IF2, and IF3), elon-
gation (EF-G, EF-Tu, EF-Ts), peptide chain
release (RF1 and RF3), ribosome recycling
(RRF), methionyl-tRNA transformylase
(MTF) for formylation of the initial Met-
tRNA, and the 20 aminoacyl-tRNA synthe-
tases (ARSs) for transfer RNA (tRNA) recy-
cling. These components are combined
with RNA polymerase (His-tagged) for
transcription, NTPs, 46 tRNAs, folinic
acid, amino acids and factors involved in
ATP regeneration. The complete composi-
tion of a PURE transcription/translation
reaction is given in Table 15.1.
The PURE system is important as it def-
initively established the minimum require-
ments for high-level expression in an E.
coli cell-free system. The benefits of this
system include the elimination of pro-
teases and nucleases, and stability of the
ATP concentration. All of these factors
make the system well-suited to fed-batch
reaction formats for robot-controlled syn-
thesis. Since most of the proteins in the
PURE system are His-tagged, they are
readily removed from the reaction by che-
lating resins, and purification of a non-
His-tagged translation product is dramati-
cally simplified.
The main drawback of the PURE system
is that preparing the reaction components
Table 15.1 Composition of the PURE cell-free systems from E. coli
Component Concentration
Ribosomes 240 pmol mL
1
Initiation Factor 1 (IF1) His-tagged 20 lg mL
1
1
1
Extension Factor G (EF-G) His-tagged 20 lg mL
1
Extension Factor Ts (EF-Ts) His-tagged 20 lg mL
1
Extension Factor Tu (EF-Tu) His-tagged 20 lg mL
1
Release Factor 1 (RF1) His-tagged 10 lg mL
1
Release Factor 3 (RF2) His-tagged 10 lg mL
1
Ribosome recycling factor (RRF) His-tagged 10 lg mL
1
20 Aminoacyl-tRNA synthetases (ARSs) His-tagged 6006000 U mL
1
Methionyl-tRNA Transformylase (MTF) His-tagged 6000 U mL
1
T7 RNA polymerase His-tagged 10 lg mL
1
Mixture of 46 tRNAs 56 A
260
U mL
1
Creatine kinase 4 lg mL
1
Myokinase 3 lg mL
1
Nucleoside-diphosphate kinase 1.08 lg mL
1
Pyrophosphatase 2 U mL
1
Creatine phosphate 10 mM
Magnesium acetate 9 mM
Potassium phosphate 5 mM
Potassium glutamate 93 mM
Ammonium chloride 5 mM
Calcium chloride 0.5 mM
Spermidine 1 mM
Putrescine 8 mM
Dithiothreitol (DTT) 1 mM
ATP and GTP 2 mM
CTP and UTP 1 mM
Folinic acid 10 lg mL
1
Each of 20 amino acids 0.1 mM
requires cloning, expression, and purifica-
tion of more than 30 proteins. In addition,
the system lacks factors endogenous to dif-
ferent cells that might enhance translation
and facilitate proper protein folding, such
as chaperones. For these reasons, most
cell-free systems are derived from crude
cell extracts containing ribosomes, transla-
tion factors, other endogenous proteins,
and tRNAs.
15.3.2
Cell Extracts
The best-characterized cell-free systems de-
rived from crude extracts are the E. coli
(prokaryotic) and wheat embryo (eukary-
otic) systems. The extract preparation pro-
tocols used are central to the productivity
of any cell-free system.
15.3.2.1 E. coli Extracts
As with in vivo expression of recombinant
proteins in E. coli, different cell lines can
be used to prepare E. coli extracts for in vi-
tro expression. A19 E. coli is commonly
used for E. coli preparation. BL21 lines ex-
hibit reduced protease activity, and are pre-
ferred when expressing proteolytically la-
bile species. Yokoyama and co-workers [13]
recommend the use of BL21 codon-plus
RIL (BL21 CP, Stratagene) E. coli that con-
tain extra copies of the argU, ileY, and
leuW tRNA genes. BL21 CP extracts are
enriched in tRNAs that recognize minor
codons for arginine (AGA and AGG), iso-
leucine (AUA), and leucine (CUA). Ex-
tracts prepared from BL21 CP can increase
expression yields up to 50% on those ob-
tained from BL21 or A19.
Many variations on the protocol for pre-
paring E. coli extract have appeared in the
literature. However, Kigawa et al. [13] re-
cently published the most comprehen-
sively documented protocol for E. coli ex-
tract preparation since that described by
Pratt [14]. The success of Yokoyamas
group in determining the structures of
more than 100 proteins prepared by cell-
free synthesis in recent years recommends
their protocol as the most reliable guide
for E. coli extract preparation.
The conditions under which E. coli are
grown and harvested have a marked im-
pact on the activity of the E. coli extract ob-
tained. The choice of growth media, tem-
perature, aeration conditions and point-of-
harvest must be optimized to suit the
choice of cell-line, and to compensate for
different laboratory conditions. Once opti-
mized, growth conditions must be strictly
adhered to in order to minimize batch-to-
batch variation. Kigawa et al. [13] described
the growth of BL21 CP strain in 500 mL
lots of 2YT medium (Table 15.2), incubated
in 2-L baffled flasks at 378C with circular
shaking at 160 r.p.m., and harvested at
OD
600
%3 by centrifugation.
Cells are washed at least three times in
S30 buffer (see Table 15.2) containing
0.5 mL L
1
2-mercaptoethanol. Pelleted cells
are then frozen in liquid nitrogen and
stored at 808C for up to 3 days. Thorough
washing of the cells is difficult under these
conditions, as E. coli tend clump together in
this buffer; however, a Polytron cell homo-
genizer (Kinematica AG, PT-MR3100) was
used to overcome this problem.
Washed cells are thawed at 48C in S30
buffer (50100 mL) containing 0.5 mL L
1
2-mercaptoethanol. Cells are again pelleted
(16000g, 30 min, 48C) and resuspended
in S30 buffer. The volume of S30 buffer
used to resuspend cells is dependent upon
the pellet mass, but should be close to
1.3 mL per gram E. coli. Cells are then
lysed in this suspension.
Several lysis methods can be used, de-
pending on the facilities available. Sonica-
tion can be used, but should be regarded
as a last resort because of poor reproduc-
ibility and problems with local sample
heating. Multiple passes through a French
press or cell crusher can be used to repro-
ducibly prepare highly active E. coli ex-
tracts, as long as every effort is made to
minimize sample heating and eliminate
RNases from the apparatus. The method
of lysis involves disruption of cells (7 g in
8.9 mL S30 buffer) with 22.7 g of glass
beads (B. Braun Melsungen, AG 854-150/
7, =0.170.18 mm) in a Multi-beads
Shocker Type MB301 (Yatsui Kikai, Japan),
switched on for three, 30-s periods sepa-
rated by 30 s at rest. The authors claim
that this method of lysis leads to greater
reproducibility than the French press
method.
The lysate is then centrifuged (30000g,
30 min, 48C) to remove glass beads and/or
cell debris. The supernatant is carefully
decanted off and subjected to further cen-
trifugation (30000g, 30 min, 48C). The
supernatant obtained is incubated with
0.3 volumes of pre-incubation buffer (see
Table 15.2) at 378C for 80 min, and dia-
lyzed four times against 50 volumes of S30
buffer (45 min each at 48C). The final E.
coli extract is the supernatant obtained
after centrifugation (4000g, 10 min, 48C)
of the dialyzed mixture. This extract is im-
mediately frozen in liquid nitrogen in
small aliquots, and stored in liquid nitro-
gen or at 808C until required (it is stable
for at least one year).
15.3.2.2 Wheat Embryo Extracts
Selected strains of E. coli can be grown un-
der controlled conditions in a highly repro-
ducible manner. Wheat embryo extracts
suffer from batch-to-batch variation due to
differences in the conditions under which
the wheat is grown, harvested, milled, and
stored. This source of variation in the ac-
tivity of wheat embryo extracts can be off-
set by the simplicity of the preparation
protocol. Small amounts of extract can be
quickly prepared from embryos obtained
from different sources, and the activity of
each assessed. Large batches of extract can
then be prepared from the best embryo
source and stored at 808C until required.
As with E. coli extract, wheat embryo ex-
tracts are stable for more than a year if
stored in liquid nitrogen, or at 808C.
Table 15.2 Composition of media and buffers for E.
coli and wheat embryo extract preparations.
Reagent Concentration
2YT Growth Media
Tryptone 16 g L
1
Yeast extract 10 g L
1
NaCl 5 g L
1
De-ionized water Make up to 1 L
5 N NaOH Adjust pH to 7.0
S30 Buffer
Potassium acetate 60 mM
Tris-acetate 10 mM
DTT 1 mM
E. coli Lysate Pre-incubation Buffer
Pyruvate kinase 6.7 U mL
1
Tris-acetate (pH 8.2) 293 mM
Magnesium acetate 9.2 mM
ATP 13.2 mM
Phosphoenol pyruvate 84 mM
DTT 4.4 mM
Each amino acid 40 lM
Wheat germ (WG) Buffer
HEPES, pH 7.6 40 mM
Potassium acetate 100 mM
Calcium chloride 2 mM
DTT 4 mM
Each amino acid 300 lM
Freshly milled raw wheat germ is readily
available from mills or granaries. Intact
embryos are separated from damaged em-
bryos and debris by solvent flotation. Car-
bon tetrachloride or chloroform is mixed
with cyclohexane in a 600: 240 mL ratio
until Schlieren mixing lines are no longer
visible. Wheat germ (40 g) is added to the
solvent mixture and stirred thoroughly to
remove endosperm from embryos. The
suspension is allowed to stand until good
separation is achieved. Floating embryos
are harvested in a Buchner funnel, air-
dried, and set aside in a fume hood. The
remaining solvent is filtered for re-use,
and the process repeated until the desired
quantity of embryos is obtained. Embryos
are left to stand overnight in a fume hood.
The protocol for extract preparation giv-
en here is slightly modified from that de-
scribed by Endo and colleagues [9]. At this
stage, the wheat embryos are covered with
endosperm which contains ribosome-inac-
tivating proteins such as RNA N-glycosi-
dase and tritin [9]. Most endosperm is re-
moved by washing the embryos three
times in 10 volumes of water. Residual en-
dosperm is removed by sonicating for
3 min in a 0.5% Nonidet P-40 solution
and for a further 3 min in RNase-free
water. Clean embryos are then ground in
liquid nitrogen in a mortar and pestle. A
portion (5 g) of the ground powder is sus-
pended in 5 mL ice-cold 2 WG Buffer
(see Table 15.2), vortexed, and centrifuged
at 30000g for at least 30 min at 48C. The
supernatant is then desalted on a G25
(fine) column and re-centrifuged
(30000g for 10 min, 48C). Endo et al.
recommend dilution of the extract with
WG Buffer to 200 A260 mL
1
, and storage
in small aliquots at 808C.
15.4
Treatment of Extracts for Synthesis
of Disulfide-bonded Proteins
The reducing environment of the E. coli
cytoplasm prevents the formation of disul-
fide (SS) bonds that are often essential to
correct folding and functioning of proteins
and complexes. As a consequence, SS-
bonded proteins can misfold when ex-
pressed in bacteria. Misfolded polypeptide
chains are proteolytically labile, susceptible
to aggregation, and are often found packed
into inclusion bodies.
Recovering misfolded proteins can be
time-consuming because of the many pa-
rameters that need to be optimized for sol-
ubilization and then refolding. Conditions
that might need to be screened in a re-fold
include temperature, protein concentra-
tion, redox potential, pH, ionic strength,
and the presence of chaperones, ligands
and cofactors. Alternatively, the protein of
interest can be expressed conjugated to a
signal peptide that will see the protein ex-
ported to the E. coli periplasm where SS
bonding is supported. However, periplas-
mic folding typically results in substan-
tially lower protein yield.
Recent attempts have seen modifications
made to E. coli and wheat embryo cell-free
systems for synthesizing proteins in an en-
vironment suitable for co-translational for-
mation of disulfides. In general, these
studies have involved translation under
conditions typically screened in refolding
experiments and assessing their effect on
total and functional protein yield. The fo-
cus has been on batch-mode reaction sys-
tems because, as with refolding, optimum
conditions for producing correctly folded
proteins vary between constructs and
batch-mode reactions are compatible with
high-throughput condition screening.
Only a limited number of reports de-
scribing the synthesis of a few model SS-
bonded proteins have appeared in the lit-
erature. However, this reflects the relative
novelty of producing significant quantities
of SS-bonded proteins by cell-free meth-
ods, rather than the potential of these
methods to serve this purpose.
15.4.1
Sulfhydryl Redox Potential
Reducing agents such as dithiothreitol
(DTT) are added at the early stages of cell-
free lysate preparation to stabilize lysate
components during preparation and stor-
age. A strongly reducing environment is
usually maintained over the course of a
translation reaction. Naturally, this envi-
ronment inhibits SS-bond formation. An
oxidizing environment is created in the
translation reaction by omitting reducing
agents from reaction mixtures, removing
reducing agents from cell lysates by gel fil-
tration, or by dialysis just prior to their
use, and by adjusting the sulfhydryl redox
potential with the glutathione redox pair.
In one example [15], genes for the heavy
and light chains of a catalytic antibody,
6D9, were simultaneously expressed in a
batch E. coli coupled transcription/transla-
tion system that was depleted of reducing
agents and buffered with oxidized and re-
duced glutathione (GSSG/GSH). The two
proteins formed a single intermolecular
disulfide bond, co-translationally, yielding
an enzymatically active Fab construct. This
treatment seems sufficient for SS-bonds
to form in cases where only one or two
disulfides are present, or where a dialysis
system is used to buffer the redox poten-
tial.
Enzymes such as glutathione reductase
and thioredoxin reductase that are respon-
sible for maintaining a reducing environ-
ment in the E. coli cytoplasm are active in
extracts prepared from standard E. coli
strains. This is demonstrated by the fact
that GSSG is rapidly reduced when incu-
bated in E. coli lysate. This makes it im-
possible to maintain a stable redox poten-
tial optimal for SS-bond formation.
Further modification of the E. coli extract
is necessary for synthesis of multiple SS-
bonded proteins in their functional forms,
particularly in batch mode.
A strategy has been devised [16] to inac-
tivate disulfide-reducing enzymes in E. coli
lysates by covalently blocking free sulfhy-
dryl groups present. This is done by incu-
bating a standard E. coli extract with 2 mM
iodoacetamide (IAM) as alkylating reagent
at room temperature for 30 minutes. Re-
sidual IAM is then removed by extensive
dialysis against S30 buffer, or by gel filtra-
tion. The same authors showed that IAM-
treated E. coli extract can maintain a 4: 1
GSSG: GSH ratio over 18 hours of incuba-
tion. Only a small loss (15%) in translation
activity of the IAM-treated E. coli extract
was observed relative to untreated extract
in batch mode. Further studies from the
Swartz laboratory indicated that the use of
a lower IAM concentration (1 mM) dis-
ables reduction pathways sufficient for the
purpose of maintaining an oxidizing sulf-
hydryl redox potential, and obviates the
need for extensive dialysis to remove ex-
cess IAM [17].
15.4.2
pH
The initial pH of a cell-free reaction
should be around 7.6 for maximum trans-
lation activity, though this may need to be
adjusted with Tris-acetate buffer. However,
the initial pH can be varied between 6.5
and 8.5 to improve the activity/solubility of
a particular translation product. The for-
mation of disulfide bonds is enhanced at
the higher end of this pH range. The se-
rine protease domain of murine urokinase,
containing six disulfides, was expressed in
an IAM-treated E. coli batch reaction [16].
The yield of functional protein was almost
doubled by increasing the pH from 7.5 to
8.2, despite a small drop in total protein
yield.
15.4.3
Chaperones
Co-expression of chaperones such as
GroEL/GroES, DnaK/DnaJ/GrpE can en-
hance the solubility of proteins expressed
by cell-free or in vivo methods. Chaperones
that enhance solubility and enzymes that
promote disulfide bond formation can be
added in a controlled way to cell-free reac-
tions. The impact of these agents on the
yield of functional protein still needs to be
determined empirically. In the case of
murine urokinase [16], the addition of
DsbC disulfide isomerase from E. coli to
an IAM-treated E. coli reaction increased
the rate of functional urokinase formation
and doubled the yield of functional pro-
tein. When protein disulfide isomerase
(PDI) from humans was added to the
same expression system, the impact on
urokinase activity was negligible. The
same group expressed a plasminogen acti-
vator (PA) protein containing nine disul-
fide bonds in a similar reaction system
containing DsbC [17]. It was found that
addition of the E. coli periplasmic chaper-
one Skp not only increased yield of active
protein, but also extended the life of the
expression reaction. Similarly, co-expres-
sion of the heavy and light chains of the
catalytic antibody 6D9 showed that addi-
tion of the glutathione redox pair and PDI
increased both yield of soluble 6D9 Fab
fragment and total expression yield [15].
15.4.4
Wheat Embryo System
Endo and colleagues [18] expressed a single-
chain antibody variable fragment (scFv) in a
wheat embryo batch-mode reaction with
DTT eliminated from the reaction mixture.
Synthesized scFv consisting of heavy and
light chain variable domains of an antibody
connected by a linker was shown to be ac-
tive, indicating the correct formation of
both intradomain disulfide bonds. The addi-
tion of PDI had a positive impact on prod-
uct solubility, showing that wheat embryo
cell-free systems can be used to prepare ac-
tive disulfide-bonded proteins in batch
mode. Unfortunately, the productivity of
the system fell by almost 50% when DTT
(usually at 4 mM) was excluded from the re-
action. This is a marked response compared
with the 15% drop in activity for the IAM-
treated, non-reducing E. coli system de-
scribed elsewhere [16]. When it is consider-
ed that wheat embryo systems are much
more stable but much less efficient than
E. coli cell-free systems, this significant fall
in activity on removing DTT suggests that
scale-up to dialysis-mode protein synthesis
in a non-reducing wheat embryo system is
likely to be problematic.
15.5
ATP Regeneration Systems
Adenosine triphosphate (ATP) is the pri-
mary energy source for cell-free synthesis
reactions. Rapid depletion of ATP leads to
the cessation of translation in under an
hour in batch mode. ATP regeneration sys-
tems are used to extend the life of transla-
tion reactions by maintaining a stable ATP
concentration. ATP is regenerated by the
enzyme-catalyzed transfer of high-energy
phosphate bonds from a secondary energy
store to ADP. The productivity of any cell-
free system relies on the efficiency and
stability of the ATP regeneration system.
Creatine phosphate (1) is most com-
monly used as the secondary energy store
in both E. coli and wheat embryo cell-free
systems. However, other secondary energy
stores (25) can be used to replace or sup-
plement the creatine phosphate:
1 creatine phosphate+ADP
creatine kinase
! creatine+ATP
2 phosphoenol pyruvate (PEP) +ADP
pyruvate kinase
! pyruvate+ATP
3 acetyl phosphate+ADP
acetate kinase
! acetate+ATP
4 ADP+ADP
myokinase
! AMP+ATP
5 CTP+ADP
CDP kinase
! CDP+ATP
Pyruvate dehydrogenase and pyruvate-for-
mate lyase are enzymes that are endoge-
nous to E. coli and which process pyruvate
into acetyl coenzyme A. Acetyl phosphate
is then produced from acetyl coenzyme A
by phosphotransacetylase, also endogenous
to E. coli. In this way, pyruvate can be pro-
cessed in E. coli to yield acetyl phosphate,
a phosphate donor (3). The activities of
these pathways in E. coli lysate were tested
[7] by supplying a batch-mode reaction
with pyruvate in the absence of an alterna-
tive secondary energy store.
Protein synthesis was observed in the
pyruvate-only system. However, addition of
cofactors involved in pyruvate processing
(0.33 mM NAD, 0.27 mM coenzyme A) en-
hanced protein synthesis to 70% of that
observed when a PEP/pyruvate kinase (2)
secondary energy store was used. Further-
more, supplying these cofactors and a PEP
synthetase inhibitor (oxalate) to a PEP/py-
ruvate kinase batch reaction substantially
improved ATP regeneration and protein
synthesis activity.
The improved PEP/pyruvate system de-
veloped by Kim and Swartz [7] yielded
350 lg mL
1
of chloramphenicol acetyl-
transferase (CAT) in the first hour of an E.
coli batch reaction. The same reaction sys-
tem yielded 750 lg mL
1
CAT in 3 hours
when periodically fed with amino acids,
PEP, and magnesium acetate [6]. This is
remarkable efficiency for a system lacking
a dialysis membrane, and is a feature that
well suits robotic handling and high-
throughput screening (HTS) strategies.
15.6
Reaction Conditions
15.6.1
E. coli Systems
Conditions for performing coupled tran-
scription/translation reactions in an E. coli
cell-free system are listed in Table 15.3.
These conditions have been described by
two prominent groups in the development
of E. coli cell-free systems [13, 16, 17].
Although most of the reaction conditions
are similar in the two examples given in
Table 15.3, they differ in three important
aspects. The conditions reported initially
[2] are for standard protein synthesis,
whereas those reported by others [17] have
been optimized for the production of di-
sulfide-bonded proteins. These systems
also differ in their choice of stabilizing re-
agents (PEG8000 or spermidine/putres-
cine) and the ATP regeneration systems.
The conditions given in Table 15.3 refer
to the cell-free reaction mixture. In dialysis
mode, this reaction mixture is dialyzed
against a buffer of the same composition
but lacking E. coli extract, RNA polymer-
ase, creatine kinase/pyruvate kinase, and
mRNA/DNA template. If a dialysis mem-
brane with a molecular weight cut-off
(MWCO) of 50 kDa is used, tRNA must be
included in the dialysis buffer or the reac-
tion will be depleted of tRNA. Dialysis
membranes with MWCOs of 15 or 10 kDa
are commonly used, and under those cir-
cumstances tRNA can be excluded from
the dialysis buffer.
Polyethylene glycol (PEG) is used as a
crowding agent, and is thought to stabilize
mRNA. Excluding PEG from the reaction
mixture can lower protein synthesis yield
by around 40% [17]. These authors found
that spermidine and putrescine could be
used in place of PEG with no loss in syn-
thesis activity, but with some enhancement
in the yield of functional disulfide-bonded
PA protein (discussed in Section 15.4.3).
One benefit of excluding PEG from the re-
action mixture is that it interferes with
analysis by sodium dodecyl sulfate-polya-
Table 15.3 Conditions for coupled transcription/translation sys-
tems from E. coli under reducing and oxidizing conditions
Reagent Concentration Units
Reducing
a)
Oxidizing
b)
HEPES/KOH 55 57.2 mM
DTT 1.7 mM
ATP 1.2 1.2 mM
CTP 0.8 0.86 mM
GTP 0.8 0.86 mM
UTP 0.8 0.86 mM
Creatine phosphate 80 mM
Creatine kinase 250 lg mL
1
PEG8000 4 % (w/v)
Spermidine 1.5 mM
Putrescine 1.0 mM
3',5'-cyclic AMP 0.64 0.64 mM
Folinic acid 68 57 lM
E. coli total tRNA 175 170.6 lg mL
1
Potassium glutamate 210 230 mM
Ammonium acetate 27.5 80 mM
Magnesium acetate 10.7 16 mM
Each amino acid 1 2 mM
T7 RNA polymerase 93 70 lg mL
1
E. coli extract 30 24 % (v/v)
Plasmid DNA 6 6.8 lg mL
1
Sodium azide 0.05 % (w/v)
Phosphoenol pyruvate 30 mM
NAD 0.33 mM
CoA 0.27 mM
Oxalic acid 2.69 mM
Reduced glutathione 1 mM
Oxidized glutathione 4 mM
a) From Ref. [13].
b) From Ref. [17].
crylamide gel electrophoresis (SDS-PAGE),
causing streaking and background stain-
ing. If included in the reaction mixture,
PEG can be removed from samples for
SDS-PAGE analysis by either trichloroace-
tic acid (TCA) or acetone precipitation.
In general, most of the components of
the cell-free system do not need to be var-
ied during expression optimization trials.
It is essential, however, to establish opti-
mum concentrations of RNA polymerase
and mRNA/DNA template for each expres-
sion construct. An optimum magnesium
concentration should only need to be es-
tablished once for each batch of extract.
Antibody detection methods can be used
to test expression conditions in batch/dia-
lysis mode, and Coomassie blue-stained
SDS-PAGE is often sufficient for an analy-
sis of dialysis mode reactions. However,
the simplest method for the rapid screen-
ing of expression conditions is to measure
radiolabeled
35
S-met/
35
S-cys incorporation
into the expressed protein. Using this
approach, unpurified cell-free reaction
mixtures are separated by SDS-PAGE,
after which the polyacrylamide gel is ex-
posed to X-ray film or a phosphor imaging
screen.
Because cell-free methods are highly ef-
ficient in their incorporation of amino
acids, only minute quantities of radiola-
beled amino acids are required for analysis
by autoradiography. Background transla-
tion of endogenous mRNA is also ex-
tremely low. This is advantageous because
expression conditions can be rapidly
screened for their effect on aggregation,
solubility, and proteolytic susceptibility of
the expressed construct. Low background
translation also means that total and solu-
ble protein yields can be analyzed by auto-
Fig. 15.1 Normalized (to maximum) expression levels of
35
S-met-labeled a-synuclein at dif-
ferent Mg
2+
and DNA template concentrations as determined by autoradiography. DNA con-
centrations (lg mL
1
) studied were 0.6 (diagonal lines), 2.7 (white solid), 10.6 (checked),
26.5 (black solid), and 53.0 (horizontal lines).
radiography of dot-blots, or by scintillation
counting of TCA-precipitable material.
The results of expression trials for
35
S-
met labeled a-synuclein at different DNA
template and Mg
2+
concentrations are
shown in Fig. 15.1. The height of each bar
in the graph represents the relative
amount of a-synuclein synthesized in
1 hour, normalized to the highest-yielding
reaction, as measured by autoradiography.
The figure demonstrates the sensitivity of
expression yields on both DNA and Mg
2+
concentrations. Similar DNA concentration
dependence at each magnesium concentra-
tion suggests that these parameters are
mutually exclusive.
15.6.2
Wheat Embryo System
The conditions for performing a wheat
embryo cell-free translation reaction from
an mRNA template are listed in Table
15.4. Points discussed for E. coli optimiza-
tion in Section 15.6.1 are relevant to ex-
pression in wheat embryo extracts. How-
ever, it is worth noting that the wheat em-
bryo system is better suited to the transla-
tion of added mRNA template, whereas
coupled transcription/translation is better
in the E. coli system. The principal reason
for this difference is that transcription
with bacteriophage RNA polymerases re-
quires a relatively high Mg
2+
concentration
(ca. 16 mM); E. coli translation-only reac-
Table 15.4 Conditions for wheat embryo translation systems under reducing and oxidizing conditions
Reagent Concentration Units
Reducing
a)
Oxidizing
b)
HEPES/KOH 24 (pH 7.8) 28 (pH 7.8) mM
DTT 2 mM
ATP 1.2 1.2 mM
GTP 0.25 0.25 mM
Creatine phosphate 16 15 mM
Creatine kinase 0.45 1.8 lg mL
1
Spermidine 0.4 0.4 mM
Wheat embryo deacylated tRNA 50 50 lg mL
1
Potassium acetate 100 53 mM
Magnesium acetate 2.5 1.6 mM
Calcium chloride 0.6 mM
Each amino acid 0.3 0.23 mM
Embryo lysate 24 24 % (v/v)
mRNA 144 200 lg mL
1
Sodium azide 0.005 % (w/v)
Nonidet P-40 0.05 % (v/v)
E-64 proteinase inhibitor 1 lM
RNasin (ribonuclease inhibitor) 0.4 U lL
1
Biotin (for biotinylation only) 19.5 lM
Biotin ligase (for biotinylation only) 19.5 lg mL
1
a) From Ref. [9].
b) From Ref. [18].
tions occur optimally at a similar Mg
2+
concentration. In contrast, wheat embryo
translation-only reactions perform best
when the Mg
2+
concentration is an order
of magnitude lower (ca. 1.6 mM). For this
reason, coupled transcription/translation
best suites analytical-scale expression,
whereas uncoupled reactions are best used
for preparative-scale synthesis.
15.6.3
Temperature
Once optimal conditions have been
achieved for total expression, yield factors
that influence folding and solubility must
be examined; these include the addition of
cofactors, ligands, and chaperones. The re-
action temperature may also have a dra-
matic effect on protein solubility; lowering
it leads to lower rates of protein synthesis,
but this can be compensated for by extend-
ing the reaction time. For example, expres-
sion of dihydrofolate reductase (DHFR) at
37
o
C in a dialysis-mode E. coli reaction
generally yields 23 mg of insoluble
DHFR in 8 hours, but the same yield of
mostly functional protein is obtained after
24 hours at 30
o
C.
15.6.4
Protease Inhibitors
Inhibition of proteolysis is best achieved
by promoting proper folding. As discussed
above, this is achieved by manipulating ex-
pression conditions such as pH, tempera-
ture, and the addition of ligands and cha-
perones. The cell-free systems discussed
here are fairly tolerant to the addition of
protease inhibitors commonly used in the
purification of proteolytically labile species.
Protease inhibitor cocktails can be used as
long as they are free of EDTA, which may
chelate magnesium ions and inhibit
mRNA/protein synthesis. EGTA can be
added to approximately 10 mM in order to
inactivate Ca
2+
-dependent enzymes,
though the Mg
2+
concentration must be
optimized under these conditions.
15.6.5
RNase Inhibitors
Every effort must be made to minimize
RNase contamination during cell extract
preparation, and in the preparation of re-
action mixtures. These efforts include
chloroform washing and subsequent bak-
ing of glassware prior to use, and the use
of diethyl pyrocarbonate (0.1%)-treated de-
ionized water in the preparation of extracts
and reaction mixtures. Gloves should be
worn at all times and, ideally, RNase-free
work areas should be dedicated to cell-free
expression. RNase inhibitors such as hu-
man placental RNase inhibitors can be in-
cluded in the reaction mixture to mini-
mize the degradation of mRNA and ribo-
somes. However, the use of these inhibi-
tors adds substantially to the cost of a
reaction, and they should be used spar-
ingly. The concentration of RNase inhibi-
tor used should be optimized in dialysis-
mode to identify the lowest concentration
required to protect a given mRNA tem-
plate.
15.7
Conclusion
An SDS-PAGE analysis of optimized dialy-
sis-mode expression (10 vols dialysis solu-
tion) in an E. coli system is shown in
Fig. 15.2 for:
Peptide chain elongation factor Ts (EF-
Ts) from E. coli at 37
o
C after 1, 3, 6,
and 12 hours in lanes 1, 2, 3, and 4, re-
spectively; and
a-synuclein at 308C after 0, 1, 3, 12, and
24 hours in lanes 5, 6, 7, 8, and 9, re-
spectively. These expression profiles cor-
relate with yields of several milligrams
of protein per mL reaction mixture.
Such excellent outcomes are a result of
careful and systematic optimization of
DNA constructs and expression conditions.
There are occasions, however, where pro-
teins fail to express, or are expressed in an
inactive form. Other proteins express well
in one expression system, but not in an-
other, and this is of course also true of all
in vivo expression systems. A true benefit
of cell-free synthesis over in vivo methods
is that the open nature of the systems
presents the researcher with unparalleled
opportunities to affect the outcome of an
expression experiment. As such, cell-free
synthesis methods represent an invaluable
addition to any laboratory investigating the
expression of biopharmaceuticals.
References
1 Zubay, G. (1973) Annu. Rev. Genet. 7: 267287.
2 Yokoyama, S., Matsuo, Y., Hirota, H, Kigawa,
T., Shirouzu, M., Kuroda, Y., Kurumizaka, H.,
Kawaguchi, S., Ito, Y., Shibata, T., Kainosho,
M., Nishimura, Y., Inou, Y., and Kuramitsu, S.
(2000) Prog. Biophys. Mol. Biol. 73: 363376.
3 Kigawa, T., Yabuki, T., Yoshida, Y., Tsutsui, M.,
Ito, Y., Shibata, T., and Yokoyama, S. (1999)
FEBS Lett. 442: 1519.
4 Spirin, A. S., Baranov, V. I., Ryabova, L. A.,
Ovodov, S. Y., and Alakhov, Y. B. (1988) Science
242: 11621164.
5 Kim, D. M. and Choi, C. Y. (1996) Biotechnol.
Prog. 12 (5): 645649.
6 Kim, D.-M. and Swartz, J. R. (2000) Biotechnol.
Prog. 16: 385390.
Bioeng. 74 (4): 309316.
8 Endo, Y. and Sawasaki, T. (2004) J. Struct.
Funct. Genomics 5: 4547.
9 Madin, K., Sawasaki, T., Ogasawara, T., and
Endo, Y. (2000) Proc. Natl. Acad. Sci. USA 97
(2): 559564.
10 Sawasaki, T., Hasegawa, Y., Tsuchimochi, M.,
Kasahara, Y., and Endo Y. (2000) Nucleic Acids
Symp. Ser. 44: 910.
11 Hendrickson, W. A., Horton, J. R., and LeMas-
ter, D. M. (1990) EMBO J. 9: 16651672.
Fig. 15.2 SDS-PAGE analysis of optimized dialy-
sis-mode expression in an E. coli cell-free system
for: (A) peptide chain elongation factor Ts (EF-Ts)
from E. coli at 37
o
C after 1, 3, 6, and 12 hours in
lanes 1, 2, 3, and 4, respectively; and (B) a-synu-
clein at 308C after 0, 1, 3, 12, and 24 hours in
lanes 5, 6, 7, 8, and 9, respectively. Expression
bands are indicated by arrows to the left of each
gel.
12 Shimizu, Y., Inoue, A., Tomari, Y., Suzuki, T.,
Yokogawa, T., Nishikawa, K., and Ueda, T.
(2001) Nature Biotechnol. 19 (8): 751755.
13 Kigawa, T., Yabuki, T., Matsuda, N., Matsuda,
T., Nakajima, R., Tanaka, A., and Yokoyama,
S. (2004) J. Struct. Funct. Genomics 5: 6368.
14 Pratt, J. M. (1984) In: Hames, B. D. and Hig-
gins, S. J. (Eds), Transcription and Translation.
IRL Press, Oxford, Washington, pp. 179209.
15 Jiang, X., Ookubo, Y., Fujii, I., Nakano, H.,
and Yamane, T. (2002) FEBS Lett. 514: 290
294.
Bioeng. 85 (2): 122129.
17 Yin, G. and Swartz, J. R. (2004) Biotechnol.
Bioeng. 86 (2): 188195.
18 Kawasaki, T., Guoda, M. D., Sawasaki T., Takai,
K., and Endo, Y. (2003) Eur. J. Biochem. 270
(23): 47804786.
References 1081
Abstract
The manufacture of monoclonal antibod-
ies or antibody fragments requires suitable
expression systems. Whilst whole antibod-
ies are currently produced in expensive
mammalian cell culture systems, for some
applications the expression of antibody
fragments in prokaryotes, yeast, or fila-
mentous fungi is an alternative. Currently,
a large portion of the worldwide fermenta-
tion capacity for the production of mono-
clonal antibodies or antibody fragments is
provided by contract manufacturing orga-
nizations (CMOs). Among the different
methods, continuous perfusion cell culture
(in suspension) is a proven method to
manufacture monoclonal antibodies and
recombinant proteins in quantities of up
to several hundred kilograms. However,
that method appears to be less common in
the industry compared to fed-batch cell
culture. Based on a large number of indus-
try interviews, this chapter highlights the
key factors of outsourcing projects. Out-
sourcing reflects a wide range of benefits
such as time-to-market, avoidance of capi-
tal expenditure, cost containment and flex-
ibility. Key issues of CMO/client contract-
ing are discussed. This chapter also pro-
vides insights into successful strategies for
the management of technology transfer
processes. The management of technology
transfer requires careful preparation and
stringent definition of all project steps and
scopes covering the process, the analytical
procedures, the technical equipment, and
the qualification and validation protocols.
The ability to meet demands for biophar-
maceutical production capacity whether
through in-house manufacturing or out-
sourced contract manufacturing carries
strategic and financial implications for bio-
pharmaceutical companies. FDA approval
policy will have a major impact on future
capacity requirements for the production
of biopharmaceuticals.
1083
ISBN: 3-527-31184-X
16
Contract Manufacturing of Biopharmaceuticals
Including Antibodies or Antibody Fragments
When Success Raises its Ugly Head
Outsourcing to Uncork the Capacity Bottleneck
Abbreviations
ATF alternating tangential filtration
CMC chemistry manufacturing and
control
CMO contract manufacturing organiza-
tion
CRO contract research organisation
DSP downstream processing
IPC in-process-control
MAB monclonal antibody
P packaging
QA/QC quality assurance/quality control
support
RTP Research Triangle Park
USP upstream processing
WCB working cell bank
16.1
Introduction
Although this chapter will focus on antibod-
ies and fragments, the content clearly also
applies to all biopharmaceuticals in the same
way. Since the famous Khler and Milstein
[1] experiment in 1975, the success story
of monoclonal antibodies was always linked
to the development of appropriate expres-
sion systems and production technologies.
Immortalized mouse cell lines secreting
one single type of antibody with a unique
binding affinity against virtually any type
of antigen (proteins, carbohydrates, or nu-
cleic acids) initiated heavy competition in
the development of faster, cheaper, and
safer procedures. In order to guarantee the
safety and efficacy of these pharmaceuti-
cally active products, all regulatory issues
need to be addressed and fulfilled.
Whilst expression rates in the 1980s
achieved product concentrations of only
about 100 mg L
1
, some ten years later sev-
eral immunoglobulins were being produced
at concentrations which were between two-
to seven-fold higher [2]. The reasons for this
ongoing improvement are linked to more
advanced production procedures and better
expression systems. Indeed, in 2005 the
hurdle of 5 g L
1
heterologous protein ex-
pressed in a high-yielding CHO cell culture
process is likely to be exceeded [3] (see also
Part IV, Chapters 1 and 4).
The growing market demand for all
kinds of monoclonal antibodies or anti-
body fragments is a key factor for the
pharmaceutical industry. Antibody-based
therapeutics cover a wide range of indica-
tions such as inflammation and oncology,
in addition to different types of infectious
diseases. High dosage requirements for
the treatment of chronic diseases in large
markets necessitate the development of
high-yielding, low-cost and large-scale
manufacturing processes that consistently
deliver high-quality product.
At present, there are more than 370 bio-
tech therapeutics undergoing clinical trials,
and as many as 125 new drugs may reach
the market during the next five to seven
years. In this respect, efficient expression
systems and biopharmaceutical manufac-
turing capacities will have to be developed.
In this chapter, novel aspects of the
manufacturing procedures of monoclonal
antibodies from a process engineering
point of view will be highlighted. The
manufacture of monoclonal antibodies in
mammalian cell culture requires certain
technologies and capacities, which will be
described within the first section of the
chapter. Technical hurdles linked to these
still-expensive procedures and aimed to-
wards the development for more economi-
cal eukaryotic or even prokaryotic manu-
facturing systems will be highlighted in
the second section.
16 Contract Manufacturing of Biopharmaceuticals Including Antibodies or Antibody Fragments 1084
Since the availability of manufacturing/
fermentation capacity at present remains a
major challenge for the pharmaceutical in-
dustry, the third section of the chapter re-
lates to the management of outsourcing
processes.
16.2
Expression Systems and Manufacturing
Procedures
The manufacture of monoclonal antibod-
ies or antibody fragments requires suitable
expression systems. Advances in molecular
biology of higher eukaryotes, together with
recent developments in the field of bioen-
gineering, have made the choice of an ap-
propriate expression system more com-
plex. Some 20 years ago there were only a
few different options, and systems incor-
porating recombinant Escherichia coli, Sac-
charomyces cerevisiae or Chinese hamster
ovary (CHO) cells have each presented
their own challenge, having been the pio-
neers of recombinant protein expression
(see also Part IV, Chapter 12).
The situation has changed during the
past few years, however, and a variety of
products that are either undergoing clinical
trials or are already on the market have
been produced by alternative systems.
Transgenic animals and plants will comple-
ment existing cell culture-based expression,
as described in Part IV, Chapters 511 (for a
recent review, see Knblein J, Biopharma-
ceuticals expressed in plants a new era
in the new Millennium. In: R. Mller and
O. Kayser (eds), Applications in Pharmaceuti-
cal Biotechnology. Wiley-VCH. ISBN 3-527-
30554-8). Manufacturing processes using
insect cell lines (see Part IV, Chapter 14)
have been developed to commercial scale.
Within the literature, optimistic projections
can also be found for transgenic animals
(see Part IV, Chapter 11), yeasts (see Part
IV, Chapter 13) and fungi.
In process development, the selection of
an appropriate expression system affects
important factors such as:
product characteristics
regulatory hurdles
cost of goods
intellectual property
Time to market the most important is-
sue when focusing on commercial success
is a direct function of the above-men-
tioned factors.
Nowadays, the systems of choice for the
production of monoclonal antibodies or
antibody fragments are either eukaryotic
cell lines or, in some cases, prokaryotes.
The following sections will highlight cer-
tain manufacturing aspects which are
linked to these expression systems.
16.2.1
Manufacture of Monoclonal Antibodies
All monoclonal antibodies currently ap-
proved by the US regulatory authority
(Food and Drug Administration, FDA) as
drug substances are prepared in cell cul-
ture (see Table 16.1). Among the biophar-
maceutical products made in cell culture,
monoclonal antibodies (MAbs) represent
the single most production capacity de-
manding group. Currently, a significant
portion of the worldwide cell-culture ca-
pacity is provided by contract manufactur-
ing organizations (CMOs).
Compared to microbial cultures, cell cul-
tures still operate at low cell densities.
While medium development and feed
strategies during fed-batch cell culture in-
creases cell densities and productivities,
the fact that unused media components
and products accumulate in the reactor re-
presents a natural limitation.
Perfusion cell culture overcomes that lim-
itation and is, therefore, particularly attrac-
tive for inhibiting or unstable products
(which is typically not the case for MAbs).
Furthermore, perfusion cell culture achieves
higher cell densities and volumetric produc-
tivities, even when effects of the media com-
ponents are not fully understood. However,
that comes at a cost of a more complex bior-
eactor system (Figs. 16.1 and 16.2).
Perfusion cell culture (in suspension) is
a proven method for the manufacture of
MAbs and recombinant proteins in quanti-
ties of up to several hundred kilograms
(e.g., ReoPro, Remicade). However, this
method appears to be less common in the
industry compared to fed-batch cell cul-
ture. Very few companies manufacture
their clinical and approved MAb products
by using perfusion cell-culture techniques.
Table 16.1 Therapeutic monoclonal antibodies approved by the
US FDA. (Products with an annual turnover of more than US$
300 million are in italics)
Trade-name (generic name) Sponsor company Approval date
Orthoclone OKT3
(muromomab-CD3) Ortho Biotech, Inc. (subsidiary of John-

son & Johnson)
June 1986
ReoPro
TM
(abciximab)
sales >300 Mill$/year
Centocor, Inc. (subsidiary of Johnson &
Johnson) and Eli Lilly and Company
December 1994
Rituxan
TM
(rituximab)
IDEC Pharmaceuticals Corp. and Genen-
tech, Inc.
November 1997
Zenapax
(daclizumab) Hoffmann-La Roche, Inc., and Protein

Design Labs
December 1997
Simulect
(basiliximab) Novartis Pharmaceutical Corp. May 1998

SYNAGIS
TM
(palivizumab)
MedImmune, Inc. June 1998
Remicade
(infliximab)
Centocor, Inc. (subsidiary of Johnson &
Johnson)
August 1998
Herceptin
(trastuzumab)
Genentech, Inc./Roche September 1998
Mylotarg
TM
(gemtuzumab ozogamicin) Celltech Pharmaceuticals and Wyeth May 2000
Campath
(alemtuzumab) Ilex Oncology, Inc., Millennium Phar-

maceuticals, Inc., and Berlex Laborato-
ries, Inc.
May 2001
Zevalin
TM
(ibritumomab tiuxetan)
IDEC Pharmaceuticals Corp./Schering
AG
February 2002
HUMIRA
TM
(adalimumab)
Cambridge Antibody Technologies and
Abbott Laboratories
December 2002
Bexxar
(Tositumomab and I-131 tosi-

tumomab)
Corixa Corp. and GlaxoSmithKline June 2003
Xolair
(Omalizumab; rec. Immuno-

globin-E) sales >100 Mill$/year
Genentech, Tanox, Inc. and Novartis
Pharmaceuticals
June 2003
Raptiva
TM
(efalizumab; selective, revers-
ible T-cell blocker [subcutaneous injec-
tion; self-administered])
Xoma, Ltd. and Genentech, Inc. October 2003
For new processes, even these compa-
nies aim to develop efficient fed-batch pro-
cesses, and if MAb concentrations of
>0.7 g L
1
could be achieved in fed-batch
systems, they would no longer use the per-
fusion cell-culture approach.
The arguments to move to fed-batch,
and to abandon perfusion cultivation as
given by various companies (e.g., Medarex
and Serono) included:
better availability of fed-batch manufactur-
ing capacity at contract organizations, and
already high concentrations (e.g.,
4.2 g L
1
, as reported by Lonza Biologics;
see Part IV, Chapter 4); and
shorter processes and therefore shorter
validation cycles.
However, there are arguments not to aban-
don perfusion cultivation:
Recently developed cell retention devices
such as alternating tangential filtration
(ATF) or an acoustic device (Biosep) can
be fitted to existing bioreactors at con-
tract organizations.
High productivities Crucell/DSM Bio-
logics reported 0.9 g L
1
per day at 120
14010
6
cells mL
1
(see also Part IV,
Chapter 3).
A proven strategy to achieve shorter vali-
dation cycles during perfusion cell cul-
ture processes is to include a short per-
fusion time (e.g., 15 days) in the regula-
tory filing, and to increase the perfusion
time to 50 or 60 days as processes
Fig. 16.1 General schematic
for fed-batch and perfusion
cultivation. X=biomass;
S=substrate; P=product;
F=flow.
Fig. 16.2 Typical fed-batch
culture, recently reported up
to 4.2 g L
1
. (Source: Lonza
Biologics; public domain)
change (that the long-term perfusion
clearly should be investigated at small
scale before entering that strategy).
16.2.2
Manufacture of Antibody Fragments
in Microbial Systems
While whole antibodies are currently pro-
duced in expensive mammalian cell culture
systems, for some applications the expres-
sion of antibody fragments in prokaryotes,
yeast or filamentous fungi is an alternative.
Tissue penetration is facilitated, and clear-
ance from the blood or kidney is faster using
antibody fragments. In addition, PEGylation
allows the modification of the half-life,
which is another advantage of antibody frag-
ments (see also Part VI, Chapter 2).
The variety of mechanisms mediating
the therapeutic effect do not always re-
quire the complete antibody glycoprotein.
For some applications the optimal struc-
ture is not necessarily the whole immuno-
globulin, but is rather a distinctive frag-
ment containing the specific antigen-bind-
ing domain. Recently, it has been shown
for example that their small size permits
them to penetrate tissues or solid tumors
more rapidly than whole antibodies [4].
However, the advantages that these frag-
ment antigen binding (Fabs) provide have
not only been limited to therapy, as reduc-
ing the requirements from the complex
immunoglobulin expression towards unde-
manding expression of antibody fragments
permits the application of simple expres-
sion systems.
For the large-scale production of anti-
body fragments, the expression system
must be:
economical,
accessible for genetic modifications,
easily scalable,
safe.
Therefore, E. coli is one of the preferred
hosts for the expression of antibody frag-
ments [5]. Genetic modification of bacterial
cell lines especially E. coli is rapid and
trouble-free, and the media and nutrient
supply is simple. The metabolic potential
of microbial cell lines allows the use of se-
rum-free, chemically defined and cheap
media. Expensive viral removal steps can
be avoided, whilst the removal of pyrogens
does not usually cause any problems.
The volumetric productivity achievable in
simple bioreactor systems when using bac-
terial expression systems is superior to that
of mammalian expression systems. Growth
rates are higher, and the ease of scale-up of
the fermentation process enables manufac-
turing at scales up to five-fold larger com-
pared to mammalian systems.
In 2005, the worldwide manufacturing
capacity for microbial fermentation is ex-
pected to exceed 250 000 m
3
, thus provid-
ing a potential solution to the problem of
limited capacity available on the market
for mammalian cultivation systems [6].
The secretion of soluble antibody frag-
ments in the periplasm allows the use of
simple down-stream technologies. Simple,
but efficient, purification steps reduce the
technical hurdles for large-scale production.
The use of E. coli as anexpressionsystemis
limited to those proteins where post-transla-
tional modification such as the glycosylation
or galactosylation of antibody fragments is
not required. Inclusion body formation
can be another disadvantage that occurs with
this prokaryotic model, as these insoluble
protein aggregates demand laborious and
cost-intensive in vitro refolding (denatura-
tionandrenaturation) andpurificationsteps.
Antibody fragments and fusion proteins
can also be produced cost-effectively in
large-scale production using yeasts or fun-
gal fermentations. An extensive knowledge
covering the upstream and downstream
processing used for the bulk production of
several other recombinant proteins is avail-
able. Filamentous fungi and also yeasts
are accessible for genetic modification,
and the protein of interest may be secreted
into the culture broth. Saccharomyces cere-
visiae, the genome of which was fully se-
quenced in 1996, has been genetically
modified to produce a wide range of pro-
teins and antibody fragments. The absence
of pyrogenic compounds, toxins or viral
contaminations offers an additional advan-
tage of these expression systems.
16.3
Outsourcing and Contract Manufacturing
Expertise and knowledge along the whole
value chain, from high-throughput screen-
ing, lead identification and process devel-
opment to clinical development, manufac-
turing and marketing is available only in
large companies.
One potential alternative to in-house ex-
pertise are partnerships perhaps between
smaller companies specialized in certain
fields such as drug development, manufac-
turing companies, or contract research or-
ganizations. This trend is usually referred
to simply as outsourcing.
Outsourcing reflects a wide range of
benefits such as time-to-market, avoidance
of capital expenditure, cost containment
and flexibility. Indeed, focusing on the out-
sourcing of manufacturing processes in-
volved in the need for production is the
major driver for such a decision. Some as-
pects of outsourcing strategy, together with
the management of these processes, will
be discussed in the following sections.
16.3.1
Integrating Contract Manufacturing
into the Manufacturing Strategy
An increasing number of companies in-
clude contract manufacturing into their
plans for developing a bio-pharmaceutical
drug (Table 16.2). Although, initially an ap-
propriate quantity of material is required
for toxicology studies (the pre-clinical
phase), much larger quantities will be re-
quired further down the development
chain as the clinical phase approaches
(Fig. 16.3).
Table 16.2 Manufacturing development plan for a particular pro-
ject in a medium-size biotech company required to implement its
own manufacturing capabilities. As a first step the company
needs enough material to characterize the protein and to further
develop the process. Thereafter, the company needs material
manufactured under GMP conditions for safety studies, clinical
studies and, within 23 years, a commercial scale manufacturing
facility to fulfill marketing requirements.
Function Phase
Preclinical I/II III Launch Commercial
Fermentation In-house Contract Contract Mixed
Purification In-house Contract Contract Mixed
Filling/Finishing Contract Contract Contract Contract
Testing Mixed Mixed Mixed Inhouse
Key reasons, to include contract manu-
facturing into the development strategy,
are to increase cash flow/minimize fixed
assets by avoiding investments into manu-
facturing facilities, and to accelerate the
drug development by creating access to ad-
ditional expertise.
During the past few years there has
been a capacity shortage in particular for
large-volume commercial manufacturing.
Since then, all major CMOs have ex-
panded their capacities, and it is therefore
now easier to identify the correct CMO.
Nevertheless, outsourcing of the majority
of the manufacturing development should
be well prepared in advance.
16.3.2
Typical Capacity Requirements
The availability of bioreactors of a certain
size is an important factor in choosing the
correct CMO, as downstream process
equipment does not vary very much from
product to product. The capacity require-
ments are dependent on typical dose
ranges, the size of the clinical studies, and
market projections. These requirements
and the productivity of the cell culture
determine the scale of operation. For ex-
ample, a successful therapeutic monoclo-
nal antibody product would require about
200300 g for clinical Phases I/II, 500 g
per indication (2 indications studied) for
clinical Phase III, and for the first year
supply more than 5 kg. Depending on the
dosage which typically is milligrams per
kilogram body weight and repeated treat-
ment, the commercial annual supply could
easily reach 80200 kg. If the antibody is
targeted at a rare disease which receives
an orphan drug status, the requirements
might be much lower; for example, for
clinical Phases I/II 100150 g, for clinical
Phase III about 250 g, and for the first
year supply about 2.5 kg. Depending on
dosage, the commercial annual supply
might reach only 5 kg.
These capacity requirements, together
with assumptions for the productivity of
Fig. 16.3 The development of a biopharmaceutical drug.
the cell culture, determine the bioreactor
sizes (Figs. 16.4 and 16.5).
16.3.3
Selecting a CMO
Once the decision to outsource the manu-
facturing of a biopharmaceutical has been
taken, a suitable partner must be identi-
fied, and for this numerous aspects must
be taken into account. Some typical con-
siderations in finding a CMO include:
What are the typical capacity requirements
to manufacture the products? On the basis
of market development and future expec-
tations, a first guess should be made.
Knowing the expression level of an appro-
priate system will then allow quantifica-
tionof the manufacturing capacity needed.
Which of the manufacturing technolo-
gies is unique and should be developed
in-house? In some cases, cost-effective
manufacturing depends on highly so-
phisticated manufacturing know-how
and expertise. The assessment of the
technology position (weak, strong, domi-
nant) will have an impact on the out-
sourcing strategy.
For which phases shall a CMO be used?
The manufacturing capacity which is
needed to support clinical trials differs
from the commercial-scale production.
During the clinical development phases
smaller amounts of the product will be
needed.
Who are the main players on the con-
tract manufacturing market for cell cul-
ture and microbial fermentation?
Fig. 16.4 Annual output using 10 perfusion bioreactors
(x-axis liquid volume of each single bioreactor) with the fol-
lowing assumptions: 50 days perfusion rate at 0.5 L per day;
overall yield 46.6%; inoculation and capture in decoupled
work-centers; 90% occupation.
What technologies do they offer?
What are their strengths and weak-
nesses? Management skills, availability
of trained and experienced technical and
production staff should be analyzed. An
excellent technical facility may not nec-
essarily implicate experienced scientific
or technical staff.
What are the prevailing price levels?
At the start of such a search a detailed
market analysis is required. The assess-
ment of the technology position is needed,
and the companys own strategy must be
defined. To help in this decision making it
is useful to collect the knowledge of the
whole organization into a decision tree on
which decisions are clearly framed and the
criteria for making any particular decision
are made clear [11]. The next section pro-
vides an overview of the major players in
the CMO market.
16.3.4
Major Players
A list of Biologics CMOs (developmental
organizations included) can easily reach
about a hundred entries (see Table 16.3)
Among the selection presented in the ta-
ble, the underlined companies can be con-
sidered major players and are fully dedi-
cated to contract manufacturing. The list
is constantly changing: a number of
CMOs were recently acquired by other
companies (e.g., Covance RTP by Dio-
synth, Collaborative Bioalliances by Dow,
Alpha Bioverfahrenstechnik by Siegfried,
Lek by Sandoz) or have even stopped busi-
ness (e.g., Accentus).
Fig. 16.5 Annual output using 10 fed-batch bioreactor trains
(x-axis liquid volume of largest single bioreactor of each train)
with the following assumptions: growth rate 1 per day; overall
yield 46.6%; capture indecoupledwork-centers; 90%occupation.
All major CMOs offer the following ben-
efits to their clients:
Experience in technology transfer.
Pilot-scale cGMP manufacturing avail-
ability to supply product for clinical
Phases I and II.
Large-scale cGMP licensed manufactur-
ing availability to supply product for
clinical Phase III and early marketing
requirements.
Compliance to worldwide cGMPs.
Some CMOs offer additional benefits such
as:
Process development capabilities.
Cell banking capabilities.
Proprietary expression systems.
Formulation development capabilities.
QC development capabilities.
Protein analytical chemistry for product
characterization.
Clinical development services (CRO).
Large-scale cGMP licensed commercial
manufacturing capacity.
Complete manufacturing process chain
including fill/finishing.
Assistance in worldwide registration.
As mentioned above, all major CMOs have
recently expanded their capacity.
16.3.5
Expression Systems for Manufacturing
Among the drivers of manufacturing costs,
productivity of the cell culture is the major
factor, while purification, yield, media, dis-
posables and virus testing costs represent
minor factors.
The productivity of the cell culture part
is a function of the expression level (spe-
cific productivity) and the cell density in
the bioreactor system. Typical specific pro-
ductivities in cell culture are in the range
of 10 pg per cellday for recombinant pro-
tein (e.g., MAb), whilst high-yield expres-
sion systems (high-producing cell lines
generated with a range of productivity-en-
hancing selection methods) achieve 15
40 pg per cellday. Typical cell densities
are 5 to 2010
6
cells mL
1
in fed-batch
culture, and 20 to 14010
6
cells mL
1
in
perfusion cell culture. Thus, a key decision
Table 16.3 Contract manufacturing organizations
(CMOs) utilized by Biologic
Abio, Singapore
Abbott, USA
Siegfried Biologics (Alpha Bioverfahrenstechnik),
Germany
Archport, UK
Avecia, UK
Avid Bioservices, USA
BioInvent, Sweden
BioReliance (Invitrogen), USA, Germany, United
Kingdom
Boehringer Ingelheim, Germany, USA
BTC, Singapore
Cambrex, USA
Chiron, USA
Cobra Therapeutics, UK
CMC, Denmark
Diosynth (now again part of Organon), The Neth-
erlands, USA
DOW, USA
DSM Biologics, The Netherlands, Canada, USA
Eurogentec, Belgium
Excell Biotech (Q-biogene), UK
Girindus, Germany, USA
Goodwin Biotech, USA
Henogen, Belgium
ICOS, USA
Innogenetics, Belgium
Lonza Biologics, Switzerland, United Kingdom
and USA
Novartis Hunigue, France
Rentschler Biotechnologie, Germany
Sandoz Kundl, Austria and Lek, Slovenia
SynCo-Biopartners, The Netherlands
Companies in italics are major players and fully
dedicated to contract manufacture.
before selecting a CMO is to determine
whether the project requires an improved
expression system.
Proprietary expression systems provided
by CMOs already possess the experience
of a standard selection and manufactur-
ing process which may shorten the project
time and enhance process performance
significantly. On the other hand, typical li-
censing terms for those expression sys-
tems include license fees and royalties,
therefore making it more difficult for
small biotech companies to sell the pro-
ject at an appropriate and competitive
price. The licensing fees increase as the
project progresses from preclinical re-
search through the clinical phases. The to-
tal licensing fees and milestone payments
for a typical seven-year development pro-
ject up to approval could easily exceed
1 10
6
. Negotiations related to royalties typi-
cally start in the range of 0.53% of net
sales; thus, license and royalty-free options
should be investigated thoroughly.
Examples of mammalian protein expres-
sion systems/cell lines currently available
at CMOs include:
BIHEX/CHO Boehringer Ingelheim
High Expression System/Chinese Ham-
ster Ovary Cell, Boehringer Ingelheim,
Germany.
GS/CHO (or GS/NS0) Glutamine
Synthetase System/Chinese Hamster
Ovary Cell, Lonza Biologics, Switzer-
land.
Several vectors, e.g., pcDNA3002Neo/
PER.C6PER.C6
human cell line tech-

nology by Crucell and manufacturing by
DSM Biologics, The Netherlands.
Several vectors/ MARtech/CHO Matrix
Attachment Region Technology by Se-
lexis and manufacturing by Diosynth,
The Netherlands and USA.
CHEF1/CHO by ICOS, USA.
16.3.6
Typical Scope
Once a general framework for the project
is determined, the to-do list will be:
Contact short-listed companies in order
to determine generic capacity load and
capabilities.
Provide general conditions of project
(time, quantity, etc.) to CMO and let
company express their interest.
Develop a scope document: a detailed
manufacturing project plan with all the
requirements (mark requirements which
can be handled internally or outsourced
independent of the main CMO as OP-
TION).
Define and rank selection criteria.
Decide on final shortlist of CMOs.
Exchange confidential information dis-
closure agreement.
Visit companies for pre-qualification
audits and discussions of a pre-qualifica-
tion questionnaire.
Issue scope document and request for
quotation to selected CMO(s).
Receive and evaluate quote.
Evaluate CMO according to predefined
selection criteria.
Enter contract negotiations.
Reevaluate CMO.
A typical scope document for a CMO
might include the following items:
Technology transfer.
Cell line development.
Purification development.
Raw materials sourcing and testing.
Manufacturing technical support.
Transfer of assays/QC development.
Assay validation.
Product specific validation.
Manufacturing.
Quality testing.
Managing of external testing laboratories.
Release of drug substance.
Fill/finishing.
Release of drug product.
Stability studies.
Inventory building.
Shipment and logistic support.
During the scope development, do not for-
get to include process improvements into
the capacity planning. Typical selection cri-
teria during a formal evaluation of a CMO
have the following priorities:
Highest priority (or weight factor)
Experience with required manufactur-
ing technology
Expression system (if applicable)
Capacity, scale
cGMP compliance
Cooperation, clients input regarding
manufacturing operations
High priority (or weight factor)
Experience with cGMP and/or com-
mercial manufacturing
QA/QC support
Regulatory support
Project schedule
Project cost
On a case by case basis items from
the section CMOs offer additional
benefits
Low priority (or weight factor)
Location
These criteria might change depending on
the scope and size of the project. However,
once the technical and quality issues are
found to be adequate, the most important
one is finding the CMO which truly be-
comes a partner.
16.3.7
Key Issues of CMO/Client Contracting
Key issues during the contract negotiations
are:
Reservation fees required by the CMO to
assure loading their GMP capacity. The
terms vary from no payment (payment
to secure a manufacturing slot only if
another client is interested) to 30% or
more of the price paid, 6 months or up
to 18 months ahead of the start of man-
ufacturing a particular batch.
License fees (and future royalties) when
the CMOs expression system is part of
the contract.
Process specific laboratory and validation
work becomes either part of the price or
will be reimbursable on a time and ma-
terial basis.
Intellectual property rights brought in or
developed during the project.
Liability CMO to comply with cGMP
and to be liable for operational risk; cli-
ent to be liable for the suitability of the
product for the intended use.
Performance criteria particularly when
the CMOs expression system is part of
the contract.
Payment schedule incentives tied to
time and performance.
Quantity loss quantity loss during re-per-
formance will trigger a price reduction.
Termination if CMO does not comply
with cGMP, general termination clause.
Change of ownership in case the CMO
is sold to another company.
Assignment to make sure the project
can be licensed out in the future.
Demarcation of pharmaceutical responsi-
bilities to document the responsibilities
towards the regulatory authorities.
As mentioned above,
Price is especially during the early clin-
ical phases just one consideration
among others. While a typical MAb man-
ufacturing project up to clinical Phase II
may cost 1 3 million, the overall cost of
clinical production, consistency lots in-
cluding activities related to the chemistry
manufacturing and control (CMC) section
of the registration dossier may exceed
1 15 million. However, clinical trials are
typically more expensive.
The absolute latest point in time to mar-
ry the CMO is after clinical Phase IIb.
As a change to another CMO will take
about three years (as post-approval
change), it is important to select a CMO
which will not only supply the material for
clinical Phase III but is willing to deliver
product for the initial three years of mar-
keting requirements. Typically, that will re-
quire scale-up of the process to larger bio-
reactor(s) and purification unit operations.
Therefore, key issues of CMO/Client
contracting for clinical Phase III and
launch requirements are:
Capacity the CMO must be capable of
supplying quantity for at least the first
year; clients commitment.
Scale-up performance criteria.
Regulatory support.
Launch requirements typically manu-
factured with consistency lots.
Transferability to transfer the process
to own facility, a strategic partner (or to
another CMO)
and for a typical long-term manufacturing
agreement.
$/gram pricing (e.g., based on ranges
for fermentation output, DSP yield and
success rate).
Raw materials cost reimbursable.
Minimum volumes.
Shared savings for yield improvements.
Contract duration 7 years+.
Royalties (if applicable).
16.3.8
Technology Transfer
Whenever process development leaves the
laboratory scale, the know-how which is
well established within the research group
must be transferred to staff in the develop-
ment department. Both groups are facing
only minor obstacles as they know each
other very well, and communication is es-
tablished between the key persons. This
know-how shift during an in-house project
is usually accompanied with a technology-
driven scale-up process. In many cases,
leaving laboratory-scale to enter the pilot
scale means re-engineering the whole pro-
cess.
These two factors the proximity of
both groups and the simultaneous devel-
opment of a pilot process often conceal
the complexity of a technology transfer
once a third party is engaged.
The transfer of a biopharmaceutical pro-
cess between a research-driven process
owner and a CMO is usually a challenging
and often time-consuming and therefore
expensive project. Underestimating the
process of technology transfer often causes
an unexpected delay of the ambitious
time-table.
Management of the technology transfer
requires careful preparation and stringent
definition of all project steps and scopes:
The process itself needs to be defined in
detail: All the relevant process informa-
tion must be described and documented.
Technology transfer of chemical pro-
cesses is rather simple compared to bio-
tech processes due to well-understood,
robust and predictable chemical reac-
tions. In contrast, technology transfers
of bio-processes often cause difficulties
due to the lack of process understand-
ing/information. These areas, where
process knowledge is rather limited, also
need to be identified up front, in order
to minimize the efforts.
The analytical procedures which have
been developed especially to control and
to monitor the efficacy of each bio-pro-
cess step need also to be defined and to
be transferred (see Part VII, Chapter 1).
Protocols must be written that describe
each in-process-control (IPC) step from
the analysis of the WCB cell line, the
(on-line) monitoring of the cultivation to
the analysis of the purification need to
be established. Due to the complexity of
bioprocesses, the need for good bio-
chemical analytical methods and capabil-
ities cannot be overemphasized.
Technical equipment: The robustness of
a bioprocess and the yield of a bioprod-
uct are often directly linked to the pro-
cess equipment used for the manufac-
turing. Reactor geometry, shear stress,
mixing and rheology have an impact on
the biology. The metabolic activity of the
cell line during cultivation not only de-
pends on standardized media but is also
a function of mass and energy transfer.
Temperature profiles are a key factor for
induction processes, and depend on the
very reaction system. Harvesting after
cultivation is also challenging and re-
quires strong expertise. Batch centri-
fuges give tight, dry pellets, whereas
continuous disk-stack separators produce
a heavy phase slurry, either carrying
over supernatant or leaving product in
the heavy phase.
Continuous-perfusion cell culture poses
additional challenges to the team.
Qualification and validation protocols.
Comparability of the product and consis-
tency of the process transferred from a
client to a CMO needs to be demon-
strated. A given reference standard must
be met under predefined protocols to
proof the ability of the CMO to manu-
facture the biopharmaceutical consis-
tently.
A consensus about all features of the
transfer process needs to be achieved be-
tween the receiving party and the process
owner. In order to guide the technology
transfer successfully through all phases, a
stringent project management structure
must be established. There are at least two
different possibilities to structure this pro-
ject organization, and Fig. 16.6 illustrates
a parallel structure between the CMO and
the technology provider. This structure fo-
cuses on a liaison (window) person or a
(liaison) window team.
Although the additional interfaces with-
in this structure can cause difficulties and
misunderstandings, this structure can of-
ten be found in smaller hierarchical com-
panies. To channel information via one
central person prevents direct communica-
tion, inhibits information flow, and causes
delay. An extended project timeline is the
consequence.
An alternative organization will be pre-
sented below. Since information flow is
the key for success, the formation of work-
ing teams is crucial. These teams, repre-
senting all technical aspects, must be
brought together in an initial kick-off
meeting. Each function needs to know
his/her counterpart:
Quality control
Quality assurance
Biology (cell culture, cell banking)
Process (the complexity of the process
operations often requires a larger team,
with appropriate technical skills)
Upstream processing
Downstream processing
Polishing/Filling/Packaging
For a well-structured project management
it is essential that both parties understand
each other as one integrated team, which
can only achieve its target together. Failure
on either side of this transfer project leads
to a failure of the whole project. This well-
established partnership is a key factor for
the success.
On top of this working level, a steering
committee representing the senior man-
agement of both partners should supervise
the whole project. It is their responsibility
to solve problems escalated from team lev-
el, to monitor progress and milestones of
Fig. 16.6 Organization of a technology transfer using a window person or a window team.
Fig. 16.7 Successful project management structure for technology transfer project.
the project, and to supervise the time-
schedule and financing. While the work-
ing team should function as one single
unit, it is essential for the steering com-
mittee to meet on a regular basis. With
this procedure the senior management is
in the position to act before reaction is
needed. Fig. 16.7 provides an example of a
successful technology transfer project orga-
nization.
Within this structure, each member of
the team communicates directly with the re-
spective partner on the other side. The
teams should meet on a regular basis to dis-
cuss progress, constraints and timelines.
This hierarchical structure facilitates
communication and ensures that the re-
quired information can flow directly from
one partner to another. These communica-
tions are particularly important to ensure
timely information which is required
across all levels of the project organization.
Delay due to long-lasting information flow
can be avoided.
16.3.9
Time to Market, and the Bottom Line
It is almost impossible to comment on the
average time that a biopharmaceutical
product needs to progress from laboratory
bench to market. For a MAb, from discov-
ery to FDA approval and commercial
launch, it costs about US$ 800 million,
and takes 1215 years [10]. However, a re-
cent FDA whitepaper reported that the re-
quired investment for one successful drug
launch (discovery through launch) is US$
1.7 billion [12]. In addition, one must be
aware of the fact that only one in 5000
candidates will make it all the way to com-
mercial success.
Biopharmaceutical development projects
generally take between 20 and 65
months to complete. This number corre-
sponds to between 12 000 and 800 000
man-hours, depending on the complexity
of the process [7]. In fact, the slowest pro-
ject took 2.7 times longer to complete and
consumed almost twice as many resources
hours compared to the fastest. An analysis
of different development projects revealed
that the efficiency plays a major role for
commercial success. More efficient compa-
nies complete two to three times more
projects during the same period of time,
and using the same resources, than do
companies with a lower efficiency [7].
The reduction in development time
leads to a significant competitive advan-
tage, especially as getting a new biophar-
maceutical to the market faster allows a
market share to be established before com-
petitors.
Fig. 16.8 illustrates an approximate mod-
el calculating the cost of the lost opportu-
nity that a company faces for each working
day a bio-therapeutic remains in process
development rather than being on the
market.
While 130 biopharmaceuticals accounted
for sales of approximately US$42.6 billion
in 2002, the average revenue potential of a
single product can be calculated at
US$328 million per year, or US$1.3 mil-
lion per day. The calculation of the daily
burn rate that a biopharmaceutical com-
pound has during its development phase
is based on certain estimates. On the basis
of the above-mentioned data, the average
number of man-hours required for a devel-
opment project is 406 000, within a period
of 42.5 months. At an average personnel
cost of US$225 per hour, the burn rate of
a biopharmaceutical product during the
development phase accounts for $107 000
per day [8, 9].
These figures illustrate very well the
need for efficient process development.
Moreover, improving collaboration between
the CMO and the process provider during
process development in order to shorten
the timeline could also significantly reduce
costs.
16.4
Summary and Outlook
The structural complexity of a specific bio-
pharmaceutical protein, together with the
amount required, determine the choice of
an appropriate expression system. While
antibody fragments can easily be produced
in yeasts, fungi, or E. coli cells, the expres-
sion of whole antibodies still requires com-
plex mammalian cell culture processes.
Although recombinant mammalian cell
culture involves expensive growth media
and complex purification steps, it is the only
production alternative for many products.
As therapy with these potent biopharmaceu-
tical drugs often requires high doses, the
manufacturing capacity can indeed be a lim-
iting factor for the process development.
Much of the expense in the biopharma-
ceutical manufacturing of antibodies or
antibody fragments derives from the large
capital investment required to build and
operate a manufacturing facility. In this re-
spect, the outsourcing of certain services is
an alternative to building in-house capaci-
ties.
Projecting market demand for specific
classes of therapeutics in development will
remain a major planning issue for bio-
pharmaceutical companies. The ability to
meet the demand for biopharmaceutical
production capacity, whether through in-
house manufacturing or outsourced con-
tract manufacturing, carries strategic and
financial implications. Also, FDA approval
policy and the development of the biogene-
rics market will have a major impact on
future capacity requirements (see Part VII,
Chapter 4).
Fig. 16.8 Process development cost versus revenue potential of biopharmaceuticals.
References
1 Milstein C, Khler G. Continuous cultures of
fused cells secreting antibody of predefined
specificity. Nature 1975, 256: 495497.
2 Cacciuttolo MA, et al. Large-Scale Production
of a Monoclonal IgM in a Hybridoma Suspen-
sion Culture. BioPharm 1998, 11(4): 2027.
3 Wurm FM. Development of High-Yielding
Cell Culture processes for medical applica-
tions: How to meet the challenge. Presented
at Scale-Up of Biopharmaceutical Products,
NL-Amsterdam, IBC Life Science UK, 2627
January, 2004.
4 Yokota T, Milenic DE, Whitlow M, and
Schlom J. Rapid tumor penetration of a sin-
gle-chain F
v
and comparison with other im-
munoglobulin forms. Cancer Res 1992, 52:
34023408.
5 Harrison JS and Keshavarz-Moore E. Produc-
tion of antibody fragments in Escherichia coli.
Ann NY Acad Sci 1996, 782: 143158.
6 Kara B. Challenges and solutions: Scale-up of
microbial production systems. Basel Switzer-
land, 24 January 2003 (IBC Life Sciences, Lon-
don, UK).
7 Pisano GP. The Development Factory: Unlock-
ing the potential of process innovation. Har-
vard Business School Presentation, Boston,
MA, 1997, pp. 1343.
8 DiMasi JA. The value of improving the pro-
ductivity of the drug development process:
Faster times and better decisions. Pharmacoe-
conomics 2002, 20 (Suppl.) 3: 110.
9 Neway JO. How the data ecosystem affects
process development and the bottom line. Bio-
Process Int 2004, 2(5): 8291.
10 DiMasi JA, Hansen RW, and Grabowski HG.
The price of innovation: New estimates of
drug development costs. J Health Econ 2003,
22: 151185.
11 Booth R. Extending your enterprise with out-
sourcing. BioProcess Int 2003, 1(10): 1418.
12 US Department of Health and Human Ser-
vices Food and Drug Administration: Chal-
lenge and Opportunity on the Critical Path to
New Medical Products, March 2004.
References 1101
ISBN: 3-527-31184-X
Part V
Biopharmaceuticals used for Diagnositics and Imaging
Abstract
Only about 30 years ago, Khler and Mil-
stein set the stage for one of the key tech-
nologies revolutionizing human life: The
invention of monoclonal antibodies
(MAbs) provided the basis for new tools in
biochemical research, and applications
spread throughout all medical areas. The
1980s were stamped by a hype that anti-
bodies as magic bullets would provide a
major breakthrough in oncology. Drug-tar-
geting, immunotoxins, and radioimmu-
notherapy were key words that led to the
foundation of research programs, expert
groups, and eventually also to production
companies. Since then, various drawbacks
have hit the sector, and antibodies have
survived in niches, as research agents, di-
agnostic tools and for highly specific indi-
cations. Today, with the maturation of the
whole biotech industry, MAbs face a signif-
icant renaissance and appear stronger than
ever. Today, MAbs represent the fastest
growing pharmaceutical market segment,
with a potential to reach worldwide sales
of US$ 20 billion by the year 2010. Since
1982, some 20 MAb-based biopharmaceuti-
cals have been approved for the treatment
of chronic and life-threatening disease,
and there are hundreds of second-genera-
tion products under clinical investigation
(see Part VII, Chapter 4). This develop-
ment is further accelerated by validated
disease targets that are becoming accessi-
ble through the human genome sequence
and many innovative research avenues.
While the commercialization of MAbs is
gathering momentum, the sector is facing
a worldwide shortfall of available biomanu-
facturing capacity that is becoming a criti-
cal strategic limitation, especially for com-
panies without established market access
(see Part IV, Chapter 16 and Part VIII,
Chapter 1). As a result of early clinical fail-
ures, a number of strategies were intro-
duced to address some of the underlying
limitations. Recombinant DNA technolo-
gies, humanized molecules with low im-
munogenic potential (see Part V, Chapter
1105
1
Thirty Years of Monoclonal Antibodies:
A Long Way to Pharmaceutical and Commercial Success
ISBN: 3-527-31184-X
From Hunter to Craftsman
Engineering Antibodies with Natures Universal Toolbox
2), individualized treatment strategies on
the basis of diagnostic tests (see Part I,
Chapters 2 and 5), and genetic profiles as
well as recent improvements in cell biol-
ogy and fermentation are now providing
the basis for pharmaceutical and commer-
cial success (see Part IV, Chapter 1). The
next generation of MAbs will be developed
on the basis of tailor-made, miniaturized
molecules with optimized pharmacoki-
netics, and given to patients most likely to
benefit from the respective therapy.
Although the basic principle of GMP man-
ufacturing is identical for all pharmaceuti-
cals, macromolecules such as MAbs are
subject to some specifics with regard to
their manufacturing costs and potency
profile. Since the products are large and
complex molecules with isoforms and mi-
croheterogeneities, the production process
itself must meet the highest requirements
with regard to consistency and reproduc-
ibility, and process changes are monitored
with a demanding comparability policy.
Improvements are due in all areas of the
pharmaceutical supply chain to manage
current manufacturing challenges, and are
therefore vital for the long-term success of
MAbs as biopharmaceuticals.
Abbreviations
ADCC antibody-dependent cellular
cytotoxicity
AML acute myeloid leukemia
BHK baby hamster kidney cells
BLA biologics license application
BRMs biological response modifiers
CAGR compound annual growth
rate
CDC complement-dependent cyto-
toxicity
CDR complementarity-determining
region
practice
CIPP capturing-intermediate purifi-
cation-polishing
CMC chemistry manufacturing and
controls
COG cost of goods
DMARDs disease-modifying anti-rheu-
matic drugs
DOE design of experiments
DTPA diethylentriaminopenta-ace-
tate
ELISA enzyme-linked immunosor-
bent assay
FACS fluorescence-activated cell
sorting
HAMA human anti mouse antibody
HTST high-temperature short-time
IBD inflammatory bowel disease
Ig immunoglobulin
IGIV immune globulin intravenous
IL interleukin
IL-1ra IL-1 receptor antagonist
IND Investigational New Drug
MAbs monoclonal antibodies
MWCB Master Working Cell Bank
NHL non-Hodgkin lymphoma
NSCLC non-small cell lung cancer
PI primary immunodeficiency
PPV porcine parvovirus
RA rheumatoid arthritis
RIA radioimmunoassay
RSV respiratory syncitial virus
SIP selectively infective phages
TAA tumor-associated antigens
YAC yeast artificial chromosome
variable domain of single an-
tibodies
1 Thirty Years of Monoclonal Antibodies: A Long Way to Pharmaceutical and Commercial Success 1106
1.1
Introduction
1.1.1
Immunoglobulins: Form Follows Function
Antibodies represent a class of flexible mo-
lecular adaptors that play a vital role in the
adaptive immune systems of vertebrates
[1]. First discovered by von Behring and
Kitasato in 1890 [2], it was not before 1960
that their basic structure was determined
by Porter and Edelmann [3]. With their in-
herent diversity and heterogeneity, antibod-
ies mediate humoral and cellular immune
responses, and directly execute biochem-
ical mechanisms such as antigen recogni-
tion, complement-dependent cytotoxicity
(CDC) and antibody-dependent cellular cy-
totoxicity (ADCC) [4]. In most higher
mammals, antibodies exist in five distinct
classes (IgG, IgA, IgM, IgD, and IgE) that
differ in both form (size, charge, amino
acid composition, carbohydrate content)
and function [5].
Immunoglobulins are bifunctional mole-
cules with a basic symmetrical structure
consisting of pairs of identical heavy (l, d,
c, e, a) and light chains (j, k) linked
through disulfide bridges (Fig. 1.1) [6].
The class and subclass of an antibody
are determined by its heavy chain type,
and are linked to different physiological
functions. The individual chains form
globular domains that are either highly
conserved between different immunoglob-
ulins (constant region) or contain se-
quence variability (variable region) and are
stabilized by intrachain disulfide bridges.
Fc-related properties such as complement
activation and lymphocyte binding reside
on domains of the constant region, and
are essential for the cellular immune re-
sponse [7, 8]. Other structural features are
Fig. 1.1 Form follows function: the domain structure of an IgG antibody.
N-linked carbohydrate moieties at Asn297
of the C
H
2 domain and a protein A bind-
ing site between C
H
2 and C
H
3 [9, 10].
The two antibody-binding sites are
formed by the interaction of six hypervari-
able loop regions that are exposed near the
N-terminus of the polypeptide chains.
These complementarity-determining re-
gions (CDRs) are surrounded by relatively
invariant framework residues (Fig. 1.2).
Their diversity (idiotypic variation) repre-
sent the central aspect of the humoral im-
mune response. Variability and thus flex-
ible and efficient molecular antigen recog-
nition arises at several levels, and is the
result of somatic mutation and recombina-
tion of multiple genetic elements during
B-cell differentiation [11].
The binding of antigen to antibody (epi-
topeparatope interaction) is a result of
multiple non-covalent interactions which,
in combination, lead to a considerable bind-
ing energy with thermodynamic affinity
constants in the nM range. Avidity the
binding strength of a multivalent antibody
is usually greater than the sum of the
monovalent affinities, and is the more rele-
vant term under physiological conditions.
Upon primary antigen stimulation, secreted
IgM is released from B cells. During ma-
turation of the immune response, isotype
switching occurs upon support through T
cells, leading to the secretion of IgG, IgA,
or IgE. A T-cell-dependent immune re-
sponse is also accompanied by an affinity
maturation which is basically the selection
of high-affinity B-cell clones that are prefer-
entially triggered to divide and differentiate,
plus the result of somatic hypermutation to
further increase the binding energy of the
immunecomplex formation.
Antigen recognition shows a high level
of specificity, and cross-reactivity can only
occur if determinants are shared between
different antigens. The natural immune
response represents a powerful mixture of
polyclonal antibodies directed to various
epitopes of the antigen, which is exploited
in the secondary antibody response after a
vaccination.
The two functions of an antibody are
therefore represented by the variable do-
main binding to the antigen, and the con-
stant Fc domain that mediates recruitment
of effector cells to establish a cellular im-
mune response.
In a life-saving therapeutic application,
plasma-derived antibodies are widely used
Fig. 1.2 The variable part of an antibody, as seen, 1984 and 2004. The CDR loops are high-
lighted. Left: Original drawing from the Nobel Prize lecture of Niels Jerne ( The Nobel
Foundation). Right: Reproduced with permission of Hoffmann/Fischer RWTH Aachen.
for the treatment of various primary im-
munodeficiency (PI) diseases [12], and
their full potential is just beginning to be
discovered in a wide range of indications
[13]. Polyclonal antibodies from human se-
rum is a multi-tonne product isolated
from approximately 3010
6
L of plasma
each year. Immune globulin intravenous
(IGIV) are now also available as second-
generation products with chromatographic
purification and a strict safety concept con-
sidering all known human pathogens [14,
15]. For a comprehensive overview of the
IGIV products available commercially, see
Ref. [16].
1.2
Making Monoclonal Antibodies
For the specific recognition of a single epi-
tope, however, polyclonal antibodies are
not the right tool. It was therefore good
news in 1975 when George F. Khler and
Csar Milstein described a method to gen-
erate monoclonal antibodies (MAbs) [17].
In Milsteins group at the MRC Laboratory
of Molecular Biology in Cambridge, Khler
had cloned an immortalized cell line se-
creting Sp1, a mouse IgM antibody. The
original process is based on the injection
of an antigen into a mouse to induce a
specific immune response. Differentiated
B lymphocytes are isolated from the
spleen and fused with immortalized mye-
loma cells. The resulting hybridoma cells
are then selected for their potential of im-
mortalized growth and screened for specif-
ic antibody production. The development
of MAbs starts from single cell clones that
after propagation are laid down in a
cell bank and characterized for routine
production in repeated manufacturing
trains. Amazingly, the full significance of
this invention was not immediately recog-
nized, and was not even patented
(Fig. 1.3).
However, together with Niels K. Jerne,
George D. Khler and Csar Milstein were
later awarded the 1984 Nobel Prize in
Medicine . . . for theories concerning the
specificity in development and control of
the immune system and the discovery of
the principle for production of monoclonal
antibodies.
The traditional method of MAb genera-
tion has a number of drawbacks and lim-
itations. Firstly, these antibodies are mur-
ine. Despite many similarities and struc-
tural homologies, the possibilities of thera-
peutic use are rather limited due to the
short serum half-life in humans. In addi-
tion, the foreign protein elicits an im-
mune response after repeated administra-
tion, with the most discouraging being the
human anti mouse antibody response
(HAMA) a possibly life-threatening im-
munological event that is directed against
the idiotype and the isotype of murine an-
tibodies [18].
New strategies to generate MAbs that
are less immunogenic and eventually fully
human have since been invented, however.
1.2.1
Generation of Human Antibodies
A variety of possibilities to reduce the risk
of immunogenicity of a therapeutic anti-
body exist, and have been employed.
Chimeric antibodies are hybrid molecules
combining the antigen-specific variable do-
main of the mouse antibody fused to the
constant regions of a human IgG molecule
(Fig. 1.4). This reduces the risk of immu-
nogenicity somewhat, and the human Fc
domain prolongs the serum half-life and is
more effective in triggering the effector
systems of complement and Fc receptors.
Fig. 1.3 Making monoclonal antibodies. An original drawing from the Nobel Prize lecture of
Csar Milstein ( The Nobel Foundation).
Further genetic engineering has led to the
development of humanized antibodies that
retain the antigen-specific CDR of the orig-
inal antibody, grafted onto a fully human ac-
ceptor framework [19]. However, this graft-
ing process is not always straightforward,
and certain framework residues have been
identified that are essential for correct pre-
sentation of the CDRs. Despite advanced
molecular modeling efforts, grafting still re-
quires the experimental validation, to arrive
at a molecule that retains the same specific-
ity and affinity as the parent antibody [20].
Further efforts have therefore gone into al-
ternative technologies that allow the genera-
tion of fully human antibodies [21].
1.2.1.1 Transgenic Mice
One powerful technology is the develop-
ment of transgenic mice that have func-
tionally replaced the mouse antibody genes
with the human equivalents. This was pos-
sible through technologies which allowed
the insertion of large yeast artificial chro-
mosome (YAC) fragments into the germ-
lines of mice [22]. As these mice raise an
immune response following immuniza-
tion, they generate fully human antibodies,
which can then be isolated with the tradi-
tional hybridoma technology [23]. Cur-
rently, Abgenix, Medarex, and Kirin exploit
this technology commercially.
1.2.1.2 Recombinant Libraries
Another approach employs large combina-
torial libraries based on human VH and
VL genes. These libraries have the advan-
tage that clonal selection against self anti-
gen does not take place, which can be a
problem when raising an antibody against
a highly conserved epitope in an animal.
With recombinant libraries, the chal-
lenge is to link the phenotype (physical
binding of the antibody protein to the anti-
Fig. 1.4 From murine to human antibodies. (Figure courtesy of Steven Chamow, Bio Process
Consultant.)
gen) to the genotype, the cDNA encoding
the protein. The most commonly em-
ployed method to isolate new binders from
a recombinant antibody library is that of
phage display [24, 25].
Phage display technology
This method has been employed for the
isolation of new scFvs from naive libraries,
as well as for affinity maturation. To select
for scFvs binding to a particular antigen,
the scFv is fused to a minor coat protein,
typically pIII (g3p) of filamentous M13
phage. During phage panning, the scFv on
the phage is bound to immobilized anti-
gen and enriched during consecutive cy-
cles of binding, elution, and amplification
after the infection of bacteria. However, a
disproportional amplification of non-specif-
ically binding phages can occur, since non-
specifically and specifically bound phages
can infect the bacterial cell. Therefore, it is
of interest to minimize non-specific ad-
sorption of the phages in the binding pro-
cedure and to specifically enrich those
phages, which display high-affinity mole-
cules over highly abundant, low-affinity
binders. The method of selectively infec-
tive phages (SIP) addresses this problem.
In the SIP procedure, the desired antigen
antibody interaction itself is essential for
restoring infectivity in an otherwise non-
infective phage displaying the scFv. The
coat protein pIII (g3p) is lacking the N-ter-
minal domain responsible for infectivity. A
fusion between the antigen and the miss-
ing domains of the coat protein restores
infectivity, thereby linking binding and in-
fectivity. In contrast to phage display,
where the interacting scFvs bind to an
antigen on a solid-phase surface (e.g., an
affinity column), no solid-phase interaction
is necessary in SIP, thus avoiding the
problem of non-specific interactions.
Furthermore, SIP is a one-step procedure
(binding and infection are coupled, and
elution becomes unnecessary) and there-
fore faster and easier to perform [26].
Phage display as well as SIP is a
powerful tool used to isolate new antibod-
ies and select for increasing antibody affin-
ity. A number of companies exploit this
powerful, but heavily patent-protected tech-
nology; these include Morphosys (see Part
V, Chapter 2), Cambridge Antibody Tech-
nology (CAT), and Dyax. Humira
is the
first phage display-derived MAb on the
market. On the other hand, about 30% of
all human antibodies currently in clinical
development have been generated using
this technology [27].
Yeast Surface Display and Ribosome Display
Alternative methods for selection of inter-
acting proteins, including antibodyanti-
gen interactions include yeast surface dis-
play and ribosome display (Fig. 1.5).
Despite its relatively late advent in the
antibody engineering field, yeast cell surface
display also presents an attractive an
powerful method to isolate affinity-ma-
tured antibodies [28]. Unlike phages, yeast
cells are large enough to be screened and
separated using flow cytometry. In the
yeast cell surface display the scFv is deliv-
ered to the yeast cell wall by fusion to the
cell surface protein Aga2p. The binding of
fluorescent or biotinylated antigen to a
specific scFv on the cell surface can be
measured using flow cytometry, after
which the individual cells are sorted by
fluorescence-activated cell sorting (FACS).
This system was used successfully for the
affinity maturation of antibodies and recep-
tors, and allows very fine discrimination be-
tween mutants of slightly different affinity.
One disadvantage of both phage and
yeast surface display-technologies lies in
the fact that the displayed peptide or pro-
tein is rather small in size compared to
the carrier (phage particle, or yeast cell).
This can lead to false positives due to un-
specific interactions. Ribosome display is an
in vitro method that avoids this display
problem, since it links the peptide directly
to the genetic information (mRNA): an
scFv cDNA library is expressed in vitro
using a transcriptiontranslation system.
The translated scFvs are stalled to the
translating ribosome by the addition of
puromycin, and thus linked to the encod-
ing mRNA. The scFv is then bound to the
immobilized antigen and unspecific ribo-
some complexes are removed by extensive
washes. The remaining complexes are
eluted and the RNA is isolated, reverse-
transcribed to cDNA, and subsequently re-
amplified by polymerase chain reaction
(PCR). The PCR product is then used for
the next cycle of enrichment, and the mu-
tations that are introduced through the er-
ror-prone reverse transcriptase can actually
contribute to affinity evolution [29].
1.3
Other Antibody Formats:
Antibody Fragments
Although full-sized IgG molecules are the
naturally occurring format of antigen-bind-
ing molecules, fragments and derivatives
Fig. 1.5 Methods to generate fully human antibodies. A) Phage display; B) ribosomal dis-
play; C) transgenic mice. (www. Esbatech.com; www.Medarex.com.)
Fig. 1.6 Antibody-fragments under evaluation.
Table 1.1 Pharmacological characteristics of different antibody formats
Parameter mAB Fab scFv
MW 150 kDa 50 kDa 27 kDa
Glomerular filtration (<5060 kDa) (+) +
Tumor/tissue penetration + ++ +++
a-elimination half-life in blood
(equilibrium half-life)
>1 h 30 min 810 min
b-elemination half-life in blood
(clearance half-life)
13 weeks 56 h 34 h
Biological effects of Fc part (e.g., ADCC; CDC) +
are also developed and employed for a
variety of applications. These include Fab
fragments, single chain antibodies (scFv),
and even single domain antibodies
(Fig. 1.6).
Advances in genetic engineering meth-
ods have enabled the generation of a vari-
ety of different antibody fragments and fu-
sion proteins, and allowed the construc-
tion of molecules, the properties of which
are tailored for the specific requirements
of the intended therapeutic application
[30]. These properties focus on efficacy
that is, the biological activity that ad-
dresses the mechanism of action. In addi-
tion, issues such as biochemical and bio-
physical stability, solubility, and production
yield which influence the cost of goods
(COG), and the pharmacokinetic properties
as well as the immunogenicity of the pro-
tein, play an important role (see Table 1.1).
As described earlier, antibody-binding to
its cognate antigen can elicit an immune
response through the binding of the Fc
part of the antibody to the Fc receptor on
various effector cells, and by activation of
the complement system. This can be an
undesired effect, in particular in inflam-
matory diseases, where generally only neu-
tralizing antigen-binding activity is re-
quired. Antibody constructs that eliminate
the constant Fc domains can therefore be
an advantageous format. A number of
antibody fragments have been developed
specifically to address the individual appli-
cations and their requirements; for a com-
prehensive overview, see Ref. [31].
The Fab fragment contains the variable
domain of the heavy and the light chain
and the adjacent constant domains C
H
1
and C
L,
linked through a disulfide bond,
that naturally links these two chains to-
gether [32,33]. Although the specificity and
selectivity of the monovalent Fab fragment
is the same as that of the parent antibody,
its avidity is lower, because it has only one
antigen-binding site. Through different
linker constructs at the hinge region of
the antibody, dimeric and even trimeric
Fab constructs have been generated [34].
The Fv fragment represents the smallest
antibody domain that retains an acceptable
affinity. It is usually very unstable, and ag-
gregates as a result of only weak, non-
covalent forces, but can be stabilized with
a linker polypeptide yielding a single chain
Fv antibody fragment (scFv) and allows ex-
pression from a single gene in a number
of hosts. It also facilitates additional genet-
ic engineering such as fusion to effector
domains [35]. Through variation in the lin-
ker size, not only multimeric variants of
the scFv such as diabodies and tribodies
but also other linear antibody fragments
can be generated that possess a higher
avidity than monovalent versions [36, 37].
A variation thereof is the generation of
bispecific antibodies [38]. In certain appli-
cations, these are designed to bind to a
cancer-specific surface molecule with one
binding domain, while the other binding
domain recruits cytotoxic T cells to the
pathogenic cell, inducing T-cell-dependent
cytotoxicity [39].
Single-chain antibodies that consist of
only a single variable domain with three
CDRs occur naturally in camels and lla-
mas species that are lacking the light
chains of antibodies with still high stability
and affinities [40, 41]. Human versions
have been isolated and are being devel-
oped a process called camelization [42].
The resulting VHHs are the smallest avail-
able intact antigen-binding fragment, with
a molecular weight of approx. 15 kDa,
118136 residues that remain highly solu-
ble and stable with affinities in the nano-
molecular range [43]. Secretion in the en-
doplasmic reticulum (ER) is enhanced
through hydrophilic amino acids.
1.3.1
Pharmacokinetics of Antibody Fragments
One problem associated with small-sized
antibody fragments is that of rapid renal
clearance from the bloodstream that leads
to a reduction in half-life to hours, rather
than weeks (as is the case for full-sized an-
tibodies).
This may be an advantage for some
acute indications such as myocardial in-
farction or acute infections. It has also
been shown, that the much smaller pro-
tein size allows faster tissue penetration,
which might be advantageous for solid tu-
mors and local applications [44]. For sys-
temic applications, however, the possibility
of avoiding rapid elimination via glomeru-
lar filtration is to increase the size of the
molecule, for example by PEGylation. The
addition of polyethylene glycol (PEG) to a
random or defined site of the protein in-
creases its apparent hydrodynamic size,
such that the protein becomes more stable
and possibly even less immunogenic, as
the large parts of the protein surface are
no longer accessible by the immune sys-
tem [45, 46] (see also Part VI, Chapter 2).
Examples of therapeutic fragment antibod-
ies are CDP 870 (an anti-TNF-alpha PEGy-
lated Fab fragment for the treatment of
Crohns disease), and the complement in-
hibitor PexelizuMAb, which is currently
under clinical development [47, 48].
The production of full-sized antibodies
is restricted to mammalian cell systems
that possess the correct production and
glycosylation machinery (see Part IV,
Chapter 1). Antibody fragments, on the
other hand, can also be expressed at much
lower cost in microbial expression sys-
tems, such as bacterial, yeast, or fungal
fermentation [49, 50]. For the high-level ex-
pression of antibody fragments, E. coli fer-
mentation provides a well-established tech-
nology basis with periplasmic expression
levels at 12 g L
1
during high cell density
fermentation [51].
1.4
Medical Application Areas for MAbs
The 30-year history of MAbs has been a
rollercoaster ride to success. From hype to
depression, and back to hype, is probably
the shortest summary of what has hap-
pened. MAbs have been rapidly introduced
into a number of applications within and
outside the medical field (Table 1.2) [52
54]. Analytical in vitro methods such as
enzyme-linked immunosorbent assay
(ELISA), radioimmunoassay (RIA), various
blotting techniques, flow-cytometry, immu-
nofluorescence, confocal imaging, and im-
munohistochemistry are each dependent
upon the use of polyclonal or monoclonal
antibodies.
Table 1.2 Application areas for monoclonal antibod-
ies
In-vitro use
Research reagent in drug discovery
Analytical tool in drug development and manu-
facturing
In-vitro diagnostics (biochemical, histological,
pathological)
Immunoaffinity chromatography
Biosensors
Catalytic antibodies
In-vivo use
Immunoscintigraphy
Isotypic immunotherapy
Idiotypic immunotherapy (antagonistic and ago-
nistic)
Drug targeting with immunotoxins and immuno-
conjugates
Radioimmunotherapy
Edible vaccines
Based on their physiological role, unmo-
dified MAbs can trigger different immuno-
logical reactions that are used in medical
applications [55]. These reactions may vary
from passive immunization through effec-
tor functions to selective cytotoxic effects
with bispecific antibody constructs that re-
cruit cells of the immune system to de-
stroy identified targets [56].
The vast majority of todays antibody
treatment concepts rely on active immuni-
zation directed towards specific disease tar-
gets. The effect can either be antagonistic
through the neutralization of a signaling
molecule or its specific receptor or agonis-
tic with rather enzyme-like support of a
physiological function [57]. Antagonistic
treatment concepts typically require high
doses in the region of >1 mg kg
1
, and
they include the treatment of inflamma-
tory disorders such as rheumatoid arthri-
tis, inflammatory bowel disease (IBD),
psoriasis, allergies, and autoimmune dis-
eases (e.g., multiple sclerosis and Crohns
disease). Soluble cytokines such as inter-
leukin (IL)-1 and tumor necrosis factor
(TNF) are targets to prevent pro-inflamma-
tory mechanisms.
Following marketing authorization of the
first therapeutic antibody OKT3 in 1986for
the treatment of acute transplant rejection,
premature hopes and unrealistic expecta-
tions were raised, and MAbs had a difficult
time in living up to their original promise
(see Introduction and Part II, Chapter 4).
During the ensuing period, antibodies sur-
vived in niches as research reagents and di-
agnostic enabling tools, and it was not until
1995 that Centocors ReoPro
(Abciximab)
a chimeric MAb fragment won approval
for the prevention of thrombotic side effects
in patients undergoing coronary artery an-
gioplasty [58]. ReoPro binds to a glycopro-
tein receptor on human platelets and pre-
vents their aggregation.
1.5
From Initial Failure to Success:
Getting the Target Right
The initial lessons learned were derived
from a number of spectacular failures dur-
ing the development of MAbs antibodies,
and these in turn led the way to success.
This intermezzo is in fact one important
reason for the first downturn in biotech-
nology when investors sent the sector into
some of its leanest times and a good learn-
ing lesson for the companies involved.
A closer look at antibodies against endo-
toxins as a cause of septic shock may serve
as an example [59]. Septic shock continues
to be a severe problem, with mortality
rates up to 70% and TNFa as a primary
mediator that can be stimulated by endo-
toxins [60, 61]. Drug treatment standards
for septic shock remain poor [62, 63]. Cur-
rently, Synergen is the leader among a
long list of biotech companies that have
suffered badly from attempts to achieve
sepsis treatment. Their drug Antril
a
recombinant form of the naturally occur-
ring 23 kDa IL-1 receptor antagonist (IL-
1ra) was the first recombinant protein
that failed to improve significantly the sur-
vival of sepsis patients in large-scale clini-
cal trials following positive intermediate
results [64]. The company built a produc-
tion facility based on the promising results
of early Phase II trials, but these could not
be reproduced in subsequent studies [65].
Monoclonal antibodies for sepsis treatment
lost favor in 1992 when the FDA refused
to approve Centoxin
(Centocor), a human
monoclonal IgM antibody binding to lipid
A type endotoxins from Gram-negative
bacteria [66]. Centoxin
was voluntarily
withdrawn from the European market in
1993 after it showed a lack of benefit in
post-approval studies in septic shock treat-
ment [67]. Besides Centoxin
, another
anti-endotoxin antibody was investigated
in large-scale clinical trials. This was E5
,
a MAb from Xoma which also failed to
demonstrate any benefit after encouraging
initial results, and was not developed
further [68]. The basic problem with anti-
endotoxin antibodies has been recognized
as being the small therapeutic window be-
fore inflammatory cytokine expression oc-
curs [69]. Other strategies therefore fo-
cused on TNFa as the central mediator of
inflammation following various stimuli,
including endotoxins. Intracellular signal
transduction occurs upon clustering of
TNFa receptors, and this can be prevented
by high-affinity neutralizing antibodies to
the circulating TNFa trimers. This
approach has been followed by Bayer with
a murine MAb that also failed to demon-
strate a significantly positive effect on pa-
tient mortality [70].
Failures such as those described above
can ruin a company financially, or at least
shatter its faith in biotechnology, as the de-
velopment costs for a new biological entity
are usually in the range of US$ 500
800 million, with late-stage clinical trials
consuming the majority of this money
(see Part IV, Chapter 16; Part VII, Chapter
4; and Part VIII, Chapter 1).
The fallout from these early cases per-
sisted for a long time, and it took the en-
tire biotech sector many years to recover
from this negative impact. As a result of
this consolidation, however, the sector is
now stronger than ever, is generating a
very positive news flow with a number of
biopharmaceuticals that entered the mar-
ket recently, and once again it is the anti-
body sector that is leading the way:
TNF-a: this is still an important disease
target, and a number of drugs are now on
the market. In fact, these biological re-
sponse modifiers (BRMs), directed towards
specific cytokines, have changed the treat-
ment of rheumatoid arthritis (RA) drasti-
cally and now represent the therapy stan-
dard for this condition [71]. Globally, some
five million people have RA, with 1.25 mil-
lion suffering from moderate-to-severe
symptoms. Biological drugs represent a
second-line therapy prescribed to the 50%
of those patients who fail disease-modify-
ing anti-rheumatic drugs (DMARDs). The
therapies approved comprise three anti-
body-based drugs that target TNF-a [En-
brel
(Etanercept); Amgen Inc.; Remi-

cade
(InflixiMAb); Centocor Inc., and

Humira
(AdalimuMAb), Abbott), and

one IL-1 receptor agonist that targets IL-1
(Kineret
(Anakinra), Amgen, Inc.) [72,

73]. (In the interim, Synergen was ac-
quired by Amgen, and Antril was
launched as Kineret to treat RA.) Interest-
ingly, the efficacy and safety profiles for all
four of these materials are different [74].
Enbrel
: (etanercept) a recombinant hu-

man fusion protein of two soluble TNF-a
type II receptor molecules and part of the
Fc portion of an IgG1-antibody to circum-
vent rapid clearance and facilitate effector
functionality. This represents an interest-
ing class of fusion proteins that carry anti-
body domains for their specific function
[75]. Inhibiting TNF-a has proven to pro-
vide a highly effective approach to the
treatment of chronic inflammatory ill-
nesses. Originally approved for RA, etaner-
cept has now gained approval for addi-
tional indications, including psoriatic ar-
thritis, juvenile rheumatoid arthritis, and
ankylosing spondylitis. Enbrel was also a
learning example as to what can happen
when demand exceeds capacity and short-
age of materials is limiting a drugs
growth potential. Without sufficient manu-
facturing capacity, Immunex Corp. (the in-
ventor of Enbrel) was forced into a take-
over by Amgen.
Another representative of this class of
molecules is Amevive
(alefacept), a fu-
sion protein consisting of the first LFA-3
extracellular domain fused to the hinge
CH2 and CH3 regions of human IgG1 de-
veloped by Biogen [76, 77]. The molecule
has immunomodulatory properties be-
cause it is targeting the co-stimulatory role
of CD2 in T-cell activation as well as cell
adhesion and the migration of lympho-
cytes into tissue.
A closer look at currently available MAbs
inhibiting the inflammatory effects of
TNF-a in diseases such as RA, Crohns dis-
ease, psoriatic arthritis, bacterial septic
shock, as well as in the prevention of allor-
eactivity and graft rejection, confirm their
high potential [78]. Strategies include neu-
tralization of the cytokine via either anti-
TNF antibodies, soluble receptors, or re-
ceptor fusion proteins. Rheumatoid arthri-
tis is currently addressed by no less than
10 antibody products that are either on the
market or in late stage clinical develop-
ment. The competitive edges for these can-
didates are certainly a low to non-existing
antigenicity profile with no neutralizing
antibodies detected in patients, and also a
convenient mode of administration such
as subcutaneous depot forms.
The anti-TNF biologics class is expected
to continue steady growth within the RA
market, and Lehman Brothers analysts
predict that it will peak at sales of US$ 5
6 billion per annum in the US. From the
pipeline, CDP-870 is the largest threat to
the established brands.
Compared with RA, the market for bio-
logics in psoriatic arthritis is in its infancy.
According to Lehman Brothers, peak an-
nual sales may reach US$ 23 billion. Many
of the same biologics from the RA market
are also now targeting psoriasis. Enbrel re-
ceived approval for this indication in the
US in early 2002, and both Humira
and
Remicade
are in pivotal Phase III trials.

Other lessons learned from the several
high-profile clinical trial failures were the
limitations of fully murine antibodies that
may eliminate the therapeutic efficacy
(neutralization, rapid clearance) or even
lead to serious side effects (serum sick-
ness, anaphylactic shock) [79]. Although a
number of murine antibodies are still
commercially available, the focus has
shifted to chimeric, humanized and even-
tually human antibodies that are now be-
coming available. Whether maximizing the
human sequence in an antibody is mini-
mizing antigenicity is questionable, how-
ever. In fact, data on the generation of
neutralizing antibodies in patients have
demonstrated that some of the marketed
humanized antibodies still exhibit signifi-
cant immunogenicity a phenomenon
that is also known for other protein thera-
peutics of human origin and is probably
related to other host cell-derived differ-
ences in the manufacturing process.
Other strategies to reduce the immuno-
genicity in therapeutic antibodies comprise
the covalent attachment of PEG to mask
antigenic sites through extensive hydration
(see Part VI, Chapter 2), veneering (re-
moval of exposed B-cell epitopes in the fra-
mework region), and removal of T-cell epi-
topes to avoid proliferation of helper T
cells and stimulation of a mature immune
response [80].
1.6
The Market Perspective
To date, a total of 22 MAbs and MAb-re-
lated proteins has been approved, and are
being produced in different annual
amounts ranging from the multi-gram to
the hundreds of kilogram scale, and with
1,6 The Market Perspective 1119
Table 1.3 Approved therapeutic monoclonal antibodies
Marketed
product
Company Antibody type Year of
first
approval
Molecular target/
main indication
Anti-cytokine/anti-AB MAbs
Remicade Centocor/Schering
Plough
Chimeric 1998 TNFa; Crohns disease,
Rheumatoid arthritis
Enbrel Amgen/Immunex TNF-receptor/Fc (IgG1)
fusion protein
1998 TNFa; Rheumatoid arthritis
Humira Abbott/CAT Human 2002 TNFa; Rheumatoid arthritis
Xolair Genentech/Novartis/
Tanox
Humanized 2003 IgE; Allergic asthma
Anti-cytokine-receptor MAbs
Herceptin Genentech/Roche Humanized 1998 EGFR2; Breast cancer
Avastin Genentech/Roche Humanized 2004 VEGF; Various cancers
Erbitux ImClone, BMS, E.
Merck
Chimeric 2004 EGFR; Various cancers
MAbs to lymphocyte surface antigens
OrtoClone
OKT3
Johnson & Johnson Murine 1986 CD3; Transplant rejection
Rituxan/
MabThera
BiogenIdec/Schering Chimeric 1997 CD20; Non-Hodgkin
Lymphoma (NHL)
Zevalin BiogenIdec/Schering Murine, linked to
90
Y 2002 CD 20; NHL
Bexxar Corixa/GSK Murine, linked to
131
I 2003 CD20; NHL
Zenapax PDL/Roche Humanized 1997 CD25 (IL-2 receptor);
Transplant rejection
Simulect Novartis Chimeric 1998 CD25; Transplant rejection
ReoPro Centocor/Eli Lilly Fab Chimeric 1995 Platelet receptor;
Anti-thrombotic
Antilfa Immunotech/SangStat Murine 1997 Anti-LFA-1a; CD11a;
Transplant rejection
Mylotarg Celltech/Wyeth Humanized; linked
to Calecheamicin
2000 CD33; AML
Campath ILEX/Schering Humanized 2001 CD52; B-CLL
Raptiva Genentech/Xoma/
Serono
Humanized 2003 CD11a; Psoriasis
Amevive BiogenIdec LFA3-IgG1 fusion
protein
2003 CD2; Psoriasis
MABs to cancer-associated antigens
Panorex Centocor/GSK Murine 1995 Anti-HEA (EpCAM);
Colorectal cancer
MAbs to pathogens
Synagis Medimmune/Abbott Humanized 1998 RSV; RSV infection1
ton requirements coming into sight (Table
1.3).
The first humanized MAb was Synagis
(Medimmune), directed against an epitope

on the surface of the respiratory syncitial
virus (RSV). This was launched in 1998,
and the first fully human MAb is now Hu-
mira
(Abbott/CAT), approved in 2002for

the treatment of RA.
Many MAbs are directed towards tumor-
associated antigens or receptors that are
up-regulated in malignancy (see Introduc-
tion and Part II, Chapter 4). In 1997, Re-
tuxiMAb (Rituxan
, MabThera
) became
the first cancer therapeutic monoclonal
antibody [81]. In 1991, rituximab was de-
veloped as a chimeric anti-CD 20 antibody
with non-Hodgkin lymphoma (NHL) as
the original indication. Clinical trials were
initiated in 1993 after Investigational New
Drug (IND) filing in 1992, and lasted no
longer than 3 years. The antibody binds to
B cells where it induces CDC, ADCC, and
apoptosis. It has been shown that Ritux-
an
functions via the recruitment of cyto-

toxic T cells to the tumor cells. Patients
with a polymorphism in the Fc receptor
that has higher affinity to the Fc part, have
significantly better responses, demonstrat-
ing the importance of this interaction for
therapeutic efficacy [82] (see Part I, Chap-
ter 2). Today, the drug is being investigated
for the treatment of various autoimmune
disorders, and it is said to have high addi-
tional potential [83].
Another target is the receptor for the
epidermal growth factor (EGF) that is in-
volved in a number of solid tumor forms.
Anti-EGF receptor antibodies are applied,
together with traditional cytotoxic drugs.
One promising new approach to tumor
treatment is the inhibition of angiogenesis
as one primary disease target to slow
down tumor growth and metastasis. Re-
leased pro-angiogenic factors trigger vascu-
larization of the primary tumor to provide
nutrients and oxygen for the rapidly prolif-
erating cancer cells. The first in class ther-
apeutic is Avastin
, a humanized MAb
that received market approval in 2004for
the treatment of first-line metastatic colo-
rectal cancer, but only after some draw-
backs in initial studies [84].
Herceptin
(trastuzumab) is a huma-
nized antibody with antitumoral activity
through the antagonistic inhibition of
EGF-receptor activation [85] (see Part I,
Chapter 5). EGF (also called HER14) re-
presents a family of cytokine-regulated, tyr-
osine kinase-type receptors regulating an
intracellular signaling cascade that can in-
itiate cell regulation or apoptosis [86, 87].
HER1 and HER2 require dimerization for
transduction signals to occur, and are
therefore targets for MAbs [88, 89]. Her-
ceptin binds to HER2 and triggers various
immunological defense mechanisms. To-
day, Herceptin is approved as the first-line
treatment for metastatic breast cancer in
combination with chemotherapy for recep-
tor-positive patients [90]. Erbitux
(cetuxi-
mab) is another representative of the
EGFR-antagonist MAbs. This is a chimeric
antibody that binds to HER1, where it pre-
vents ligand-fixation, homodimerization
and intracellular signal transduction,
which in turn leads to cell death and tu-
mor regression [91].
In addition, with Zevalin
, Bexxar
,
and Mylotarg
, the first drug targeting de-

vices gained market access for the specific
treatment of neoplastic diseases such as
NHL and acute myeloid leukemia (AML)
[92].
Some of the now commercially available
antibody-based products have blockbus-
ter potential, and the overall development
of the sector is very strong. MAbs repre-
sent the fastest growing segment within
biopharmaceuticals, and are outperform-
1,6 The Market Perspective 1121
ing recombinant proteins with a com-
pound annual growth rate (CAGR) of 20%
[93] (see Part VIII, Chapter 1). In addition,
there are hundreds of second- and third-
generation antibody-based products in pre-
clinical and clinical development, and even
under the assumption of growing attrition
rates, substitution pressure and margin
squeeze, MAbs will probably reach a stable
plateau of US$ 20 billion within 10 years
from now, and the required amounts of
material will rise accordingly [94].
Among the MAbs in late-stage clinical
investigation, or queuing for regulatory ap-
proval, are Antegren
(Elan, Biogen MS),

Tarceva
(Genentech, OSI Pharmaceuti-

cals, and Roche) and CPD 870 (Celltech).
Antegren is a monoclonal antibody for the
treatment of multiple sclerosis and Crohns
disease, and received fast track review sta-
tus in the US after only one year of Phase
III human trials. Antegren
attaches to al-
pha-4-integrin and renders it unable to
stick to VLA-4, thereby preventing chemo-
taxis of helper T cells into the endotheli-
um. Tarceva
blocks the EGF-receptor,

and is under evaluation for the treatment
of advanced non-small cell lung cancer
(NSCLC). The drug has also undergone
early-stage trials for a range of other tu-
mor types, such as ovarian, colorectal,
head and neck, glioma and gastrointestinal
cancers. CPD 870 is an anti-TNFa anti-
body fragment with the indication for RA
and Crohns disease (both in Phase III
trials).
1.7
Drug Targeting: The Next Generation
in Cancer Treatment
Despite impressive advances in the treat-
ment of specific cancer forms, the therapy
of most solid tumor forms is still missing
a major breakthrough, and is in a striking
misbalance to the in-depth knowledge on
carcinogenesis, tumor development and
Fig. 1.7 Annual sales and manufacturing demands for marketed MAbs. (Data see Ref. [93].)
metastasis [95]. In traditional chemother-
apy, physiological mechanisms (and the
fact that the overall metabolism and prolif-
eration of cancer cells are increased) are
used for selective treatment. Selectivity,
however, is rather poor because tumor-spe-
cific differences are quantitative rather
than qualitative. This in turn leads to a
therapeutic window that is very small, in-
cluding the well-known side effects on all
proliferating tissue such as bone marrow,
lymphatic tissue, and gastrointestinal
epithelium. Malignant growth is generated
through transformation of the bodys own
cells. During malignant transformation,
cells lose their differentiated status as well
as their physiological role, including the
basis for intracellular communication.
Among various other changes, the expres-
sion of cell surface receptors is typically
up- or down- regulated, and tumor-asso-
ciated antigens (TAA) such as oncofetal
antigens are presented [96]. These anti-
gens differ in their copy number or struc-
ture, with glycosylation being the predomi-
nant change a phenomenon that is also
building the basis for the organotrophy of
metastasis [97]. TAA may, on the other
hand, provide molecular targets that are
suitable docking stations for the binding
and internalization of antibodydrug con-
jugates and immunotoxins [98]. Since the
first appearance of MAbs, a large number
have been developed as specific diagnostic
agents for the detection as well as the ser-
ological, biochemical, and histological
characterization of cancer cells and tissues
[99]. Immunoscintigraphy with radiola-
beled MAbs and antibody fragments is an
appropriate tool for the evaluation of their
biodistribution, and is a standard diagnos-
tic method for the detection of primary tu-
mor and metastasis with tumor markers
such as the carcinoembryonic antigen
(CEA) [100] (see Part V, Chapter 6). An ac-
ceptable compromise of rapid penetration
and thus localization and clearance rate
must be found for every application. The
radioisotopes used provide a rapid decay
rate (e.g.,
99m
Tc, t
1/2
=6.1 h;
111
In, t
1/2
=2.8 days), and are covalently linked to
the antibody with complexing agents such
as diethylentriaminopenta-acetate (DTPA)
[101]. Typically, less than 1% of the applied
radioactivity is localized at the tumor after
68 hours. In that respect, immunoscinti-
graphy is a good tool for investigating the
possibilities, as well as the limitations, of a
therapeutic use [102, 103] (see Part V,
Chapter 4).
Drug targeting with MAbs has always
been a fascinating theoretical approach to
the rational treatment of acquired diseases
such as cancer, according to the concept of
Magic Bullets a term created by Paul
Ehrlich who suggested the coupling of
toxic natural compounds to specific vectors
that are able to find their way to the dis-
ease and thereby increase the pharmaceu-
tical index [104].
Specificity, dosimetric considerations as
well as various pharmacological and toxico-
logical problems, have so far limited the ap-
plication of immunotoxins and radioimmu-
notherapy to life-threatening diseases where
other treatment regimes have failed [105].
In the payload approach, different classes
of cytotoxic agents have been immobilized
to MAbs (Table 1.4) [106, 107]. Their toxic
potential varies widely, and this must be
considered in dosimetric calculations. With
a stoichiometric mode of action, sufficient
antibody-conjugates must recognize and en-
ter malignant cells to facilitate cell killing,
through whatever mechanism [108]. Tumor
localization and especially diffusion into
solid tumors is however relatively ineffi-
cient for large molecules, and huge doses
must be administered in order to generate
a therapeutic effect. Protein toxins are
usually very potent inhibitors of ribosomal
protein translation and due to their enzy-
matic mechanism of action a single mole-
cule can kill a cell [109]. Limitations arise
from the fact that the slightest systemic
cross-reaction has drastic consequences
[110, 111]. In addition, protein toxins are
highly immunogenic.
The ideal cytotoxic drug is a small mole-
cule that is still very potent. A promising
class of cytotoxic agents is a group of
radiomimetic substances called ene-
diyenes; these are secondary metabolites
from Streptomyces that are minor groove
DNA-binders and introduce double-strand
breaks that may lead to apoptosis [112
114] (Fig. 1.8). Their anti-tumor potential
has been investigated in a number of ex-
perimental studies [115117].
The ideal antibody vector recognizes a tu-
mor-specific antigen that facilitates recep-
tor-mediated endocytotic uptake, and re-
leases the cytotoxic drug inside the cell
upon cleavage of the chemical linkage. Even
if in vitro experiments demonstrate impres-
sive selective toxicity, a number of addi-
tional drawbacks develop from a variety of
pharmacological problems. Malignant tis-
sue is not monoclonal (tumor cell heteroge-
neity), and cells from different parts of the
tumor may escape damage through lack of
antigen, or for anatomical reasons. In addi-
tion, certain antigens shed from the tu-
mor cell surface into the bloodstream
can neutralize immunoconjugates any-
where in the body. Serum stability is re-
duced due to proteolytic degradation.
The ideal linker for the covalent attach-
ment of toxins to antibodies is a device
that is stable in the bloodstream and clea-
vable upon uptake into the cellular com-
partment, where action is thought to take
Table 1.4 Cytotoxic drugs and toxins linked to monoclonal anti-
bodies in drug-targeting studies
Substance Origin
Small-molecule toxins and antimetabolites:
Methotrexate (MTX) Synthetic
Cytosin-1 b-D-arabinoside (ARA-C) Synthetic
Daunomycin (DNM) Streptomcyces sp.
Mitomycin C (MMC) Streptomyces sp.
Protein toxins (plant-derived):
Ricin A chain Ricinus communis
Abrin Abrus precatorius
Gelonin Gelonium multiflorum
Momordin Momordica charantia
Pokeweed Antiviral Protein (PAP) Pikeweed
Protein toxin (bacterial-derived):
Diphteria Toxin Cornyebacterium diphtheriac
Pseudomonas Exotoxin Pseudomonas
Radiomimetic drugs
Bleomycin (BLM) Streptomyces verticillus
Neocarcinostratin (NCS) Streptomyces carcinostaticus
Calicheamicin Micromonospora echinospora
Esperamicin Actinomadura verrucosospora
place. A number of approaches have been
used based on heterobifunctional coupling
agents; among these, an acid-labile hydra-
zone chemistry-based linker has shown
the most promise. This principle repre-
sents the basis for Mylotarg
, a first-in-
class conjugate of calicheamicin and a
anti-CD33 antibody for the treatment of
AML [118].
First-generation immunotoxins were hy-
brid molecules of bacterial toxins cova-
lently linked to antibodies. In a number of
studies, these turned out to be insufficient,
with dose-limiting toxicities such as vascu-
lar leak syndrome, thrombocytopenia, and
organ damage. Modern immunotoxins are
recombinant fusion products of, for exam-
ple, the variable domain (Fv) of a MAb,
and a truncated bacterial toxin [121]. Re-
combinant immunotoxins are currently
being developed to target a number of
known hematologic and solid tumor anti-
gens [122]. Various new strategies have
been investigated, such as attacking tumor
vascularization with immunotoxins [123].
Studies with radioactive isotopes have
been performed to facilitate a therapeutic ef-
fect a strategy called radioimmunotherapy
[124]. Radiolabels include b-emitters such
as
131
I,
47
Sc, and
90
Y [125] (see Part V, Chap-
ters 4 and 5). When compared to immuno-
scintigraphy, the applied doses are higher in
order to establish a cytotoxic effect. Even
cells within the tumor that are inaccessible
to the antibody can be killed without prior
internalization (the bystander effect).
The number of additional approaches
using MAbs in experimental cancer thera-
py is very high, including bispecific anti-
bodies, pro-drugs, and different forms of
fragments and fusion proteins. In the pro-
drug approach, a non-toxic precursor mol-
Fig. 1.8 (a) Structural DNA-inactivating motives in radiomi-
metic drugs; (b) Bergmann reaction of dehydrobenzenes
[119, 120].
ecule is administered systemically and is
converted into a toxic principle with the
use of a pre-targeting antibody that is lo-
calized at the tumor site [126].
1.8
Developing a Manufacturing Process
for MAbs
Biomanufacturing is a risky and fixed-cost-
driven business, with the COG following a
clear relationship with annual production
demand. For a MAb with an yearly materi-
al requirement of approximately 100 kg,
that results in an overall cost range in the
area of 100 US$ per gram. Approximately
75% of this is related to fixed costs, and
the remainder accounts for variable costs
derived from process consumables [127]
Process development follows an overall
development plan that coordinates all the
various disciplines involved. It is typically
performed in three cycles to allocate re-
sources and thus manage product pipelines
efficiently, and involves a scale-up of up to
1000-fold and more [128]. During the ex-
ploratory phase, a research method is used
to generate small amounts of the respective
MAbs to investigate target-specific effects in
vitro. Important technical goals at this stage
are feasibility and speed.
For the manufacture of clinical material,
an IND-enabling method must be estab-
lished under current Good Manufacturing
Practice (cGMP) with validation of some
critical elements. The whole process
should be designed for robustness and
scalability very early on.
A Biologics License Application (BLA)-
enabling process is typically used for the
manufacture of clinical Phase III material
and the full-scale market supply after
launch out of the final plant. This process
contains the full validation package for the
Chemistry Manufacturing and Controls
(CMC) part of the application dossier, in-
cluding a virus safety concept (see Part I,
Chapter 6). With transfer into the process
scale, there is a shift of emphasis towards
quality and process economy, and this re-
sults in an integrated, robust, cGMP-com-
pliant manufacturing process. For macro-
molecules such as MAbs and recombinant
proteins, reliable manufacturing including
rational acceptance criteria is extremely
important. Since the products are large
and complex molecules with isoforms and
microheterogeneities, the production pro-
cess itself must be designed in a way that
it meets the highest requirements with re-
gard to consistency and reproducibility re-
quested by regulatory authorities (see Part
VII, Chapter 4).
In modern process development, process
trains are compiled with generic modules
that are predeveloped at scale rather than
de-novo design for every new antibody.
State-of-the-art tools such as design of ex-
periments (DOE), process modeling and
simulation are employed to adapt and opti-
mize the individual steps, and to identify
potential bottlenecks in the overall process
which is vital for the best engineering so-
lution. Process development is, however, a
highly complex field that involves a num-
ber of different disciplines. Clear tasks and
project structures are necessary to manage
the various interfaces, including a ratio-
nale change control system towards the
end of the development cycle. Given the
complex nature of an antibody product,
most process changes within an author-
ized manufacturing process are major
ones leading to biochemical comparability
and clinical bridging studies, if not to a
new license application.
1.9
Routine Manufacture of MAbs
Monoclonal antibodies as biopharmaceuti-
cals, and their manufacturing processes,
are subject to the same basic requirements
as any other drug, and the assessment of a
licensing application focuses on the quali-
ty, safety, efficacy, and environmental risk
of the product. Process-related questions
such as the characterization of raw materi-
als, in-process controls, change policies as
well as process and test validation, are very
demanding [129, 130]. As with other bio-
logical products, antibodies have consider-
ably larger sizes and complexities when
compared to chemicals, and an analysis of
the finished product is not sufficient to
control their quality and safety. A suitable
process management, in-process control of
all critical parameters identified within
process validation are decisive factors (see
Part VII, Chapter 1).
The routine production of MAbs under
cGMP became feasible with the imple-
mentation of cell culture and downstream
processing in industrial biotechnology.
From the economic point of view, proce-
dures for the manufacture of MAbs must
be scalable and efficient and, at the same
time, simple and robust.
Traditional methods such as the genera-
tion of mouse ascites are no longer ac-
cepted for quality reasons, and nowadays
are banned in most countries. Antibody-se-
creting mammalian cells are cultivated in
bioreactors under optimized conditions,
with various parameters monitored on a
continuous basis. The expression of re-
combinant MAbs in Chinese hamster
Fig. 1.9 Development of MAb expression rate in mammalian cell culture. (Data adapted from Ref. [131].)
ovary (CHO) cells is the industry standard
[131] (see Part IV, Chapter 1). Other com-
monly used immortalized cell lines are the
mouse myeloma-derived NS0, baby ham-
ster kidney cells (BHK), and PerC6 derived
from human kidney [132] (see Part IV,
Chapters 1, 2, 3, 4, and 12). Fed-batch fer-
mentation with expression rates in the
gram MAb per L range is today the bench-
mark, and much higher levels are in sight
(Fig. 1.9) [133]. Compared to the original
processes, this represents a 1000-fold im-
provement in volumetric productivity, and
in less time. Recent developments in cell
culture have led to both higher cell densi-
ties all well as higher specific production
rates. Despite the enormous advances that
have been made in this field, there is still
room for improvement, and developments
towards higher biomass and more efficient
gene transfer using the whole molecular
genetic repertoire have only just begun
[134].
In the downstream arena, process trains
with the orthogonal combination of prede-
veloped generic modules are state of the
art. Basic principles such as the Captur-
ing-Intermediate Purification-Polishing
(CIPP) strategy, as well as overall process
design issues, are key in order to optimize
quality, yield, and overall productivity
[135]. From high dilution to high purity is
the principle of a good downstream pro-
cess, with the priority to concentrate and
thus stabilize the product very early on
and to remove critical impurities through-
out the process. In this way, the focus is
on selectivity and resolution towards the fi-
nal steps of bulk manufacturing (drug
substance). Although antibodies even from
the same class and subclass are consider-
ably different, the basic elements of the
purification strategy have a generic charac-
ter and allow for a rationale approach in
process development (Fig. 1.10). Typically,
every process step addresses a certain task
in the overall strategy, and redundancy is
avoided wherever possible.
Even closely related antibody molecules
vary in their solubility and their stability to
chemical and physical influences. Isoelec-
tric points are typically between 5 and 9,
mainly due to the different sialic acid con-
tents. Purification techniques involved are
not considerably different from the biose-
paration of other proteins as they focus
mainly on bioaffinity, charge, hydrophobi-
city and size (for an extensive review, see
Ref. [136]). In modern bioseparation pro-
cesses, initial recovery (capturing) and pol-
ishing are the most critical phases as they
address the key features of biopharmaceuti-
cals. A high dilution of the target molecule
after biosynthesis is directly translating into
higher cost of goods as compared to chem-
ical drugs. In modern bioseparation pro-
cesses, this is addressed by a rapid and effi-
cient isolation of the product in a robust
and productive capturing step to facilitate
both rapid volume reduction as well as sep-
aration, and therefore stabilization of the
antibody. High throughput and high dy-
namic capacity demands are driving the ap-
Table 1.5 Physical properties of monoclonal antibod-
ies compared to typical process-derived contami-
nants
Protein Molecular weight Isoelectric point
Monoclonal
antibody
150000 7.78.3
BSA 65000 4.54.8
Transferrin 76000 5.8
Insulin 5700 5.3
Protein A 42000 5.1
DNA 10001000000 Highly negative
Pyrogens 1000001000000 Highly negative
Viruses <7 for most
strains
BSA: Bovine serum albumin.
plications towards innovative high-end
technology on the one hand, but also to-
wards revisiting robust technology of the
low-end type on the other hand.
In a typical procedure for the manufac-
ture of a MAb, the first step is batch fer-
mentation of mammalian cells from a
comprehensively characterized Master
Working Cell Bank (MWCB). The cells
are harvested for further processing of the
cell-free feed stream. The UF-TCF is ap-
plied to a capture step. From a proces-
sing standpoint, the harvest is a highly di-
luted feed stream that causes many han-
dling issues as long as its volume is not
reduced by a productive, high-throughput
technology. Modern capturing supports
thus combine some selectivity for initial
purification with excellent flow properties
(low back-pressure at high linear flow
rates), and can be used for feed streams of
up to 500 column volumes (cv) and with
linear velocities far above 1000 cm h
1
.
Concepts in the initial recovery of MAbs
are numerous, but none of these has yet
provided a major breakthrough.
Early on or as part of the intermediate
purification, an affinity chromatography
step is always the workhorse that provides
very high selectivity for the target molecule
[137, 138].
The benchmark for the purification of
most IgG species is Protein A, a 42 kDa mo-
lecular weight protein derived from a strain
of Staphylococcus aureus. The natural mole-
cule is located on the outer membrane sur-
face of Staphylococcus aureus, and is linked
to the cell surface through its C-terminal re-
gion [139]. This allows the pathogen to bind
the Fc part of the IgG and re-direct the anti-
gen-binding domain. This prevents interac-
tion of the Fc part of the antibody with effec-
tor cells, thereby circumventing the activa-
tion of the immune system.
Fig. 1.10 A generic manufacturing strategy for monoclonal antibodies.
Protein A consists of a single polypep-
tide chain which is structurally and func-
tionally composed of two parts. The N-ter-
minal region contains four homogeneous
subregions, each of which can bind hu-
man IgG molecules, with a total of two ac-
tive binding sites at a time (approx.
K
a
=10
8
M
1
) [140].
Today, however, a recombinant form of
Protein A which is a genetically truncated
version with a molecular mass of ca.
32 kDa has been developed. Non-essential
regions were removed from the C-termi-
nus, resulting in a protein containing 301
amino acids, 28 of which are lysines. It ex-
hibits the same affinity for IgG molecules
as the native protein, but has considerably
lower non-specific binding properties.
Originally, the main step in the isolation
process of natural Protein A was an affini-
ty purification on immobilized virus-inacti-
vated human IgG. With recombinant Pro-
tein A, this is no longer required, and has
led to an increased virus safety of the
product as compared to the conventional
material.
Immobilized Protein A has been used
extensively as an affinity support for the
purification of a wide variety of IgG mole-
cules from many different species of
mammals [141].
For chromatographic use as an affinity
ligand, Protein A is chemically immobi-
lized to a base matrix through NH
2
- or
SH-groups of the protein [142]. The pro-
tein is genetically engineered, allowing for
a site-directed immobilization through in-
troduced, N-terminal bridging groups.
Protein A as a biological reagent is still
raising concerns as to the high cost of the
matrix, the problem of leakage into the
product, the cleanability after repeated use,
and especially the sensitivity to caustic so-
lutions that are widely used as bacterio-
static agents [143]. Most recent variants of
protein A supports focus on the long-term
stability in caustic cleaning agents such as
0.1 M NaOH and increased dynamic ca-
pacity [144, 145].
A number of other immunoglobulin-
binding proteins have also been identified
and characterized, including Protein G
from Streptococcus pyogenes and Protein L
from Peptostreptococcus magnus, and the ge-
netically engineered Protein Z which offer
different specificities and can be used for
other antibody classes and also fragments
[146148]. However, protein-affinity sor-
bents form the basis of a significant part
of the variable costs in manufacturing,
and a number of strategies have been eval-
uated to identify chemical pseudoaffinity
ligands from which MAbs can be bound
and eluted selectively [149151]. Various
biomimetics directed to the Protein A
binding domain in the C
H
2-domain based
on peptide libraries, combinatorial chemis-
try and rational ligand design have been
synthesized and studied for MAb purifica-
tion, and also to increase their visibility
[152].
For the polishing step, it is vital to
achieve final purification. The biomanufac-
turing environment provides an excellent
Table 1.6 Typical release specifications for monoclo-
nal antibodies [28]
Host cell and media proteins (HCP)
ppm-level
Process-related impurities (e.g., Protein A)
ppm-level
Nucleic acids
10100 pg/dose
Endotoxin (LPS)
5.0 EU kg
1
body weight h
1
Microorganisms
absence of detectable bacteria, fungi, yeast and
mycoplasma
basis for the growth of all kinds of organ-
isms and their metabolic products such as
viruses, DNA, host cell proteins and endo-
toxins, as well as process-derived contami-
nants and impurities. In addition, the
products are complex macromolecules
with isoforms and microheterogeneities
that require state-of-the-art production pro-
cesses to meet consistency and reproduc-
ibility demands. Reliable polishing tools
must provide a generic platform type of
tool that works flexibly, reliably, and ac-
cording to highest safety standards.
Polishing can be realized with high reso-
lution; for example, in the removal of iso-
forms and other microheterogeneic vari-
ants, or through selective removal for host
cell-derived or process-related contami-
nants such as host cell proteins (HCP),
DNA, RNA, endogenous viruses, adventi-
tious viruses, media components, endotox-
ins, Protein A, microorganisms, prions,
and other human pathogenic agents [153].
Most contaminants have an acidic PI, and
can be removed through either binding
the antibody on a cation-exchange (CEX)
support and further downstream in a flow
through mode of an anion-exchange (AEX)
chromatography at very high linear veloci-
ties (Fig. 1.11) [154].
As chromatographic supports are diffu-
sion-limited, and flow-through columns
must be designed based on the flow rate
rather then on binding capacity, membrane
adsorbers with a more open structure and
virtually no diffusion limitation provide a
robust alternative to remove a whole set of
contaminants in one step. In this respect,
a number of concepts have been described
(Fig. 1.11) [155]. The maximum separation
power with membrane chromatography is
achieved at high dilution of the target mol-
ecules and thus in capturing and polishing
[156]. In addition, these membrane-based
tools can be discarded after a single use; this
makes them of great interest from a clean-
ability, as well as from a process economy,
standpoint [157].
For final product analysis, evidence
must be provided with validated bioanalyti-
cal quality control methods that, besides
correct identity and homogeneity, critical
impurities have been reduced below speci-
fied limits [158, 159] (see Part I, Chapter 6
and Part VII, Chapter 1). For a validated,
product-related HCP assay, mock fermen-
Fig. 1.11 Structural features of membrane adsorbers versus resin-based media. (From Ref. [154].)
tation and purification runs are performed
to obtain an antigen for the production of
an antiserum with a sensitivity in the 10
100 p.p.m. range [160].
Monoclonal antibodies derived from
mammalian cell culture require a reliable
virus safety concept that comprises three
levels:
Characterization of host cell line and
raw material sources.
Testing of product from various manu-
facturing stages.
Validation of the virus clearance capabil-
ity of the production process.
Since most immortalized cell lines have
been shown to express endogenous retro-
virus-like particles that may or may not be
replication-competent and infectious, and
other virus classes may also be present in
mammalian cells, the risk for a virus con-
tamination of the end product is inherent.
In addition, adventitious viruses may come
into contact with the product during proces-
sing. In any case, cell culture processes must
demonstrate the robust and reliable ability
to eliminate viruses in a risk-based approach
[162] (see Part IV, Chapter 1). Virus valida-
tion studies employ model viruses that are
relevant in representing known risks from
the sources involved [163, 164].
For the clearance of enveloped and non-
enveloped viruses, todays requirements
ask for an orthogonal combination of
methods that are based on the different
physical principles of removal and inacti-
vation, and are complementary to each
other [165]. Virus filtration and solvent/de-
tergent treatment are state of the art for
removal and inactivation [166, 167]. Parti-
tioning steps are considered less robust, as
they are somewhat influenced by the ac-
tual process conditions. In any case,
scaled-down models must be designed
carefully to represent the process condi-
tions; moreover, aging of the support dur-
ing column lifetime must be considered
appropriately [168].
Among the different methods for virus
inactivation, high-temperature short-time
(HTST) has been used for retrovirus clear-
ance, whereas ultra-violet irradiation is
most powerful for the elimination of
small, non-enveloped and otherwise very
resistant viruses such as porcine parvo-
virus (PPV) [169171].
1.10
Glycosylation and Other Post-translational
Modifications
Immunoglobulins are glycoproteins that
contain 312% carbohydrates [172]. In an
IgG molecule, the sugar part is N-linked
to a highly conserved site at Asn297 in the
C
H
2 domain of both heavy chains [173].
The complex carbohydrate has a biantenn-
ary structure with a pentasaccharide core
and variable sugar residues (fucose, man-
nose, GlcNAc, sialic acid) [174] (see Part
IV, Chapters 2 and 7). Although N-linked
glycosylation does not interfere with anti-
gen recognition, a number of implications
are linked to this functionality such as sta-
bility, pharmacokinetics, antigenicity, Fc-re-
lated effector functions, and serum stabili-
ty of antibodies [175]. For example, the re-
moval of sialic acid variants results in a
drastically reduced half-life and increased
liver uptake through the asialoglycoprotein
receptor which is responsible for the recy-
cling of mature glycoconjugates. High-
mannose variants lead to rapid clearance
in vivo.
Different post-translational modification
may lead to the synthesis of different gly-
cosylation variants within a cell clone, and
even within a single cell. The glycosylation
pattern of a MAb is therefore a sensitive
tool to demonstrate the effective control of
critical parameters within a manufacturing
process, and is also a central element in
biochemical comparability studies during
scale-up from laboratory scale to 10000-L
fermentation runs (Fig. 1.12) [176, 177].
This is further driven by modern protein
analytical tools, as well as the requirement
for multi-site production of large amounts
of MAbs to supply the growing markets.
Glycosylation as a complex biosynthetic
event is largely influenced by cell culture
conditions such as the cell bank, the fer-
mentation process with its various control
parameters, and the medium composition
[178]. Other intrinsic differences arise
from the endogenous properties of the ex-
pression system related to the type of gly-
coenzymes and sialic acids they are using.
Even within a validated manufacturing
process, microheterogeneities in the glyco-
sylation pattern with truncated forms and
different overall monosaccharide composi-
tions must be accepted, as long as prede-
termined acceptance criteria with specified
ranges are met. It is vital that these speci-
fications are valid throughout the clinical
development and remain unchanged after
the market launch of a biopharmaceutical.
This requirement usually excludes major
changes in the manufacturing process,
such as a different expression system or
cultivation method, unless new preclinical
Fig. 1.12 Glycosylation pattern of an human IgG-
antibody. (AC) The Asn297 N-linked oligosac-
charide is a biantennary core complex (GlcNac2-
Man3GlNac) with variations such as a third bi-
secting GlNac, different numbers of terminal ga-
lactose residues, and attachment of a terminal
sialic acid that influence the effector functions
and pharmacological properties of the antibody
[179]. Plant core glycosylation is shown for com-
parison (D).
and clinical studies are performed which
basically results in a new product develop-
ment.
Whether a full-length antibody (includ-
ing the full carbohydrate functionality) is
required for the biological function of an
antibody should eventually determine the
decision for the appropriate expression
and manufacturing system. In its aglycosy-
lated form, an antibody can support all an-
tagonistic functions that are required,
without the need for a human-like glycosy-
lation. Mutants can easily by expressed in
a number of host cells, without the need
for additional post-translational modifica-
tions. If, however, the carbohydrate func-
tionality is required, hybridoma or myelo-
ma expression systems (for genetically en-
gineered antibodies) are the expression
systems of choice for full-length antibodies
[180]. Most of these cell lines are of rodent
origin, and this accounts for various host
cell-related differences in the glycan pat-
tern, with the use of different forms of sia-
lic acid being the predominant factor. To-
day, CHO cells represent the industry stan-
dard for the production of therapeutic pro-
teins, as this cell line is able to perform
most glycosylation steps comparable to the
human profile. Recent progress in fermen-
tation development has reported the possi-
bility of routine antibody production in the
35 g L
1
range, with further room for im-
provement. Increase of cell growth and
yield optimization is interfering with the
overall metabolism, including the induc-
tion of glycosyltransferases and oligosac-
charide formation. Even the use of human
cell lines cannot guarantee a reproducible
pattern because of the high influence of
the manufacturing process parameters.
One aspect of modern cell line develop-
ment focuses on the glycosylation pattern,
as overexpressed glycoprotein tend to con-
tain truncated oligosaccharide chains due
to incomplete post-translational modifica-
tion [181]. Cell lines with overexpressed
glycosyltransferases have the potential to
generate a greater homogeneity through
the reduction of terminal GlcNAc and in-
crease of sialylation [182].
1.11
Emerging Issues in MAb Production
Where do we go from here in the manu-
facture of MAbs, and what are the major
trends? Full pipelines further driven by
genomic research and still unmet medical
needs, limited GMP manufacturing capaci-
ty, growing competition between compa-
nies and products, economic problems of
the healthcare systems, higher quality de-
mands: these are the driving forces to-
wards higher efficiency and productivity,
and this also holds true for MAbs [183,
184]. As the biotech industry is maturing
and facing a significant consolidation, effi-
cient development and use of technology
is an important factor in both upstream
and downstream processing. Furthermore,
the complexity of biotech products is re-
flected in the rapid progress and growing
complexity of analytical development that
allows for comparability studies between
different production versions, scales and
manufacturing sites, and this will even-
tually build the basis for biogeneric anti-
bodies. Product development and revenue
generation is a major driver for the growth
of biopharmaceutical companies. Low cost
of goods, higher expression rates, opti-
mized product yields and robust, scalable
manufacturing operations without compro-
mising product quality are the key aspects.
The current situation in biomanufactur-
ing is characterized by the fact that fer-
mentation development is setting the pace
in terms of productivity a fact that will
be further accentuated with the use of
transgenic organisms. Innovative down-
stream processing technology that has the
potential to accommodate these improve-
ments is desperately needed, and trends
are in sight to tackle the current backlog
in bioseparation. Since process economics
matters more and more, it can no longer
be ignored that downstream processing
costs account for up to 80% of the overall
production costs [185].
A number of drivers force the down-
stream technology development towards in-
novative high-end applications on the one
hand, but also towards revisiting robust
technology of the low-end type that works
well with small molecules and commodity
proteins on the other hand. Examples are
precipitation, crystallization, and filtration
to name but a few. A robust downstream
process must be simple, with quality, yield,
productivity, and overall economics as the
primary goals and an integrated design with
compatible unit operations.
In addition, whole-process design with
the integration of upstream and down-
stream processing is becoming the pre-
ferred scenario. The adoption of simple
generic platform technology modules
rather than the de-novo design of enabling
processes provides the basis for economi-
cally viable biomanufacturing, and will
eventually be vital for the success of MAbs
and recombinant proteins.
At present, the sector is lacking the in-
frastructure for the manufacture of anti-
bodies to come. Despite major investments
however, the high demands and long lead
times prohibit a short-term solution other
than a significant increase in overall pro-
ductivity to close the current gaps in bio-
pharmaceutical supply chains.
Alternative expression systems for anti-
bodies and fragments may provide the ba-
sis for high productivity and very large-
scale applications [186]. For antibody frag-
ments, microbial expression in Escherichia
coli, Pichia pastoris and Hansenula polymor-
pha is state of the art [187190]. Trans-
genic animals and plants combine the ad-
vantage of full-length expression of MAbs
with very high biomass generation, and
thus a cell density which is two orders of
magnitude higher than that in any bior-
eactor [191, 192]. Some of these concepts
were introduced years ago, but in particu-
lar animals suffer from certain unsolved
issues such as pathogen threat and com-
plex matrices that are difficult to process
[193]. As a result, no biopharmaceutical
obtained from a transgenic source is yet
available commercially [194], though this
will change as soon as ATryn
is approved
Species-specific glycosylation is another
problem with transgenic production of hu-
man MAbs where Fc-related properties are
required. Plant-derived glycoproteins con-
tain beta-linked xylose sugars that may
have some immunogenic potential [195].
With growing volumetric demands for an-
tagonistic antibodies and economic con-
cerns, plant-based expression is certainly
an option, as discussed in detail in a re-
cent overviews [196, 197]. Full-length ex-
pression of aglycosylated MAbs in non-ed-
ible plants such as alfalfa, tobacco, lemna,
safflower, and moss will therefore become
a vital alternative to other production sys-
tems [198, 199]. Industrial-scale culture of
plant cells is another option [200]. Very
large-scale applications for MAbs with an-
nual productions of more than 1000 kg are
in sight for the chronic treatment of auto-
immune diseases, and also for topical and
oral administration as vaccines and anti-in-
fective agents. Uses are also likely outside
the medical field, in industrial areas such
as catalytic and immunopurification re-
agents [201]. It is unlikely that the mam-
malian cell culture which serves as the
standard technology of today is capable of
overcoming some of the inherent limita-
tions such as the scale-up, long culture
periods, high capital costs and sterility re-
quirements of these very large-scale appli-
cations.
Another remarkable paradigm shift in
biomanufacturing has been a clear trend
towards disposable manufacturing [202].
Change-over procedures and cleaning vali-
dation is an area of concern for auditors,
and it is not surprising that approximately
40% of all 483 citations fall into this cate-
gory (see Part VII, Chapter 4). Several ad-
ditional issues are also apparent, with reu-
sable equipment such as upfront hardware
investment on the basis of preliminary
data, inflexibility in the use of space, han-
dling, downtime during validation and
change-over, and utility consumption with
WFI being the greatest cost drivers [203].
The use of disposables such as bags (see
Part IV, Chapter 14) instead of tanks, plas-
tic tubing instead of hard piping, and
membrane adsorber capsules instead of
chromatography hardware, is not the only
answer to todays downstream problems,
but it may be a life-saving step for compa-
nies that cannot afford to lock-up their re-
sources into long-term hardware invest-
ments.
1.12
The Future of MAbs
What the future will hold for MAbs is
open to speculation. At present, MAbs are
under investigation in a variety of tumor
forms, allergic disorders, and in diseases
characterized by chronic inflammation. In
addition, a number of strategies are focus-
ing on the prevention of infectious dis-
eases.
Current limitations in the therapeutic
use of MAbs are derived from the fact that
they are pathogen-specific and that, for the
time being, they require systemic, parente-
ral, and sometimes chronic application. In
addition, antagonistic principles require
rather large doses in excess of 1 mg kg
1
body weight and this results in treatment
costs of up to US$ 10000 per patient and
therapy. For example, Avastin
treatment
is priced at US$ 4400 per month.
Although efficacy data relating to bio-
pharmaceuticals are typically fewer in
number than for small molecules, MAbs
address chronic and life-threatening dis-
eases where the treatment standard is of-
ten poor, and even small benefits can be
of high added value to the patient.
A number of areas are beginning to con-
tribute to the overall success story however,
among which are new strategies emerging
from pharmacogenomics that allow for the
redefinition of disease targets and a focus
on subsets of patients for optimized benefit.
Further progress in the treatment of multi-
factorial diseases will be possible through
the consideration of individual genetic pro-
files in the design of clinical studies. To-
gether with diagnostic tests to screen for
an overexpressed target receptor, rational
therapies can be directed towards patient
populations, and with a high response rate
(see Part I, Chapters 2 and 5).
The engineering of antibody affinity of-
fers another approach to reduce the amount
of MAb administered systemically to gener-
ate a therapeutic effect (see Part V, Chapter
2). Complement activation, Fc-receptor
binding, and antigen recognition are the ba-
sic functions to be optimized. A number of
affinity maturation methods have been de-
scribed to generate MAbs with sub-nano-
molar affinities to optimize their neutraliz-
ing activity or their biological function,
although a direct correlation does not exist
in any case [204, 205]. Cytotoxic antitumor
MAbs such as Herceptin
(see Part I, Chap-

ter 5) and Rituxan
rely on a complex inter-

action with different Fcc receptor subtypes
[206]. Selective improvement of these effec-
tor functions involves the engineering of
the peptide backbone in the Fc-part, as well
as the glycosylation pattern of the Fc region
[207, 208]. A number of experimental stud-
ies have been conducted to influence com-
plement binding and ADCC through specif-
ic engineering of the carbohydrate moiety
in order to investigate the efficacy of thera-
peutic antibodies [209211].
In future, many antibody-derived mole-
cules will be seen in many indications, in
addition to single chain Fv fragments
(scFv). Dimeric and tetrameric mini-anti-
bodies, diabodies, single antibody do-
mains, Fv fragments fused to toxins, cyto-
kines, antibiotics or radionuclides, protein
scaffolds with hypervariable antibody do-
mains, and related molecules such as an-
ticalines will emerge to complement the
toolbox of affinity-based agents and en-
hance their therapeutic efficacy.
Moreover, development towards applica-
tion routes, formulation and dosing offer
additional therapeutic potential, and will
also improve tolerability and decrease
treatment costs.
With further maturation of the antibody
sector, the areas of indication will increas-
ingly involve non life-threatening disor-
ders. In particular, disease areas with a
high level of unmet medical needs and the
potential for billion dollar markets are at-
tractive for second- and third-generation
products. On the basis of robust large-
scale manufacturing processes, MAbs can
be produced in bulk quantities, and have
the potential eventually to be used as ther-
apeutic agents in commodities such as
toothpastes and shampoos [212]. In overall
terms, antibodies have a bright future and
are the mainstay of modern biopharma-
ceuticals.
References
1 Jerne, K. J., The Immune System. Sci. Am.
1973, 229, 5260.
2 Von Behring E, Kitasato S, Deutsche Med. Wo-
chenschr. 1890, 16, 1113.
3 Edelmann, G. M, Antibody Structure and Mo-
lecular Immunology. Science 1973, 180, 830
834.
4 Roitt, L., Rabson, A., Essential immunology.
Blackwell Publishing, UK, 2000.
5 Tonegawa, S, The molecules of the immune
system. Sci. Am. 1985, 253, 104116.
6 Padlan, E. A, Anatomy of the antibody mole-
cule. Mol. Immunol. 1994, 31, 169217.
7 Hulett, M. D, Hogarth, P. M, Molecular basis
of Fc receptor function. Adv. Immunol. 1994,
57, 1127.
8 Radaev, S., Motyka, S., Fridmann, W. H.,
Sautes-Fridmann, C., Sun, P. D., The structure
of a human type III Fcgamma receptor in
complex with Fc. J. Biol. Chem. 2001, 276,
1646916477.
9 Jefferis, R., Lund, J., Glycosylation of antibody
molecules. Structure and functional signifi-
cance. Chem. Immunol. 1997, 65, 111128.
10 Suralia, A., Interaction of Protein A with the
domains of the Fc-Fragment. Trends Biochem.
Sci. 1982, 7, 318323.
11 Tonegawa, S., Somatic generation of antibody
diversity. Nature 1983, 302, 573.
12 Ballow, M., Berger, M., Bonilla, F. A., Buckley,
R. H., Cunningham-Rundles, C. H., Fireman,
P., Kaliner, M., Ochs, H. D., Skoda-Smith,
S., Sweetser, M. T., Taki, H., Lathia, C., Phar-
macokinetics and tolerability of a new intrave-
nous immunoglobulin preparation, IGIV-C,
10%, Gamunex, 10%. Vox Sanguinis 2003, 3,
202210.
13 Bayry, J., Misra, N., Latry, V., Prost, F., De-
lignat, S., Lacroix-Desmazes, S., Kazatchkine,
M. D., Kaveri, S. V., Mechanisms of action of
intravenous immunoglobulin in autoimmune
and inflammatory diseases. Transfusion Clini-
que et Biologique 2003, 10(3), 165169.
14 Lebing, W., Remington, K. M., Schreiner, C.,
Paul, H.-I., Properties of a new intravenous
immunoglobulin, IGIV-C, 10%, produced by
References 1137
virus inactivation with caprylate and column
chromatography. Vox Sanguinis 2003, 84(3),
193201.
15 Trejo, S. R., Hotta, J. A., Lebing, W., Stenland,
C., Storms, R. E., Lee, D. C., Li, H., Petteway,
S., Remington, K. M., Evaluation of virus and
prion reduction in a new intravenous immu-
noglobulin manufacturing process. Vox San-
guinis 2003, 84(3), 176187.
16 http://www.gamunex.com/web_docs/
IGIV%20Comp%20Table%2Epdf
17 Khler, G., Milstein, C., Continuous cultures
of fused cells secreting antibody of predefined
specificity. Nature 1975, 256, 495497.
18 Shawler, D. L., Bartholomew, R. M., Smith,
L.M., Human immune response to multiple
injections of murine monoclonal IgG. J. Im-
munol. 1985, 135, 112.
19 Ewert, S., Honegger, A., Pluckthun, A., Stabil-
ity improvement of antibodies for extracellular
and intracellular applications, CDR grafting to
stable frameworks and structure-based frame-
work engineering. Methods 2004, 34(2), 184
199.
20 Clark, M., Antibody humanization, A case of
the Emperors new clothes? Immunol. Today
2000, 21(8), 397402.
21 Little, M., Kipriyanov, S. M., Le Gall, F., Mol-
denhauer, G., Of mice and men, hybridoma
and recombinant antibodies. Immunol. Today
2000, 21(8), 364370.
22 Lonberg, N., Taylor, L. D., Harding, F. A.,
Trounstine, M., Higgins, K. M., Schramm,
S. R., Kuo, C. C., Mashayekh, R., Wymore, K.,
McCabe, J. G., Antigen-specific human anti-
bodies from mice comprising four distinct ge-
netic modifications. Nature 1994, 368, 856
859.
23 Green, L. L, Antibody engineering via genetic
engineering of the mouse, XenoMouse strains
are a vehicle for the facile generation of thera-
peutic human monoclonal antibodies. J. Im-
munol. Methods 1999, 231(1-2), 1123.
24 Winter, G., Griffith, A. D., Hawkins, R. E.,
Hoogenboom, H.R., Making antibodies by
phage display technology. Annu. Rev. Immu-
nol. 1994, 12, 433455.
25 Hoogenboom, H. R., Overview of antibody
phage-display technology and its applications.
26 Arndt, K. M., Jung, S., Krebber, C., Pluckthun,
A., Selectively infective phage technology.
Methods Enzymol. 2000, 328, 364388.
27 Ostendorp, R., Frisch, C., Urban, M., Genera-
tion, engineering and production of human
monoclonal antibodies using HuCAL
. In:
Antibodies, Volume 2, Novel Technologies and
Therapeutic Use, G. Subramanian (Ed.),
Kluwer Academic New York, 2004.
28 Feldhaus, M. J., Siegel, R. W., Yeast display of
antibody fragments, a discovery and character-
ization platform. J. Immunol. Methods 2004,
290(1-2), 6980.
29 Schaffitzel, C., Hanes, J., Jermutus, L.,
Pluckthun, A., Ribosome display, an in vitro
method for selection and evolution of antibod-
ies from libraries. J. Immunol. Methods 1999,
231(1-2), 119135.
30 Stockwin, L. H., Holmes, S., The role of thera-
peutic antibodies in drug discovery. Biochem.
Soc. Trans. 2003, 31(2), 433436.
31 Glover, D., Humphreys, D., Antibody Frag-
ments. Production, purification and format-
ting for therapeutic applications. In: Antibod-
ies, Volume 1, Production and Purification, G.
Subramanian (Ed.), Kluwer Academic New
York, 2004.
32 Hudson, P. J., Souriau, C., Engineered anti-
bodies. Nat. Med. 2003, 9(1), 12934.
33 Fersht, A., Winter, G., Protein engineering.
Trends Biochem. Sci. 1992, 17(8), 292295.
34 Weir, A. N., Nesbitt, A., Chapman, A. P., Pop-
plewell, A. G., Antoniw, P., Lawson, A. D., For-
matting antibody fragments to mediate specif-
ic therapeutic functions. Biochem. Soc. Trans.
2002, August 30, 512516
35 Wrn, A., Plckthun, A., Stability engineering
of antibody single-chain Fv fragments. J. Mol.
Biol. 2001, 305, 9891010.
36 Kortt, A. A., Dolezal, O., Power, B. E., Hudson,
P. J., Dimeric and trimeric antibodies, high
avidity scFvs for cancer targeting. Biomol. Eng.
2001, 18, 95108.
37 Zapata, G., Ridgway, J. B. B., Mordenti, J., Osa-
ka, G., Wong, W. L. T., Bennett, G. L., Carter,
P., Engineering linear F(ab')
2
fragments for ef-
ficient production in Escherichia coli and en-
hanced antiproliferative activity. Prot. Eng.
1995, 8, 10571062.
38 Kriangkum, J., Xu, B., Nagata, L. P., Fulton,
R. E., Suresh, M. R., Bispecific and bifunc-
tional single chain recombinant antibodies.
Biomol. Eng. 2001, 18, 3140.
39 Kufer, P., Lutterbse, R., Baeuerle, P. A., A re-
vival of bispecific antibodies. Trends Biotechnol.
2004, 22, 238244.
40 Hamers-Castermann, C., Atarhouch, T., Muyl-
dermans, S., Robinson, G., Hamers, C., Son-
ga, E. B., Bendahman, N., Hamers, R., Natu-
rally occurring antibodies devoid of light
chains. Nature 1993, 363, 446448.
41 Nguyen, V. K., Desmyter, A., Muyldermanns,
D., Functional heavy-chain antibodies in Ca-
melidae. Adv. Immunol. 2001, 79, 261296.
42 Muyldermanns, S., Single domain camel anti-
bodies: current status. J. Biotechnol. 2001, 74,
277302.
43 Holt, L. J., Herring, C., Jespers, L. S., Woolven,
B. P., Tomlinson, I. M., Domain antibodies,
proteins for therapy. Trends Biotechnol. 2003,
21, 484490.
44 Batra, S. K., Jain, M., Wittel, U. A., Chauhan,
S. C., Colcher, D., Pharmacokinetics and bio-
distribution of genetically engineered antibod-
ies. Curr. Opin. Biotechnol. 2002, 13, 603608.
45 Chapman, A. P., Antoniw, P., Spitali, M.,
West, S., Spehens, S., King, D. J., Therapeutic
antibody fragments with prolonged in vivo
half-lives. Nat. Biotechnol. 1999, 17, 780783.
46 Chapman, A. P., PEGylated antibodies and
antibody fragments for improved therapy: a re-
view. Adv. Drug Deliv. Rev. 2002, 54, 531545.
47 Rose-John, S., Schooltink, H., CDP-870. Cell-
tech/Pfizer. Curr. Opin. Investig. Drugs 2003, 4,
588592.
48 Whiss, P. A., PexelizuMAb Alexion. Curr.
Opin. Investig. Drugs 2002, 3, 870877.
49 Verma, R., Boletti, E., George, A. J. T., Anti-
body engineering: comparison of bacterial,
yeast, insect and mammalian expression sys-
tems. J. Immunol. Methods 1998, 216, 165
181.
50 Fischer, R., Drossard, J., Emans, N., Comman-
deur, U., Hellwig, S., Towards molecular farm-
ing in the future: Pichia pastoris-based produc-
tion of single-chain antibody fragments. Bio-
technol. Appl. Biochem. 1999, 30, 117120.
51 Humphreys, D. P., Production of antibodies
and antibody fragments in Escherichia coli and
comparison of their functions, uses and modi-
fications. Curr. Opin. Drug Discov. Dev. 2003,
6, 188196.
52 Borrebaeck, C. A., Antibodies in diagnostics
from immunoassays to protein chips. Immu-
nol. Today 2000, 21, 379382.
53 Holt, L. J., Enever, C., de Wildt, R. M., Tomlin-
son, I.M., The use of recombinant antibodies
in proteomics. Curr. Opin. Biotechnol. 2000,
11, 445449.
54 Wade, H., Scanlan, T. S., The structural and
functional basis of antibody catalysis. Annu.
Rev. Biophys. Biomol. Struct. 1997, 26, 461
493.
55 Groner, B., Hartmann, C., Wels, W., Thera-
peutic antibodies. Curr. Molec. Med. 2004,
4(5), 539547.
56 Jung, C., Mller Eberhard, H. J., An in vitro
model for tumour immunotherapy with anti-
body heteroconjugates. Immunol. Today 1988,
9, 257265.
57 Stahel, R., Pass, M., Zangemeister-Wittke, U.,
Cell surface antigens as therapeutic targets.
Lung Cancer 1997, 18, 818.
58 Proimos, G., Platelet aggregation inhibition
with Glycoprotein IIbIIIa inhibitors. J.
Thrombosis Thrombolysis 2001, 25, 99110.
59 Wenzel, R., Anti-endotoxin monoclonal anti-
bodies a second look. N. Engl. J. Med. 1992,
326, 11511153.
60 Parillo, J. E., Management of septic shock,
present and future. Ann. Intern. Med. 1991,
115, 491493.
61 Selby, P., Hobbs, S., Viner, C., Tumour necro-
sis factor in man, clinical and biological obser-
vations. Br. J. Cancer 1987, 56, 803808.
62 Oh, H. M. L., Emerging therapies for sepsis
and septic shock. Ann. Acad. Med. Singapore
1998, 27, 738743.
63 Shulmann, R., Current drug treatment of sep-
sis. Hospital Pharmacist 2002, 9, 97101.
64 Ohlsoon, K., Bjork, P., Bergenfeldt, M., Hage-
man, R., Thompson, R. C., Interleukin-1 re-
ceptor antagonist reduces mortality from en-
dotoxin shock. Nature 1990, 348, 550552.
65 Fischer, C. J., Chainaut, J. F., Opal, S. M., Balk,
R. A., Slotman, G. J., Iberti, T. J., Recombinant
human interleukin-1 receptor antagonist in
the treatment of patients with sepsis syn-
drome, results from a randomised, double-
blind, placebo-controlled trial. JAMA 1994,
271, 18361843.
66 Ziegler, E. J., Fisher, C. J., Sprung, C. L.,
Straube, R. C., Sadoff, J. C., Foulke, G. E., Wor-
tel, C. H., Fink, M. P., Dellinger, R. P., Teng,
N. N., Treatment of gram-negative bacteremia
and septic shock with HA-1A human mono-
clonal antibody against endotoxin. A random-
ized, double-blind, placebo-controlled trial.
The HA-1A Sepsis Study Group. N. Engl. J.
Med. 1991, 324, 429436.
67 Medicines Control Agency, Drug Alert EL, 93,
A/6. London, MCA, 1993.
References 1139
68 Bone, R. C., Balk, R. A., Fein, A. M., Perl,
T. M., Wenzel, R. P., Reines, H. D., A second
large controlled clinical trial of E5, a monoclo-
nal antibody to endotoxin. Results of a pro-
spective, multicentre, randomised, controlled
trial. Crit. Care Med. 1995, 23, 9941005.
69 Cross, A. S., Antiendotoxin Antibodies A
dead end? Ann. Intern. Med. 1994, 121, 5860.
70 Abraham, E., Wunderink, R., Silvermann, H.,
Perl, T. M., Nasraway, S., Levy, H., Efficacy
and safety of monoclonal antibody to human
tumour necrosis factor in patients with septic
syndrome. JAMA, 1995, 273, 934941.
71 Biologic Battlefields, Changes in the US Mar-
kets for Rheumatoid Arthritis and Psoriatic
Arthritis Pharmaceutical & Diagnostic Innova-
tion, 2004, vol. 1, no. 3, pp. 1620.
72 Dayer, J.-M., The pivotal role of interleukin-1
in the clinical manifestations of rheumatoid
arthritis. Rheumatology 2003, 42 (Suppl. 2),
ii3ii10.
73 Fleischmann, R., Stern, R., Iqbal, I., Anakin-
ra, an inhibitor of IL-1for the treatment of
rheumatoid arthritis. Exp. Opin. Biol. Ther.
2004, 4(8), 13331344.
74 Fleischmann, R. M., Stern, R., Iqbal, I., Con-
siderations with the use of biological therapy
in the treatment of rheumatoid arthritis. Exp.
Opin. Drug Safety 2004, 3, 391397.
75 Nanda, S., Bathon, J. M., Etanercept, a clinical
review of current and emerging indications.
Exp. Opin. Pharmacother. 2004, 5(5), 11751186.
76 Amevive: First to Enter Psoriasis Market
Competitor Products Closing In Quickly. Phar-
maceutical & Diagnostic Innovation 2003, 1(2),
2326.
77 Feldman, S. R., Menter, A., Koo, J. Y., Im-
proved health-related quality of life following
a randomized controlled trial of alefacept
treatment in patients with chronic plaque
psoriasis. Br. J. Dermatol. 2004, 150, 317326.
78 Asche, G. V., Rutgeerts, P., Anti-TNF agents in
Crohns disease. Exp. Opin. Investig. Drugs
2000, 9(1), 103111.
Adair, F., Monoclonal antibodies magic bul-
lets or a shot in the dark? Drug Discovery
World 2002, Summer, 5359.
79 Padlan, E. A., A possible procedure for reducing
the immunogenicity of antibody variable do-
mains while preserving their ligand-binding
properties. Molec. Immunol. 1991, 28, 489498.
80 Grillo-Lpez, A. J., Rituximab. Clinical devel-
opment of the first therapeutic antibody for
cancer. In: Pharmaceutical Biotechnology Drug
Discovery and Clinical Applications. O. Kayser,
R. H. Mller (Eds). Wiley-VCH, 2003.
81 Cartron, G., Dacheux, L., Salles, G., Solal-Ce-
ligny, P., Bardos, P., Colombat, P., Watier, H.,
Therapeutic activity of humanized anti-CD20
monoclonal antibody and polymorphism in
IgG Fc receptor FcgammaRIIIa gene. Blood
2002, 99, 754758.
82 Tobinai, K., RituxiMAb and other emerging
monoclonal antibody therapies for lymphoma.
Exp. Opin. Emerging Drugs 2002, 7(2), 289302.
83 Glade-Bender, J., Kandel, J. J., Yamashiro, D. J.,
VEGF blocking therapy in the treatment of
cancer. Exp. Opin. Biol. Ther. 2003, 3, 263276.
84 Carter, P., Presta, L., Gorman, C.M., Ridgway,
J. B., Henner, D., Wong, W. I., Rowland, A. M.,
Kotts, C., Carver, M. E., Shepard, H. M., Hu-
manization of an anti-p185HER2 antibody for
human cancer therapy. Proc. Natl. Acad. Sci.
USA 1992, 89, 42854289.
85 Carpenter, G., Cohen, S., Epidermal growth
factor. J. Biol. Chem. 1990, 265, 77097712.
86 Yarden, Y., Ullrich, A., Growth factor receptor
tyrosine kinases. Annu. Rev. Biochem. 1988,
57, 443478.
87 Yarden, Y., The EGFR family and its ligands
in human cancer. Signalling mechanisms and
therapeutic opportunities. Eur. J. Cancer 2001,
37, 38.
88 Cardiello, F., Tortora, G., A novel approach in
the treatment of cancer, targeting the epider-
mal growth factor receptor. Clin. Cancer Res.
2001, 7, 29582970.
89 Slamon, D. J., Leyland-Jones, B., Shak, S.,
Fuchs, H., Paton, V., Bajamonde, A., Fleming,
T., Eiermann, W., Wolter, J., Pegram, M.,
Baselga, J., Norton, L., Use of chemotherapy
plus a monoclonal antibody against HER2for
metastatic breast cancer that overexpresses
HER2. N. Engl. J. Med. 2001, 344, 783792.
90 Huang, S. M., Bock, J. M., Harari, P. M., Epi-
dermal growth factor blockade with C225
modulates proliferation, apoptosis, and radio-
sensitivity in squamous cell carcinomas of the
head and neck. Cancer Res. 1999, 59, 1925
1940.
91 White, C. A., Rituxan
immunotherapy and
Zevalin
radioimmunotherapy in the treat-

ment of non-Hodgkins lymphoma. Curr.
Pharm. Biotechnol. 2003, 4, 221238.
92 Robinson, K., An industry comes of age. Bio-
Pharm Int. 2002, 15, 2024.
93 Milroy, D., Auchincloss, C., Monoclonal An-
tibodies on the Crest of a Wave. Horizons
2003, 6.
94 Cohen, M. M., Diamond, J. M., Are we losing
the war on cancer? Nature 1986, 323, 488
492.
95 Sell, S., Cancer Marker, Past, present and fu-
ture. In: Monoclonal Antibodies and Cancer
Therapy. R. A. Reisfeld (Ed.), Alan R. Liss,
New York, 1985.
96 Hakomori, S. I., Tumour associated carbohy-
drate antigens. Annu. Rev. Immunol. 1984, 2,
103111.
97 Lloyed, K. O., Human tumour antigens. Tar-
gets for mouse monoclonal antibodies. In:
Immunoconjugates. C.W. Vogel (Ed.), Oxford
University Press, New York, Oxford, 1987.
98 Stamp, G. W. H., The rational use of mono-
clonal antibodies in tumour diagnosis. Arch.
Geschwulstforsch. 1990, 60, 7178.
99 Mach, J. P., Carrel, S., Forni, J., Tumor local-
ization of radiolabeled antibodies against car-
cinoembryonic antigen in patients with car-
cinoma. N. Engl. J. Med. 1980, 303, 1384
1390.
100 Vogel, C. W., Immunoconjugates. Antibody
conjugates in radioimaging and therapy of can-
cer. Oxford University Press, New York, Ox-
ford, 1987.
101 Souriau, C., Hudson, P. J., Recombinant an-
tibodies for cancer diagnosis and therapy.
Exp. Opin. Biol. Ther. 2003, 3(2), 305318.
102 Goldenberg, D. M., Deland, F. H., History
and status of tumour imaging with radiola-
beled monoclonal antibodies. J. Nucl. Med.
1985, 25, 538549.
103 Ehrlich, P., Chemotherapie. In: The collected
papers of Paul Ehrlich. F. Himmelweit (Ed.),
Pergamon Press, London, 1969.
104 Hudson, P. J., Recombinant antibody con-
structs in cancer therapy. Curr. Opin. Immu-
nol. 1999, 11, 548557.
105 Vogel, C. W., Immunoconjugates, Antibody
conjugates in radioimaging and therapy of can-
cer. Oxford University Press, New York,
1987.
106 Blair, A. H. Ghose, T. I. Linkage of cytotoxic
agents to immunoglobulins. J. Immunol.
Methods 1983, 59, 129155.
107 Garnett, M. C., Targeted drug conjugates,
principles and progress. Adv. Drug Deliv. Rev.
2001, 53, 171216.
108 Olsnes, S., Sandvig, K., How protein toxins
enter and kill cells. In: Immunotoxins. A. E.
Frankel (Ed.), Martinus Nijhoff Publishing,
Boston, 1984, pp. 148.
109 FitzGerald, D. J., Kreitman, R., Wilson, W.,
Squires, D., Pastan, I., Recombinant immu-
notoxins for treating cancer. Int. J. Med. Mi-
crobiol. 2004, 293, 577582.
110 Ng, H. C., Khoo, H. E., Cancer homing tox-
ins. Curr. Pharmaceut. Design 2002, 8, 1973
1985.
111 Meienhofer, J. H., Maeda, H., Gaser, C. B.,
Czomboz, J., Kuromizu, K., Primary struc-
ture of neocarzinostatin, an antitumor pro-
tein. Science 1972, 178, 875878.
112 Poon, R., Beerman, T. A., Goldberg, I.H.,
Characterization of the DNA-strand-breakage
in vitro by the antitumor protein neocarzi-
nostatin. Biochemistry 1977, 16, 486492.
113 Zein, N., Sinha, A. M., McGahren, W. J., El-
lestad, G. A., Calicheamicin gamma1I. An
antitumor antibiotic that cleaves double-
stranded DNA site specifically. Science 1988,
240, 11981192.
114 Maeda, H., Neocarzinostatin in cancer che-
motherapy. Anticancer Res. 1981, 1, 175189.
115 Gottschalk, U., Garnett, M. C., Ward, R. K.,
Maibcher, A., Khnlein, W., Increased se-
rum stability and prolonged biological half-
life of Neocarzinostatin covalently bound to
monoclonal antibodies. J. Antibiotics 1991,
44, 11481154.
116 Nitin, K., Damle, E., Tumour-targeted che-
motherapy with immunoconjugates of cali-
cheamicin. Exp. Opin. Biol. Ther. 2004, 4(9),
14451452.
117 Hamann, P. R., Hinman, L. M., Beyer, C. F.,
Lindh, D., Upeslacis, J., Flowers, D.A., Bern-
stein, I., An anti-CD33 antibody calicheami-
cin conjugate for treatment of acute myeloid
leukaemia. Choice of Linker. Bioconj. Chem.
2002, 13, 4046.
118 Nagata, R., Yamanaka, H., Okazaki, E., Sai-
to, I., Biradical formation from acyclic conju-
gated eneyne-allene system related to Neo-
carzinostatin and Esperamicin-calicheami-
cin. Tetrahedron Lett. 1989, 30, 49954998.
119 Lockhart, T. P., Bergman, R. G., Evidence for
the reactive spin state of 1,4-dehydroben-
zenes. J. Am. Chem. Soc. 1981, 103, 4091
4096.
120 Reiter, Y., Brinkmann, U., Lee, B., Pastan, I.,
Engineering antibody Fv fragments for can-
References 1141
cer detection and therapy, disulfide-stabilized
Fv-fragments. Nat. Biotechnol. 1996, 14,
12391245.
121 Choo, A. B. H., Dunn, R. D., Broady, K. W.,
Raison, R. L., Soluble expression of a func-
tional recombinant cytolytic immunotoxin in
insect cells. Protein Express. Purif. 2002,
24(3), 338334.
122 Yoshioka, Y., Tsutsumi, Y., Nakagawa, S.,
Mayumi, T., Recent progress on tumor mis-
sile therapy and tumor vascular targeting
therapy as a new approach. Curr. Vasc. Phar-
macol. 2004, 12, 259270.
123 Goldenberg, D. M., Targeted therapy of can-
cer with radiolabeled antibodies. J. Nucl.
Med. 2002, 43, 693713.
124 Cobb, L. M., Humm, J. L., Radioimmu-
notherapy of malignancy using antibody tar-
geted radionuclides. Br. J. Cancer 1986, 54,
863868.
125 Xu, G., McLeod, H. L., Strategies for en-
zyme/prodrug cancer therapy. Clin. Cancer
Res. 2001, 7, 33143324.
126 Steiner, U., The Business Case for Plant Fac-
tories. Plant Conference, Quebec, 2003.
127 Rathore, A., Velayudhan, A., Guidelines for
optimisation and scale up in preparative
chromatography. BioPharm 2003, January,
3442.
128 Federici, M. M., The quality control of bio-
technology products. Biologicals 1994, 22,
151159.
129 Office of Biological Research and Review,
US Food and Drug Administration. 1985.
Points to consider in the production and
testing of new drugs and biologicals pro-
duced by recombinant DNA technology.
130 Little, M., Kipriyanov, S. M., Le Gall, F., Mol-
denhauer, G., Of mice and men, hybridoma
and recombinant antibodies. Immunol. Today
2000, 21, 364370.
131 Wurm, F. M., Production of recombinant
protein therapeutics in cultivated mammalian
cells. Nature Biotechnol. 2004, 22, 13931398.
132 Wurm, F. M., Griffith, J., Mammalian Cell
Culture. In: The Encyclopedia of Physical
Science and Technology, 3rd edn. R. A. Meyers
(Ed.), Academic Press, 2002, 9, pp. 3147.
133 Wurm, F. M., Jordan, M., Gene transfer and
gene amplification in mammalian cells. In:
Gene Transfer and Expression in Mammalian
Cells. S. C. Makrides (Ed.), New Comprehen-
sive Biochemistry, 2003, 38, pp. 309335.
134 Grund, E., Advances in Downstream Proces-
sing. Presented at Scaling up of Biophar-
maceutical Proteins, IBC Life Sciences,
Basel, 23rd January 2003.
135 Jungbauer, A., Preparative chromatography
of biomolecules. J. Chromatogr. 1993, 639, 3
16.
136 Wilchek, M., Chaiken, I., An overview of af-
finity chromatography. Methods Molec. Biol.
2000, 147, 16.
137 Lowe, C. R., Lowe, A. R., Gupta, G., New de-
velopments in affinity chromatography with
potential application in the production of
biopharmaceuticals. J. Biochem. Biophys.
Methods 2001, 49, 561574.
138 Sjquist, J., Movitz, J., Johansson, I.-B., Lo-
calization of protein A in Staphylococcus aur-
eus. Eur. J. Biochem. 1972, 30, 190.
139 Suralia, Interaction of Protein A with the do-
mains of the Fc-Fragment. Trends Biochem.
Sci. 1982, 7, 318.
140 Huse, K., Bhme, H. J., Scholz, G. H., Purifi-
cation of antibodies by affinity chromatogra-
phy. J. Biochem. Biophys. Methods 2002, 51,
217231.
141 Mechanism of activation of Sepharose and
Sephadex by cyanogen bromide. Enzyme Mi-
crob. Technol. 1982, 4, 161.
142 Iyer, H., Henderson, F., Cunningham, E.,
Webb, J., Hanson, J., Bork, C., Conley, L.,
Considerations during development of a pro-
tein A-based antibody purification process.
Biopharm 2002, 20, 1420.
143 Mottaqui-Tabar, A., Birath, K., Johanson,
H. J., A base-tolerant resin for Monoclonal
Antibody purification. 3rd International
Symposium on Downstream Processing of
Genetically Engineered Antibodies and Re-
lated Molecules, 2004, Nice, France.
144 Bergander, T., Chirica, L., Ljunglf, A.,
Malmquist, G., Novel high capacity Protein
A affinity chromatography media. 3rd Inter-
national Symposium on Downstream Pro-
cessing of Genetically Engineered Antibod-
ies and Related Molecules, 2004, Nice,
France.
145 Raeder, R., Boyle, M. D., Analysis of immu-
noglobulin G-binding-protein expression by
invasive isolates of Streptococcus pyogenes.
Clin. Diagn. Lab. Immunol. 1995, 2, 484486.
146 Bjorck, L., Protein, L.. A novel bacterial cell
wall protein with affinity for IgL chains. J.
Immunol. 1988, 140, 11941197.
147 Gulich, S., Uhlen, M., Hober, S., Protein en-
gineering of an IgG-binding domain allows
milder conditions during affinity chromato-
graphy. J. Biotechnol. 2000, 76, 233244.
148 Curling. J., Affinity Chromatography from
Textile Dyes to Synthetic Ligands by Design,
Parts I and II. BioPharm Int. 2004, July/Au-
gust.
149 Lowe, C. R., Lowe, A. R., Gupta, G., New de-
velopments in affinity chromatography with
potential application in the production of
biopharmaceuticals. J. Biochem. Biophys.
2001, 49, 561574.
150 Boschetti, E., The use of thiophilic chroma-
tography for antibody purification: a review.
J. Biochem. Biophys. Methods 2001, 49, 361
389.
151 Li, R., Dowd, V., Stewart, D. J., Burton, S. J.,
Lowe, C. R., Design, synthesis, and applica-
tion of a Protein A mimetic. Nature Biotech-
nol. 1998, 16, 190195.
152 Fish, B., Concepts in development of manu-
facturing strategies for monoclonal antibod-
ies. In: Antibodies, Volume 1, Production and
Purification. G. Subramanian (Ed.), Kluwer
Academic New York, 2004.
153 Hanna, L. S., Pine, P., Reuzinsky, G., Nigam,
S., Omstead, D.R., Removing specific cell
culture contaminants in a MAb purification
process. BioPharm 1991, 4, 3337.
154 Gosh, R., Protein separation using mem-
brane chromatography, opportunities and
challenges. J. Chromatogr. 2002, 952, 1327.
155 Gottschalk, U., Fischer-Fruehholz, S., Reif,
O., Membrane adsorbers. A cutting edge
process technology at the threshold. BioPro-
cess Int. 2004, 2, 5665.
156 Warner, T. N., Nochumsnon, S., Rethinking
the economics of chromatography. BioPharm
2003, 16, 5860.
157 Food and Drug Administration, USA, Points
to consider in the manufacture and testing
of monoclonal antibody products for human
use. 1997.
158 Food and Drug Administration, USA, Guid-
ance for industry, Bioanalytical method vali-
dation. 2001.
159 Hoffmann, K., Strategies for host cell pro-
tein analysis. BioPharm 2000, 13, 3845.
160 Richter, A., Jostameling, M., Mller, K.,
Herrmann, A., Pitschke, M., Quality control
of antibodies for human use. In: Antibodies
Production and Purification. G. Subrama-
nian (Ed.), Kluwer Academic, New York,
2004.
161 Committee for Proprietary Medicinal Prod-
ucts (CPMP), ICH Q5a. Note for Guidance
on Quality of Biotechnology Products, Viral
Safety Evaluation of Biotechnology Products
Derived from Cell Lines of Human or Ani-
mal Origin (CPMP) ICH/295/95), 1997.
162 Committee for Proprietary Medicinal Prod-
ucts (CPMP). Note for Guidance on Virus
Validation Studies, The design, contribution
and interpretation of studies validating the
inactivation and removal of viruses. CPMP/
BWP/268/95, 1996.
163 Sofer, G., Virus Inactivation in the 1990s
Part 4, culture media, biotechnology prod-
ucts, and vaccines. BioPharm 2003, 16, 50
57.
164 Walter, J. K., Nothelfer, F., Werz, W., Valida-
tion of viral safety for pharmaceutical pro-
teins. In: Bioseparation and Bioprocessing,
Vol. 1. G. Subramanian (Ed.), Wiley-VCH
Weinheim, 1998, pp. 465496.
165 Levy, R. V., Phillips, M. W., Lutz, H., Filtra-
tion and the removal of viruses from bio-
pharmaceuticals. In: Filtration in the biophar-
maceutical industry. T. Meltzer, M. Jornitz
(Eds.), Marcel Dekker Inc., New York, Basel,
Hong Kong, 1998, pp. 619646.
166 Sofer, G., Sofer, D. C., Boose, J. A., Inactiva-
tion methods grouped by virus. Virus inacti-
vation in the 1990s and into the 21st cen-
tury. BioPharm Int. 2003, Suppl. S37S42.
167 Sofer, G., Validation: Ensuring the accuracy
of scaled-down chromatography models. Bio-
Pharm 1996, October, 5154.
168 Plavsic, A. M., Bolin, S., Resistance of por-
cine circovirus to gamma irradiation. Bio-
Pharm 2001, 14, 3236.
169 Wang, J., Mauser, A., Chao, S.-F., Reming-
ton, K., Treckmann, R., Kaiser, K., Pifat, D.,
Hotta, J., Virus inactivation and protein re-
covery in a novel ultraviolet-C reactor. Vox
Sanguinis 2004, 86, 230238.
170 Reif, O., Mora, J., Gottschalk, U., An inte-
grated concept for reliable removal of viruses
and process-related contaminants in bioma-
nufacturing. IBC BioProcess International
Conference October 47, Boston, USA.
171 Leibiger, H., Hansen, A., Schoenherr, G.,
Seifert, M., Wustner, D., Stigler, R., Marx,
U., Glycosylation analysis of a polyreactive
human monoclonal IgG antibody derived
References 1143
from a human-mouse heterohybridoma.
Mol. Immunol. 1995, 32, 595602.
172 Deng, X. K., Raju, T. S., Morrow, K. J.,
Achieving appropriate glycosylation during
the scale up of antibody production. In: An-
tibodies Novel Technologies and Therapeutic
Use. G. Subramanian (Ed.), Kluwer Aca-
demic, New York, 2004.
173 Wright, A., Morrison, S. L., Effect of glycosy-
lation on antibody function. Implications for
genetic engineering. Trends Biotechnol. 1997,
15, 2632.
174 Jefferis, R., Lund, J., Glycosylation of anti-
body molecules. Structure and functional
significance. Chem. Immunol. 1997, 65, 111
128.
175 Jenkins, N., Parekh, R. B., James, D. C., Get-
ting the glycosylation right, implications for
the biotechnology industry. Nat. Biotechnol.
1996, 14, 975981.
176 Schenermann, M. A., Hope, J. N., Kletke, C.,
Singh, J. K., Kimura, R., Tsao, E. I., Folena-
Wassermann, G., Comparability testing of a
humanized monoclonal antibody, Synagis, to
support cell line stability, process validation,
and scale-up for manufacturing. Biologicals
1999, 27, 203215.
177 Patel, T. P., Parekh, R. B., Moellering, B. J.,
Prior, C. P., Different culture methods lead
to differences in glycosylation of a murine
IgG monoclonal antibody. Biochem. J. 1992,
285, 839845.
178 Siberil, S., Teillaud, J. L., Future prospects in
antibody engineering and therapy. In: Anti-
bodies, Volume 2, Novel Technologies and Ther-
apeutic Use. G. Subramanian (Ed.), Kluwer
Academic, New York, 2004.
179 Yoo, E. M., Koteswara, R., Chintalacharuvu,
M.L., Morrison, S.L., Myeloma expression
systems. J. Immunol. Methods 2002, 261, 120.
180 Stanley, P., Glycosylation engineering. Glyco-
biology 1992, 2, 99107.
181 Weikert, S., Papac, D., Briggs, J., Cowfer, D.,
Tom, S., Gawlitzek, M., Lowgren, J., Metha,
S., Chisholm, V., Modi, N., Eppler, S., Car-
roll, K., Chamow, S., Peers, D., Berman, P.,
Engineering Chinese hamster ovary cells to
maximize sialic acid content of recombinant
glycoproteins. Nat. Biotechnol. 1993, 17,
11161121.
182 Sinclair, A., Biomanufacturing capacity, Cur-
rent and future requirements. J. Commercial
Biotechnol. 2001, 8, 4350.
183 Gottschalk, U., Biotech manufacturing is
coming of age. BioProcess Int. 2004, 4, 5461.
184 Sadana, A., Beelaram, A., Efficiency and eco-
nomics of bioseparation, some case studies.
Bioseparation 1994, 4, 221235.
185 Verma, R., Boleti, E., George, A. J., Antibody
engineering, comparison of bacterial, yeast,
insect and mammalian expression systems.
186 Plckthun, A., Escherichia coli producing re-
combinant antibodies. Bioprocess Technol.
1994, 19, 233252.
187 Grant, S. D., Cupit, P. M., Learmonth, D.,
Byrne, F. R., Graham, B. M., Porter, A. J. M.,
Harris, W. J., Expression of monovalent and
bivalent antibody fragments in Escherichia
coli. J. Hematotherapy 1995, 4, 383388.
188 Eldin, P., Pauza, M. E., Hieda, Y., Lin, G.,
Murtauch, M. P., Pentel, P. R., Pennell, C. A.,
High level secretion of two antibody single
chain Fv fragments by Pichia pastoris. J. Im-
munol. Methods 1997, 201, 6775.
189 Joosten, V., Lokman, C., Hondel, C., Punt,
P. J., The production of antibody fragments
and antibody fusion proteins by yeasts and
filamentous fungi. Microbial Cell Factories
2003, 2, 115.
190 Young, M. W., Meade, H., Curling, J. M., Zio-
mek, C. A., Harvey, M., Production of recom-
binant antibodies in the milk of transgenic
animals. Res. Immunol. 1998, 149, 609610.
191 Fischer, R., Eman, N., Molecular Farming of
pharmaceutical proteins. Transgenic Res.
2000, 9, 279299.
192 Pollock, D. P., Kutzko, J. P., Brick-Wilson, E.,
Williams, J. L., Echelard, Y., Meade, H. M.,
Transgenic milk as a method for the produc-
tion of recombinant antibodies. J. Immunol.
Methods 1999, 231, 147157.
193 Sensoli, S., Transgenic technology for mono-
clonal antibody production. In: Antibodies
Novel Technologies and Therapeutic Use. G.
Subramanian (Ed.), Kluwer Academic, New
York, 2004.
194 Hiatt, G., Monoclonal antibody engineering
in plants. FEBS Lett. 1992, 307, 7175.
195 Knblein, J., Biotech: A New Era In The
New Millennium Fermentation and Ex-
pression of Biopharmaceuticals in Plants.
SCREENING Trends Drug Disc. 2003, 4,
1416.
196 Knblein, J., Biopharmaceuticals expressed
in plants a new era in the new Millen-
nium. In: Applications in Pharmaceutical Bio-
technology. R. Mller, O. Kayser (Eds.), Wiley-
VCH, 2004.
197 Valdes, R., Reyes, B., Alvarez, T., Garcia, J.,
Montero, J. A., Figueroa, A., Gomez, L., Pa-
dilla, S., Geada, D., Abrantes, M. C., Dorta,
L., Fernandez, D., Mendoza, O., Ramirez,
N., Rodriguez, M., Pujol, M., Borroto, C.,
Brito, J., Hepatitis B surface antigen immu-
nopurification using a plant-derived specific
antibody produced in large scale. Biochem.
Biophys. Res. Commun. 2003, 310, 742747.
198 Peeters, K., De Wilde, C., De Jaeger, G., An-
genon, G., Depicker, A., Production of anti-
bodies and antibody fragments in plants.
Vaccine 2001, 19, 27562761.
199 Hellwig, S., Drossard, J., Twyman, R. M.,
Fischer, R., Plant cell cultures for the pro-
duction of recombinant proteins. Nature Bio-
technol. 2004, 22, 13931398.
200 Abiko, Y., Passive immunization against
dental caries and periodontic disease, devel-
opment of recombinant and human mono-
clonal antibodies. Crit. Rev. Oral Biol. Med.
2000, 11, 140158.
201 Hodge, G., Disposable components enable a
new approach to biopharmaceutical manu-
facturing. BioPharm Int. 2004, 15, 3849.
202 Warner, T. N., Nochumsnon, S., Rethinking
the economics of chromatography. BioPharm
2003, 16, 5860.
203 Hawkins, R. E., Russell, S. J., Winter, G., Se-
lection of phage antibodies by binding affini-
ty, mimicking affinity maturation. J. Mol.
Biol. 1992, 226, 889896.
204 Hoogenboom, H. R., Designing and optimis-
ing library selection strategies for generating
high-affinity antibodies. Trends Biotechnol.
1997, 15, 6270.
205 Clynes, R. A., Towers, T. L., Presta, L. G., Ra-
vetch, J. V., Inhibitory Fc receptors modulate
in vivo cytotoxicity against tumor targets.
Nature Med. 2000, 6, 443446.
206 Shields, R. L., Namenuk, A. K., Hong, K.,
Meng, Y. G., Rae, J., Briggs, J., Xie, D., Lai,
J., Stadlen, A., Li, B., Fox, J. A., Presta, L. G.,
High resolution mapping of the binding site
on human IgG for Fcgamma RI, Fcgamma
RII, Fcgamma RIII, and FcRn and design of
IgG1 variants with improved binding to the
Fcgamma R. J. Biol. Chem. 2001, 276, 6591
6604.
207 Radaev, S., Sun, P. D., Recognition of IgG by
Fcgamma receptor. The role of Fc glycosyla-
tion and the binding of peptide inhibitors. J.
Biol. Chem. 2001, 276, 1647816483.
208 Umana, P., Jean-Mairet, J., Moudry, R., Am-
stutz, H., Bailey, J. E., Engineered glyco-
forms of an antineuroblastoma IgG1 with
optimized antibody-dependent cellular cyto-
toxic activity. Nat. Biotechnol. 1999, 17, 176
180.
209 Boyd, P. N., Lines, A. C., Patel, A. K., The ef-
fect of the removal of sialic acid, galactose
and total carbohydrate on the functional ac-
tivity of Campath-1H. Mol. Immunol. 1995,
32, 13111318.
210 Cartron, G., Dacheux, L., Salles, G., Solal-Ce-
ligny, P., Bardos, P., Coloms, P., Watier, H.,
Therapeutic activity of humanized anti-CD20
monoclonal antibody and polymorphism in
IgG Fc receptor FccRIIIa gene. Blood 2002,
99, 754758.
211 Joosten, V., Lokman, C., Hondel, C., Punt,
P.J., The production of antibody fragments
and antibody fusion proteins by yeasts and
filamentous fungi. Microbial Cell Factories
2003, 2, 115.
References 1145
Abstract
Antibodies are now the mainstream of bio-
pharmaceuticals. By the end of 2003, 17
marketed therapeutic antibodies generated
over $5 billion in combined annual sales,
with market growth at 30%. Ten years ear-
lier, this class of biopharmaceutical drugs
was almost written off, based on disap-
pointments experienced with the first gen-
eration of murine monoclonal antibodies.
This chapter will look at how new technol-
ogies have provided solutions to problems
that hampered early efforts to develop ef-
fective antibody therapeutics and trans-
formed the market for antibody drugs.
This includes the generation of fully hu-
man antibodies, their affinity maturation
and the selection of antibodies to bind to
particular epitopes on disease-relevant tar-
gets. The chapter will also highlight what
distinguishes a therapeutic from a simple
binding molecule different modes of ac-
tions of antibodies in different molecular
and cellular settings will be compared. Fi-
nally, some of the available formats of the
antibody and their effect on molecular/
pharmacological properties will be dis-
cussed.
Abbreviations
cytotoxicity
CDC complement-directed cytotoxicity
CDR complementarity-determining
region
CTL cytotoxic T lymphocyte
DOX doxorubicin
EBV EpsteinBarr virus
FcRn neonatal Fc receptor
GlcNAc N-acetylglucosamine
GM-CSF granulocyte macrophage colony-
stimulating factor
HIV human immunodeficiency
IL interleukin
NK natural killer
siglec sialic acid-binding, immuno-
globulin-like lectins
VEGF vascular endothelial growth
factor
2.1
Introduction
The initial promise of antibody-based bio-
pharmaceuticals has taken a long time to
1147
2
Modern Antibody Technology: The Impact on Drug Development
ISBN: 3-527-31184-X
be realized. The breakthrough that led to
the routine generation of monoclonal anti-
bodies (mAbs) [1] created expectations that
antibodies would become a major class of
drugs. That it has taken over 20 years for
the potential of therapeutic antibodies to
be translated into commercial success is
attributable to the time needed to solve
problems associated with the first genera-
tion of antibodies. While clinical efficacy
and safety depend critically on the target
against which the antibody is directed, as
well as the exact binding epitope, the mo-
lecular properties of the therapeutic itself
are equally important in a successful drug
(see also Part IV, Chapter 16 and Part V,
Chapter 1).
While some challenges have been large-
ly solved, others remain. Over the next de-
cade it is likely that additional improve-
ments in the molecular properties of anti-
bodies will be made, further increasing
their importance as biopharmaceuticals.
The main factors that limit the clinical
utility of antibodies are:
Immunogenicity.
Inability to reach a disease-relevant tar-
get in sufficient concentration.
Inability to trigger a particular biological
effect which translates into modification
of the disease process.
Technological approaches to reduce prob-
lems in each of these areas have met with
varying degrees of success and are dis-
cussed in detail in the following sections.
It is useful to discuss in turn each of the
molecular requirements for making a ther-
apeutic antibody and we will begin with
immunogenicity.
2.2
Immunogenicity
Early clinical applications of murine mAbs
quickly encountered the problem of im-
munogenicity in humans. While the in-
sight into cellular mechanisms that has
been enabled by the use of mAbs in basic
discovery research has been remarkable,
the problems with early attempts to use
antibodies in therapy cast a shadow of
doubt over whether this class of molecule
would ever be clinically useful. After the
first clinical disappointments with mAbs
in the early and mid 1980s, many gave up
on their promise as therapeutic agents.
The technological developments that have
led to a reduction in the difficulties posed
by immunogenic murine antibodies, i.e.,
the development of methods to make chi-
meric, then humanized and, finally, fully
human antibodies (Fig. 2.1), count among
the major success stories of the modern
biotechnology era.
It is clear now that the immune reaction
against the original murine format was a
major factor limiting the therapeutic use
of antibodies. A relatively small number of
academic groups and biotechnology com-
panies tackled these problems, and devel-
oped methods for solving them. Only with-
in the last decade have the resulting tech-
nological solutions led to the creation of
the successful class of drugs that antibod-
ies today represent. As a result, enthu-
siasm for this class of drugs in the phar-
maceutical industry is a very recent phe-
nomenon.
Immunogenicity is undesirable because
it can be the source of a number of safety
concerns, such as hypersensitivity and al-
lergic reactions, thrombocytopenia, ane-
mia, etc. Very problematic with recombi-
nant therapeutic proteins, but fortunately
extremely unlikely with antibodies, are
cases in which the therapeutic protein eli-
cits an immune response against one of
the bodys own proteins.
In addition to these safety concerns,
there is still the problem of a loss of effi-
cacy when the therapeutic molecule is re-
moved by the immune system. This could
be an issue in chronic indications, when
the antibody has to be given repeatedly, as
well as when the efficacy critically depends
on half-life, as this may be reduced by an
immune response against the therapeutic
antibody. In certain other cases (for some
examples, see below and Section 2.3.1), an
immune response is elicited, but causes
no major clinical effect.
Thus, the immunogenicity of antibodies
(or any protein, for that matter) in humans
is a very important parameter, but predict-
ing it remains an inexact science. To date,
no predictive scheme has emerged that
can obviate the need for a clinical trial.
The prediction of T cell epitopes from pep-
tide sequence has been attempted and there
are a number of websites available (http://
www.imtech.res.in/raghava/propred, http://
mif.dfci.harvard.edu/Tools/rankpep.html and
http://www.jenner.ac.uk/MHCPred) [24]
where this can be undertaken. Additionally,
some antibody manufacturers test Tcell epi-
topes experimentally in rather simple as-
says of T cell stimulation [5] by using a se-
ries of overlapping peptides covering the
whole protein.
The rationale for these experiments is
that the major human MHC alleles all re-
quire characteristic anchor residues in the
peptides they bind. To elicit an immune
reaction against a foreign protein, a T
helper cell response is needed, which in
turn requires that the protein is degraded
to peptides and this also shows sequence
specificity. A part of the protein is then
presented in MHC class II on antigen-pre-
senting cells. Peptides from the bodys
own proteins are also presented in MHC
class II, but T cells that would recognize
them do not normally exist, as they are
eliminated in the thymus. In order to
make a foreign protein invisible to T
cells, none of its peptides must bind to
MHC II, as T cells recognizing them will
exist. Using available crystal structures
and empirical data on peptide binding,
profiles can be formulated for likely an-
chor residues [6]. While the prediction of
the major antigenic epitope within a pro-
tein (or the absence of a clear hit) may be
possible by exploiting the available struc-
ture and sequence information, it is still
less clear whether such predictions can be
extrapolated for engineering purposes. A
Fig. 2.1 Diagram showing the proportion of human (green)
and murine (red) sequences in mouse, chimeric, humanized
and human antibody structures, as exemplified by HuCAL.
sufficient number of clinical trials will be
required to show whether proteins can be
engineered by point mutations to com-
pletely evade any MHC binding and thus
T cell surveillance without losing their
folding and function.
Derivatization with polyethylene glycol,
or PEGylation (see also Part VI, Chapter
3 and Part VI, Chapter 1) [79], while pri-
marily regarded as a means of increasing
serum half-life of small antibody frag-
ments (see Section 2.4.3), can also be used
to decrease the immunogenicity of foreign
proteins. As antibodies of fully human
composition can be now obtained, PEGyla-
tion, which introduces additional manufac-
turing problems, might be more appropri-
ate for modifying potential nonhuman ef-
fector domains, such as toxins (see Section
2.5.5). Nevertheless, clinical data will be
needed for each individual case.
Table 2.1 summarizes data on the immu-
nogenicity of a number of therapeutic anti-
bodies currently on the market. As is imme-
diately apparent from this and a wealth of
data on mouse monoclonals, antibodies of
more human composition are, in general,
less immunogenic than those of murine
origin. However, drawing firm conclusions
is difficult for a number of reasons. First,
immunogenicity depends on a number of
factors unrelated to the molecular composi-
tion of the drug, including dose, route of ad-
ministration, type of formulation and im-
munocompetence of the patient. Second,
the strict demarcation of antibody struc-
tures into the categories chimeric, huma-
nized or human, terms reflective of the
way the antibodies were made, diverts atten-
tion from the key issue of sequence homol-
ogy at the amino acid level. As has been
pointed out [10], since mouse and human
antibodies are rather homologous, the clo-
seness of a sequence to the nearest human
germline gene is perhaps a more important
determinant of immunogenicity and it may
thus be advantageous to create antibodies
with this property. Similarly, any human
protein that has been mutated, e.g., by di-
versifying a region, is potentially immuno-
genic.
That a number of the chimeric, human-
ized and human biopharmaceuticals in Ta-
ble 2.1 are highly successful drugs proves
that the movement away from murine
monoclonals towards antibodies of more
human composition has paid dividends in
the clinic. The conclusion for drug devel-
opment is that antibodies that are predom-
inantly human in their composition are
less likely to encounter problems of im-
munogenicity than murine antibodies.
This difference between murine antibod-
ies and those comprising some human
content is also evident in overall develop-
mental success rates. Data from the Tufts
Centre for the Study of Drug Development
[11] show that the probabilities of chi-
meric, humanized or fully human antibod-
ies progressing from entry into clinical
trials to the market are 26, 18 and 14%, re-
spectively, while for murine antibodies the
corresponding probability is only 4.5%.
Caution should be used in drawing any
conclusions from the apparently higher
success rates in developing chimeric over
humanized and human antibodies since
the sample size is limited to those 17 anti-
bodies which had reached the market at
the time the study was performed.
Although the advent of technologies that
can provide fully human antibodies (see
Section 2.3) would appear to have solved
the problem of immunogenicity, it is to be
expected that the solution to this problem
is not complete. Some human antibodies
are known to be immunogenic, typically
through anti-idiotypic or anti-allotypic ef-
fects. For example, the fully human anti-
body adalimumab (Humira), which was
T
a
b
l
e
2
.
1
I
m
m
u
n
o
g
e
n
i
c
i
t
y
o
f
a
n
u
m
b
e
r
o
f
t
h
e
r
a
p
e
u
t
i
c
a
n
t
i
b
o
d
i
e
s
c
u
r
r
e
n
t
l
y
o
n
t
h
e
m
a
r
k
e
t
C
o
m
p
o
u
n
d
n
a
m
e
T
r
a
d
e
n
a
m
e
C
o
m
p
a
n
y
A
n
t
i
b
o
d
y
f
o
r
m
a
t
A
n
t
i
g
e
n
A
p
p
r
o
v
e
d
a
p
p
l
i
c
a
t
i
o
n
Y
e
a
r
o
f
f
i
r
s
t
a
p
p
r
o
v
a
l
I
s
o
t
y
p
e
I
m
m
u
n
o
-
g
e
n
i
c
i
t
y
a
)
A
n
t
i
b
o
d
i
e
s
a
f
f
e
c
t
i
n
g
t
h
e
i
m
m
u
n
e
s
y
s
t
e
m
(
i
n
f
l
a
m
m
a
t
i
o
n
,
a
l
l
e
r
g
y
,
t
r
a
n
s
p
l
a
n
t
a
t
i
o
n
)
M
u
r
o
m
o
n
a
b
O
r
t
h
o
c
l
o
n
e
O
K
T
3
J
o
h
n
s
o
n
a
n
d
J
o
h
n
s
o
n
m
u
r
i
n
e
m
A
b
C
D
3
o
r
g
a
n
t
r
a
n
s
p
l
a
n
t
r
e
j
e
c
t
i
o
n
1
9
8
6
I
g
G
2
a
*
8
0
%
D
a
c
l
i
z
u
m
a
b
Z
e
n
a
p
a
x
H
o
f
f
m
a
n
n
L
a
R
o
c
h
e
h
u
m
a
n
i
z
e
d
m
A
b
C
D
2
5
k
i
d
n
e
y
t
r
a
n
s
p
l
a
n
t
r
e
j
e
c
t
i
o
n
1
9
9
7
I
g
G
1
1
4
%
B
a
s
i
l
i
x
i
m
a
b
S
i
m
u
l
e
c
t
N
o
v
a
r
t
i
s
c
h
i
m
e
r
i
c
m
A
b
C
D
2
5
k
i
d
n
e
y
t
r
a
n
s
p
l
a
n
t
r
e
j
e
c
t
i
o
n
1
9
9
8
I
g
G
1
<
2
%
I
n
f
l
i
x
i
m
a
b
R
e
m
i
c
a
d
e
J
o
h
n
s
o
n
a
n
d
J
o
h
n
s
o
n
c
h
i
m
e
r
i
c
m
A
b
T
N
F
-
a
r
h
e
u
m
a
t
o
i
d
a
r
t
h
r
i
t
i
s
,
C
r
o
h
n
s
d
i
s
e
a
s
e
,
a
n
k
y
-
l
o
s
i
n
g
s
p
o
n
d
y
l
i
t
i
s
1
9
9
8
,
1
9
9
9
I
g
G
1
1
3
%
E
f
a
l
i
z
u
m
a
b
R
a
p
t
i
v
a
G
e
n
e
n
t
e
c
h
,
S
e
r
o
n
o
h
u
m
a
n
i
z
e
d
m
A
b
C
D
1
1
a
p
s
o
r
i
a
s
i
s
2
0
0
3
I
g
G
1
6
%
O
m
a
l
i
z
u
m
a
b
X
o
l
a
i
r
G
e
n
e
n
t
e
c
h
,
N
o
v
a
r
t
i
s
h
u
m
a
n
i
z
e
d
m
A
b
I
g
E
a
l
l
e
r
g
i
c
a
s
t
h
m
a
2
0
0
3
I
g
G
1
<
0
.
1
%
A
d
a
l
i
m
u
m
a
b
H
u
m
i
r
a
A
b
b
o
t
t
f
u
l
l
y
h
u
m
a
n
m
A
b
T
N
F
-
a
r
h
e
u
m
a
t
o
i
d
a
r
t
h
r
i
t
i
s
2
0
0
3
I
g
G
1
5
%
A
n
t
i
b
o
d
i
e
s
i
n
o
n
c
o
l
o
g
y
R
i
t
u
x
i
m
a
b
R
i
t
u
x
a
n
G
e
n
e
n
t
e
c
h
,
B
i
o
g
e
n
,
I
D
E
C
,
H
o
f
f
m
a
n
n
L
a
R
o
c
h
e
c
h
i
m
e
r
i
c
m
A
b
C
D
2
0
n
o
n
-
H
o
d
g
k
i
n
s
l
y
m
p
h
o
m
a
1
9
9
7
I
g
G
1
<
1
%
T
r
a
s
t
u
z
u
m
a
b
H
e
r
c
e
p
t
i
n
G
e
n
e
n
t
e
c
h
,
H
o
f
f
-
m
a
n
n
L
a
R
o
c
h
e
h
u
m
a
n
i
z
e
d
m
A
b
H
E
R
2
/
n
e
u
b
r
e
a
s
t
c
a
n
c
e
r
1
9
9
8
I
g
G
1
<
0
.
1
%
G
e
m
t
u
z
u
m
a
b
o
z
o
g
a
m
i
c
i
n
M
y
l
o
t
a
r
g
W
y
e
t
h
h
u
m
a
n
i
z
e
d
m
A
b
/
c
a
l
i
c
h
e
a
m
i
c
i
n
C
D
3
3
a
c
u
t
e
m
y
e
l
o
i
d
l
e
u
k
e
m
i
a
2
0
0
0
I
g
G
4
0
%
A
l
e
m
t
u
z
u
m
a
b
C
a
m
p
a
t
h
-
1
H
S
c
h
e
r
i
n
g
h
u
m
a
n
i
z
e
d
m
A
b
C
D
5
2
c
h
r
o
n
i
c
l
y
m
p
h
o
c
y
t
i
c
l
e
u
k
e
m
i
a
2
0
0
1
I
g
G
1
2
%
I
b
r
i
t
u
m
o
m
a
b
t
i
u
x
e
t
a
n
Z
e
v
a
l
i
n
B
i
o
g
e
n
,
I
D
E
C
m
u
r
i
n
e
m
A
b
/
9
0
Y
C
D
2
0
n
o
n
-
H
o
d
g
k
i
n
s
l
y
m
p
h
o
m
a
2
0
0
2
I
g
G
1
4
%
T
o
s
i
t
u
m
o
m
a
b
B
e
x
x
a
r
G
S
K
m
u
r
i
n
e
m
A
b
/
1
3
1
I
a
n
d
m
u
r
i
n
e
m
A
b
C
D
2
0
n
o
n
-
H
o
d
g
k
i
n
s
l
y
m
p
h
o
m
a
2
0
0
3
I
g
G
2
a
>
7
0
%
T
a
b
l
e
2
.
1
(
c
o
n
t
i
n
u
e
d
)
C
o
m
p
o
u
n
d
n
a
m
e
T
r
a
d
e
n
a
m
e
C
o
m
p
a
n
y
A
n
t
i
b
o
d
y
f
o
r
m
a
t
A
n
t
i
g
e
n
A
p
p
r
o
v
e
d
a
p
p
l
i
c
a
t
i
o
n
Y
e
a
r
o
f
f
i
r
s
t
a
p
p
r
o
v
a
l
I
s
o
t
y
p
e
I
m
m
u
n
o
-
g
e
n
i
c
i
t
y
a
)
C
e
t
u
x
i
m
a
b
E
r
b
i
t
u
x
I
m
C
l
o
n
e
,
M
e
r
c
k
K
G
a
A
c
h
i
m
e
r
i
c
m
A
b
e
p
i
d
e
r
m
a
l
g
r
o
w
t
h
f
a
c
t
o
r
r
e
c
e
p
t
o
r
c
o
l
o
r
e
c
t
a
l
c
a
n
c
e
r
2
0
0
3
I
g
G
1
5
%
o
f
e
v
a
l
u
a
b
l
e
p
a
t
i
e
n
t
s
B
e
v
a
c
i
z
u
m
a
b
A
v
a
s
t
i
n
G
e
n
e
n
t
e
c
h
,
H
o
f
f
m
a
n
n
L
a
R
o
c
h
e
h
u
m
a
n
i
z
e
d
m
A
b
V
E
G
F
c
o
l
o
r
e
c
t
a
l
c
a
n
c
e
r
2
0
0
4
I
g
G
1
0
%
A
n
t
i
b
o
d
i
e
s
i
n
i
n
f
e
c
t
i
o
u
s
d
i
s
e
a
s
e
s
P
a
l
i
v
i
z
u
m
a
b
S
y
n
a
g
i
s
M
e
d
i
m
m
u
n
e
,
A
b
b
o
t
t
h
u
m
a
n
i
z
e
d
m
A
b
F
p
r
o
t
e
i
n
o
f
r
e
s
p
i
r
a
t
o
r
y
s
y
n
c
y
t
i
a
l
v
i
r
u
s
l
o
w
e
r
r
e
s
p
i
r
a
t
o
r
y
t
r
a
c
t
d
i
s
e
a
s
e
i
n
i
n
f
a
n
t
s
1
9
9
8
I
g
G
1
<
1
%
A
n
t
i
b
o
d
i
e
s
i
n
p
e
r
i
o
p
e
r
a
t
i
v
e
c
a
r
e
A
b
c
i
x
i
m
a
b
R
e
o
P
r
o
C
e
n
t
o
c
o
r
,
E
l
i
L
i
l
l
y
c
h
i
m
e
r
i
c
F
a
b
G
P
I
I
b
/
I
I
I
a
p
e
r
c
u
t
a
n
e
o
u
s
c
o
r
o
n
a
r
y
i
n
t
e
r
-
v
e
n
t
i
o
n
1
9
9
5
F
a
b
6
%
A
n
t
i
b
o
d
i
e
s
i
n
c
e
n
t
r
a
l
n
e
r
v
o
u
s
s
y
s
t
e
m
d
i
s
e
a
s
e
s
N
a
t
a
l
i
z
u
m
a
b
T
y
s
a
b
r
i
B
i
o
g
e
n
I
d
e
c
/
E
l
a
n
h
u
m
a
n
i
z
e
d
m
A
b
V
L
A
4
m
u
l
t
i
p
l
e
s
c
l
e
r
o
s
i
s
2
0
0
4
I
g
G
4
6
1
0
%
a
)
P
e
r
c
e
n
t
a
g
e
o
f
p
a
t
i
e
n
t
s
w
i
t
h
a
n
t
i
g
e
n
-
a
n
t
i
b
o
d
y
t
i
t
e
r
.
F
r
o
m
F
D
A
-
a
p
p
r
o
v
e
d
p
r
o
d
u
c
t
l
a
b
e
l
o
r
o
t
h
e
r
F
D
A
s
u
b
m
i
s
s
i
o
n
s
.
approved at the end of 2002 for the treat-
ment of rheumatoid arthritis, is immuno-
genic in a significant number of patients.
However, this has not hampered its suc-
cessful use in the clinic.
As we learn more about the precise mo-
lecular features that contribute to immu-
nogenicity, further improvements will be
made in antibody composition to reduce
this effect. However, for the foreseeable fu-
ture, only clinical data in humans will al-
low determination of the severity of the ef-
fect and whether it has a negative impact
on clinical utility. As a result of the reli-
ance on human clinical data to determine
immunogenicity, we may expect further
progress in this area to be slow. It remains
to be seen whether the molecular under-
standing of this complex process will ad-
vance to a point that immunogenicity can
be engineered out at the amino acid level.
It is also possible that a complete evasion
of the immune system will simply not be
possible in all cases.
2.3
Technology
The therapeutic antibodies currently on the
market were developed at a time when mo-
lecular engineering had not progressed to
its current level. Thus, because of the long
development times typical of pharmaceuti-
cal development, all antibodies that are
now on the market are derived in some
way from mouse mAbs. Nevertheless, as
will be clear from the following subsections,
molecular engineering has now progressed
to a point that therapeutic antibodies can
be obtained without having an animal-de-
rived mAb as a starting point. The great
number of antibodies and derivatives in var-
ious phases of clinical trials that are derived
from libraries bear witness to this point.
2.3.1
Chimeric and Humanized Antibodies
In the early and mid 1980s, as the first
murine mAbs were being tested in the
clinic, it was quickly recognized that in or-
der to avoid the problems associated with
their immunogenicity, antibodies of more
human composition would be needed.
While various methods have been investi-
gated, such as immortalization of human
B cells by Epstein-Barr virus (EBV trans-
formation) (see Section 2.3.3), most suc-
cess was achieved by starting with a mur-
ine mAb and engineering it to have a
more human composition. We now sum-
marize the different ways of achieving this
goal.
Historically, the first generation of hy-
brid antibodies (part mouse/part human)
comprised the entire murine variable do-
mains of the original mAb, with the re-
mainder of the IgG (constant C
L
domain,
usually j, plus C
H
1, hinge, C
H
2 and C
H
3
domains) coming from a human antibody
(Fig. 2.1). Thus, in these so-called chimeric
antibodies, four out of 12 domains in the
IgG remain of murine origin (two V
L
and
two V
H
). In total, approximately two-thirds
of the sequence is of human origin, the re-
maining one-third being murine.
Currently, five chimeric antibodies are
approved for human therapeutic use (Ta-
ble 2.1). The immunogenicity of two chi-
meric antibodies, abciximab (ReoPro) and
infliximab (Remicade) has recently been
evaluated in some detail [12]. ReoPro is a
chimeric Fab fragment that binds to the
a
IIb
b
3
integrin, also called platelet mem-
brane glycoprotein GPIIb-IIIa [13]. GPIIb-
IIIa is an adhesion receptor for fibrinogen
and the von Willebrand factor, both of
which carry multiple binding motifs for
the integrin and thus mediate platelet ag-
gregation. ReoPro is approved for use in
2.3 Technology 1153
percutaneous coronary intervention to pre-
vent cardiac ischemic complications. In or-
der to exert its effect by inhibiting throm-
bus formation, it must block a large frac-
tion of its target integrins. At least 6% of
patients develop an immune response to
the antibody, but re-administration of the
Fab fragment remains possible. Infliximab
is a chimeric IgG1 specific for human tu-
mor necrosis factor (TNF)-a, and is ap-
proved for the acute treatment of the signs
and symptoms of Crohns disease (an in-
flammation of the small intestine), and for
the chronic treatment of rheumatoid ar-
thritis. An immune response occurs in at
least 10% of patients, although there does
not seem to be a reduction in clinical effi-
cacy.
The next improvement was humaniza-
tion the grafting of the complementarity-
determining regions (CDRs) of a mouse
antibody onto a human framework [14].
For this purpose, a human framework is
chosen from the database of human genes
for V
H
, another for V
L
(j or k), and an
alignment of the murine and human se-
quences is made. In addition to the CDRs
themselves, differences in the framework
must be taken into account (Fig. 2.2). For
example, the so-called outer loop (some-
Fig. 2.2 Antibodies binding haptens, oligopep-
tides and oligosaccharides or proteins. Super-
posed crystal structures of the variable regions
were sorted into three classes according to the
type of antigen. Structurally variable residues with-
in the CDRs of the antibodies are shown in green;
those at the N-terminus, to the N-terminal side of
CDR-1 and within the outer loop in cyan. Residues
within the inner (dimer interface) b-sheet of V
L
and V
H
whose side-chains contribute both to the
dimer interface and to antigen binding if it
reaches deep into this pocket are shown in yellow,
orange and red, depending on depth. Note that
these residues formally belong to the framework.
The structurally least-variable residues whose Ca
positions were used for the least-squares superpo-
sition of the antibody fragments are shaded gray.
times called CDR4) influences the confor-
mation of CDR3 and a number of frame-
work residues near the pseudo-2-fold axis
(relating V
H
and V
L
by a rotation) are im-
portant for the binding of hydrophobic
side-chains in a cavity in the binding site
[15]. Such residues, even though formally
from the murine donor, frequently need to
be maintained. This involvement of frame-
work residues is also the reason why the
immune system uses different variable do-
main subtypes to bind antigens of many
shapes and compositions (Fig. 2.3).
A structural model is usually built, ex-
ploiting the great number of three-dimen-
sional structures of antibodies available to-
day (http://www.biochem.unizh.ch/antibody)
(over 300 structures of independent V
H
,
220 independent V
j
and 40 independent
V
k
sequences). Then a loop-grafted ver-
sion of the antibody, carrying additional
framework mutations as necessary, can be
created with similar affinity to that of the
original mouse monoclonal. While this
goal can usually be achieved, albeit with
significant effort in cloning, engineering,
and assay development, the greatest limita-
tion of this technology is that a mouse
mAb with the desired specificity is needed
as a starting point.
A related approach to humanization,
i.e., antibody resurfacing [1618], relies on
2.3 Technology 1155
Fig. 2.3 Complexes of antibodies binding haptens,
oligopeptides and oligosaccharides or proteins.
The antigens are colored pink; parts of the anti-
body are colored as in Fig. 2.2. It can be seen that
there is extensive binding to residues which for-
mally belong to the framework, either in the dimer
interface region for hapten binders or binders of
peptides (which frequently use a side-chain in a
hapten-like binding mode), and to the outer loop
in protein binders. Hapten binders commonly
form a deep, funnel-shaped binding pocket en-
larged by a long CDR-L1 and open CDR-H3 con-
formation, while protein binders preferentially uti-
lize a relatively flat antigen binding surface charac-
terized by a short CDR-L1 and a closed CDR-H3
conformation. This is one of the main reasons
why the immune system uses different frame-
works, rather than different CDRs all on the same
framework. It also highlights the points to consid-
er in loop grafting. More details can be found
elsewhere [15].
making point mutations in surface resi-
dues of the murine antibody, converting
amino acids to those found in human fra-
meworks. This process requires alignment
of the sequences of the original murine
antibody with various human congeners
that fulfill requirements of sequence com-
patibility with the antibody being modi-
fied. One such antibody is now in a phase
I clinical trial [19, 20].
The first humanized antibody to enter
the clinic was alemtuzumab (CAMPATH-
1H) which is directed against CD52, a gly-
cosylphosphatidylinositol-anchored glyco-
protein with unknown function, present
on lymphocytes. CD52 is abundantly ex-
pressed on B and T cells, macrophages,
monocytes, and eosinophils. Alemtuzumab
is used for the treatment of non-Hodgkins
lymphoma [21] and was approved in 2001
as first-line treatment for chronic lympho-
cytic leukemia. It has been proposed [22]
that alemtuzumab inhibits the growth of
B and T cells by cross-linking of CD52. It
has also been suggested that the antibody
works entirely through its effector func-
tion, using both antibody-dependent cellu-
lar cytotoxicity (ADCC) and complement-
directed cytotoxicity (CDC) [23]. Alemtuzu-
mab has also been tested as an immuno-
suppressive reagent in transplantation and
autoimmune diseases. When the human-
ized antibody was administered i.v., no im-
mune response was elicited (but infusion
site reactions were observed), while a s.c.
injection did elicit an anti-idiotypic im-
mune response in two of 32 patients [24].
Immunogenicity data for some other hu-
manized antibodies are summarized in Ta-
ble 2.1.
2.3.2
The Limitations Imposed by Immunization
The use of immunization to generate anti-
bodies in animals is one of the oldest tech-
niques in biology. The generation of mAbs
[1] uses an elegant cellular cloning tech-
nique to obtain (usually murine) antibod-
ies with single specificities, but still must
start with an immunized animal. Chimeri-
zation and humanization are modern mo-
lecular engineering methods, but they also
require immunization to provide a mAb as
a starting point.
Even adalimumab (Humira), a fully hu-
man antibody directed against TNF-a, was
obtained using a pre-existing mouse
monoclonal as a starting point and a pro-
cess termed guided selection [25]. In this
approach, one chain of the murine anti-
body, in recombinant form, was used in
phage display together with a library of
human antibodies encoding the other
chain, followed by the reciprocal experi-
ment, thereby generating an equivalent
human antibody in several steps. Adalimu-
mab is on its way to becoming a highly
successful biopharmaceutical drug for the
treatment for rheumatoid arthritis and
possibly also psoriasis.
Notwithstanding this spectacular suc-
cess, the reliance on immunization is a
major limitation in all antibody generation
methods that use this step. For example,
the possibility of directing the response to
a particular binding epitope during immu-
nization is very limited and usually re-
stricted to screening hybridomas [26]. It
may well be that the epitope desired for
the biological effect is not particularly fa-
vored, and not one which leads to high-af-
finity antibodies. Furthermore, the animal
repertoire is limited to those variations
that the immune system introduces dur-
ing somatic mutation, and these are only a
small subset of the total theoretical reper-
toire. Therefore, the process of somatic
mutation does not provide an exhaustive
screen for the highest affinity or even for
the optimal biological function. In con-
trast, modern library-based methods (see
Section 2.3.3.1) provide the opportunity to
select for antibodies with defined affinities,
for particular epitope specificities, and
even for pre-defined cross-reactivity pat-
terns [2729].
2.3.3
Fully Human Antibodies
Despite the outstanding achievements
made with the techniques of chimerization
and humanization, a means of routinely
accessing fully human antibodies has al-
ways been a goal for developers of thera-
peutic antibodies. Historically, the first
method for making human IgGs was the
immortalization of human B cells with
EBV. Since this method is rather ineffi-
cient, it has not been widely used. Re-
cently, however, a new method has been
introduced that dramatically increased the
efficiency of transformation [30]. This of-
fers the opportunity to immortalize mem-
ory B cells from patients after an infection,
and potentially even from cancer patients,
and thus complements cloning of such an-
tibodies from patients and recovery of anti-
bodies by display technologies [31]. Today,
the most widely used technologies for
making fully human antibodies are either
library-based methods or transgenic mouse
approaches. The fact that over 30 antibod-
ies based on these technologies are cur-
rently in clinical trials indicates how well
established they have become.
2.3.3.1 Library-based Methods
In the late 1980s, work was commenced
on constructing antibody libraries. This de-
velopment took antibody generation in a
new direction, for the first time obviating
the need for immunization of an animal.
Library methods for antibody generation
brought together several new technological
developments, including construction of
the library itself, but equally importantly,
methods of screening the resulting library.
This section summarizes both library con-
struction and screening methods.
Library construction The first antibody li-
braries that had not been obtained from
immunized animals were based on human
genetic material isolated from natural
sources [32, 33]. The most convenient
sources of the appropriate genetic material
are DNA or mRNA from human peripher-
al B cells, bone marrow B cells or tonsil B
cells. The resulting library reflects the
make-up of the human repertoire of the
particular donor(s) and will to some de-
gree reflect amplifications from recent in-
fection events. Often, blood from many do-
nors is pooled for this reason [33]. This
bias can also be exploited if peripheral
blood lymphocytes from a patient with an
infectious disease are used, such a library
is an excellent source of antibodies against
the infectious agent [31]. It is also possible
to use non-rearranged genomic DNA as a
starting point to make a library [34]. As
the antibody genes in their germline con-
figuration do not contain the V, D or J seg-
ments (and would thus be lacking CDR3
and the last b-strand of the domain), these
elements need to be added in a subse-
quent PCR assembly.
An alternative approach, giving the high-
est level of control over the process, is to
synthesize the antibody genes completely
[35]. In this case the frameworks can be cho-
2.3 Technology 1157
sen to represent the optimal diversity of the
antibody repertoire. In the case of HuCAL
GOLD (unpublished), a fully synthetic hu-
man combinatorial library, seven frame-
works for V
H
and seven for V
L
are used, giv-
ing 49 V
H
V
L
combinations (see Fig. 2.4).
The system also comprises pre-synthesized
libraries of cassettes for all six CDRs, each
of which reflects the composition of the cor-
responding human CDRs.
To maximize diversity yet maintain
structural integrity, the CDRs are not sim-
Fig. 2.4 The schematic structure of the genes underlying the
HuCAL library, showing the modularity of the CDRs and the
pre-assembled cassettes used during optimization.
ply randomized, but are diversified accord-
ing to structural criteria, keeping a few
key anchor residues constant and limiting
the diversity of others to those observed in
nature or compatible with the loop struc-
ture. Furthermore, the diversity of CDR1
and CDR2 is close to what is observed for
a particular subtype, maximizing similarity
to the germline. Conversely, the great
length variation normally observed in
CDR-H3 is also recaptured in HuCAL, al-
lowing the resulting antibodies to bind to
epitopes with a great variety of shapes and
composition. Using this strategy, key struc-
tural residues are maintained and the vari-
ation observed in rearranged human anti-
body sequences, reflecting the process of
affinity maturation and subsequent selec-
tion as documented in sequences of hu-
man antibodies, is well mimicked.
The HuCAL GOLD library also incorpo-
rates unique restriction sites bracketing
the CDRs (Fig. 2.4), a feature made possi-
ble by the use of chemical synthesis for
construction of the encoding genes. The
ends of the CDR cassettes match the re-
striction sites bracketing their positions in
the HuCAL library. A fully modular sys-
tem is the result. The benefit of such a
system is that antibody optimization can
be rapidly and systematically carried out
by replacing CDRs in turn to create new
sublibraries based on one or more hits
from a first screening. Multiple examples
have shown that this is a reliable means of
generating antibodies with predefined
properties, while still retaining 100% hu-
manness in the sequence. Such systemat-
ic optimization of antibodies builds on
their inherent affinity and specificity to
create substances with drug-like character-
istics, including predefined cross-reactivity
patterns as well as the ability to activate,
deactivate and/or block certain biological
processes.
The question sometimes arises as to
why multiple frameworks are desirable in
antibody libraries. There is a good reason
why nature uses more than one frame-
work in the antibody repertoire. As already
mentioned above, antigens not only con-
tact the CDRs, but also framework resi-
dues (Fig. 2.3). For example, large antigens
make additional contact to the outer loop,
or CDR4 mentioned above, while small
haptens, but also side chains of peptide
antigens, often bind in a deep cavity near
the pseudo-2-fold axis of the antibody [15].
These residues formally belong to the
framework and thus vary between sub-
groups. Therefore, in order to capture the
diversity of the human repertoire, a simi-
lar diversity of frameworks is necessary.
Single-framework libraries have also
been made, and with a large enough diver-
sity, high-affinity binders against many tar-
gets can be obtained [36, 37]. It must be
remembered, however, that in the case of
making therapeutic antibodies, usually
only a very small subset of epitopes is of
any utility. Therefore, obtaining high-affin-
ity binders is a necessary, but not suffi-
cient, condition for making a therapeuti-
cally active antibody. In general it will thus
be necessary to have a technology available
that can generate high-affinity binders to
any epitope, so that cellular assays and
preclinical experiments in animals can be
used to identify the binders with highest
potency. For example, a therapeutic anti-
body may need to block an interaction. In
this case the antibody must bind to that
part of the target engaged in the interac-
tion, no matter whether it is protruding,
recessed or flat. It is immediately obvious
why a good antibody repertoire must
therefore be able to bind any shape on any
target molecule.
Libraries need to comprise diversity in
the variable region if they are to effectively
2.3 Technology 1159
mimic the natural immune repertoire. The
molecular arrangement of the library (scFv
fragment or Fab fragments being the
dominant formats) is, however, mostly dic-
tated by the selection strategy used (dis-
cussed below). For example, in phage dis-
play both scFv and Fab libraries can be
used, as they can be in yeast display. In
contrast, in ribosome display only single
polypeptides can be used, as in scFv frag-
ments.
Screening technologies An essential tech-
nological adjunct to the creation of an
antibody library is an effective means of
screening it for binders having the desired
characteristics. Over recent years, several
technologies have been developed that al-
low the selection of proteins, including an-
tibodies, from large repertoires. Some
methods also allow directed molecular evo-
lution [38]. All of them have in common
that the fact that the phenotype of the pro-
tein (in the case of an antibody, its binding
specificity) must be connected to the geno-
type (the sequence encoding the antibody
in question). A schematic overview of
these concepts is shown in Fig. 2.5. On af-
finity enrichment of the antibody, its
DNA sequence is thereby enriched as well,
allowing the identification of the precise
molecular composition that gave the de-
sired binding phenotype.
While selections in model systems and
from complex antibody libraries have been
described using a number of these tech-
nologies, we restrict the discussion here to
a brief description of the three methods
with which the greatest experience in the
field of antibodies has been obtained to
date.
Historically, the first technology for the
selection of polypeptides from repertoires
was display using filamentous phages
(Fig. 2.5a), which was first demonstrated
for peptides [39] and subsequently adapted
to antibodies [40, 41]. Escherichia coli are
transformed with the library of interest in
the form of an expression plasmid that en-
codes the antibody (either as an scFv frag-
ment or as a Fab fragment) in fusion with
a minor coat protein (usually g3p or its C-
terminal domain) of the phage. Note that
each E. coli cell receives, in general, only
one plasmid molecule and will thus en-
code only one molecular species of anti-
body. When a helper phage is added to
these bacteria, new phage particles are pro-
duced that incorporate the fusion protein
into their coat and thus display the par-
ticular antibody, while packaging the plas-
mid, which contains the coding informa-
tion of the antibody. The phage population
is then added to an immobilized target
and those phages that do not bind specifi-
cally to the target are removed by washing.
As a result, phage-carrying antibodies with
the desired binding properties can be en-
riched and, in principle, recovered by an
appropriate elution step. Successful proof-
of-principle experiments were initially re-
ported for scFv and Fab fragments [40, 41],
and many sources of libraries have been
used with this selection technology since
then. While space limitations do not per-
mit a full account of the many permuta-
tions of this fundamental concept, we
want to stress that it has found broad utili-
ty in selecting antibodies against purified
targets, selected epitopes, but also antigens
on whole cells [42]. Most antibody selec-
tions have been carried out with this tech-
nology.
A variant on this method uses a disul-
fide bond to link the antibody to the coat
protein of the phage, in which a pendant
cysteine residue is introduced [43]. The ad-
vantage of this method is that the elution
step can be replaced by mild reduction of
the disulfide bond, resulting in reliable re-
covery of enriched phage. In contrast, elu-
tion that relies on a change of pH or a
competing ligand may not necessarily lead
to recovery of the most interesting antibod-
ies, i.e., those that bind to the target with
the highest affinity.
A second technology that has been used
more widely recently is yeast (Saccharo-
myces cerevisiae) display [44] (Fig. 2.5b). In
this case, yeast cells are transformed with
a plasmid library, which encodes a fusion
with yeast a-agglutinin (Aga1p/Aga2p).
Again, this has been carried out both for
scFv [45] and Fab [46] libraries. While
transformation of yeast cells is somewhat
more difficult than that of E. coli, meth-
odologies do exist to achieve this. With an-
tibodies displayed on the surface of yeast,
not only a mechanical enrichment (e.g.,
with magnetic beads) is possible, but also
the use of cell sorters together with a
fluorescently labeled target. Therefore,
thresholds for affinity can be defined and
the affinity can be measured directly on
2.3 Technology 1161
Fig. 2.5 The linking of phenotype and genotype in
selection systems for antibodies. For more details
on (ac), see text. (a) In phage display, the anti-
body is fused to the minor coat protein g3p of
filamentous phage, while the DNA is on the inside
of the phage. (b) In yeast or bacterial cell surface
display, the antibody is displayed on the outer sur-
face of the cell, while the genetic information is
encoded on a plasmid inside the cell. (c) In ribo-
some display, mRNA and protein product are
linked by the ribosome, and the selection takes
place in an in vitro translation system. Addition-
ally, antibodies can be screened and selected in-
tracellularly (d), and selection for growth puts
high demands on the folding efficiency of the anti-
body, while the selection for affinity is not very
stringent. Intracellular screening is achieved by
fusing the antigen and antibody to two protein
halves which, when brought together by the anti-
bodyantigen interaction, allow cellular growth by
allowing transcription of a selectable marker
(yeast two-hybrid system) [154] or reconstitution
of a selectable enzyme activity (protein-fragment
complementation assay) [155].
the yeast cells by titration. It has been re-
ported that the yeast libraries can be am-
plified without changing their composi-
tion, i.e., particular clones do not seem to
be enriched or depleted upon copying the
library [45].
A third display technology that has been
used to screen antibody libraries is ribo-
some display [4750] (Fig. 2.5c). In con-
trast to the other methods, this works en-
tirely in vitro, without using any cells at
all. The library has to be in the scFv for-
mat, as only a single polypeptide chain
can be displayed at a time. A library of
PCR fragments is used, encoding a pro-
moter and the open reading frame of the
scFv fragment, fused to a spacer, which
runs to the physical end of the fragment
and does not encode a stop codon. The
function of the spacer is to allow the scFv
fragment to exit from the ribosomal tun-
nel and fold into its correct three-dimen-
sional structure. An in vitro translation is
thus carried out with a quantity of ribo-
somes stoichiometric to mRNA. The
mRNA is translated to the end and re-
mains connected to the tRNA within the
ribosome, the scFv protein thereby remain-
ing connected to the ribosome, which is
also still attached to the encoding mRNA.
Thereby, the antibody and its encoding
mRNA remain linked. The main advan-
tage of this method is that very large li-
braries can be used (typically 10
12
different
variants), as unlike in the other technolo-
gies no diversity is lost in the transforma-
tion step of E. coli or yeast. Furthermore,
by using polymerases without proof-read-
ing capability or even, deliberately, error-
prone polymerase chain reaction or other
methods to increase diversity, combined
with a stringent selection for affinity, bind-
ers with picomolar affinity can be selected,
thereby mimicking somatic mutation in vi-
tro [49, 50].
2.3.3.2 Transgenic Mice
A scientifically elegant development was
the creation of transgenic mice, in which
part of the human antibody repertoire is
inserted into the mouse genome (see also
Part III, Chapter 4) [5158]. This has been
achieved in a variety of ways, e.g., with
yeast artificial chromosomes or pieces of
human chromosomes and homologous re-
combination (see also Part III, Chapter 2).
In order to get an efficient response of hu-
man antibodies, the mouse repertoire
needs to be inactivated. This has been
done by targeted deletion of the J
H
and J
j
region (together with the constant j re-
gion), to prevent V(D)J rearrangement of
the murine antibody genes. This strategy
permits well-established methods for the
generation of mouse mAbs to be used to
produce fully human antibodies. Histori-
cally, most therapeutic applications have
used whole IgGs and the transgenic
mouse approach produces them directly.
In the case of the library approaches,
which use antibody fragments during se-
lection, a conversion to IgGs involves an
additional (very straightforward) step. It
can be expected, however, that a greater
variety of formats for therapeutic antibod-
ies will be employed in the future (see also
Part V, Chapters 1 and 6), where the li-
brary technologies would have an addi-
tional advantage, since when transgenic
mice are used, the antibody genes have
first to be isolated by molecular cloning.
While a great scientific achievement, the
disadvantages mentioned above in connec-
tion with immunization pertain here: lack
of full control over the target epitope dur-
ing the immunization process and inabil-
ity to pre-determine cross-reactivity and af-
finity. In addition, and not unexpectedly,
proteins that are highly conserved between
man and mouse may not be immunogenic
in this system.
2.4
Reaching the Target: The Importance
of Specificity, Affinity and Format
An advantage of antibodies as potential
therapeutics is their inherent affinity and
specificity for their binding partner. These
properties are a prerequisite for their ther-
apeutic effectiveness. A third property is
also crucial to the performance of a thera-
peutic antibody, i.e., its format. The impor-
tance of these molecular properties for the
application of antibodies as biopharmaceu-
ticals is considered here.
2.4.1
Epitope Specificity
Specificity for target is one of the main
properties that distinguishes antibodies
from other bioactive molecules. The first
factor to consider is where exactly the anti-
body binds on the target. Some cases are
easy to understand at the molecular level,
such as in the case of blocking the action
of a cytokine, such as TNF-a. This inflam-
matory cytokine is produced too abun-
dantly in a number of diseases, such as in-
flammatory bowel disease (Crohns disease
and ulcerative colitis), rheumatoid arthritis
and psoriasis. The therapeutic strategy
thus consists of preventing the binding of
this soluble, homo-trimeric molecule to its
receptor. The antibody, obviously, must
bind in a way that it overlaps with the
binding interface to the receptor it must
bind to a neutralizing epitope.
In some cases, such as the one men-
tioned, such binders are quite straightfor-
ward to select, as the receptor contact sur-
face is large. Indeed, TNF-a is a popular
target for antibody and other protein-based
therapeutic approaches, developers being
encouraged by the success of etanercept
(Enbrel, a soluble receptorFc fusion pro-
tein), infliximab (Remicade) and adalimu-
mab (Humira) for the treatment of arthri-
tis, psoriasis and Crohns disease [59]. All
three molecules are therapeutically active
in rheumatoid arthritis. Interestingly how-
ever, the soluble receptor etanercept shows
no activity in Crohns disease, while inflixi-
mab does. A possible reason has been pro-
posed [59]: to be effective, the transmem-
brane form of TNF-a must be targeted,
which is present on T cells, where it may
have a slightly different conformation, and
this conformation is not recognized by the
receptor, but by the antibody infliximab.
Only the latter thus helps controlling in-
flammation by inducing apoptosis in T
cells.
The antibody-binding site does not have
to be identical to that of the receptor, it
only has to overlap in such a way as to
prevent simultaneous binding of the target
to its receptor. In cases other than the
ones described, this can be more difficult
to achieve, e.g., if the binding site is small
and not favored for binding. In such a
case, using antibody library technologies it
is often possible to guide the selection to
the relevant epitope. This can be done in a
variety of ways, such as using the real
partner (e.g., the soluble receptor) as a
competitor or using mutants of the soluble
molecule to pre-bind all antibodies that are
not directed to the desired epitope. The in-
terested reader is directed to publications
in which the technical approaches are dis-
cussed in more detail (see, e.g., [60]).
It is almost always necessary to test the
binders so obtained in cell-based assays, in
order to verify that they react with the anti-
gen in its proper context on the cell. In
some cases, on the other hand, it may not
even be possible to obtain any soluble ver-
sion of the protein of interest. In this case,
selections can be carried out on whole
cells and many of the same strategies ap-
plied. Again, the technical approaches and
remaining challenges have been described
[2628], but space limitations do not per-
mit us to discuss this in detail.
Blocking (or neutralizing) epitopes are
easy to conceptualize. It should be noted,
however, that in many cases the rationale
why only binders to a particular epitope
give a biological response is not at all
clear. This is illustrated with two antibod-
ies against HER2, a member of the epider-
mal growth factor receptor family overex-
pressed in about 2530% of women with
breast cancer and correlated with a poor
prognosis (see also Part I, Chapter 5). The
recently described crystal structure of 2C4
(Pertuzumab, Omnitarg) in complex with
HER2 shows that this antibody inhibits
the homodimerization of HER2 and also
its heterodimerization with other members
of the family, and thus signaling [61].
However, the antibody trastuzumab (Her-
ceptin, 4D5) also inhibits signaling, yet
without inhibiting dimerization [62, 63],
and it binds close to the membrane, as
shown in the crystal structure of its com-
plex with HER2 [64]. The mechanistic rea-
son for its inhibitory action, which is lim-
ited to tumors with high levels of homodi-
mers of HER2, is still not entirely clear,
and it is likely that internalization and/or
proteolytic shedding are two of the factors
influenced by trastuzumab binding, which
then decrease signaling [62, 63]. Addition-
ally, there is evidence [65] that Fc receptors
on natural killer (NK) cells are recruited
by the exposed Fc part of trastuzumab
while bound to HER2 and this action is
providing part of the therapeutic effect of
this antibody. It is almost certain that the
combined effect of all these factors is what
gives Herceptin its efficacy.
With all the excitement surrounding this
biopharmaceutical, it should not be forgot-
ten that, for example, with Herceptin as
the sole treatment in metastatic breast can-
cer, only eight complete responses
amongst 222 women were seen in a phase
III trial [66]. Similarly, only 323% of all
patients receiving Rituximab for relapsed
or indolent refractory B cell non-Hodgkins
lymphoma showed a complete response
[67]. These two examples, both concerning
FDA-approved antibodies (see also Part
VII, Chapter 4), underline the urgent need
for further work in understanding the ac-
tion and subsequent improvement of
such antibodies, as well as the biological
function of potential target antigens.
In cases where the structure and func-
tion of the target are known, presenting
the relevant epitope in order to generate
binders of the desired specificity is still the
key challenge. A prime example is the dif-
ficulty in generating protective antibodies
against HIV [68] (see also Part II, Chapter
7). The surface proteins of the virus,
gp120 and gp41, are highly immunogenic,
regardless of whether they are presented
in the context of the virus particle (in in-
fected patients) or as the soluble protein
(as shed protein in infected patients). Anti-
bodies can also be obtained readily, using
immunization or display technologies.
However, the great majority of antibodies
that result are directed against non-protec-
tive surface epitopes. Indeed, the isolated
protein does not even present the protec-
tive epitope, which is recessed between
trimeric subunits when arranged as pres-
ent on the virus. However, the virus parti-
cle normally does not elicit protective anti-
bodies either, and thus it has been so far
impossible to obtain an AIDS vaccine.
Since HIV biology is well studied, a num-
ber of strategies are currently under way
to overcome this challenge [69] (see also
Part II, Chapter 8). Very few broadly neu-
tralizing antibodies have been cloned from
infected patients, but the mechanism of
neutralization is not clear for all of them
[70].
The difficulty of generating broadly neu-
tralizing antibodies against HIV may have
attracted much attention, but it is likely
that inaccessible neutralizing epitopes, and
epitopes that are only formed after an ini-
tial binding event, are much more com-
mon also in other biological systems and
not restricted to infective agents.
In some cases, even the precise molecu-
lar definition of the desired epitope is un-
clear. Accordingly, the preclinical observa-
tion will frequently be that only very few
antibodies will show a biological effect,
even though many others bind to the
same target with high affinity. It will
usually be of great benefit to attempt to
understand, at the molecular and structur-
al level, the key features distinguishing the
biologically active antibodies from the
others. In the case of cellular targets, this
may involve differences in receptor multi-
merization (the antibody either inducing
or preventing it), in the ensuing receptor
internalization, proteolytic shedding, block-
ing (or enhancing) the binding of an exter-
nal ligand, or making the antibody Fc part
accessible while binding to the surface
target to macrophages, neutrophils,
monocytes or NK cells carrying Fc recep-
tors. It immediately follows that the type
of Fc desired (or its desired absence) is an-
other parameter important for engineering
and it requires an understanding of the
mode of action of the antibody.
A field of medicine where this lack of me-
chanistic understanding at the level of mo-
lecular structure is particularly notable and
hampering progress is the treatment of can-
cer (see also Part II, Chapters 5 and 6).
There are great challenges in targeting solid
tumors, brought about by the enormous dif-
ficulty inherent in obtaining significant and
selective enrichment of the antibody at the
tumor site. As a consequence, the majority
of antibodies approved today are directed
against leukemia, myeloma and lymphoma
(see also Part V, Chapters 5 and 6). In these
cases, two factors favor clinically successful
treatment with antibodies: (1) the target
cells are easily accessible in the bone mar-
row, lymph nodes or blood, and (2) the tu-
mors respond well to radiation and che-
motherapy. Furthermore, they can be selec-
tively targeted via several cell-lineage-specif-
ic markers. For example, CD20 is a marker
specific for B cells (a more detailed descrip-
tion is given in Section 2.5.2). In this case,
the antigen is not restricted to the diseased
cell, but the redundancy and the self-regen-
eration of the immune system can sustain
the temporary depletion of B cells. It is in
general much more difficult to identify se-
lective markers for solid tumors [71]. The
number of such tumor markers that are sui-
ted for targeted therapy is small: despite
massive attempts using a variety of tech-
niques, including screening of healthy and
diseased tissues with antibody libraries,
only very few new tumor-associated surface
proteins have been added to the list over the
years [7274].
2.4.2
Affinity
An important factor determining therapeu-
tic efficacy is affinity. If the goal is to block
the action of a soluble target such as a cy-
tokine, then as little as possible of the cy-
tokine should remain in an active form.
The affinity directly determines the
amount of cytokine that will be free at
equilibrium. In general, the affinity should
thus be as high as possible for such appli-
cations. It should be noted that in many
cases the soluble protein to be inhibited is
a monomer (trimeric TNF-a and its homo-
logs being more unusual in this respect),
so that the true monovalent thermody-
namic affinity, i.e., the affinity of a single
binding site, is the property of interest.
Another example where the importance
of affinity has been clearly highlighted is
the protective function of antibodies
against toxic or infectious agents. For ex-
ample, post-challenge protection against
the anthrax toxin, a tripartite protein, cor-
related well with the dissociation constant
of the antibody, all other properties being
equal [75].
If the antibody is to be used for cellular
targeting, however, more complicated rela-
tionships apply. A number of investiga-
tions in tumor targeting have uncovered
some of these trends [7678]. From these
and other studies, it appears that tumor
targeting in general improves with affinity,
but seems to reach a plateau at affinities
around 10
9
M. The steady-state concentra-
tion at the tumor does not get higher with
higher affinity, as the dissociation rate
from the tumor antigen is no longer limit-
ing steady-state concentration once disso-
ciation is very slow. Instead, cellular up-
take and bulk flow become dominant pa-
rameters, and these are influenced by the
format of the molecule. One should draw
such conclusions on the importance of af-
finity only from comparisons of molecules
which are point mutants of each other, but
have otherwise exactly the same format
as different epitopes (some eliciting anti-
gen internalization, others preventing di-
merization), formats, molecular size, etc.,
will change targeting efficacy for different
reasons (see Section 2.4.3).
Affinity is increased in the immune sys-
tem by somatic mutation [7981]. Space
limitations do not allow us to discuss in
detail strategies for the affinity maturation
of antibodies using the various display
technologies, and how the selection of
high-affinity binders can be favored. This
is, however, now possible in a variety of
ways and the interested reader is referred
to a number of articles [8287].
2.4.3
Format and its Impact on Pharmacokinetics
An important factor to consider is also the
format of the final therapeutic molecule.
The availability of bivalent immunoglobu-
lins as well as monovalent Fab and scFv
fragments makes valency an important
consideration in drug design. If the pro-
tein to be blocked is soluble and contains
only one copy of the epitope, as is, for ex-
ample, the case in many cytokines and
protein hormones, no interaction strength
is gained by having IgG molecules over
Fab fragments or scFv fragments. This is
also true for targets on the cell surface, if
they are so far apart that the two arms of
an IgG cannot reach two identical epi-
topes. In this case again, the binding affin-
ity and thus the blocking affinity of a
monovalent antibody fragment will be
identical to that of a bivalent one.
Nevertheless, the longer half-life of IgG
molecules may be advantageous as it guar-
antees a longer duration of the blocking
function. The half-lives of therapeutic anti-
bodies have been reviewed [88] and were
mentioned in Section 2.2 in the context of
immunogenicity (see also Part VI, Chap-
ters 1 and 2). The longer half-life of an
IgG is caused not only by the higher mo-
lecular weight of the IgG and thus the in-
ability to be cleared through the kidney,
but mostly because of a particular mecha-
nism selectively protecting IgG from nor-
mal serum protein catabolism [89]. Over 5
days, the human vascular endothelium en-
gulfs all serum by endocytosis. The con-
tent is processed through a complex net-
work of endosomes and tubules with de-
creasing pH. The neonatal Fc receptor
(FcRn; also termed Brambell receptor after
its discoverer) is expressed in hepatocytes,
endothelial cells and phagocytic cells of
the reticuloendothelial system, the main
location of protein catabolism. The recep-
tor binds the Fc part of IgGs using
charged histidines and thus prevents anti-
bodies from ending up in lysosomes. In-
stead, FcRn with the bound IgG recycles
to the same cell surface, releasing the in-
tact IgG at the serum pH of 7.4. By this
mechanism, the half-life of IgG is in-
creased by a factor of 10 compared to the
absence of this receptor in transgenic ani-
mals [89, 90]. In short, the use of whole
IgG guarantees a long blocking function
through its long half-life, even if it binds
only monovalently to its target.
Murine IgG does not bind to human
FcRn, and this explains the shorter half-
life of murine antibodies in human pa-
tients, typically 12 to 48 h [91]. The half-
lives of endogenous human IgG isotypes
have been well studied and they do differ
3 weeks for IgG1, IgG2 and IgG4, while
IgG3 has a half-life of 1 week [88, 92].
Abciximab (ReoPro), mentioned in Sec-
tion 2.3.1 as an example of a chimeric anti-
body, is unique among therapeutic antibod-
ies marketed thus far in being a Fab frag-
ment. Its half-life in plasma is only 20
30 min [93], but when interacting with
platelets, this rises to 4 h. It is now clear
that the antibody is dynamically redistribut-
ed between individual target molecules and
platelets in less than 1 h. Thus, while the
half-life of dissociation from each individual
integrin molecule is rather fast, the high lo-
cal concentration of integrin molecules on
the platelets provides the drug with a long
platelet-bound half-life, with antibody de-
tected on platelets as long as 2 weeks after
therapy. This leads to a prolonged inhibition
of platelet function in response to shear
stress for 72 h to 1 week [94].
In the targeting of solid tumors, the sit-
uation is again far more complicated.
Large IgG molecules, because of their long
life time in serum, maintain a very high
steady-state concentration. From this reser-
voir, levels at the tumor can reach very
high percentages (2030% of the injected
dose per gram), but tumor to blood ratios
are very small. The problem is that diffu-
sion of large proteins such as IgG through
a solid tumor is very slow and inefficient,
because the antibody is in competition
with removal by bulk flow. In addition, the
tumor has a high hydrostatic interstitial
pressure, is heterogeneous in composition
and density of antigen expression, and has
reduced vasculature. Furthermore, because
of the slow accumulation and the long
half-life, the antibody will at no time be
truly selectively enriched at the tumor
(when expressed as the percentage of the
injected dose per gram of tissue or blood).
This essentially constitutes a limitation
on the use of toxic molecules or radioac-
tive isotopes (see Section 2.5.7) being con-
jugated to the antibody in many applica-
tions: at a dose approaching the toxicity
limit, the tumoricidal effect is often not
yet reached [95] (see also Part V, Chapters
1 and 6). The dose-limiting organ is usual-
ly determined by the action of the free
drug, as some amount of drug can be
cleaved non-specifically. Since IgGs are
mainly degraded in the liver, there may be
concerns for liver toxicity as well for anti-
bodytoxin conjugates. In contrast, for
radionuclides, bone marrow is usually the
dose-limiting organ (see below and Section
2.5.7), as some radionuclides (e.g., yttri-
um) can be incorporated in bone marrow
[96] (see also Part V, Chapters 4 and 5).
Radioimmunotherapy provides an exam-
ple of a setting in which a shorter half-life
can be advantageous. Currently, two anti-
bodyradioisotope conjugates are on the
market, both for the treatment of non-
Hodgkins lymphoma, i.e., ibritumomab
tiuxetan (Zevalin) and tositumomab (Bex-
xar) (see also Part V, Chapter 7). Zevalin
and Bexxar carry
90
Y and
131
I, respectively,
but both are mouse antibodies. As detailed
above, in the radioimmunotherapy setting,
one of the main challenges is maximizing
the dosage of radioactivity reaching the tar-
geted tumor cells without delivering danger-
ous levels of non-specific radiation to or-
gans, notably the bone marrow, the organ
where hematopoietic stem cells are gener-
ated, the precursors of all blood cells. This
balancing act requires that the antibody
has a relatively short half-life, which is the
reason why murine antibodies have been fa-
vored in this setting. The half-lives of the
two marketed products illustrate this point:
both ibritumomab tiuxetan [97, 98] and tosi-
tumomab [99] have half-lives of 13 days. To
safely use rituximab in radioimmuno-
therapy, a chimeric antibody whose antigen
(CD20) is the same as that of tositumomab
would require dosimetry of individual pa-
tients [100].
In other words, the degree of human-
ness of an antibody, or in molecular terms,
the lower affinity to the FcRn and thus the
lack of selective prevention of degradation,
can be used to achieve a half-life that is re-
quired for a particular therapeutic window.
It is likely, however, that in the future fully
human IgGs will be engineered with de-
creased FcRn receptor binding to obtain a
particular half-life.
Smaller protein molecules (Fab frag-
ments, scFv fragments) localize to a solid
tumor much faster and also diffuse better
through tumor tissue, but because of their
faster excretion rate the steady-state level
reached in serum is much lower. Particu-
larly below about 2550 kDa, clearance
through the kidney becomes possible [101,
102]. It should be noted that this also
shifts safety concerns for antibodies conju-
gated with a toxic moiety from liver toxici-
ty (for large proteins) to kidney toxicity
(for smaller proteins), as a fraction of the
recombinant molecules can be taken up by
kidney parenchyma cells.
The increase in functional affinity to cell
surfaces by having multiple binding sites
previously a hallmark of IgGs has now
been engineered into scFv and Fab frag-
ments as well [103]. This avidity effect
strongly increases the residence time on a
surface-bound target molecule, provided
that the antibody can reach two epitopes
simultaneously. In combination with site-
specific PEGylation remote from the anti-
gen-binding site, the hydrodynamic prop-
erties and number of binding sites can
now be engineered independently to
achieve a compromise in the quest to opti-
mize tumor targeting [79].
From these considerations it is clear that
the optimal format of the antibody is de-
pendent on the exact mode of action, its
location, the effector mechanism, whether
the antibody has been fused or conjugated
with a toxin and what kind of toxin or
toxic radioisotope is used. The use of an
IgG is thus only one of many options. The
advent of molecular engineering is thus
pivotal to the further development of these
therapeutic modalities.
2.5
Exerting an Effect at the Target
Ever since the first attempts to create anti-
bodies for human therapy it has been rec-
ognized that, in most cases, mere binding
to the target is necessary, but may not be
sufficient. There may be additional pre-
requisites for an antibody to be a biophar-
maceutical, including the blockade of a
particular interaction and/or cell killing.
The largest number of therapeutic anti-
bodies in development is in the area of
cancer. This area also provides the most
examples of the variety of ways in which
antibodies can be used in medicine.
Whereas in other diseases blockade of a
particular interaction may be sufficient to
have a therapeutic effect, the objective in
cancer is to kill tumor cells, which usually
requires some form of direct cytotoxicity.
It is generally assumed [104, 105] that ef-
fective tumor killing by a naked antibody
will use one or a combination of (1) block-
ing a growth signal, (2) delivering an inhi-
bitory signal, (3) inducing apoptosis and
(4) eliciting an immune response against
the tumor. The relative importance of
these factors depends on the tumor type
and the targeted antigen. This section con-
siders several such approaches in the con-
text of anticancer drug development.
2.5.1
Blockade
Bevacizumab (Avastin) is a humanized
antibody that is approved for the treatment
of metastatic colorectal cancer [106] and is
directed against vascular endothelial cell
growth factor (VEGF), a molecule that
stimulates angiogenesis. The antibody
binds VEGF and thereby inhibits vasculari-
zation of tumors that overexpress the
growth factor; preclinical studies show
clear inhibition of tumor growth in a
mouse xenograft model [107]. Bevacizu-
mab is therefore a rare example of an anti-
cancer antibody which exerts its effect by
blocking a growth factor which is impor-
tant for tumor cell proliferation, but with-
out interacting directly with the cancer
cell.
2.5.2
Naked Antibodies that Trigger Cell Killing
From a drug development perspective, an
antibody that exerts its therapeutic effect
without the requirement for further modi-
fication has a major advantage. Steps that
require conjugation, chemical modification
or new production methodology all add
complexity, cost and risk to the develop-
ment of a therapeutic antibody. Naked an-
tibodies, which can be made using estab-
lished, well-characterized cell lines and
production/purification methods avoid
these difficulties.
Several antibodies have been suggested
to kill tumor cells by ADCC or CDC, or a
combination of these effects. As noted
above, CAMPATH is an example of an an-
ticancer antibody that seems to work via
such effects. Rituximab (Rituxan) is a chi-
meric antibody against CD20 that is ap-
proved for the treatment of non-Hodgkins
lymphoma [108]. The function of CD20, a
well-known marker for B cell activation, is
not precisely known, but it has been sug-
gested to be involved in Ca
2+
influx as a
tetrameric molecule. Direct effects of Ri-
tuximab, including growth inhibition and
apoptosis, have been shown in vitro and a
cross-linking of lipid rafts by the antibody
has been proposed, with a trans-activation
of src kinases eventually leading to apopto-
sis [109]. However, it is unclear whether
this contributes to the clinical benefit ob-
served. In fact, the published data suggest
that the predominant effector mechanism
is ADCC, with a minor role played by
complement [110]. This is the same target
as that against which two radionuclide-
conjugated antibodies (Zevalin and Bexxar;
see Section 2.5.7) are directed, where the
therapeutic mechanism is thought to be at
least partially dependent on radiation dam-
age.
A recent example of an anticancer anti-
body in development that functions via
apoptosis is 1D09C3, a HuCAL-derived
antibody specific for human leukocyte
antigen (HLA)-DR [111, 112], which is
highly expressed in B and T cell lympho-
mas. At the time of writing, this antibody
is about to enter clinical trials for B cell
lymphomas. The antibody causes cell kill-
ing selectively in activated tumor cells
without the need for exogenous effector
cells, via a mechanism that is caspase in-
dependent.
Safety concerns with drug- or radionu-
clide-conjugated antibodies are mostly due
to the systemic effect of the toxic effector
molecules. This problem is augmented by
the high dose required, which in turn is
due to the insufficient localization of the
antibody, notably to solid tumors.
In contrast, naked antibodies are not
systemically toxic per se. Of course, a par-
ticular antibody can be toxic by its biologi-
cal effect on the chosen target either at
the intended tumor site or at another site
where the target is expressed. For example,
Herceptin shows incidence of cardiotoxici-
ty, especially in combination with anthra-
cyclines [113]. This is caused by HER2
being expressed in cardiac myoblasts,
where it is involved in muscle spindle
maintenance.
2.5.3
Modifying the Fc Portion
to Enhance Effector Cell Recruitment
As antibodies are often able to trigger cell
killing via Fc-mediated effector functions
such as ADCC and CDC, much effort has
been put into increasing this effect by
modification of the Fc portion. A crucial
experiment underlining the importance of
the different receptors involved in mediat-
ing ADCC was reported recently [65].
While activator receptors, such as
FccRIIIA on NK cells, are responsible for
arresting tumor growth by Herceptin in a
transgenic mouse model, FccRIIB recep-
tors were found to be inhibitory to the kill-
ing action mice deficient in this receptor
showed much more pronounced ADCC. It
is therefore tempting to modulate the rela-
tive binding to the activating and inhibi-
tory receptor by engineering the Fc part.
Using a systematic series of point mu-
tants in the Fc part, Presta and colleagues
[114] identified residues which discrimi-
nate between binding to FccRI, FccRIIB
and FccRIIIA. A further complication is
introduced by the presence of an allelic
variant in human FccRIIIA, which binds
differentially to the mutants. Using a triple
mutant IgG (S298A/E333A/K334A), im-
proved binding to FccRIIIA as well as
more potent cytotoxicity in ADCC assays
was observed [115]. Interestingly, in the re-
cent crystal structure of the IgG1Fc/
FccRIIIA complex, only one of these resi-
dues (Ser298) was found to make a direct
contact to the receptor.
Using the crystal structure of the Fc/Fc
receptor complex as a guide, other muta-
tions in the protein sequence of the Fc
part have been designed that were pre-
dicted to increase the interaction with the
receptor (Dahiat et al. unpublished; Xen-
cor).
The Fc part is glycosylated on an aspara-
gine residue (Asn297), and while the re-
cent crystal structure of a complex of an
Fc with the FccRIII and FccRIIA [116120]
shows that the oligosaccharide makes only
minimal contact to the receptor, it has
been known for a long time that efficient
binding to the receptors requires glycosyla-
tion [121]. Structurally, the oligosacchar-
ides attached to the conserved Asn297 of
IgG are a biantennary type with a core
heptasaccharide that consists of four N-
acetylglucosamines (GlcNAc) and three
mannoses, and variable fucose addition to
the core at the first GlcNAc residue [121].
From the investigation of the crystal struc-
ture of the Fc/FcR complex and a series of
glycosylation truncation mutants, it be-
came clear that the sugars act both to in-
crease the distance (with the wild-type
being optimal) and to decrease mobility of
the receptor-interacting segments of C
H
2
domains [122].
One approach to improve FccRIII inter-
actions has been to add a bisecting
GlcNAc sugar, by using engineered Chi-
nese hamster ovary (CHO) cells expressing
a glycosyl transferase [b(1,4)-N-acetylgluco-
saminyltransferase III] (see also Part IV,
Chapters 2, 5 and 7). The resulting anti-
body killed neuroblastoma cells at 10- to
20-fold lower concentrations than when
this bisecting GlcNAc was not present
[123].
Recently, another strategy of glyco-engi-
neering was reported to lead to an in-
creased binding to FccRIIIA and thereby
to enhanced ADCC [124], i.e., the produc-
tion of the antibody in host cells that do
not add fucose. In a direct comparison,
this strategy was more effective than add-
ing the bisecting GlcNAc.
In an independent series of experiments
using a CHO-derived cell line (lec 13) that
is unable to add fucose, this lack of fucose
was shown to have no effect on binding to
FccRI (CD64) and only a marginal effect
on binding to FccRIIA and FccRIIB
(CD32), but led to a significant increase in
the binding to FccRIIIA (CD16) [125] and
even led to a further increase of binding
of the engineered triple mutant IgG1
S298A/E333A/K334A, which indeed trans-
lated to improved ADCC in vitro [126]. In-
terestingly, the lack of fucose correlates
with an increased receptor on-rate, sug-
gesting Fc stabilization in the active con-
formation, while the point mutations lead
to a decreased off-rate, suggesting higher
interaction strength [127].
In contrast, the presence and absence of
fucose had no effect on binding to FcRn
(which controls half-life) or C1q (which
controls complement activation). The bind-
ing of the Fc part to the neonatal receptor,
FcRn, can also be influenced by mutations
[128, 129]. These mutations must, how-
ever, be designed to differentially affect the
binding to the receptor in a pH dependent
manner, such that recycling of the IgG
works properly. Encouraging mouse ex-
periments have been reported (summa-
rized in Presta [125]), but clinical trials in
patients have not yet been carried out.
The role of carbohydrate in the clearance
rate of IgG remains unclear. Binding of
aglycosyl IgG to the FcRn appears identical
to that of the glycosylated form. Neverthe-
less, there is some disagreement in the lit-
erature about the influence of glycosyla-
tion on the rate of clearance. While some
studies detected a difference, others did
not (summarized in Shields et al. [114]).
2.5.4
Low-molecular-weight Drug Conjugates
New effector functions may be needed that
exceed the efficacy of the Fc part itself,
since, in a particular tumor setting, ADCC
and CDC may either not be elicited well
or even not at all or not be effective enough
by themselves. For many years, antibodies
have been conjugated to cytotoxic agents
with the intention that the antibody tar-
gets the desired payload to the tumor, thus
providing the proverbial magic bullet.
This has been a long and winding road,
with the therapeutic window frequently
not opening wide enough between systemic
toxicity and lack of efficacy. Nevertheless,
improved molecular understanding, and
consequently better molecular design, is
likely to give these approaches a renewed
chance. While totally selective targeting
may never be achieved, engineering of the
antibody with regard to target epitope, affin-
ity, selectivity, molecular size and thus phar-
macokinetics, with the consequences of de-
gradation of the non-localized drug-loaded
antibody firmly in mind, may increase tu-
mor selectivity above the threshold of thera-
peutic utility, such that more toxic conju-
gates can be used as drugs. IgGs may not
be the preferred molecules for such ap-
proaches and the advent of recombinant
technologies has greatly increased the num-
ber of possibilities regarding the molecular
format.
The conjugation of antibodies to low-mo-
lecular-weight cytotoxic agents has been in-
vestigated for many years. The target in
these cases should be an internalizing sur-
face protein, as most small molecule drugs
act as inhibitors of cell replication and
therefore need to reach the cytoplasm or nu-
cleus to exert their effect [130]. A case in
point is gemtuzumab ozogamicin (Mylo-
targ), the only antibody-based drug based
on this approach to reach the market. Mylo-
targ is a chimeric anti-CD33 antibody conju-
gated to the highly potent enediyne drug ca-
licheamicin and is approved for the treat-
ment of acute myeloid leukemia [131].
CD33 belongs to a growing family of sialic
acid-binding, immunoglobulin-like lectins
(siglecs). It appears to be an inhibitory re-
ceptor in myeloid lineage development
[132] and is highly expressed in myeloid leu-
kemia. The conjugation of the drug to the
antibody gives a more favorable therapeutic
window compared to the drug alone, de-
spite the fact that the marketed preparation
is heterogeneous, comprising a significant
fraction of unconjugated antibody.
The antibody BR96, which recognizes an
extended form of the Lewis Y carbohydrate
antigen present on many carcinomas [133,
134], has shown some promise as a conju-
gate with different small molecule drugs.
Early attempts with doxorubicin (DOX), a
drug of the anthracycline family, conju-
gated to BR96 provided excellent preclini-
cal data. Nevertheless, phase II clinical
trials of BR96DOX for the treatment of
metastatic breast cancer or gastric carcino-
ma showed limited efficacy, with elevated
gastrointestinal toxicity, probably a conse-
quence of the target being also expressed
in healthy gastric epithelial cells [135, 136].
More recently, synergistic antitumor activ-
ity in animal models has been demon-
strated for BR96DOX in combination
with the taxanes docetaxel and paclitaxel
[137]. The conjugate is currently in clinical
development for the treatment of non-
small cell lung carcinoma in combination
with docetaxel.
That factors beyond the antibody and cy-
totoxic agent are crucial in creating an ef-
fective biopharmaceutical is demonstrated
by recent work with BR96. Conjugates
comprising BR96 linked, via two different
chemistries, to cytotoxic auristatin deriva-
tives have shown promise in animal stud-
ies [138]. Clear superiority in this study
was achieved by incorporating an enzyme-
cleavable linker between antibody and cy-
totoxic agent, thereby increasing the effi-
ciency and specificity of release of drug at
the target.
In general, antibodydrug conjugates
should be based on very highly potent
small molecule drugs (see also Part V,
Chapter 6). In addition to the cytotoxic
agents mentioned above, examples of
other small molecule drugs being investi-
gated include tubulin polymerization inhi-
bitors such as the maytansanoids, CC1065
and taxoids [130, 139]. As mentioned
above, the linker chemistry, being either
acid labile or enzymatically cleavable, pos-
sibly with matrix metalloproteinases as tu-
mor-specific release mechanisms, are also
factors being studied. While preclinical
data are impressive, it remains to be seen
whether a useful therapeutic window can
be found for broad application of toxin im-
muno-conjugates in oncology.
2.5.5
Protein Toxin Conjugates
A conceptually similar approach is the
conjugation of protein toxins to antibodies.
Such toxins, typically from plants or bacte-
ria, are enzymes that catalytically inactivate
essential cellular processes such as transla-
tion. By covalently modifying a translation
factor or the ribosome itself in an enzy-
matic process, a single enzyme molecule
is sufficient to kill a cell [140142]. The
best clinically studied members of this
group are Pseudomonas exotoxin A, a tri-
partite protein that enzymatically ADP-ri-
bosylates elongation factor 2, and ricin, de-
rived from the plant Ricinus communis,
which modifies a critical nucleotide in eu-
karyotic ribosomal RNA. The natural tox-
ins are produced with their own, unspecif-
ic uptake mechanism that allows them to
infect any cell, exploiting receptor mole-
cules ubiquitously expressed on mamma-
lian cells. By deleting these cell-binding
domains and replacing them by an inter-
nalizing antibody, typically in the single-
chain format, tumor-selective killing can
be achieved. The antibody thus mediates
uptake of the enzyme by tumor cells.
Nevertheless, toxicity remains an issue
for systemic applications of these immu-
notoxins, as does immunogenicity of the
toxin part, which limits repeated dosing.
In order to reach therapeutic levels in a
solid tumor, high initial doses have to be
used, but liver toxicity was observed in a
trial against HER2 [143] (see also Part I,
Chapter 5). Therefore, more recent work
has focused on applications in leukemia
and lymphoma. Encouraging results were
observed with an immunotoxin against
CD22 in patients with hairy cell leukemia
and chronic lymphocytic leukemia [144].
CD22 is a member of the siglec family,
and serves as a receptor for sialic acid-
bearing ligands expressed on erythrocytes
and all leukocyte classes. CD22 appears to
be primarily involved in the generation of
mature B cells within the bone marrow,
blood and marginal zones of lymphoid tis-
sues [145].
A combined phase I/II trial for ricin
conjugates with antibodies against the
lymphocyte activation markers CD25 (the
IL-2 receptor a-chain) or CD30 (a member
of the TNF receptor superfamily, possibly
involved among others in memory T cell
development) for patients with Hodgkins
lymphoma showed some promise [146].
2.5.6
Cytokine Fusions
Since many tumors do elicit an immune
response, albeit one which may be unable
to eradicate the tumor, an attractive strat-
egy would appear to be to enhance this re-
sponse with immunostimulatory cytokines.
In order to localize the cytokine to the tu-
mor, fusion proteins with antibodies have
been made. Constructs investigated in-
clude interleukin (IL)-2, IL-12, granulocyte
(GM-CSF) and members of the TNF
superfamily [147, 148] (see also Part V,
Chapters 1 and 6). In a recently reported
phase I trial of a humanized mAb directed
against the GD2 disialoganglioside, revers-
ible clinical toxicities were reported to-
gether with the desired immune stimula-
tion [149]. As is the case with bispecific
antibodies (see below), the main challenge
in these approaches will be to prevent sys-
temic engagement of the cytokine receptor
by the cytokine part of the conjugate in
the absence of the antibody binding to the
tumor, as this is the most likely source of
side-reactions. The severity of the problem
will depend on the complex interplay of
pharmacokinetics of the antibody reaching
the tumor or the cytokine receptor.
2.5.7
Antibody-Radioisotope Conjugates
Some aspects of radioimmunotherapy
have already been discussed above in the
context of half-life. The conjugation of an
antibody to a radioactive element that
should cause radiation damage at the tu-
mor site is an idea that has been pursued
for a number of years [95] (see also Part
II, Chapter 5, and Part V, Chapters 5 and
7). Again, progress with solid tumors has
been modest, while encouraging results
are obtained in the treatment of hemato-
poietic neoplasms. As noted above, the fact
that [
131
I]tositumomab (Bexxar) and
[
90
Y]ibritumomab tiuxetan (Zevalin), the
only FDA-approved radiolabeled therapeu-
tic antibodies, are both of murine origin
contributes significantly to a shorter half-
life, which is desirable in this setting, but
also creates a considerable immune re-
sponse. This, however, is diminished in
patients with hematopoietic disorders or
prior chemotherapy. Interestingly, large
quantities of the unlabeled antibody must
be administered prior to or concomitantly
with the radioconjugate to improve target-
ing. The relatively low dose that is suffi-
cient for treating hematopoietic malignan-
cies reduces adverse side-effects and may
be the reason why a therapeutic window
can be found in this case.
Improvements for solid tumors may po-
tentially come from the use of pre-target-
ing strategies [95]. In this case the radio-
nuclide is not directly coupled to the anti-
body. Instead, an antibody Fab fragment
with a hapten binding function (e.g., a bi-
specific antitumor antihapten construct)
is injected first and allowed to concentrate
at the tumor. Once the majority has left
the circulation, the radionuclide, coupled
to a monomeric or dimeric hapten [95] is
injected. The hapten derivatives extremely
short half-life, combined with the newly
generated binding sites by the noninternal-
izing Fab fragment on the tumor, if pres-
ent on a nonshedding surface antigen, al-
low excellent tumor selectivity. Neverthe-
less, it remains to be seen whether a use-
ful therapeutic window can be obtained,
with the concern of potentially new dose-
limiting mechanisms for uptake of the
radionuclide.
2.5.8
Bispecific Antibodies
Attempts have also been made to use bi-
specific antibodies to recruit effector cells.
A number of challenges need to be over-
come in this field. First, a robust method
must be found by which such proteins can
be produced. Initially, the co-expression of
two antibodies in one cell and the separa-
tion of the one desired out of the 10 con-
ceivable molecular forms did not seem an
attractive proposition. However, a multi-
tude of methods have been reported over
the last few years [103, 150] to create bi-
specific formats of the antibody with a de-
fined molecular composition. These in-
clude (1) the co-expression of heavy chains
that have been engineered to allow only
the desired pairing, (2) the direct chemical
linkage of two different Fab fragments, (3)
a number of different bispecific recombi-
nant antibody constructs based on scFv
fragments fused to heterodimerizing pep-
tides and proteins or (4) the direct en-
forced pairing of the domains in so-called
diabodies (see also Part IV, Chapter 16 and
Part V, Chapter 1). It would be too early to
favor one form over the others.
A second challenge is derived from the
fact that binding with only one of the arms,
e.g., the one engaging the effector cells, is a
likely intermediate in the reaction: no sys-
temic toxicity should result in such a case
so as to avoid safety concerns. It is very likely
that the binding to a solid tumor is much
slower than binding to cells of the immune
system found in the serum. The third chal-
lenge is the converse the antibody may
eventually bind to the tumor, but never
reach an effector cell, because there is none
there, or, for geometric reasons, its receptor
cannot be reached or activated.
Factors such as these translate into prac-
tical limitations. Typically, excessively high
concentrations of bispecific antibody are
needed to see an effect in vivo. In addition,
particularly in solid tumors, the effector
cells are often ineffective in the absence of
a local co-stimulatory signal, which usually
requires the addition of an exogenous fac-
tor, a serious drawback for a viable thera-
peutic strategy.
A number of strategies have been devel-
oped to use bispecific antibodies to recruit
different types of effector cells. Much of
the earlier work sought to recruit effector
cells via the IgA receptor FcaRI or the IgG
receptors FccRI or FccRIII. In a phase I/II
trial in 16 patients, a bispecific anti-
CD30anti-FccRIII construct led to one
complete and three partial remissions plus
four cases of stable disease [151]. Pretreat-
ment with IL-2 cytokine resulted in augmen-
ted antitumor activity, possibly by an addi-
tional mechanism of activation of NK cells.
In another trial, a bispecific anti-CD30an-
ti-FccRI construct was tested in 10 Hodg-
kins lymphoma patients [149]: one complete
and three partial remissions, plus four cases
of stable disease were reported.
Quite in contrast to the situation with
lymphoma, no responses were seen with
solid tumors, using a bispecific anti-
HER2anti-FccRI antibody in combination
with interferon-c or GM-CSF [152] (see
also Part VIII, Chapter 3).
Attention has also focused on recruiting
cytotoxic T lymphocytes (CTLs) using CD3
as the trigger. Limitations of the type men-
tioned above have again hampered pro-
gress in this field: to date, no clinical effi-
cacy has been observed on systemic ad-
ministration of anti-CD3 based bispecific
antibodies [150].
A newer technology, which seeks to over-
come some of the disadvantages of previous
bispecific CTL approaches, is unique in
comprising two single-chain Fv fragments
linked in tandem [150]. These types of mol-
ecule have been termed bispecific Tcell en-
gager or BiTE constructs. It also has an
anti-CD3 recruitment arm and a second
specificity directed against a tumor marker.
The molecules tested seem to have two ad-
vantages over other CTL-recruiting bispeci-
fic antibodies described previously: (1) they
do not appear to require co-stimulation of
Tcells and (2) they appear to catalyze killing
of multiple target cells by a single T cell.
These advantages, if they should translate
into the clinic, might offer significant dos-
ing and cost-of-goods benefits for therapeu-
tic antibodies of this type.
2.6
Antibodies in their Natural Habitat:
Infectious Diseases
There is one field of medicine where the
antibody in its classical format may indeed
constitute the optimal molecular design
in the defense against infectious agents,
i.e., the normal function of an antibody.
Palivizumab (Synagis) is a humanized
mAb that prevents lower respiratory tract
infection by respiratory syncytial virus, and
it is used in pre-term infants and other
young children at risk, such as with
cardiopulmonary disease or immunosup-
pression [153].
In general, however, passive immuno-
therapy has not been the focus of many de-
velopment projects, as small molecule anti-
viral and antibacterial agents would clearly
be advantageous for ease of administration
provided they are available. However, the
increasing spread of resistant strains of
viruses and bacteria not least caused by
the indiscriminate use of antibiotics may
become one of the severe medical problems
of the future. The logistics of passive immu-
notherapy, having available large doses of
the required specificities in due time, are
daunting, but perhaps not insurmountable
with better production techniques, the use
of well-designed cocktails of specificity and
concentration on infections of great risk to
global health. The multiple cases of out-
break of novel deadly diseases over the last
few years, AIDS, SARS and Ebola fever, to
name just a few, illustrate the need for rapid
development of novel containment strate-
gies. It is likely that recombinant antibodies
will still have a role to play in this respect, as
it may be faster in some cases to develop
protective antibodies than a vaccine (see
also Part II, Chapters 7 and 8).
2.7
Opportunities for New Therapeutic
Applications Provided
by Synthetic Antibodies
The advent of synthetic antibodies now al-
lows a number of the previous key limita-
tions in using antibodies for therapy to be
addressed. In addition to solving the prob-
lem of immunogenicity, which may be a
factor preventing therapeutic utility, as ex-
plained above, the new technologies for
tailor-making the antibody molecule per-
mit new approaches to improve efficacy.
First and foremost, the relevant epitope
can be more easily targeted. If its location
is known at the molecular level, selections
for such binders are possible, as has been
delineated above, and this will often be the
decisive factor in determining whether a
particular antibody has any in vivo potency
at all. Secondly, affinity can be addressed
independently. Again, as described above,
high affinity is almost always a benefit. In
the traditional immunizations of animals,
even when a large number of different
mAbs is obtained, there is no guarantee
that high-affinity antibodies from the pan-
el are directed against the relevant epitope.
As a rule, other places on the large surface
of a protein will give many opportunities
for high-affinity binding, such that it
would be rather unexpected that the high-
est affinity antibodies happen to be direct-
ed against the epitope of choice.
It is this particular situation, where a
synthetic library technology for antibody
generation such as HuCAL can play out to
its full advantage: affinity maturation strat-
egies can be applied to the antibodies tar-
geting the relevant epitope. Therefore, hav-
ing any lead compound will usually allow
the generation of a molecule, which main-
tains the binding at the desired location
but achieves an affinity commensurate
with the desired mode of action.
Because of the modularity of the HuCAL
design (see above), it is possible to tailor
not only one, but several antibodies that
have been identified as having the desired
binding specificity. Since identical restric-
tion sites in all members of HuCAL flank
the CDR cassettes, they can be exchanged
in several antibodies at once, and those
with higher affinity can be selected. There-
by, it is possible to walk across the whole
antigen-binding site and replace each CDR
in turn. Since the cassettes are not ran-
dom, but instead reflect the composition
of the human repertoire, key structural
properties of the antibody can be main-
tained during optimization. By separating
the selection of efficacy (or binding epi-
tope) from affinity, which can be subse-
quently improved for any hit, antibodies
can be made to almost any specification.
The anti-HLA-DR antibody 1D09C3
mentioned above [111, 112] provides a
good example of how antibody engineer-
ing can be applied to optimize antibody
properties. The antibody was derived from
a screen of a HuCAL library in the single-
chain Fv format, which yielded a number
of binders that, although of moderate af-
finity, efficiently killed tumor cells. The af-
finity of the initial lead antibodies was in-
creased by sequential replacement of the
L-CDR3 and L-CDR1, utilizing the modu-
larity of the HuCAL gene library (see
above). This process enables retention of
the epitope specificity of the initial lead
antibodies, which are vital for their cell-
killing properties, while achieving the de-
sired level of target affinity.
In practice, the ability to generate anti-
bodies having particular properties from li-
braries is limited by the screen that is em-
ployed. Initial screens for affinity and/or
specificity are used to reduce the number
of potential hits from the initial library of
10
10
to 10
3
10
5
, which must then be
screened for a particular biological func-
tion. As delineated above, the (usually few)
binders with a biological effect can then
be affinity matured, such that affinity does
not have to be an initial selection criterion.
HuCAL thus provides a systematic means
of screening large parts of sequence and
structure space for antibodies with prede-
fined properties. Importantly, this can be
done while retaining the intrinsic human
composition of antibodies emerging from
primary screens of the library.
2.8
Future Directions and Concluding
Statements
Antibody engineering has clearly helped to
solve a number of problems that have
hampered early attempts in successfully
using mAbs as biopharmaceuticals. Be-
cause of the lengthy development times
typical of drug development, many thera-
peutic antibodies entering the market re-
cently have still been made with earlier
technologies and did not even fully profit
from the possibilities available today.
A pivotal development of the last few
years is that it is now possible to make a
human antibody to practically any specifi-
cation regarding epitope, specificity,
mode of binding, affinity, format and any
molecule that might be linked with it.
While robust manufacturing of more com-
plicated molecules still needs to be im-
proved, the main processes for manufac-
turing recombinant IgGs and fragments
thereof are established.
Encouraging progress has been made in
the area of neoplastic diseases of the hema-
topoietic system (lymphomas and leuke-
mias). These tumors are characterized by
good accessibility to the drug and the bodys
immune defense, high antigen density and
perhaps a more homogeneous tumor cell
population. In contrast, progress in the area
of solid tumors has been only incremental.
This is a huge need for society, and thus a
great opportunity for science and medicine
and the industry. It is apparent that several
therapeutic antibodies are directed not only
against closely related diseases, but also
against the same target (TNF-a, CD20 and
CD25), while the medical need in many
other areas is unmet.
The key challenge for the future is to
back up todays molecular engineering cap-
abilities with a much better molecular un-
derstanding of disease and the conse-
quences of the application of particular
molecules. A more detailed understanding
of the exact molecular effect required for a
particular target (blocking, dimerization,
its prevention, exposure needed) will allow
much better engineering of molecular
properties. In particular, preclinical models
must become more relevant and predic-
tive. This is a difficult topic in complex
diseases such as cancer, as not only the in-
teraction between the therapeutic antibody
and its target is of importance, but also
the multitude of interactions with other
cells and their surface proteins, the com-
plex pharmacokinetics, and the detailed
metabolism. For example, in many mouse
models with tumors of human origin, the
antigen is selectively expressed on the tu-
mor, but not in the murine tissues. To bet-
ter model the human situation, systematic
approaches are dearly needed.
The use of antibodies as biopharmaceuti-
cals to treat some of the most serious dis-
eases affecting mankind today arises direct-
ly from impressive technological develop-
ments that have been made in this field
over the last 20 years. This has been one
of the most significant achievements in
the field of modern biotechnology. The de-
velopments described here promise that
this class of modern biopharmaceuticals
will play an even more important role in
the clinicians armamentarium for the fore-
seeable future. Molecular engineering holds
the promise that the remaining problems,
many of them due to incomplete molecular
understanding of the most important dis-
eases, will be addressable in the future.
Eventually, this will further lead to the ulti-
mate biopharmaceutical, which, for exam-
ple, targets the desired payload to the tu-
mor, thus fully realizing Paul Ehrlichs vi-
sion of the proverbial magic bullet.
Acknowledgments
We would like to thank Drs. Marlies
Sproll, Uwe Zangemeister-Wittke and Mi-
chael Stumpp for critical reading of the
manuscript, and Dr. Annemarie Honegger
for Figs. 2.2 and 2.3.
References
1 Khler, G., Milstein, C. Continuous cultures
of fused cells secreting antibody of predefined
specificity. Nature 1975, 256, 495497.
2 Schirle, M., Weinschenk, T., Stevanovic, S.
Combining computer algorithms with experi-
mental approaches permits the rapid and ac-
curate identification of T cell epitopes from
defined antigens. J. Immunol. Methods 2001,
257, 116.
3 Flower, D. R. Towards in silico prediction of
immunogenic epitopes. Trends Immunol. 2003,
24, 667674.
4 Flower, D. R. Databases and data mining for
computational vaccinology. Curr. Opin. Drug
Discov. Dev. 2003, 6, 396400.
5 Koren, E., Zuckerman, L. A., Mire-Sluis, A. R.
Immune responses to therapeutic proteins in
humans clinical significance, assessment
and prediction. Curr. Pharm. Biotechnol. 2002,
3, 349360.
6 Sturniolo, T., Bono, E., Ding, J., Raddrizzani,
L., Tuereci, O., Sahin, U., Braxenthaler, M.,
Gallazzi, F., Protti, M. P., Sinigaglia, F., Ham-
mer, J. Generation of tissue-specific and pro-
miscuous HLA ligand databases using DNA
microarrays and virtual HLA class II matrices.
Nat. Biotechnol. 1999, 17, 555561.
7 Chapman, A. P., Antoniw, P., Spitali, M.,
West, S., Stephens, S., King, D. J. Therapeutic
antibody fragments with prolonged in vivo
half-lives. Nat. Biotechnol. 1999, 17, 780783.
8 Greenwald, R. B., Choe, Y. H., McGuire, J.,
Conover, C. D. Effective drug delivery by PEG-
ylated drug conjugates. Adv. Drug Deliv. Rev.
2003, 55, 217250.
9 Yang, K., Basu, A., Wang, M., Chintala, R.,
Hsieh, M. C., Liu, S., Hua, J., Zhang, Z.,
Zhou, J., Li, M., Phyu, H., Petti, G., Mendez,
M., Janjua, H., Peng, P., Longley, C., Borows-
ki, V., Mehlig, M., Filpula, D. Tailoring struc-
turefunction and pharmacokinetic properties
of single-chain Fv proteins by site-specific
PEGylation. Protein Eng. 2003, 16, 761770.
10 Clark, M. Antibody humanization: a case of
the Emperors new clothes? Immunol. Today
2000, 21, 397402.
11 Reichert, J., Pavlou, A., Monoclonal antibodies
market. Nat. Rev. Drug Discov. 2004, 3, 383
384.
12 Wagner, C. L., Schantz, A., Barnathan, E., Ol-
son, A., Mascelli, M. A., Ford, J., Damaraju, L.,
Schaible, T., Maini, R. N., Tcheng, J. E. Conse-
quences of immunogenicity to the therapeutic
monoclonal antibodies ReoPro and Remicade.
Dev. Biol. 2003, 112, 3753.
13 Nurden, A. T., Nurden, P. GPIIb/IIIa antago-
nists and other anti-integrins. Semin. Vascular
Med. 2003, 3, 123130.
14 Jones, P. T., Dear, P. H., Foote, J., Neuberger,
M. S., Winter, G. Replacing the complementar-
ity-determining regions in a human antibody
with those from a mouse. Nature 1986, 321,
522525.
15 Ewert, S., Honegger, A., Plckthun, A. Stabili-
ty improvement of antibodies for extracellular
and intracellular applications: CDR grafting to
stable frameworks and structure-based frame-
work engineering. Methods 2004, 34, 184199.
16 Padlan, E. A. A possible procedure for reduc-
ing the immunogenicity of antibody variable
domains while preserving their ligand-binding
properties. Mol. Immunol. 1991, 28, 489498.
17 Roguska, M. A., Pedersen, J. T., Keddy, C. A.,
Henry, A. H., Searle, S. J., Lambert, J. M.,
Goldmacher, V. S., Blattler, W. A., Rees, A. R.,
Guild, B. C. Humanization of murine mono-
clonal antibodies through variable domain re-
surfacing. Proc. Natl Acad. Sci. USA 1994, 91,
969973.
18 Roguska, M. A., Pedersen, J. T., Henry, A. H.,
Searle, S. M., Roja, C. M., Avery, B., Hoffee,
M., Cook, S., Lambert, J. M., Blattler, W. A.,
Rees, A. R., Guild, B. C. A comparison of two
murine monoclonal antibodies humanized by
CDR-grafting and variable domain resurfac-
ing. Protein Eng. 1996, 9, 895904.
19 Helft, P. R., Schilsky, R. L., Hoke, F. J., Wil-
liams, D., Kindler, H. L., Sprague, E., DeWitte,
M., Martino, H. K., Erickson, J., Pandite, L.,
Russo, M., Lambert, J. M., Howard, M., Ra-
tain, M. J. A phase I study of cantuzumab
mertansine administered as a single intrave-
nous infusion once weekly in patients with
advanced solid tumors. Clin. Cancer Res. 2004,
10, 43634368.
20 Tolcher, A. W., Ochoa, L., Hammond, L. A.,
Patnaik, A., Edwards, T., Takimoto, C., Smith,
L., de Bono, J., Schwartz, G., Mays, T., Jonak,
Z. L., Johnson, R., DeWitte, M., Martino, H.,
Audette, C., Maes, K., Chari, R. V., Lambert,
J. M., Rowinsky, E. K. Cantuzumab mertan-
sine, a maytansinoid immunoconjugate direc-
ted to the CanAg antigen: a phase I, pharma-
cokinetic, and biologic correlative study. J.
Clin. Oncol. 2003, 21, 211222.
21 Hale, G., Dyer, M. J., Clark, M. R., Phillips,
J. M., Marcus, R., Riechmann, L., Winter, G.,
Waldmann, H. Remission induction in non-
Hodgkin lymphoma with reshaped human
monoclonal antibody CAMPATH-1H. Lancet
1988, 2, 13941399.
22 Rowan, W., Tite, J., Topley, P., Brett, S. J.
Cross-linking of the CAMPATH-1 antigen
(CD52) mediates growth inhibition in human
B- and T-lymphoma cell lines, and subsequent
emergence of CD52-deficient cells. Immunol-
ogy 1998, 95, 427436.
23 Dyer, M. J. The role of CAMPATH-1 antibod-
ies in the treatment of lymphoid malignan-
cies. Semin. Oncol. 1999, 26, 5257.
24 Hale, G., Rebello, P., Brettman, L. R., Fegan,
C., Kennedy, B., Kimby, E., Leach, M., Lundin,
J., Mellstedt, H., Moreton, P., Rawstron, A. C.,
Waldmann, H., Osterborg, A., Hillmen, P.
Blood concentrations of alemtuzumab and
antiglobulin responses in patients with
chronic lymphocytic leukemia following intra-
venous or subcutaneous routes of administra-
tion. Blood 2004, 104, 948955.
25 Jespers, L. S., Roberts, A., Mahler, S. M., Win-
ter, G., Hoogenboom, H. R. Guiding the selec-
tion of human antibodies from phage display
repertoires to a single epitope of an antigen.
Biotechnology 1994, 12, 899903.
26 Jia, X. C., Raya, R., Zhang, L., Foord, O.,
Walker, W. L., Gallo, M. L., Haak-Frendscho,
M., Green, L. L., Davis, C. G. A novel method
References 1179
of multiplexed competitive antibody binning
for the characterization of monoclonal anti-
bodies. J. Immunol. Methods 2004, 288, 9198.
27 Hoogenboom, H. R., Winter, G. By-passing
immunisation. Human antibodies from syn-
thetic repertoires of germline V
H
gene seg-
ments rearranged in vitro. J. Mol. Biol. 1992,
227, 381388.
28 Hoogenboom, H. R., Lutgerink, J. T., Pelsers,
M. M., Rousch, M. J., Coote, J., Van Neer, N.,
De Bruine, A., Van Nieuwenhoven, F. A.,
Glatz, J. F., Arends, J. W. Selection-dominant
and nonaccessible epitopes on cell-surface re-
ceptors revealed by cell-panning with a large
phage antibody library. Eur. J. Biochem. 1999,
260, 774784.
29 Mutuberria, R., Hoogenboom, H. R., van der
Linden, E., de Bruine, A. P., Roovers, R. C.
Model systems to study the parameters deter-
mining the success of phage antibody selec-
tions on complex antigens. J. Immunol. Meth-
ods 1999, 231, 6581.
30 Traggiai, E., Becker, S., Subbarao, K., Kolesni-
kova, L., Uematsu, Y., Gismondo, M. R., Mur-
phy, B. R., Rappuoli, R., Lanzavecchia, A. An
efficient method to make human monoclonal
antibodies from memory B cells: potent neu-
tralization of SARS coronavirus. Nat. Med.
2004, 10, 871875.
31 Burton, D. R., Barbas, C. F., 3rd, Persson,
M. A., Koenig, S., Chanock, R. M., Lerner, R.
A. A large array of human monoclonal anti-
bodies to type 1 human immunodeficiency
virus from combinatorial libraries of asympto-
matic seropositive individuals. Proc. Natl Acad.
Sci. USA 1991, 88, 1013410137.
32 Marks, J. D., Hoogenboom, H. R., Bonnert,
T. P., McCafferty, J., Griffiths, A. D., Winter, G.
By-passing immunization. Human antibodies
from V-gene libraries displayed on phage. J.
Mol. Biol. 1991, 222, 581597.
33 Vaughan, T. J., Williams, A. J., Pritchard, K.,
Osbourn, J. K., Pope, A. R., Earnshaw, J. C.,
McCafferty, J., Hodits, R. A., Wilton, J., John-
son, K. S. Human antibodies with sub-nano-
molar affinities isolated from a large non-im-
munized phage display library. Nat. Biotechnol.
1996, 14, 309314.
34 Hoogenboom, H. R. Overview of antibody
phage-display technology and its applications.
35 Knappik, A., Ge, L., Honegger, A., Pack, P.,
Fischer, M., Wellnhofer, G., Hoess, A., Wlle,
J., Plckthun, A., Virneks, B. Fully synthetic
human combinatorial antibody libraries (Hu-
CAL) based on modular consensus frame-
works and CDRs randomized with trinucleo-
tides. J. Mol. Biol. 2000, 296, 5786.
36 Sderlind, E., Strandberg, L., Jirholt, P., Ko-
bayashi, N., Alexeiva, V., Aberg, A. M., Nils-
son, A., Jansson, B., Ohlin, M., Wingren, C.,
Danielsson, L., Carlsson, R., Borrebaeck, C. A.
Recombining germline-derived CDR se-
quences for creating diverse single-framework
antibody libraries. Nat. Biotechnol. 2000, 18,
852856.
37 Pini, A., Viti, F., Santucci, A., Carnemolla, B.,
Zardi, L., Neri, P., Neri, D. Design and use of
a phage display library. Human antibodies
with subnanomolar affinity against a marker
of angiogenesis eluted from a two-dimen-
sional gel. J. Biol. Chem. 1998, 263, 21769
21776.
38 Golemis, E. (ed.). ProteinProtein Interactions:
A Molecular Cloning Manual. Cold Spring Har-
bor Laboratory Press, Cold Spring Harbor, NY,
2002.
39 Smith, G. P. Filamentous fusion phage: novel
expression vectors that display cloned antigens
on the virion surface. Science 1985, 228, 1315
1317.
40 McCafferty, J., Griffiths, A. D., Winter, G.,
Chiswell, D. J. Phage antibodies: filamentous
phage displaying antibody variable domains.
Nature 1990, 348, 552554.
41 Bass, S., Greene, R., Wells, J. A. Hormone
phage: an enrichment method for variant pro-
teins with altered binding properties. Proteins
1990, 8, 309314.
42 Barbas, C. F., 3rd, Burton, D. R., Scott, J. K.,
Silvermann, G. J. Phage Display: A Laboratory
Manual. Cold Spring Harbor Laboratory Press,
Cold Spring Harbor, NY, 2001.
43 Ostendorp, R., Frisch, C., Urban, M. Genera-
tion, engineering and production of human
antibodies using HuCAL. In Antibodies, Vol. 2:
Novel Technologies and Therapeutic Use, Subra-
manian, G. (ed.). Kluwer/Plenum, New York,
2004.
44 Boder, E. T., Wittrup, K. D. Yeast surface dis-
play for screening combinatorial polypeptide
libraries. Nat. Biotechnol. 1997, 15, 553557.
45 Feldhaus, M. J., Siegel, R. W., Opresko, L. K.,
Coleman, J. R., Feldhaus, J. M., Yeung, Y. A.,
Cochran, J. R., Heinzelman, P., Colby, D.,
Swers, J., Graff, C., Wiley, H. S., Wittrup, K. D.
Flow-cytometric isolation of human antibodies
from a nonimmune Saccharomyces cerevisiae
surface display library. Nat. Biotechnol. 2003,
21, 163170.
46 Weaver-Feldhaus, J. M., Lou, J., Coleman, J. R.,
Siegel, R. W., Marks, J. D., Feldhaus, M. J.
Yeast mating for combinatorial Fab library
generation and surface display. FEBS Lett.
2004, 564, 2434.
47 Hanes, J., Plckthun, A. In vitro selection and
evolution of functional proteins by using ribo-
some display. Proc. Natl Acad. Sci. USA 1997,
94, 49374942.
48 Hanes, J., Jermutus, L., Weber-Bornhauser, S.,
Bosshard, H. R., Plckthun, A. Ribosome dis-
play efficiently selects and evolves high-affinity
antibodies in vitro from immune libraries.
Proc. Natl Acad. Sci. USA 1998, 95, 14130
14135.
49 Jermutus, L., Honegger, A., Schwesinger, F.,
Hanes, J., Plckthun, A. Tailoring in vitro evo-
lution for protein affinity or stability. Proc.
Natl Acad. Sci. USA 2001, 98, 7580.
50 Hanes, J., Schaffitzel, C., Knappik, A., Plck-
thun, A. Picomolar affinity antibodies from a
fully synthetic naive library selected and
evolved by ribosome display. Nat. Biotechnol.
2000, 18, 12871292.
51 Fishwild, D. M., ODonnell, S. L., Bengoechea,
T., Hudson, D. V., Harding, F., Bernhard, S. L.,
Jones, D., Kay, R. M., Higgins, K. M.,
Schramm, S. R., Lonberg, N. High-avidity hu-
man IgG kappa monoclonal antibodies from a
novel strain of minilocus transgenic mice.
Nat. Biotechnol. 1996, 14, 845851.
52 Mendez, M. J., Green, L. L., Corvalan, J. R., Jia,
X. C., Maynard-Currie, C. E., Yang, X. D., Gallo,
M. L., Louie, D. M., Lee, D. V., Erickson, K. L.,
Luna, J., Roy, C. M., Abderrahim, H.,
Kirschenbaum, F., Noguchi, M., Smith, D. H.,
Fukushima, A., Hales, J. F., Klapholz, S., Fi-
ner, M. H., Davis, C. G., Zsebo, K. M., Jakobo-
vits, A. Functional transplant of megabase hu-
man immunoglobulin loci recapitulates hu-
man antibody response in mice. Nat. Genet.
1997, 15, 146156.
53 Tomizuka, K., Yoshida, H., Uejima, H., Ku-
goh, H., Sato, K., Ohguma, A., Hayasaka, M.,
Hanaoka, K., Oshimura, M., Ishida, I. Func-
tional expression and germline transmission
of a human chromosome fragment in chimae-
ric mice. Nat. Genet. 1997, 16, 133143.
54 Green, L. L., Jakobovits, A. Regulation of B cell
development by variable gene complexity in
mice reconstituted with human immunoglob-
ulin yeast artificial chromosomes. J. Exp. Med.
1998, 188, 483495.
55 Nicholson, I. C., Zou, X., Popov, A. V., Cook,
G. P., Corps, E. M., Humphries, S., Ayling, C.,
Goyenechea, B., Xian, J., Taussig, M. J., Neu-
berger, M. S., Brggemann, M. Antibody re-
pertoires of four- and five-feature translocus
mice carrying human immunoglobulin heavy
chain and kappa and lambda light chain yeast
artificial chromosomes. J. Immunol. 1999, 163,
68986906.
56 Gallo, M. L., Ivanov, V. E., Jakobovits, A., Davis,
C. G. The human immunoglobulin loci intro-
duced into mice: V, (D) and J gene segment
usage similar to that of adult humans. Eur. J.
Immunol. 2000, 30, 534540.
57 Magadan, S., Valladares, M., Suarez, E., San-
juan, I., Molina, A., Ayling, C., Davies, S. L.,
Zou, X., Williams, G. T., Neuberger, M. S.,
Bruggemann, M., Gambon, F., Diaz-Espada,
F., Gonzalez-Fernandez, A. Production of anti-
gen-specific human monoclonal antibodies:
comparison of mice carrying IgH/kappa or
IgH/kappa/lambda transloci. Biotechniques
2002, 33, 680684.
58 Ishida, I., Tomizuka, K., Yoshida, H., Tahara,
T., Takahashi, N., Ohguma, A., Tanaka, S.,
Umehashi, M., Maeda, H., Nozaki, C., Halk,
E., Lonberg, N. Production of human mono-
clonal and polyclonal antibodies in TransChro-
mo animals. Cloning Stem Cells 2002, 4, 91
102.
59 Van den Brande, J. M., Braat, H., van den
Brink, G. R., Versteeg, H. H., Bauer, C. A.,
Hoedemaeker, I., van Montfrans, C.,
Hommes, D. W., Peppelenbosch, M. P., van
Deventer, S. J. Infliximab but not etanercept
induces apoptosis in lamina propria T-lym-
phocytes from patients with Crohns disease.
Gastroenterology 2003, 124, 17741785.
60 Parsons, H. L., Earnshaw, J. C., Wilton, J.,
Johnson, K. S., Schueler, P. A., Mahoney, W.,
McCafferty, J. Directing phage selections to-
wards specific epitopes. Protein Eng. 1996, 9,
10431049.
61 Franklin, M. C., Carey, K. D., Vajdos, F. F., Lea-
hy, D. J., de Vos, A. M., Sliwkowski, M. X. In-
sights into ErbB signaling from the structure
of the ErbB2pertuzumab complex. Cancer
Cell 2004, 5, 317328.
References 1181
62 Agus, D. B., Akita, R. W., Fox, W. D., Lewis,
G. D., Higgins, B., Pisacane, P. I., Lofgren,
J. A., Tindell, C., Evans, D. P., Maiese, K.,
Scher, H. I., Sliwkowski, M. X. Targeting li-
gand-activated ErbB2 signaling inhibits breast
and prostate tumor growth. Cancer Cell 2002,
2, 127137.
63 Badache, A., Hynes, N. E. A new therapeutic
antibody masks ErbB2 to its partners. Cancer
Cell 2004, 5, 299301.
64 Cho, H. S., Mason, K., Ramyar, K. X., Stanley,
A. M., Gabelli, S. B., Denney, D. W., Jr., Leahy,
D. J. Structure of the extracellular region of
HER2 alone and in complex with the Hercep-
tin Fab. Nature 2003, 421, 756760.
65 Clynes, R. A., Towers, T. L., Presta, L. G., Ra-
vetch, J. V. Inhibitory Fc receptors modulate in
vivo cytotoxicity against tumor targets. Nat.
Med. 2000, 6, 443446.
66 Cobleigh, M. A., Vogel, C. L., Tripathy, D., Ro-
bert, N. J., Scholl, S., Fehrenbacher, L., Wolter,
J. M., Paton, V., Shak, S., Lieberman, G., Sla-
mon, D. J. Multinational study of the efficacy
and safety of humanized anti-HER2 monoclo-
nal antibody in women who have HER2-over-
expressing metastatic breast cancer that has
progressed after chemotherapy for metastatic
disease. J. Clin. Oncol. 1999, 17, 26392648.
67 Plosker, G. L., Figgitt, D. P. Rituximab: a re-
view of its use in non-Hodgkins lymphoma
and chronic lymphocytic leukaemia. Drugs
2003, 63, 803843.
68 Burton, D. R. Antibodies, viruses and vaccines.
Nat. Rev. Immunol. 2002, 2, 706713.
69 Pantophlet, R., Burton, D. R. Immunofocus-
ing: antigen engineering to promote the in-
duction of HIV-neutralizing antibodies. Trends
Mol. Med. 2003, 9, 468473.
70 Burton, D. R., Desrosiers, R. C., Doms, R. W.,
Koff, W. C., Kwong, P. D., Moore, J. P., Nabel,
G. J., Sodroski, J., Wilson, I. A., Wyatt, R. T.
HIV vaccine design and the neutralizing anti-
body problem. Nat. Immunol. 2004, 5, 233
236.
71 Urban, J. L., Schreiber, H. Tumor antigens.
Annu. Rev. Immunol. 1992, 10, 617644.
72 Agnantis, N. J., Goussia, A. C., Stefanou, D.
Tumor markers. An update approach for their
prognostic significance. Part I. In Vivo 2003,
17, 609618.
73 Guillemard, V., Saragovi, H. U. Novel
approaches for targeted cancer therapy. Curr.
Cancer Drug Targets 2004, 4, 313326.
74 Roselli, M., Guadagni, F., Buonomo, O., Belar-
di, A., Ferroni, P., Diodati, A., Anselmi, D.,
Cipriani, C., Casciani, C. U., Greiner, J.,
Schlom, J. Tumor markers as targets for selec-
tive diagnostic and therapeutic procedures.
Anticancer Res. 1996, 16, 21872192.
75 Maynard, J. A., Maassen, C. B., Leppla, S. H.,
Brasky, K., Patterson, J. L., Iverson, B. L., Geor-
giou, G. Protection against anthrax toxin by
recombinant antibody fragments correlates
with antigen affinity. Nat. Biotechnol. 2002, 20,
597601.
76 Adams, G. P., Schier, R., McCall, A. M., Sim-
mons, H. H., Horak, E. M., Alpaugh, R. K.,
Marks, J. D., Weiner, L. M. High affinity re-
stricts the localization and tumor penetration
of single-chain Fv antibody molecules. Cancer
Res. 2001, 61, 47504755.
77 Adams, G. P., Schier, R. Generating improved
single-chain Fv molecules for tumor targeting.
78 Verel, I., Heider, K.-H., Siegmund, M., Oster-
mann, E., Patzelt, E., Sproll, M., Snow, G. B.,
Adolf, G. R., van Dongen, G. A. M. S. Tumor
targeting properties of monoclonal antibodies
with different affinity for target antigen
CD44V6 in nude mice bearing head-and-neck
cancer xenografts. Int. J. Cancer 2002, 99,
396402.
79 Neuberger, M. S., Harris, R. S., Di Noia, J., Pe-
tersen-Mahrt, S. K. Immunity through DNA
deamination. Trends Biochem. Sci. 2003, 28,
305312.
80 Besmer, E., Gourzi, P., Papavasiliou, F. N. The
regulation of somatic hypermutation. Curr.
Opin. Immunol. 2004, 16, 241245.
81 Sale, J. E., Bemark, M., Williams, G. T., Jolly,
C. J., Ehrenstein, M. R., Rada, C., Milstein, C.,
Neuberger, M. S. In vivo and in vitro studies of
immunoglobulin gene somatic hypermuta-
tion. Philos. Trans. R. Soc. London B Biol. Sci.
2001, 356, 2128.
82 Gram, H., Marconi, L. A., Barbas, C. F., 3rd,
Collet, T. A., Lerner, R. A., Kang, A. S. In vitro
selection and affinity maturation of antibodies
from a naive combinatorial immunoglobulin
library. Proc. Natl Acad. Sci. USA 1992, 89,
35763580.
83 Hawkins, R. E., Russell, S. J., Winter, G. Selec-
tion of phage antibodies by binding affinity.
Mimicking affinity maturation. J. Mol. Biol.
1992, 226, 889896.
84 Schier, R., Bye, J., Apell, G., McCall, A.,
Adams, G. P., Malmqvist, M., Weiner, L. M.,
Marks, J. D. Isolation of high-affinity mono-
meric human anti-c-erbB-2 single chain Fv
using affinity-driven selection. J. Mol. Biol.
1996, 255, 2843.
85 Chen, Y., Wiesmann, C., Fuh, G., Li, B.,
Christinger, H. W., McKay, P., de Vos, A. M.,
Lowman, H. B. Selection and analysis of an
optimized anti-VEGF antibody: crystal struc-
ture of an affinity-matured Fab in complex
with antigen. J. Mol. Biol. 1999, 293, 865881.
86 Plckthun, A., Schaffitzel, C., Hanes, J., Jer-
mutus, L. In vitro selection and evolution of
proteins. Adv. Protein Chem. 2000, 55, 367
403.
87 Zahnd, C., Spinelli, S., Luginbhl, B., Am-
stutz, P., Cambillau, C., Plckthun, A. Direc-
ted in vitro evolution and crystallographic
analysis of a peptide-binding single chain anti-
body fragment (scFv) with low picomolar af-
finity. J. Biol. Chem. 2004, 279, 1887018877.
88 Roskos, L. K., Davis, C. G., Schwab, G. M. The
clinical pharmacology of therapeutic monoclo-
nal antibodies. Drug Dev. Res. 2004, 61, 108
120.
89 Junghans, R. P. Finally! The Brambell receptor
(FcRB). Mediator of transmission of immunity
and protection from catabolism for IgG. Im-
munol. Res. 1997, 16, 2957.
90 Telleman, P., Junghans, R. P. The role of the
Brambell receptor (FcRB) in liver: protection
of endocytosed immunoglobulin G (IgG) from
catabolism in hepatocytes rather than trans-
port of IgG to bile. Immunology 2000, 100,
245251.
91 Trang, J. M. Pharmacokinetics and metabolism
of therapeutic antibodies. In Protein Pharma-
cokinetics and Metabolism, Ferraiolo, B. L.,
Mohler, M. A., Gloff, C. A. (eds). Plenum, New
York, 1992.
92 Waldmann, T. A., Strober, W. Metabolism of
immunoglobulins. Prog. Allergy 1969, 13, 1
110.
93 Schror, K., Weber, A. A. Comparative pharma-
cology of GP IIb/IIIa antagonists. J. Thromb.
Thrombolysis 2003, 15, 7180.
94 Tam, S. H., Sassoli, P. M., Jordan, R. E., Naka-
da, M. T. Abciximab (ReoPro, chimeric 7E3
Fab) demonstrates equivalent affinity and
functional blockade of glycoprotein IIb/IIIa
and a
v
b
3
integrins. Circulation 1998, 98, 1085
1091.
95 Goldenberg, D. M. Advancing role of radiola-
beled antibodies in the therapy of cancer.
Cancer Immunol. Immunother. 2003, 52, 281
296.
96 Wilder, R. B., DeNardo, G. L., DeNardo, S. J.
Radioimmunotherapy: recent results and fu-
ture directions. J. Clin. Oncol. 1996, 14,
13831400.
97 Bischof Delaloye, A. The role of nuclear
medicine in the treatment of non-Hodgkins
lymphoma (NHL). Leuk. Lymphoma 2003, 44
(Suppl. 4), S29S36.
98 Wiseman, G. A., Kornmehl, E., Leigh, B., Er-
win, W. D., Podoloff, D. A., Spies, S., Sparks,
R. B., Stabin, M. G., Witzig, T., White, C. A.
Radiation dosimetry results and safety corre-
lations from
90
Y-ibritumomab tiuxetan radio-
immunotherapy for relapsed or refractory
non-Hodgkins lymphoma: combined data
from 4 clinical trials. J. Nucl. Med. 2003, 44,
465474.
99 Kaminski, M. S., Estes, J., Zasadny, K. R.,
Francis, I. R., Ross, C. W., Tuck, M., Regan,
D., Fisher, S., Gutierrez, J., Kroll, S., Stagg,
R., Tidmarsh, G., and Wahl, R. L. Radioim-
munotherapy with iodine
131
I tositumomab
for relapsed or refractory B-cell non-Hodgkin
lymphoma: updated results and long-term
follow-up of the University of Michigan ex-
perience. Blood 2000, 96, 12591266.
100 Scheidhauer, K., Wolf, I., Baumgartl, H. J.,
Von Schilling, C., Schmidt, B., Reidel, G.,
Peschel, C., Schwaiger, M. Biodistribution
and kinetics of
131
I-labelled anti-CD20 MAB
IDEC-C2B8 (rituximab) in relapsed non-
Hodgkins lymphoma. Eur. J. Nucl. Med.
Mol. Imaging 2002, 29, 12761282.
101 Maack, T. Renal handling of proteins and
polypeptides. In Handbook of Physiology, Sec-
tion 8: Renal Physiology, Windhager, E. E.
(ed.). APS, Washington, DC, 1992.
102 Renkin, E. M., Gilmore, J. P. Glomerular fil-
tration. In Handbook of Physiology: Renal
Physiology, Orloff, J., Berliner, R. W. (eds.).
APS, Washington, DC, 1973.
103 Plckthun, A., Pack, P. New protein engi-
neering approaches to multivalent and bi-
specific antibody fragments. Immunotechnol-
ogy 1997, 3, 83105.
104 Carter, P. Improving the efficacy of anti-
body-based cancer therapies. Nat. Rev. Can-
cer 2001, 1, 118129.
References 1183
105 Glennie, M. J., Johnson, P. W. Clinical trials
of antibody therapy. Immunol. Today 2000,
21, 403410.
106 Zondor, S. D., Medina, P. J. Bevacizumab: an
angiogenesis inhibitor with efficacy in colo-
rectal and other malignancies. New Drug
Dev. 2004, 38, 12581264.
107 Kim, K. J., Li, B. Inhibition of vascular endo-
thelial growth factor-induced angiogenesis
suppresses tumour growth in vivo. Nature
1993, 362, 841844.
108 Cartron, G., Watier, H., Golay, J., Solal-Ce-
ligny, P. From the bench to the bedside:
ways to improve rituximab efficacy. Blood
2004, 104, 26352642.
109 Deans, J. P., Li, H., Polyak, M. J. CD20-
mediated apoptosis: signalling through lipid
rafts. Immunology 2002, 107, 176182.
110 Maloney, D. G., Smith, B., Rose, A. Rituxi-
mab: mechanism of action and resistance.
Semin. Oncol. 2002, 29, 29.
111 Nagy, Z. A., Mooney, N. A. A novel, alterna-
tive pathway of apoptosis triggered through
class II major histocompatibility complex
molecules. J. Mol. Med. 2003, 81, 757765.
112 Nagy, Z. A., Hubner, B., Lhning, C., Rau-
chenberger, R., Reiffert, S., Thomassen-Wolf,
E., Zahn, S., Leyer, S., Schier, E. M., Zahrad-
nik, A., Brunner, C., Lobenwein, K., Rattel,
B., Stanglmaier, M., Hallek, M., Wing, M.,
Anderson, S., Dunn, M., Kretzschmar, T.,
Tesar, M. Fully human, HLA-DR-specific
monoclonal antibodies efficiently induce
programmed death of malignant lymphoid
cells. Nat. Med. 2002, 8, 801807.
113 Cook-Bruns, N. Retrospective analysis of the
safety of Herceptin immunotherapy in meta-
static breast cancer. Oncology 2001, 61
(Suppl. 2), 5866.
114 Shields, R. L., Namenuk, A. K., Hong, K.,
Meng, Y. G., Rae, J., Briggs, J., Xie, D., Lai,
J., Stadlen, A., Li, B., Fox, J. A., Presta, L. G.
High resolution mapping of the binding site
on human IgG1 for FccRI, FccRII, FccRIII,
and FcRn and design of IgG1 variants with
improved binding to the FccR. J. Biol. Chem.
2001, 276, 65916604.
115 Presta, L. G. Engineering antibodies for ther-
apy. Curr. Pharm. Biotechnol. 2002, 3, 237
256.
116 Maxwell, K. F., Powell, M. S., Hulett, M. D.,
Barton, P. A., McKenzie, I. F., Garrett, T. P.,
Hogarth, P. M. Crystal structure of the hu-
man leukocyte Fc receptor, FccRIIa. Nat.
Struct. Biol. 1999, 6, 437442.
117 Radaev, S., Motyka, S., Fridman, W. H.,
Sautes-Fridman, C., Sun, P. D. The structure
of a human type III Fcc receptor in complex
with Fc. J. Biol. Chem. 2001, 276, 16469
16477.
118 Sondermann, P., Huber, R., Oosthuizen, V.,
Jacob, U. The 3.2- crystal structure of the
human IgG1 Fc fragmentFccRIII complex.
Nature 2000, 406, 267273.
119 Sondermann, P., Kaiser, J., Jacob, U. Molec-
ular basis for immune complex recognition:
a comparison of Fc-receptor structures. J.
Mol. Biol. 2001, 309, 737749.
120 Zhang, Y., Boesen, C. C., Radaev, S., Brooks,
A. G., Fridman, W. H., Sautes-Fridman, C.,
Sun, P. D. Crystal structure of the extracellu-
lar domain of a human FccRIII. Immunity
2000, 13, 387395.
121 Jefferis, R., Lund, J., Pound, J. D. IgG-Fc-
mediated effector functions: molecular defi-
nition of interaction sites for effector ligands
and the role of glycosylation. Immunol. Rev.
1998, 163, 5976.
122 Krapp, S., Mimura, Y., Jefferis, R., Huber,
R., Sondermann, P. Structural analysis of
human IgG-Fc glycoforms reveals a correla-
tion between glycosylation and structural in-
tegrity. J. Mol. Biol. 2003, 325, 979989.
123 Umana, P., Jean-Mairet, J., Moudry, R., Am-
stutz, H., Bailey, J. E. Engineered glycoforms
of an antineuroblastoma IgG1 with opti-
mized antibody-dependent cellular cytotoxic
activity. Nat. Biotechnol. 1999, 17, 176180.
124 Shinkawa, T., Nakamura, K., Yamane, N.,
Shoji-Hosaka, E., Kanda, Y., Sakurada, M.,
Uchida, K., Anazawa, H., Satoh, M., Yama-
saki, M., Hanai, N., Shitara, K. The absence
of fucose but not the presence of galactose
or bisecting N-acetylglucosamine of human
IgG1 complex-type oligosaccharides shows
the critical role of enhancing antibody-de-
pendent cellular cytotoxicity. J. Biol. Chem.
2003, 278, 34663473.
125 Presta, L. G. Antibody engineering for thera-
peutics. Curr. Opin. Struct. Biol. 2003, 13,
519525.
126 Shields, R. L., Lai, J., Keck, R., OConnell,
L. Y., Hong, K., Meng, Y. G., Weikert, S. H.,
Presta, L. G. Lack of fucose on human IgG1
N-linked oligosaccharide improves binding
to human Fcc RIII and antibody-dependent
cellular toxicity. J. Biol. Chem. 2002, 277,
2673326740.
127 Okazaki, A., Shoji-Hosaka, E., Nakamura,
K., Wakitani, M., Uchida, K., Kakita, S., Tsu-
moto, K., Kumagai, I., Shitara, K. Fucose de-
pletion from human IgG1 oligosaccharide
enhances binding enthalpy and association
rate between IgG1 and FccRIIIa. J. Mol. Biol.
2004, 336, 12391249.
128 DallAcqua, W. F., Woods, R. M., Ward, E. S.,
Palaszynski, S. R., Patel, N. K., Brewah, Y. A.,
Wu, H., Kiener, P. A., Langermann, S. In-
creasing the affinity of a human IgG1, for
the neonatal Fc receptor: biological conse-
quences. J. Immunol. 2002, 169, 51715180.
129 Hinton, P. R., Johlfs, M. G., Xiong, J. M., Ha-
nestad, K., Ong, K. C., Bullock, C., Keller, S.,
Tang, M. T., Tso, J. Y., Vasquez, M., Tsurushi-
ta, N. Engineered human IgG antibodies
with longer serum half-lives in primates. J.
Biol. Chem. 2004, 279, 62136216.
130 Trail, P. A., King, H. D., Dubowchik, G. M.
Monoclonal antibody drug immunoconju-
gates for targeted treatment of cancer. Can-
cer Immunol. Immunother. 2003, 52, 328337.
131 Voutsadakis, I. A. Gemtuzumab ozogamicin
(CMA-676, Mylotarg) for the treatment of
CD33
+
acute myeloid leukemia. Anticancer
Drugs 2002, 13, 685692.
132 Paul, S. P., Taylor, L. S., Stansbury, E. K.,
McVicar, D. W. Myeloid specific human
CD33 is an inhibitory receptor with differen-
tial ITIM function in recruiting the phos-
phatases SHP-1 and SHP-2. Blood 2000, 96,
483490.
133 Hellstrm, I., Garrigues, H. J., Garrigues,
U., Hellstrm, K. E. Highly tumor-reactive,
internalizing, mouse monoclonal antibodies
to Le
y
-related cell surface antigens. Cancer
Res. 1990, 50, 21832190.
134 Saleh, M. N., Sugarman, S., Murray, J., Os-
troff, J. B., Healey, D., Jones, D., Daniel,
C. R., LeBherz, D., Brewer, H., Onetto, N.,
LoBuglio, A. F. Phase I trial of the anti-Lewis
Y drug immunoconjugate BR96doxorubicin
in patients with lewis Y-expressing epithelial
tumors. J. Clin. Oncol. 2000, 18, 22822292.
135 Ajani, J. A., Kelsen, D. P., Haller, D., Har-
graves, K., Healey, D. A multi-institutional
phase II study of BMS-182248-01 (BR96
doxorubicin conjugate) administered every
21 days in patients with advanced gastric
adenocarcinoma. Cancer J. 2000, 6, 7881.
136 Tolcher, A. W., Sugarman, S., Gelmon, K. A.,
Cohen, R., Saleh, M., Isaacs, C., Young, L.,
Healey, D., Onetto, N., Slichenmyer, W. Ran-
domized phase II study of BR96-doxorubicin
conjugate in patients with metastatic breast
cancer. J. Clin. Oncol. 1999, 17, 478484.
137 Wahl, A. F., Donaldson, K. L., Mixan, B. J.,
Trail P. A., Siegall, C. B. Selective tumor sen-
sitization to taxanes with the mAbdrug con-
jugate cBR96doxorubicin. Int. J. Cancer
2001, 93, 590600.
138 Doronina, S. O., Toki, B. E., Torgov, M. Y.,
Mendelsohn, B. A., Cerveny, C. G., Chace,
D. F., DeBlanc, R. L., Gearing, R. P., Bovee,
T. D., Siegall, C. B., Francisco, J. A., Wahl,
A. F., Meyer, D. L., Senter, P. D. Development
of potent monoclonal antibody auristatin
conjugates for cancer therapy. Nat. Biotech-
nol. 2003, 7, 778784.
139 Prinz, H. Recent advances in the field of tu-
bulin polymerization inhibitors. Expert Rev.
Anticancer Ther. 2002, 2, 695708.
140 Falnes, P. O., Sandvig, K. Penetration of pro-
tein toxins into cells. Curr. Opin. Cell Biol.
2000, 12, 407413.
141 Perentesis, J. P., Miller, S. P., Bodley, J. W.
Protein toxin inhibitors of protein synthesis.
Biofactors 1992, 3, 173184.
142 Stirpe, F. Ribosome-inactivating proteins.
Toxicon 2004, 44, 371383.
143 Pai-Scherf, L. H., Villa, J., Pearson, D., Wat-
son, T., Liu, E., Willingham, M. C., Pastan, I.
Hepatotoxicity in cancer patients receiving
erb-38, a recombinant immunotoxin that tar-
gets the erbB2 receptor. Clin. Cancer Res.
1999, 5, 23112315.
144 Pastan, I. Immunotoxins containing Pseudo-
monas exotoxin A: a short history. Cancer
Immunol. Immunother. 2003, 52, 338341.
145 Poe, J. C., Fujimoto, Y., Hasegawa, M., Haas,
K. M., Miller, A. S., Sanford, I. G., Bock,
C. B., Fujimoto, M., Tedder, T. F. CD22 regu-
lates B lymphocyte function in vivo through
both ligand-dependent and ligand-indepen-
dent mechanisms. Nat. Immunol. 2004, 5,
10781087.
146 Schnell, R., Borchmann, P., Staak, J. O.,
Schindler, J., Ghetie, V., Vitetta, E. S., Engert,
A. Clinical evaluation of ricin A-chain im-
munotoxins in patients with Hodgkins lym-
phoma. Ann. Oncol. 2003, 14, 729736.
147 Helguera, G., Morrison, S. L., Penichet, M. L.
Antibodycytokine fusion proteins: harness-
References 1185
ing the combined power of cytokines and
antibodies for cancer therapy. Clin. Immunol.
2002, 105, 233246.
148 Gillies, S. D., Lan, Y., Brunkhorst, B., Wong,
W. K., Li, Y., Lo, K. M. Bi-functional cytokine
fusion proteins for gene therapy and anti-
body-targeted treatment of cancer. Cancer
Immunol. Immunother. 2002, 51, 449460.
149 Borchmann, P., Schnell, R., Fuss, I., Man-
zke, O., Davis, T., Lewis, L. D., Behnke, D.,
Wickenhauser, C., Schiller, P., Diehl, V., En-
gert, A. Phase 1 trial of the novel bispecific
molecule H22Ki-4 in patients with refrac-
tory Hodgkin lymphoma. Blood 2002, 100,
31013107.
150 Kufer, P., Lutterbuse, R., Baeuerle, P. A. A
revival of bispecific antibodies. Trends Bio-
technol. 2004, 22, 238244.
151 Hartmann, F., Renner, C., Jung, W., da
Costa, L., Tembrink, S., Held, G., Sek, A.,
Konig, J., Bauer, S., Kloft, M., Pfreundschuh,
M. Anti-CD16/CD30 bispecific antibody
treatment for Hodgkins disease: role of infu-
sion schedule and costimulation with cyto-
kines. Clin. Cancer Res. 2001, 7, 18731881.
152 James, N. D., Atherton, P. J., Jones, J., Ho-
wie, A. J., Tchekmedyian, S., Curnow, R. T.
A phase II study of the bispecific antibody
MDX-H210 (anti-HER2CD64) with GM-
CSF in HER2
+
advanced prostate cancer. Br.
J. Cancer 2001, 85, 152156.
153 Pollack, P., Groothuis, J. R. Development
and use of palivizumab (Synagis): a passive
immunoprophylactic agent for RSV. J. Infec-
tion Chemother. 2002, 8, 201206.
154 Visintin, M., Tse, E., Axelson, H., Rabbitts,
T. H., Cattaneo, A. Selection of antibodies for
intracellular function using a two-hybrid in
vivo system. Proc. Natl Acad. Sci. USA 1999,
96, 1172311728.
155 Mssner, E., Koch, H., Plckthun, A. Fast
selection of antibodies without antigen puri-
fication: adaption of the protein fragment
complementation assay to select antigen
antibody pairs. J. Mol. Biol. 2001, 308, 115
122.
Abstract
The production of antibodies directed at
self-proteins is a hallmark of many auto-
immune diseases, and the detection of
specific autoantibody responses is used in-
creasingly to aid diagnosis of various auto-
immune diseases. Molecular characteriza-
tion of autoantibodyantigen interaction
sites may help to identify subsets of pa-
tients with certain clinical features or prog-
nostic outcomes (personalized medicine)
(see Part I, Chapter 2 and 5). Moreover, it
can facilitate the development of immu-
noassays that use recombinant or synthetic
antigens as substrates for autoantibody de-
tection. The vast majority of autoantibody
epitopes is conformational that is, com-
prised of amino acids from distant parts of
the target protein sequence that cluster in
the folded protein. When linear sequence
stretches make dominant contributions to
antibody binding, these regions may be
identified by peptide mapping. In addition,
many investigators are now making use of
the growing number of known three-di-
mensional structures of autoantigens to
guide mapping and mutagenesis studies.
However, detailed information about
strictly conformational epitopes can only
be obtained by crystallographic studies
which are so far confined to the structure
of IgG
4
Fc complexed with the Fab of an
IgM rheumatoid factor and the recently
determined structure of the multiple
sclerosis autoantigen myelin oligodendro-
cyte glycoprotein (MOG) complexed with
the Fab of the pathogenic autoantibody 8-
18C5. The MOG-(8-18C5) crystal structure
identified a highly discontinuous epitope
centered about MOG residues 101108.
These residues encompass a strained tight
turn that is kept in its conformation by
the protein environment; this explains the
failure to detect this antigenic region by
conventional peptide mapping. Interest-
ingly, the immunodominant 8-18C5 epi-
tope on MOG, sequestered behind the
bloodbrain barrier, is composed of resi-
dues that are least conserved between
MOG and its homologues that are ex-
pressed outside the CNS and induce B cell
tolerance; this point will be discussed in
the chapter.
Abbreviations
ANAs anti-nuclear antibodies
ANCAs anti-neutrophil cytoplasmic anti-
bodies
APF antiperinuclear factor
BP bullous pemphigoid
BTN butyrophilin
1187
3
Molecular Characterization of Autoantibody Responses
in Autoimmune Diseases:
Implications for Diagnosis and Understanding of Autoimmunity
ISBN: 3-527-31184-X
CDRs complementarity determining
regions
CENP-C centromer protein C
CNS central nervous system
DDC DOPA decarboxylase
EAE experimental autoimmune
encephalomyelitis
EIA enzyme immunoassay
ERMAP erythroid membrane-associated
protein
GABA gamma-aminobutyric acid
GBM glomerular basement mem-
brane
GP Goodpasture disease
GST glutathione-S-transferase
IIF indirect immunofluorescence
LADA latent autoimmune diabetes in
adults
MBP myelin basic protein
MOG myelin oligodendrocyte glyco-
protein
PLP pyridoxal-5'-phosphate
snRNP small nuclear ribonucleoprotein
particles
TSH thyroid-stimulating hormone
WG Wegeners granulomatosis
3.1
Autoantibodies in Autoimmune Diseases
Autoimmune diseases form a heteroge-
neous group of chronic diseases in which
the immune system erroneously attacks
self-molecules (autoantigens) leading to
the destruction of organs, tissue and cells.
More than 80 clinically distinct autoim-
mune diseases have been identified [1], in-
cluding systemic disorders such as rheu-
matoid arthritis and systemic lupus erythe-
matosus (SLE), or organ-specific disorders
such as multiple sclerosis (MS), type I dia-
betes mellitus, and autoimmune thyroid
diseases. Though many autoimmune dis-
eases are individually rare, they collectively
affect an estimated 58% of the United
States population [1] and, presumably, a
similar percentage of the population else-
where in the industrialized world. Affect-
ing women disproportionately, autoim-
mune diseases are among the ten leading
causes of death for young and middle-aged
women [2].
3.1.1
Pathogenic Autoantibodies
The mechanisms that disrupt immunolo-
gical self-tolerance and lead to autoim-
mune disease are largely unknown; envi-
ronmental factors including infectious
agents, chemicals and diet are all sus-
pected to trigger or modulate autoimmune
responses in genetically predisposed indi-
viduals. Cellular damage in autoimmune
diseases is mediated by immune effector
mechanisms triggered by autoaggressive T
cells, high-affinity autoantibodies or a
combination of both. A direct pathogenic
effect mediated by autoantibodies is the
hallmark of many of the best-characterized
autoimmune diseases including myasthe-
nia gravis and Graves disease, where auto-
antibodies interfere with essential protein
functions. Myasthenia gravis is caused by
autoantibodies that bind the nicotinic ace-
tylcholine receptor (AChR) expressed on
skeletal muscle cells at the neuromuscular
junction. Antibody binding leads to the re-
duction of functional AChR, and this re-
sults in fatigue and muscular weakness.
The clinical signs of myasthenia gravis can
be induced in experimental animals by the
passive transfer of serum or purified anti-
bodies from myasthenic patients, an ex-
perimental approach that defined the
pathogenic role of autoantibodies in this
disease. The majority of autoantibodies
bind to one defined region exposed at the
extracellular part of the a subunit of the
3 Molecular Characterization of Autoantibody Responses in Autoimmune Diseases 1188
pentameric transmembrane AChR. By
cross-linking receptors they stimulate the
physiological process of internalization
and intracellular degradation, reducing the
half-life of AChR to one-third. In addition,
AChR-specific antibodies binding to the
neuromuscular junction activate the com-
plement cascade; this leads to focal de-
struction of the postsynaptic membrane
and further internalization of AChR, prob-
ably due to disruption of the cytoskeleton
[3, 4]. Graves disease is the most common
form of hyperthyroidism caused by auto-
antibodies directed against the thyroid-
stimulating hormone (TSH) receptor.
When activated by TSH, the TSH receptor
stimulates the synthesis of thyroid hor-
mones. In Graves disease, autoantibody
binding activates the TSH receptor that
results in the production of overlarge
amounts of thyroid hormones, as activa-
tion of the TSH receptor becomes inde-
pendent of the physiological feedback reg-
ulation that operates by inhibiting the
TSH synthesis. Another type of autoanti-
body-mediated dysfunction is due to the
formation of large amounts of immune
complexes as seen in SLE. Autoantibodies
in SLE bind to various ubiquitous self-anti-
gens such as DNA and nucleoprotein par-
ticles. The resulting immune complexes
accumulate in small blood vessels of var-
ious organs causing severe inflammatory
damage [5].
3.1.2
Autoantibodies as Markers
for Autoimmune Disease
Irrespective of their role in disease patho-
genesis, the detection of serum autoanti-
bodies is becoming increasingly important
for the diagnosis, prognosis and monitor-
ing of autoimmune diseases. The use of
autoantibody tests to identify individuals at
risk of developing a specific autoimmune
condition is based on the observation that
serum autoantibodies often appear long
before the onset of clinical disease [6].
Clinical manifestation of type I diabetes
mellitus, for instance, is preceded by an
asymptomatic prodromal period character-
ized by the appearance of autoantibodies
directed at several antigens of the pancrea-
tic islet cells. Numerous prediction studies
have analyzed islet-associated antibodies in
relatives of patients with autoimmune dia-
betes and genetically predisposed indivi-
duals, and the presence of antibodies
against two or more specific islet autoanti-
gens was found to be highly predictive for
the development of clinical diabetes [7, 8].
Testing the serum of blood donors for IgM
rheumatoid factor and anti-cyclic citrullin-
ated peptide antibodies, both markers for
rheumatoid arthritis (RA), showed that
nearly half of the patients with RA were
positive for one or both autoantibodies a
medium of 4.5 years before onset of symp-
toms, compared to about 1% of false-posi-
tives in control samples [9]. Similarly, anal-
ysis of serum samples from members of
the US Armed Forces with SLE, collected
before onset of disease, identified autoanti-
bodies characteristic for SLE that accumu-
lated gradually before the development of
the first clinical symptoms in 88% of the
individuals [10].
The use of autoantibodies for diagnosis
can be advantageous when diagnosis is
complicated by the lack of specific symp-
toms in early autoimmune disease. Early
SLE, for example, is associated with mal-
aise, arthralgia, and fatigue [11]; MS pa-
tients can exhibit symptoms similar to
those of infectious diseases, trauma or ma-
lignancy [1]. Vice versa, individuals with
the same autoimmune disease can show
different clinical phenotypes and distinct
disease courses. Numerous autoantigens
associated with autoimmune diseases have
been identified so far, and tests for autoan-
tibody responses directed against these
antigens are currently applied to: 1) aid di-
agnosis; 2) monitor the degree of disease,
the present immunologic activity or the re-
sponse to therapy; or 3) provide prognostic
information about the course and severity
of the disease. An elevated level of anti-nu-
clear antibodies (ANAs), for example, con-
stitutes one of the American College of
Rheumatology criteria for the diagnosis of
SLE. The production of ANAs is character-
istic for various autoimmune connective
tissue disorders including SLE, scleroder-
ma, mixed connective tissue disease and
Sjgrens syndrome. ANAs are detected by
enzyme immunoassay tests (EIA) or by in-
direct immunofluorescence (IIF) on cul-
tured cell lines that exhibit distinct fluores-
cence patterns depending on the particular
disease or disease subset. Thus, nucleoli-
positive IIF is associated with scleroderma,
homogeneous fluorescence of cell nuclei
with SLE. This diagnosis can be supported
or further differentiated by enzyme-linked
immunosorbent assays (ELISAs) detecting
anti-Scl70 or anti-Sm antibodies that are
highly specific for systemic sclerosis and
SLE, respectively. Moreover, as the titer
of SLE-specific autoantibodies directed
against nuclear DNA correlates with dis-
ease activity, anti-nDNA ELISA can be
used to monitor disease activity.
Eased by well-established methods of
molecular biology and the accessibility of
human genomic sequence data, autoanti-
gens are currently being characterized in
detail and new autoantigens are continu-
ously identified. Recombinant protein ex-
pression and the insertion of protein tags
allows for the production of large amounts
of pure autoantigen for usage in immu-
noassays [12]. In concert with the shift
from the microscopic cellular level to the
molecular protein level, current efforts are
directed at identifying particular B-cell epi-
topes on autoantigens. The determination
of fine specificities of autoantibody bind-
ing is aimed at gaining a deeper under-
standing of the development and diversifi-
cation of the autoantibody response, as
well as at investigating a potential correla-
tion of specific epitopes with a particular
disease, a disease subtype, and the stage
or the future course of the disease. In ad-
dition, epitope mapping studies could ide-
ally result in the definition of synthetic
peptidic antigens as substrates in immu-
noassays, presenting a lower cost, well
reproducible alternative to the usage of
whole proteins.
3.2
Autoantibody Epitopes
3.2.1
Structural Aspects of AntigenAntibody
Interaction
Extensive structural and biochemical stud-
ies of antigenantibody complexes a
prime example of molecular recognition
have shown the substantial diversity of
antigen binding but have also revealed sev-
eral general features characteristic for anti-
genantibody interfaces [1315].
The antibody combining site (paratope) is
mainly composed of residues that belong to
the six complementarity determining re-
gions (CDRs) formed by three loops at the
tip of the variable domain of the light chain
(L1L3) and the three corresponding loops
of the heavy chain (H1H3). In large inter-
faces (as they occur in proteinantibody
complexes), residues of the antibody frame-
work sometimes contribute to binding. In
most cases less than six CDRs are involved
in binding the antigen, the heavy chain
CDRs (and of them especially H3) often
dominate the interaction. Many paratopes
are enriched in the aromatic residues tyro-
sine and tryptophan as well as the charged
residue arginine that represent thermody-
namic hotspots, probably due to their chem-
ical composition that enables them to form
multiple interactions via hydrophobic and
polar contacts.
In general, complex formation buries
800200
2
of the solvent-accessible sur-
face of each the antibody and the antigenic
protein which is in the typical range of
proteinprotein interactions. The chemical
composition of the antigens contact area
(epitope) resembles that of the overall pro-
tein surface, as in other non-permanent
protein complexes, with a slight prepon-
derance of polar residues. The shape com-
plementarity of the contact surfaces is
good, but less perfect than for example
that of protease inhibitor complexes; this
is consistent with the lack of evolutionary
optimization of the interaction between
antibody and antigen. Slight packing im-
perfections are often compensated for by
the introduction of water molecules that
form hydrogen bonds to unpaired polar
atoms and fill cavities of the interface.
In contrast to peptide or DNAantibody
complexes that exhibit a groove-like bind-
ing site, the combining site of antibodies
that bind proteinantigens is generally
quite planar. Structural characterization of
proteinantibody complexes reveals that
about 1522 residues of each binding part-
ner contribute to the interaction, corre-
sponding to two to five separate segments
of the polypeptide chain of the antigenic
protein [16]. However, the functional epi-
tope defined as those residues which ac-
count for a large fraction of the overall
binding energy can be much smaller
[17], as is also known for proteinprotein
interfaces in general [18, 19].
3.2.2
Classification of Epitopes
Epitopes (epitope stands for B-cell epi-
tope throughout this text) are often di-
vided into continuous (linear) and discon-
tinuous epitopes, meaning that either resi-
dues contiguous in the polypeptide chain
constitute the binding site or that the epi-
tope is composed of residues separated in
the sequence but located in close spatial
proximity in the folded protein [16, 20,
21]. As in all proteinantibody complexes
studied by X-ray crystallography so far,
more than one segment of the antigen is
involved in antibody binding, a continuous
epitope more accurately denotes a se-
quence stretch that, when displayed by a
peptide fragment, can bind antibodies
raised against the whole antigenic protein.
While continuous epitopes can still be rec-
ognized after denaturation of the protein,
recognition of discontinuous or conforma-
tional epitopes requires the correctly
folded native antigen.
A further classification of epitopes estab-
lished in the 1960s to differentiate be-
tween assembled virions and separate coat
protein subunits introduces cryptotopes
and neotopes that are also found on auto-
antigens. Cryptotopes only become accessi-
ble after depolymerization, denaturation or
fragmentation of the antigen; neotopes are
newly created for instance by the assembly
of protein subunits and depend, for exam-
ple, on a particular conformation of a com-
plex component adopted only after com-
plex formation or on the contribution of
residues from several subunits. Referring
to autoimmunity, a cryptotope on the scler-
oderma autoantigen centromer protein C
(CENP-C), revealed after cleavage by the
apoptotic enzyme granzyme B, represents
the preferred target for autoantibodies in a
subset of patients with limited systemic
sclerosis [22]. The 70-kDa subunit of the U1
small nuclear ribonucleoprotein particle
recognized by autoantibodies of patients
with SLE or SLE-overlap syndromes exhibits
cryptotopes and neotopes after apoptotic
cleavage and oxidative modification, respec-
tively, that are associated with different clin-
ical disease manifestations [23, 24].
Post-translationally modified epitopes
are of growing interest in autoimmune re-
search [25]. Enzymatic processes including
phosphorylation, glycosylation, methyla-
tion, or citrullination can modify proteins,
and (particularly in stressed and apoptotic
cells) deamidation, isomerization or oxida-
tive damage can occur spontaneously.
Spontaneous modifications that create nov-
el epitopes accumulate with age as well as
in cases of defects in protein repair and
degradation systems. That these modified
autoantigens can become targets of auto-
immune attack has been shown for in-
stance in SLE, rheumatoid arthritis and
type I diabetes mellitus. Recently, autoanti-
gens that contain enzymatically modified
arginine residues were found to be specifi-
cally targeted by autoantibodies [26]. Anti-
Sm-antibodies, for instance, are anti-nucle-
ar antibodies which are found in more
than 30% of patients with SLE and that
are highly specific for SLE. The seven ho-
mologous Sm proteins form the common
core of the small nuclear ribonucleopro-
tein particles (snRNP) U1, U2, U4/U6 and
U5 that are essential cofactors for pre-
mRNA splicing in eukaryotes. One of the
epitopes on the Sm proteins was mapped
to the C-terminal glycine-arginine-rich re-
gion of the two Sm proteins D1 and D3
and was shown specifically to depend on
the methylation of both aldimino groups
of all arginines to form symmetrically di-
methylated arginines [27].
Similarly, the antiperinuclear factor
(APF; the first autoantibody discovered to
be specific for rheumatoid arthritis [28])
and the related antikeratin antibody both
bind to an epitope on the epithelial protein
fillagrin that contains citrulline residues
formed by the enzymatic deimination of
arginines [29]. While the protein recog-
nized by this antibody response in in-
flamed joints is still unknown, this epitope
could be successfully mimicked by short
citrulline-containing peptides. These re-
sults initiated the development of highly
specific ELISAs for the diagnosis of rheu-
matoid arthritis that use cyclic citrullinated
peptides as antigens [30].
3.2.3
Methods of Epitope Mapping
3.2.3.1 Continuous Epitopes
One common method for the determina-
tion of continuous epitopes is that of
peptide mapping, overlapping synthetic or
phage-displayed peptides (see Part IV,
Chapter 16 and Part V, Chapters 1, 2, and
6) that cover the complete sequence of the
antigenic protein are tested for binding an-
tibodies specific for the antigen [21]. This
procedure is often used to identify pep-
tides that might induce antibodies cross-
reactive with the antigenic protein, for in-
stance aimed to develop a peptidic vaccine
that is able to induce cross-reactive anti-
bodies. For both vaccine and autoimmune
research it is essential to verify that anti-
bodies specific for certain peptides are
really cross-reactive and thus also bind to
the native protein [16]. It is important to
realize that a linear epitope identified by
peptide mapping may in reality be a mi-
motope present on a different protein (see
Section 3.2.3.2). Immunization protocols,
as used to induce experimental autoim-
mune diseases, can cause partial denatura-
tion resulting in an antibody response that
recognizes epitopes present on the highly
immunogenic denatured antigen, but not
on the correctly folded protein. In autoim-
mune disorders, antibody responses
against linear epitopes not present on the
native autoantigen are very common, and
may represent responses to degradation
products generated by proteolysis during
necrosis and apoptosis [31]. Antibodies spe-
cific for linear epitopes of myelin oligoden-
drocyte glycoprotein (MOG) occur at high
frequency in MS patients and were shown
to be associated with myelin debris in ac-
tively demyelinating MS lesions [32]. How-
ever, as the antibody response against native
MOG that induces severe demyelination in
the experimental animal model of MS is fo-
cused on conformational epitopes [3336],
these MOG peptide-specific antibodies are
probably binding to degraded MOG gener-
ated during myelin destruction and are in-
capable of triggering primary demyelina-
tion.
Epitopes determined by peptide map-
ping and associated with antibodies that
recognize the folded antigen are often
composed of polypeptide termini or sur-
face-exposed, flexible loops and turns, as
these can be most easily mimicked by lin-
ear peptides. Unlike helices and sheets,
loops at the protein surface present a con-
tinuous stretch of accessible amino acid
side chains that can be mirrored by a pep-
tide. The same holds true for anti-peptide
antibodies that cross-react with the whole
protein [37]. In this case, the peptide con-
formation recognized by the antibody can
be imposed on the flexible protein loop
upon binding to the antibody. Further-
more, a few examples of continuous heli-
cal epitopes on autoantigens have been re-
ported that could be mapped by the corre-
sponding peptides. One such epitope com-
prises amino acids 231245 of PM/Scl-100,
one component of the multisubunit auto-
antigen of the polymyositisscleroderma
overlap syndrome (PM/Scl) [38]. Muta-
tional analysis of the 15mer peptide re-
vealed that only the amino acids 234, 237,
240, and 241 are needed for autoantibody
binding residues that would be located
at the same site of an a-helical structure.
Moreover, secondary structure prediction
and a protein structure containing a re-
lated sequence support the existence of an
a-helix formed by amino acids 231245
that can be mimicked by the correspond-
ing peptide.
In general, the contact residues identi-
fied by peptide mapping form only part of
a larger discontinuous epitope that contrib-
utes to the lower affinity of the antibody to
the peptide compared to the corresponding
antibodyprotein interaction. Strategies to
improve cross-reactivity between peptides
and antigenic proteins include the intro-
duction of conformational constraints by
cyclization [39] as well as mutagenesis of
specific amino acids [40], or the use of pro-
tein carriers aimed to shift the thermody-
namic equilibrium of peptide conforma-
tions to the conformation most similar to
the proteinaceous epitope. Peptides fused
to glutathione-S-transferase (GST), for in-
stance, were used to map the epitopes of
the major bullous pemphigoid antigen 180
(BP180, type XVII collagen). Bullous pem-
phigoid (BP) is a subepidermal blistering
disease characterized by pathogenic anti-
bodies against the 155 kDa transmem-
brane protein BP180 that contributes to
adherence of the epidermis to the base-
ment membrane. The main target on
BP180 is the extracellular 16th non-col-
lagenous domain NC16a that is recognized
by the large majority of sera of BP patients
[41], and which is currently being tested as
a substrate for diagnostic assays of BP [42,
43]. The analysis of sera immunoadsorbed
with peptide fragments of NC16a coupled
to GST revealed that the reactivity of al-
most all sera was restricted to the 40 N-ter-
minal amino acids of NC16a [41, 44]. GST
was shown to increase the level of ordered
secondary structure for one of the peptides
[45] that might contribute to the observed
strong reactivity of the constructs.
3.2.3.2 Discontinuous Epitopes
In contrast to the cases presented above,
the vast majority of clinically relevant auto-
antibody epitopes is supposed to be highly
discontinuous. Epitopes per se are not pre-
ferentially composed of flexible loops or a
particular helix, but can comprise any sur-
face patch of the antigen, as shown for
hen egg white lysozyme in biochemical
and X-ray studies [46, 47]. The full charac-
terization of a discontinuous epitope re-
quires the structure determination of the
antigen in complex with the antigen-bind-
ing domains (Fab or Fv) of the correspond-
ing monoclonal antibody, ideally combined
with mutagenesis studies of epitope resi-
dues to determine or to confirm the local-
ization of hot spots in the binding inter-
face. As this method is time-consuming
and requires large amounts of sample, the
three-dimensional structures of only two
autoantigen Fab complexes have been de-
termined so far [48, 49]. Some information
about discontinuous epitopes, however,
can also be obtained by testing if proteins
homologous to the particular antigen
maintain antibody binding. Similarly, if
the antigen can be produced recombi-
nantly, parts of the protein can be replaced
by the corresponding fragments of a re-
lated protein or single residues can be
changed by site-directed mutagenesis
[50, 51]. Changes in the binding affinity of
the resultant locally altered antigen point
towards the participation of the particular
region in antibody binding. It must be
taken into account, however, that muta-
tions of epitope residues can be compen-
sated for by the incorporation of water or
the flexibility of the antibody-combining
site that can even lead to increased bind-
ing affinity of the mutated antigen. Be-
sides, mutations can provoke conforma-
tional changes distant from the site of mu-
tagenesis, possibly within the epitope, that
then result in the observed decreased bind-
ing affinity. Instead of abrogating antibody
binding, Henriksson and colleagues re-
stored it by introducing humanizing gain
of function mutations into the homolo-
gous U1-70k autoantigen of Drosophila
that led to the identification of a major
conformational epitope [50].
Moreover, the ability of antibodies to
protect the epitope of the bound antigen
against chemical modifications [52, 53]
and limited proteolytic degradation has
been used to map discontinuous epitopes.
As antibodies are highly resistant to enzy-
matic digestion [54], the epitope of the
antigen bound to the immobilized anti-
body can be excised by endoproteolytic
enzymes. After removal of the cleaved
fragments of the antigen by a washing
step, the individual segments of the epi-
tope still bound to the antibody can be di-
rectly analyzed by laser desorption/ioniza-
tion-based mass spectrometry [5557]. Yet,
unless the separate segments of the epi-
tope exhibit a very high affinity to the anti-
body, limited proteolysis only results in
the identification of a large protein frag-
ment that cannot be cleaved further with-
out dissociation.
Another approach to the elucidation of
discontinuous epitopes is based on the
identification of mimotopes [58], origi-
nally defined by Mario Geysen as . . . mol-
ecule(s) able to bind to the antigen com-
bining site of an antibody molecule, not
necessarily identical with the epitope in-
ducing the antibody, but an acceptable
mimic of the essential features of the epi-
tope [59]. These putative mimics can be
obtained from a library of random phage-
displayed peptides after enrichment of
binding peptides by several cycles of bio-
panning. Though initially not expected
[58], this method succeeded in identifying
peptides that competed for antibody bind-
ing against discontinuous protein-epitopes
and even DNA and carbohydrate antigens.
Moreover, true mimotopes were shown to
induce cross-reactive antibodies in vivo.
The utilization of mimotopes in epitope
mapping requires that: 1) mimics do not
bind to an antibody subsite different from
the site recognized by the antigen; and 2)
mimicry is based on the side chains of
amino acids and not confined to the ar-
rangement of atoms. Supported by compu-
ter algorithms, consensus sequences of
peptide subgroups can then be identified
and mapped on the three-dimensional
structure or possibly even on the primary
structure to yield the discontinuous epi-
tope [6065].
3.3
Visualization of Epitopes
Knowledge of the molecular structure of
autoantigens is of great value for the inter-
pretation of existing mapping results (as
shown for several examples below), as well
as for the choice of peptides and muta-
tions for future studies. The number of
solved three-dimensional structures has in-
creased tremendously within the past de-
cade. In 2004, the coordinates of more
than 27 000 proteins were available in the
Protein Data Bank, among them a grow-
ing number of autoantigens.
3.3.1
Peptide Mapping of Epitopes
on Proteinase 3
Antibodies specific for cytoplasmic anti-
gens of neutrophils (anti-neutrophil cyto-
plasmic antibodies (ANCAs)) are typically
found in patients with small-vessel vasculi-
tis [6668]. ANCAs can be differentiated
by immunofluoresence assays where bind-
ing of ANCAs to ethanol-fixed neutrophils
can produce a diffuse granular cytoplasmic
(c-ANCAs) or a perinuclear (p-ANCAs)
staining pattern. c-ANCAs that are primar-
ily directed against proteinase 3 (PR3),
found in azurophilic granules of neutro-
phils and granules of monocytes, are
strongly associated with Wegeners granu-
lomatosis (WG). A combination of IIF and
ELISA to detect PR3-specific ANCAs that
was tested in large multicenter studies
showed high specificity and sensitivity for
patients with active (99%, 91%) and inac-
tive (99.5%, 63%) WG. In addition, some
studies indicated that relapses in patients
with PR3-ANCA-associated vasculitis were
preceded by a rise of PR3-ANCA titers. In-
hibition studies with PR3-specific ANCA
sera and monoclonal antibodies revealed
that PR3 presents a limited number of epi-
tope areas that vary among patients and
change during disease course. Putative lin-
ear epitopes on PR3 have been determined
by three studies that used different detec-
tion methods and peptide constructs [69
71]. By mapping these epitopes onto the
surface of the three-dimensional structure
of PR3 [72], those parts of the peptide se-
quences that are surface-exposed in the
structure of PR3 can be determined. More-
over, linear epitopes that combine in space
to form a discontinuous epitope are de-
tected. In addition, consistencies and dis-
crepancies of the three studies can be visu-
alized. Three regions are identified consis-
F
i
g
.
3
.
1
L
i
n
e
a
r
e
p
i
t
o
p
e
s
o
n
p
r
o
t
e
i
n
a
s
e
3
.
C
o
l
o
r
e
d
r
e
g
i
o
n
s
c
o
r
r
e
s
p
o
n
d
t
o
e
p
i
t
o
p
e
r
e
s
i
d
u
e
s
d
e
s
c
r
i
b
e
d
b
y
W
i
l
l
i
a
m
s
e
t
a
l
.
(
y
e
l
l
o
w
)
[
6
9
]
,
G
r
i
f
f
i
t
h
e
t
a
l
.
(
r
e
d
)
[
7
0
]
a
n
d
v
a
n
d
e
r
G
e
l
d
e
t
a
l
.
(
b
l
u
e
)
[
7
1
]
;
r
e
s
i
d
u
e
s
i
d
e
n
t
i
f
i
e
d
b
y
m
o
r
e
t
h
a
n
o
n
e
s
t
u
d
y
a
r
e
d
e
p
i
c
t
e
d
i
n
t
h
e
r
e
s
p
e
c
t
i
v
e
c
o
m
b
i
n
a
t
i
o
n
c
o
l
o
r
s
.
tently in all three studies (Fig. 3.1). These
regions all encompass surface-exposed pro-
truding loops, two of them very flexible in
the crystal structure, and are located in
spatial proximity to each other (see Fig.
3.1), indicative of a possible immunodomi-
nant epitope region. The four main
sequence stretches described by van der
Geld et al. combine to form one surface
patch adjacent to the active site. Alto-
gether, the linear epitopes concentrate near
the active site, this being consistent with
reports that PR3-ANCA can interfere with
the enzymatic activity of PR3 and with the
binding of its physiological inhibitor-1
antitrypsin.
3.3.2
Cryptic Epitopes in Goodpasture Disease
Goodpasture disease (GP) is a prototype
autoantibody-mediated, organ-specific auto-
immune disease characterized by rapidly
progressive glomerulonephritis with or
without lung hemorrhage. The detection
of pathogenic antibodies directed against
the glomerular basement membrane
(GBM) is used for the diagnosis of GP. Im-
munopathogenicity of anti-GBM antibodies
can be inferred from the ability to induce
the disease in primates by transfer of kid-
ney-bound antibodies from patients with
GP, and from the positive effect of plasma
exchange in early disease. The major target
of anti-GBM antibodies is the C-terminal
non-collagenous (NC1) domain of the col-
lagen a3(IV) chain, one of six homologous
chains that constitute type IV collagen. a3
chains assemble with a4 and a5 chains to
form triple helical protomers that generate
collagenous networks by interaction of their
N- and C-termini. Mapping of the purely
discontinuous epitopes was performed
using gain of function chimeras that con-
sisted of the non- immunoreactive a1NC1,
in which sequence stretches most divergent
among the a(IV) chains were replaced by
the corresponding a3 sequence [73, 74].
When tested for reactivity with GP sera,
the chimeras revealed a major (Ea, amino
acids 1731) and a minor (Eb, amino acids
127141) epitope on a3NC1. Both epitopes
were fully accessible only after dissociation
of the hexamers, formed by two C-terminal
NC1 trimers at the junction of two triple he-
lices, into monomers. The recent elucida-
tion of the crystal structure of the highly ho-
mologous a1a1a2 NC1 complex [75, 76] and
investigation of the quaternary assembly of
the a3a4a5 complex [77] allows localization
of the cryptic epitopes on the homology
model of a3a4a5 NC1. The major epitope
forms a U-shaped structure composed of a
b-strand that is partially hidden in the
a3a5 NC1 interface confirming the ob-
served sequestered nature of the epitope
and of a loop followed by a coiled region
near the interface. This loop is more easily
accessible by antibodies after dissociation
of the complex (Fig. 3.2a). In the trimer,
pairs of neighboring monomers each form
a shared b-sheet with one monomer con-
tributing four b-strands, the other one con-
tributing two. Disruption of this b-sheet
upon dissociation of the complex probably
causes conformational changes that may
add to the higher accessibility of the Ea re-
gion in the monomer. Knowledge of the
molecular structure of the complex will
facilitate a detailed determination of the
immunodominant region containing the
Ea sequence that, regarding a recently re-
ported correlation between unfavorable re-
nal prognosis and antibodies binding to
the Ea region [78], may be used to develop
more specific diagnostic assays.
3.3.3
Modeling of Glutamate Decarboxylase 65
One of the islet cell autoantigens attacked
in type I diabetes is the 65 kDa isoform of
the pyridoxal-5'-phosphate (PLP)-depen-
dent glutamate decarboxylase (GAD65)
that decarboxylates glutamate, yielding the
major inhibitory neurotransmitter c-amino-
butyric acid (GABA). Competition studies
using recombinant Fab domains directed
against different discontinuous epitopes of
GAD65 revealed specific recognition pat-
terns in patients with type I diabetes com-
pared to GAD65 antibody-positive patients
with latent autoimmune diabetes in adults
(LADA), first-degree relatives and healthy
donors [79]. While precise structural map-
ping of the largely conformational epitopes
is hampered by the lack of a three-dimen-
sional structure of GAD65, some informa-
tion can be inferred from the molecular
structure of the related DOPA decarboxy-
lase (DDC) [80]. Each monomer of the di-
meric DDC consists of three domains, the
large middle domain composed of eight
helices surrounding a central seven-
stranded b-sheet that harbors the PLP
Fig. 3.2 Visualization of conformational epitopes.
(a) The cryptic epitopes Ea and Eb of the trimeric NC1
are positioned at the interface of two monomers.
(b) Stereo view of the model of the dimeric glutamate
decarboxylase 65, composed of a homology model of
the middle domain, based on coordinates of the DOPA
decarboxylase, and of the N- and C-terminal domain of
DOPA decarboxylase. Selected epitope residues at the
protein surface are marked.
a
b
binding site, a C-terminal four-stranded b-
sheet with three helices packed against it,
both typical of the a-family of PLP-depen-
dent enzymes, and in addition a small a-
helical domain at the N-terminus. Based
on the DDC structure, a tentative model of
the middle domain of GAD65 that shows
28% identity (46% similarity) to DDC can
be built (Fig. 3.2b) which allows the ap-
proximate positioning of amino acids
found to be critical for antibody binding
[81]. Evaluation of the accessibility of puta-
tive epitopes requires knowledge of the
complete quaternary enzyme structure. Re-
garding the observed conservation of do-
main and dimerization structure in decar-
boxylases of the a-family, the overall struc-
ture of the family member GAD65 is likely
to be similar to that of DDC. In this case,
the PEVKEK region (amino acids 260265)
of GAD65, which is assumed to be a
dominant epitope in type I diabetes [82,
83], is located in an accessible, surface-ex-
posed loop at the dimer interface near the
corresponding loop of the second mono-
mer.
3.4
Structural Characterization of Autoantibody
Autoantigen Complexes
The most complete and reliable character-
ization of epitopes is provided by three-di-
mensional complex structures between
antigen and Fab domain of the antibody.
The only complex structures involving
autoantigens that have been determined so
far are that of IgG
4
Fc complexed with the
Fab of an IgM rheumatoid factor [48, 84]
and that of the MS autoantigen MOG in
complex with the Fab of the pathogenic
autoantibody 8-18C5 [49]. These structures
allowed for a precise analysis of the epi-
topes, and both exhibited interesting char-
acteristics that were not expected from
prior mapping studies, providing insight
into formation and mechanism of the anti-
body response to these autoantigens. It
must be remembered however that in each
case a single antigenantibody complex
was analyzed, and additional studies are
required to verify that the interactions are
truly representative for the whole pool of
antibodies specific for the autoantigen [84].
3.4.1
Antibody Responses against MOG
MOG is a 25 kDa, quantitatively minor
myelin glycoprotein, which is highly con-
served across species and expressed exclu-
sively in the central nervous system
(CNS), where it is sequestered from the
immune system behind the bloodbrain
barrier [85]. It consists of an extracellular
N-terminal immunoglobulin (Ig) -like do-
main, a transmembrane helix and a cyto-
plasmic domain that contains a second
hydrophobic sequence, probably embedded
in the intracellular half of the myelin
membrane, and a short C-terminal tail.
The physiological function of MOG is un-
known; MOG-deficient knock-out mice
develop normally and exhibit no detectable
clinical, structural or pathological pheno-
type. Antibody binding to the Ig domain
of MOG leads to the depolymerization of
oligodendrocyte microtubuli in vitro, sug-
gesting a role for MOG as a cell surface
receptor that is involved in maintaining
myelin stability. This concept is supported
by the observation that MOG associates
with lipid rafts, known as sites for the as-
sembly of signaling complexes. Alterna-
tively, MOG has been implicated in adhe-
sion functions, in accord with the HNK1
glycan epitope, a marker for many cell ad-
hesion proteins, which is also present on a
subset of MOG molecules.
MOG was originally characterized two
decades ago as the immunodominant tar-
get of demyelinating autoantibodies in the
animal model of MS, experimental auto-
immune encephalomyelitis (EAE) [86, 87].
EAE is an inflammatory autoimmune dis-
ease of the CNS, induced in susceptible
laboratory animals by sensitization against
various myelin proteins of the CNS that
reproduces many of the clinical and patho-
logical features of MS [88, 89]. While mye-
lin-specific T cells initiate CNS inflamma-
tion in EAE, demyelination is autoanti-
body-mediated in rats and primates. In
contrast to other encephalitogenic myelin
autoantigens that are buried in the com-
pact multilamellar myelin sheath, MOG is
located on the outermost surface of the
myelin sheath and is therefore easily ac-
cessible to antibodies present in the extra-
cellular milieu. Immunization with the in-
tracellular myelin antigen myelin basic
protein (MBP) causes purely inflammatory
EAE in rats and primates. The administra-
tion of the MOG-specific monoclonal anti-
body 8-18C5 at disease onset, however,
leads to extensive demyelination and ex-
acerbates the clinical disease dramatically,
clearly demonstrating the pathogenic role
of MOG-specific antibodies in EAE. The
structural and immunopathological simi-
larities between MOG-induced demyelinat-
ing lesions in EAE and the type of plaques
seen in the majority of MS patients sug-
gests that MOG also acts as autoantigen in
MS.
MS is the most common chronic inflam-
matory demyelinating disease, affecting
about one million people worldwide.
Although MS is considered a primarily T-
cell-dependent disorder, increasing evi-
dence indicates that B cells and autoanti-
bodies are also involved in immunopatho-
genesis, though so far no target autoanti-
gen has been identified with certainty. The
classification of MOG as candidate MS
autoantigen is supported by its role in
EAE, and by the identification of MOG-
specific antibodies associated with myelin
debris in actively demyelinating MS le-
sions. Additionally, several studies have
shown that MOG-specific T- and B-cell lev-
els are elevated in MS patients, although
similar enhancements have also been ob-
served in patients with other neurological
disorders [90]. In a selected group of pa-
tients with a clinically isolated demyelinat-
ing syndrome, the presence of anti-MOG
antibodies was prognostic for early conver-
sion to clinically definite MS and rapid
disease progression [91]. However, several
groups report no significant differences in
the frequency of MOG-specific antibody re-
sponses in MS patients compared to
healthy donors. These controversial results
most probably reflect differences in the se-
lection of patients, the assay design or the
antigen preparations used by various in-
vestigators.
Recent studies revealed that two sets of
antibodies specific for the extracellular do-
main of MOG (MOG
ex
) can be distin-
guished in EAE: antibodies specific for
MOG-derived peptides and antibodies that
recognize purely discontinuous, conforma-
tion-dependent epitopes on MOG [34, 36].
While MOG-peptide-specific antibodies
were unable to bind native MOG on the
cell surface and induced no or little de-
myelination in animals with EAE, injec-
tion of conformation-dependent MOG-spe-
cific antibodies caused extensive demyeli-
nation and reproduced the immunopathol-
ogy and distribution of demyelinating
lesions, typically seen in MS. The genera-
tion of large amounts of peptide-specific
antibodies, that are not reactive with native
MOG, following immunization with
MOG
ex
produced recombinantly in Escheri-
chia coli, is probably due to the fact that
the antigen is denatured. This assumption
is supported by the observation that ani-
mals immunized with refolded MOG
ex
or
vaccinated with MOG DNA all show a
strong bias towards conformation-depen-
dent MOG-specific antibodies [92]. A simi-
lar correlation between epitope specificity
and demyelinating potential of MOG-spe-
cific antibodies might also exist in MS, as
heterogeneous MOG-peptide specific anti-
body responses are observed in many MS
patients as well as in healthy controls. An-
tibodies reactive to the MOG-peptide 21
40 were also identified in MS lesions [32],
putatively binding to MOG epitopes newly
exposed by protein degradation during
myelin destruction. However, knowledge
about the frequency of conformation-de-
pendent MOG-specific antibodies in MS is
very limited: Fabs directed at discontinu-
ous MOG epitopes were shown to compete
with antibodies of sera from MS patients
[35], one study specifically identified con-
formation-dependent MOG-specific anti-
bodies in the serum of one patient who
benefited strikingly from plasmapheresis
[34]. MS is a heterogeneous disease with
an extremely variable course, possibly
based on different pathogenetic mecha-
nisms. The development of sensitive as-
says that differentiate between antibodies
capable of binding to native MOG, and
pathologically irrelevant antibodies recog-
nizing linear peptide epitopes, is clearly
essential to evaluate their importance in
MS and to identify patients who may be
responsive to therapies targeting patho-
genic autoantibodies. The solution of the
molecular structures of MOG
ex
and
MOG
ex
bound to a Fab derived from the
conformation-dependent antibody 8-18C5
is an important step towards achieving
this goal. These data not only aid the inter-
pretation of present epitope mapping re-
sults, but also provide the basis to design
mutant proteins to characterize the confor-
mation-dependent MOG-specific antibody
responses in EAE and MS.
3.4.2
The Three-dimensional Structure of MOG
The extracellular domains of rat and
mouse MOG [49, 93] that have ~90% se-
quence identity to the human protein
adopt an immunoglobulin (Ig) fold with
features characteristic for IgV-like domains
resulting in a sandwich of two antiparallel
b-sheets comprised by strands A'GFCC'C''
and ABED (Fig. 3.3a). The Ca atoms of
the two b-sheets align very well with those
of other IgV-like proteins, such as various
variable antibody domains, the cd-T-cell re-
ceptor or their nearest structural relatives
B7-2 and sialoadhesin (root mean square
deviation <1.5 for ~100 aligned Cas). Of
interest, however, is the compactness of
the MOG IgV fold that lacks long flexible
loops but rather exhibits short connections
between strands and a multitude of hydro-
gen bonds in the loop regions that tightly
fix more than two-thirds of the loop resi-
dues. This may account for the lack of lin-
ear epitopes on the correctly folded MOG.
One exception are peptides encompassing
amino acids 6387 that bind weakly to a
subset of monoclonal mouse antibodies
[33]. The corresponding region on MOG
contains the protruding DE-loop (7280)
that is positioned at the top of MOG (is
easily accessible to antibodies); this is con-
sistent with these residues contributing to
a partially linear epitope (Fig. 3.3b). The
immunization of Lewis rats with MOG
peptide 3555 was shown to cause demye-
lination, possibly due to cross-reactivity of
the peptide with whole-length native MOG
[94]. In the MOG structure, amino acids
3555 encompass the internal, buried C
and C strands and the exposed CC'-loop
(4146) that, devoid of any crystallographic
contacts, is disordered in the rat MOG
ex
structure due to high flexibility, and thus
is well-suited to provoke peptideprotein
cross-reactivity.
3.4.3
The MOG
ex
(8-18C5)Fab Complex
The monoclonal mouse antibody 8-18C5,
used for this structural analysis, induces de-
myelination in vitro and in vivo, and prob-
ably recognizes an immunodominant epi-
tope on MOG, as indicated by its ability to
inhibit binding of 80% of a set of MOG-spe-
cific monoclonal mouse antibodies. The
structure of MOG
ex
complexed with the
Fab derived from 8-18C5 shows that 8-
18C5 binds to the upper membrane-distal
part of MOG, covering about 815
2
of sol-
vent-accessible MOG surface. As expected,
the epitope is highly discontinuous, and
consists of the N-terminus and the three
upper loops of MOG, that correspond to
the CDRs in the related variable antibody
domains, with additional contributions of
three separate residues (Fig. 3.4). All epi-
tope residues are totally conserved between
human, rat and mouse MOG, consistent
with the observed cross-reactivity of 8-
18C5 between species. 8-18C5 was raised
against native MOG and recognizes both
the glycosylated and unglycosylated protein,
indicating that the single glycosylation site
Asn31 on MOG does not contribute to the
epitope. This is supported by the structure
of the complex in which Asn31, though di-
rectly adjoining the epitope, does not inter-
act with the antibody.
The MOG
ex
(8-18C5)Fab complex repre-
sents a rather typical antigenantibody
structure. MOG interacts centrally with the
paratope of 8-18C5 formed by five of the
six CDRs, with dominant contributions
being made by the three heavy chain
CDRs (Fig. 3.4). Several buried water mol-
ecules are visible in the interface. Aro-
matic and arginine residues common
Fig. 3.3 Three-dimensional structure of myelin oligodendrocyte
glycoprotein (MOG). (a) Overall structure. (b) Putative linear
epitopes on MOG encompassing the CC-loop and the DE-loop,
respectively.
paratope amino acids form important in-
teractions with MOG. The light chain, for
instance, contributes two tyrosines to the
paratope that fix the FG-loop of MOG at
one side by strong hydrogen bonds. A
tryptophan residue of the heavy chain
CDR1 (H1-Trp33) forms a hydrogen bond
to Gln106 and extensive stacking interac-
tions with Arg101 of MOG; finally, two ar-
ginines of H2 that are the result of so-
matic mutations during maturation of the
antibody response interact specifically with
two MOG glutamate residues (Fig. 3.5).
The light chain residues of the paratope
correspond to their germline sequence; re-
markable for CDR1, however, is its length
of 17 amino acids the maximum ob-
served for Vj-CDR1 sequences that
enables CDR1 to interact with MOG.
Although the epitope of MOG is very dis-
continuous, amino acids 101108 that cor-
respond to the FG-loop and neighboring
strand residues dominate the interaction
by contributing 10 of the 12 hydrogen
bonds and about 65% of the total contact
area to the interaction. This raises the
question of why this sequence has never
been identified by epitope mapping using
linear peptides (Fig. 3.5). The FG-loop
(102105) forms a tight turn, classified as
a type II' b-turn, that usually exhibits a
glycine in the second position in order to
avoid a sterically unfavorable conforma-
tion. In MOG, this position is occupied by
His103, which results in a strained confor-
mation that is stabilized by the protein en-
vironment but is highly unlikely to be
adopted by linear peptides.
Based on the structure of the complex,
mutagenesis studies were performed to
Fig. 3.4 Stereo view of the MOG (8-18C5)Fab complex. MOG residues of the discontinuous
epitope are colored in red.
evaluate the importance of the observed
epitope. Interestingly, exchanging two ami-
no acids of the FG-loop in the center of
the epitope eliminated binding of nine out
of 10 murine monoclonal antibodies,
clearly demonstrating immunodominance
of the 8-18C5 epitope in mice. Provided
that a similarly biased reactivity exists in
humans, assaying the differential binding
of antibodies to mutant and wild-type
Fig. 3.5 Stereo view of MOG residues 101108 (yellow, green)
interacting with the 8-18C5 CDRs (heavy chain: blue, light chain: red).
Fig. 3.6 The 8-18C5 epitope on MOG. (a) Molecular surface
of MOG with epitope residues colored from orange to green
according to their position in the MOG sequence.
(b) Conservation between MOG and butyrophilin mapped
onto the molecular surface of MOG.
MOG may yield an easy and sensitive test
for pathogenic MOG-specific antibodies in
MS. Whereas MOG is a classical seques-
tered autoantigen proteins homologous to
MOG, such as butyrophilin (BTN), BTN-
like proteins or erythroid membrane asso-
ciated protein (ERMAP) that are ~50%
identical to MOG
ex
are expressed outside
the CNS, and able to induce tolerance. In-
triguingly, mapping the degree of conser-
vation onto the molecular surface of MOG
reveals a striking concordance of the epi-
tope region and residues unique to MOG
(Fig. 3.6). This correlation offers a simple
rationale for the observed specificity of
anti-MOG antibodies that might also be
applicable to other sequestered autoanti-
gens.
3.5
Conclusions
Autoantibody epitopes are currently being
elucidated in a wide range of autoimmune
diseases [95], with some of them having al-
ready become accepted tools in diagnosis
[30, 96, 97]. The advent of miniaturized
multiplex arrays in autoimmune research
provides the opportunity to perform large-
scale assays to detect specific antibody epi-
tope profiles in autoimmune disorders [98,
99] (see also Part I, Chapter 3 and Part V,
Chapter 8). The combined use of different
classic and novel epitope mapping meth-
ods, supported by information derived
from three-dimensional structures of auto-
antigens or their complexes with antibody
domains, will further increase the quality
and quantity of deciphered epitopes in or-
der to support diagnosis and, in the long
term, to guide the development of biophar-
maceuticals aimed at removing or neutra-
lizing pathogenic autoantibodies.
Acknowledgments
The author thanks Christopher Linington
and Uwe Jacob for their critical reading of
the manuscript.
References
1 Autoimmune diseases coordinating commit-
tee. (2003). Autoimmune Diseases Research
Plan, National Institute of Allergy and Infec-
tious Diseases, National Institutes of Health.
2 S. Walsh and L. Rau, (2000). Autoimmune
diseases: a leading cause of death among
young and middle-aged woman in the United
States. Am. J. Public Health 90, 14631466.
3 A. Vincent, O. Lily and J. Palace, (1999).
Pathogenic autoantibodies to neuronal pro-
teins in neurological disorders. J. Neuroimmu-
nol. 100, 169180.
4 M. De Baets and M. H. W. Stassen, (2002).
The role of antibodies in myasthenia gravis.
J. Neurol. Sci. 202, 511.
5 M. G. Robson and M. J. Walport, (2001). Patho-
genesis of systemic lupus erythematosus
(SLE). Clin. Lab. Med. 31, 678685.
6 R. H. Scofield, (2004). Autoantibodies as pre-
dictors of disease. Lancet 363, 15441546.
7 D. Devendra, L. P. Yu and G. S. Eisenbarth,
(2004). Endocrine autoantibodies. Clin. Lab.
Med. 24, 275.
8 M. Knip, (2002). Natural course of preclinical
type 1 diabetes. Hormone Res. 57, 611.
9 M. M. J. Nielen, et al., (2004). Specific autoan-
tibodies precede the symptoms of rheumatoid
arthritis A study of serial measurements in
blood donors. Arthritis Rheum. 50, 380386.
10 M. R. Arbuckle, et al., (2003). Development of
autoantibodies before the clinical onset of sys-
temic lupus erythematosus. N. Engl. J. Med.
349, 15261533.
11 D. DCruz, (2002). Testing for autoimmunity
in humans. Toxicol. Lett. 127, 93100.
12 J. Schmitt, (2003). Recombinant autoantigens
for diagnosis and therapy of autoimmune dis-
eases. Biomed. Pharmacother. 57, 261268.
13 S. Jones and J. M. Thornton, (1996). Principles
of proteinprotein interactions. Proc. Natl.
Acad. Sci. USA 93, 1320.
14 S. J. Wodak and J. Janin, (2002). Structural
basis of macromolecular recognition. Adv. Pro-
tein Chem. 61, 973.
15 E. J. Sundberg and R. A. Mariuzza, (2002). Mo-
lecular recognition in antibodyantigen com-
plexes. Adv. Protein Chem. 61, 119160.
16 M. H. V. Van Regenmortel, (1999). Molecular
dissection or protein antigens and the predic-
tion of epitopes. In: Synthetic Peptides as
Antigens (S. Pillai and P. C. van der Vliet,
Eds.), Vol. 28, pp. 179. Elsevier, Amsterdam.
17 L. Jin and J. A. Wells, (1994). Dissecting the
Energetics of an AntibodyAntigen Interface
by Alanine Shaving and Molecular Grafting.
Protein Sci. 3, 23512357.
18 G. A. Weiss, C. K. Watanabe, A. Zhong, A.
Goddard and S. S. Sidhu, (2000). Rapid map-
ping of protein functional epitopes by combi-
natorial alanine scanning. Proc. Natl. Acad.
Sci. USA 97, 89508954.
19 T. Clackson and J. A. Wells, (1995). A Hot
Spot of Binding Energy in a Hormone-Recep-
tor Interface. Science 267, 383386.
20 M. Z. Atassi and J. A. Smith, (1978). A propo-
sal for the nomenclature of antigenic sites in
peptides and proteins. Immunochemistry 15,
609610.
21 S. Muller (1999). Peptides in diagnosis of
autoimmune diseases. In: Synthetic Peptides
as Antigens (S. Pillai and P. C. van der Vliet,
Eds.), Vol. 28, pp. 247280. Elsevier, Amster-
dam.
22 L. Schachna, et al., (2002). Recognition of
Granzyme B-generated autoantigen fragments
in scleroderma patients with ischemic digital
loss.[see comment]. Arthritis Rheum. 46, 1873
1884.
23 E. L. Greidinger, L. Casciola-Rosen, S. M. Mor-
ris, R. W. Hoffman and A. Rosen, (2000). Au-
toantibody recognition of distinctly modified
forms of the U1-70-kd antigen is associated
with different clinical disease manifestations.
Arthritis Rheum. 43, 881888.
24 K. C. R. Malmegrim, G. J. M. Pruijn and W. J.
van Venrooij, (2002). The fate of the U1
snRNP autoantigen during apoptosis: Implica-
tions for systemic autoimmunity. Israel Med.
Assoc. J. 4, 706712.
25 P. A. C. Cloos and S. Christgau, (2004). Post-
translational modifications of proteins: impli-
cations for aging, antigen recognition, and
autoimmunity. Biogerontology 5, 139158.
26 M. A. van Boekel and W. J. van Venrooij,
(2003). Modifications of arginines and their
role in autoimmunity. Autoimmunity Rev. 2,
5762.
27 H. Brahms, et al., (2000). The C-terminal RG
dipeptide repeats of the spliceosomal Sm pro-
teins D1 and D3 contain symmetrical di-
methylarginines, which form a major B-cell
epitope for anti-Sm autoantibodies. J. Biol.
Chem. 275, 1712217129.
28 R. L. F. Nienhuis and E. A. Mandema, (1964).
A new serum factor in patients with rheuma-
toid arthritis: the antiperinuclear factor. Ann.
Rheum. Dis. 23, 302305.
29 G. A. Schellekens, B. A. de Jong, F. H. van den
Hoogen, L. B. van de Putte and W. J. van Ven-
rooij, (1998). Citrulline is an essential consti-
tuent of antigenic determinants recognized by
rheumatoid arthritis-specific autoantibodies.
J. Clin. Invest. 101, 273281.
30 G. A. Schellekens, et al., (2000). The diagnos-
tic properties of rheumatoid arthritis antibod-
ies recognizing a cyclic citrullinated peptide.
Arthritis Rheum. 43, 155163.
31 E. L. Greidinger, M. F. Foecking, S. Ranatunga
and R. W. Hoffman, (2002). Apoptotic U1-70
kd is antigenically distinct from the intact
form of the U1-70-kd molecule. Arthritis
Rheum. 46, 12641269.
32 C. P. Genain, B. Cannella, S. L. Hauser and
C. S. Raine, (1999). Identification of autoanti-
bodies associated with myelin damage in mul-
tiple sclerosis. Nat. Med. 5, 170175.
33 U. Brehm, S. J. Piddlesden, M. V. Gardinier
and C. Linington, (1999). Epitope specificity of
demyelinating monoclonal autoantibodies di-
rected against the human myelin oligodendro-
cyte glycoprotein (MOG). J. Neuroimmunol.
97, 915.
34 C. G. Haase, et al., (2001). The fine specificity
of the myelin oligodendrocyte glycoprotein au-
toantibody response in patients with multiple
sclerosis and normal healthy controls. J. Neu-
roimmunol. 114, 220225.
35 H. C. von Budingen, et al., (2002). Molecular
characterization of antibody specificities
against myelin/oligodendrocyte glycoprotein
in autoimmune demyelination. Proc. Natl.
Acad. Sci. USA 99, 82078212.
36 H. C. von Budingen, et al., (2004). Frontline:
Epitope recognition on the myelin/oligoden-
drocyte glycoprotein differentially influences
disease phenotype and antibody effector func-
tions in autoimmune demyelination. Eur. J.
Immunol. 34, 20722083.
37 L. Craig, et al., (1998). The role of structure in
antibody cross-reactivity between peptides and
folded proteins. J. Mol. Biol. 281, 183201.
38 M. Bluthner, M. Mahler, D. B. Muller, H.
Dunzl and F. A. Bautz, (2000). Identification
of an alpha-helical epitope region on the PM/
Scl-100 autoantigen with structural homology
to a region on the heterochromatin p25beta
autoantigen using immobilized overlapping
synthetic peptides. J. Molec. Med. 78, 4754.
39 Y. Tian, et al., (2002). Structureaffinity rela-
tionships in the gp41 ELDKWA epitope for
the HIV-(1) neutralizing monoclonal antibody
F-2(5): effects of side-chain and backbone
modifications and conformational constraints.
J. Peptide Res. 59, 264276.
40 R. C. Landry, et al., (2001). Antibody recogni-
tion of a conformational epitope in a peptide
antigen: Fv-peptide complex of an antibody
fragment specific for the mutant EGF recep-
tor, EGFRvIII. J. Mol. Biol. 308, 883893.
41 D. Zillikens, et al., (1997). Tight Clustering of
Extracellular Bp180 Epitopes Recognized by
Bullous Pemphigoid Autoantibodies. J. Invest.
Dermatol. 109, 573579.
42 E. Schmidt, et al., (2002). A highly sensitive
and simple assay for the detection of circulat-
ing autoantibodies against full-length bullous
pemphigoid antigen 180. J. Autoimmunity 18,
299309.
43 D. Zillikens, et al., (1997). A highly sensitive
enzyme-linked immunosorbent assay for the
detection of circulating anti-BP180 autoanti-
bodies in patients with bullous pemphigoid.
J. Invest. Dermatol. 109, 679683.
44 A. Kromminga, et al., (2002). Cicatricial pem-
phigoid differs from bullous pemphigoid and
pemphigoid gestations regarding the fine
specificity of autoantibodies to the
BP180NC16A domain. J. Dermatol. Sci. 28,
6875.
45 I. Laczk, et al., (2000). Conformational conse-
quences of coupling bullous pemphigoid anti-
genic peptides to glutathione-S-transferase
and their diagnostic significance. J. Peptide
Sci. 6, 378386.
46 D. R. Davies and G. H. Cohen, (1996). Interac-
tions of protein antigens with antibodies. Proc.
47 M. A. Newman, C. R. Mainhart, C. P. Mallett,
T. B. Lavoie and S. J. Smith-Gill, (1992).
Patterns of antibody specificity during the
BALB/c immune response to hen eggwhite
lysozyme. J. Immunol. 149, 32603272.
48 A. L. Corper, et al., (1997). Structure of human
IgM rheumatoid factor Fab bound to its auto-
antigen IgG Fc reveals a novel topology of
antibodyantigen interaction. Nature Struct.
Biol. 4, 374381.
49 C. Breithaupt, et al., (2003). Structural in-
sights into the antigenicity of myelin oligo-
dendrocyte glycoprotein. Proc. Natl. Acad. Sci.
USA 100, 94469451.
50 E. W. Henriksson, M. Wahren-Herlenius, I.
Lundberg, E. Mellquist and I. Pettersson,
(1999). Key residues revealed in a major con-
formational epitope of the U1-70K protein.
51 Y. Ma, et al., (2002). Key residues of a major
cytochrome P4502D6 epitope are located on
the surface of the molecule. J. Immunol. 169,
277285.
52 A. Burnens, S. Demotz, G. Corradin, H. Binz
and H. R. Bosshard, (1987). Epitope mapping
by chemical modification of free and anti-
body-bound protein antigen. Science 235, 780
783.
53 W. Fiedler, C. Borchers, M. Macht, S. O. Dei-
ninger and M. Przybylski, (1998). Molecular
characterization of a conformational epitope
of hen egg white lysozyme by differential
chemical modification of immune complexes
and mass spectrometric peptide mapping. Bio-
conj. Chem. 9, 236241.
54 R. Jemmerson and Y. Paterson, (1986). Map-
ping epitopes on a protein antigen by the pro-
teolysis of antigenantibody complexes.
Science 232, 10011004.
55 D. Suckau, et al., (1990). Molecular epitope
identification by limited proteolysis of an im-
mobilized antigenantibody complex and
mass spectrometric peptide mapping. Proc.
56 V. Legros, C. Jolivet-Reynaud, N. Battail-Poirot,
C. Saint-Pierre and E. Forest, (2000). Charac-
terization of an anti-Borrelia burgdorferi OspA
conformational epitope by limited proteolysis
of monoclonal antibody-bound antigen and
mass spectrometric peptide mapping. Protein
Sci. 9, 10021010.
57 D. I. R. Spencer, et al., (2002). A strategy for
mapping and neutralizing conformational im-
munogenic sites on protein therapeutics. Pro-
teomics 2, 271279.
58 R. H. Meloen, W. C. Puijk and J. W. Slootstra,
(2000). Mimotopes: realization of an unlikely
concept. J. Mol. Rec. 13, 352359.
59 H. M. Geysen, S. J. Rodda and T. J. Mason,
(1986). A priori delineation of a peptide which
mimics a discontinuous antigenic determi-
nant. Mol. Immunol. 23, 709715.
60 D. Enshell-Seijffers, et al., (2003). The map-
ping and reconstitution of a conformational
discontinuous B-cell epitope of HIV-1. J. Mol.
Biol. 334, 87101.
61 J. M. Davies, et al., (1999). Multiple alignment
and sorting of peptides derived from phage-
displayed random peptide libraries with poly-
clonal sera allows discrimination of relevant
phagotopes. Mol. Immunol. 36, 659667.
62 M. J. Rowley, et al., (2000). Prediction of the
immunodominant epitope of the pyruvate de-
hydrogenase complex E2 in primary biliary
cirrhosis using phage display. J. Immunol.
164, 34133419.
63 K. Beland, P. Lapierre, G. Marceau and F. Al-
varez, (2004). Anti-LC1 autoantibodies in pa-
tients with chronic hepatitis C virus infection.
J. Autoimmunity 22, 159166.
64 B. W. Bailey, et al., (2003). Constraints on the
conformation of the cytoplasmic face of dark-
adapted and light-excited rhodopsin inferred
from antirhodopsin antibody imprints. Protein
Sci. 12, 24532475.
65 D. Bresson, et al., (2003). Localization of the
discontinuous immunodominant region rec-
ognized by human anti-thyroperoxidase auto-
antibodies in autoimmune thyroid diseases. J.
Biol. Chem. 278, 95609569.
66 A. Wiik, (2003). Autoantibodies in vasculitis.
Arthritis Res. Ther. 5, 147152.
67 J. Bartunkova, V. Tesar and A. Sediva, (2003).
Diagnostic and pathogenetic role of antineu-
trophil cytoplasmic autoantibodies. Clin. Im-
munol. 106, 7382.
68 Y. M. Van der Geld, C. A. Stegeman and C. G.
M. Kallenberg, (2004). B cell epitope specifici-
ty in ANCA-associated vasculitis: does it mat-
ter? Clin. Exp. Immunol. 137, 451459.
69 R. C. Williams, et al., (1994). Epitopes on Pro-
teinase-3 Recognized by Antibodies from Pa-
tients with Wegeners Granulomatosis. J. Im-
munol. 152, 47224737.
70 M. E. Griffith, A. Coulthart, S. Pemberton,
A. J. George and C. D. Pusey, (2001). Anti-neu-
trophil cytoplasmic antibodies (ANCA) from
patients with systemic vasculitis recognize re-
stricted epitopes of proteinase 3 involving the
catalytic site. Clin. Exp. Immunol. 123, 170
177.
71 Y. M. van der Geld, et al., (2001). Antineutro-
phil cytoplasmic antibodies to proteinase 3 in
Wegeners granulomatosis: Epitope analysis
using synthetic peptides. Kidney Int. 59, 147
159.
72 M. Fujinaga, M. M. Chernaia, R. Halenbeck,
K. Koths and M. N. James, (1996). The crystal
structure of PR3, a neutrophil serine pro-
teinase antigen of Wegeners granulomatosis
antibodies. J. Mol. Biol. 261, 267278.
73 K. O. Netzer, et al., (1999). The Goodpasture
autoantigen Mapping the major conforma-
tional epitope(s) of alpha 3(IV) collagen to re-
sidues 17-31 and 127-141 of the NC1 domain.
J. Biol. Chem. 274, 1126711274.
74 T. Hellmark, H. Burkhardt and J. Wieslander,
(1999). Goodpasture disease Characteriza-
tion of a single conformational epitope as the
target of pathogenic autoantibodies. J. Biol.
Chem. 274, 2586225868.
75 M. E. Than, et al., (2002). The 1.9-angstrom
crystal structure of the noncollagenous (NC1)
domain of human placenta collagen IV shows
stabilization via a novel type of covalent Met-
Lys cross-link. Proc. Natl. Acad. Sci. USA 99,
66076612.
76 M. Sundaramoorthy, M. Meiyappan, P. Todd
and B. G. Hudson, (2002). Crystal structure of
NC1 domains Structural basis for type IV
collagen assembly in basement membranes.
J. Biol. Chem. 277, 3114231153.
77 D. B. Borza, et al., (2002). Quaternary organi-
zation of the Goodpasture autoantigen, the al-
pha 3(IV) collagen chain Sequestration of
two cryptic autoepitopes by intraprotomer in-
teractions with the alpha 4 and alpha 5 NC1
domains. J. Biol. Chem. 277, 4007540083.
78 M. Segelmark, T. Hellmark and J. Wieslander,
(2003). The prognostic significance in Good-
pastures disease of specificity, titre and affini-
ty of anti-glomerular-basement-membrane an-
tibodies. Nephron Clin. Pract. 94, C59C68.
79 C. J. Padoa, et al., (2003). Recombinant Fabs
of human monoclonal antibodies specific to
the middle epitope of GAD65 inhibit type 1
diabetes-specific GAD65Abs. Diabetes 52,
26892695.
80 P. Burkhard, P. Dominici, C. Borri-Voltattorni,
J. N. Jansonius and V. N. Malashkevich,
(2001). Structural insight into Parkinsons dis-
ease treatment from drug-inhibited DOPA
decarboxylase. Nature Struct. Biol. 8, 963967.
81 H. L. Schwartz, et al., (1999). High-resolution
autoreactive epitope mapping and structural
modeling of the 65 kDa form of human gluta-
mic acid decarboxylase. J. Mol. Biol. 287, 983
999.
82 J. C. Tong, M. A. Myers, I. R. Mackay, P. Z.
Zimmet and M. J. Rowley, (2002). The PEV-
KEK region of the pyridoxal phosphate bind-
ing domain of GAD65 expresses a dominant
B cell epitope for type 1 diabetes sera. Ann. N.
Y. Acad. Sci. 958, 182189.
83 M. A. Myers, et al., (2000). Conformational
epitopes on the diabetes autoantigen GAD65
identified by peptide phage display and molec-
ular modeling. J. Immunol. 165, 38303838.
84 B. Sutton, A. Corper, V. Bonagura and M.
Taussig, (2000). The structure and origin of
rheumatoid factors. Immunol. Today 21, 177
183.
85 T. G. Johns and C. C. A. Bernard, (1999). The
structure and function of myelin oligodendro-
cyte glycoprotein. J. Neurochem. 72, 19.
86 R. Lebar, C. Lubetzki, C. Vincent, P. Lombrail
and J. M. Boutry, (1986). The M2 autoantigen
of central nervous system myelin, a glycopro-
tein present in oligodendrocyte membrane.
Clin. Exp. Immunol. 66, 423434.
87 C. Linnington, M. Webb and P. L. Woodhams,
(1984). A novel myelin-associated glycoprotein
defined by a mouse monoclonal antibody. J.
Neuroimmunol. 6, 387396.
88 A. Iglesias, J. Bauer, T. Litzenburger, A. Schu-
bart and C. Linington, (2001). T- and B-cell re-
sponses to myelin oligodendrocyte glycopro-
tein in experimental autoimmune encephalo-
myelitis and multiple sclerosis. Glia 36, 220
234.
89 H. C. Von Budingen, et al., (2001). Immune
responses against the myelin/oligodendrocyte
glycoprotein in experimental autoimmune de-
myelination. J. Clin. Immunol. 21, 155170.
90 E. Mathey, C. Breithaupt, A. S. Schubart and
C. Linington, (2004). Commentary: Sorting
the wheat from the chaff: identifying demyeli-
nating components of the myelin oligodendro-
cyte glycoprotein (MOG)-specific autoantibody
repertoire. Eur. J. Immunol. 34, 20652071.
91 T. Berger, et al., (2003). Antimyelin antibodies
as a predictor of clinically definite multiple
sclerosis after a first demyelinating event. N.
Engl. J. Med. 349, 139145.
92 C. Linington, personal communication.
93 C. S. Clements, et al., (2003). The crystal struc-
ture of myelin oligodendrocyte glycoprotein, a
key autoantigen in multiple sclerosis. Proc.
Natl. Acad. Sci. USA 100, 1105911064.
94 M. Ichikawa, T. G. Johns, J. L. Liu and C. C. A.
Bernard, (1996). Analysis of the Fine B Cell
Specificity During the Chronic/Relapsing
Course of a Multiple Sclerosis-Like Disease in
Lewis Rats Injected with the Encephalitogenic
Myelin Oligodendrocyte Glycoprotein Peptide
35-55. J. Immunol. 157, 919926.
95 M. Mahler, M. Bluthner and K. M. Pollard,
(2003). Advances in B-cell epitope analysis of
autoantigens in connective tissue diseases.
Clin. Immunol. 107, 6579.
96 B. G. Hudson, K. Tryggvason, M. Sundara-
moorthy and E. G. Neilson, (2003). Alports
syndrome, Goodpastures syndrome, and type
IV collagen. N. Engl. J. Med. 348, 25432556.
97 J. G. Routsias, A. G. Tzioufas and H. M. Mout-
sopoulos, (2004). The clinical value of intracel-
lular autoantigens B-cell epitopes in systemic
rheumatic diseases. Clin. Chim. Acta 340, 1
25.
98 W. Hueber, P. J. Utz, L. Steinman and W. H.
Robinson, (2002). Autoantibody profiling for
the study and treatment of autoimmune dis-
ease. Arthritis Res. 4, 290295.
99 W. H. Robinson, et al., (2002). Autoantigen
microarrays for multiplex characterization of
autoantibody responses. Nat. Med. 8, 295301.
Abstract
Classically, medical imaging provides struc-
tural information of the patients body, and
is used mainly for diagnostic purposes.
The recent advances in the development of
contrast agents that can highlight mole-
cules or molecular structures and pathways
allow researchers and physicians to obtain
ever more detailed information. This field
of molecular imaging promises to improve
dramatically the future of healthcare, shift-
ing the emphasis toward much earlier diag-
nosis and treatment. The combination of
molecular imaging with the advent of de-
vices for the imaging of small animals ren-
ders it increasingly interesting for use in
drug discovery and drug development. This
chapter provides an introduction to the dif-
ferent imaging techniques available, to-
gether with examples of contrast agents
and applications for molecular imaging. Fo-
cus is then turned to the potential and im-
plications of the technique for drug discov-
ery and the development of modern bio-
pharmaceuticals.
Abbreviations
CEST chemical exchange-dependent
saturation transfer
CT computed tomography
DOT diffuse optical tomography
EAE experimental autoimmune
encephalomyelitis
ED-B-FN extra-domain B-fibronectin
FDG fluorodeoxyglucose
FID free induction decay
Gd-DTPA gadolinium diethylenetri-
aminepenta-acetic acid
ICAM-1 intercellular adhesion mole-
cule-1
ICG indocyanine green
MMP matrix metalloproteinase
MP microparticle
NIH National Institute of Health
NIR near infrared
NMR nuclear magnetic resonance
PET positron emission tomogra-
phy
1211
ISBN: 3-527-31184-X
4
Molecular Imaging and Applications for Pharmaceutical R&D
Find, Fight, and Follow
Target-specific Troika from Mother Natures Pharmacopoiea
SPECT single photon emission com-
puted tomography
SPIO superparamagnetic iron oxide
USPIO ultra-small superparamag-
netic iron oxide
factor
VEGFR2 vascular endothelial growth
factor receptor-2
4.1
Introduction
In medical diagnostics, several techniques
are available to obtain an image of a part
or the whole of a patients body. The use
of imaging techniques, also called modali-
ties, has grown exponentially during the
past 30 years. Imaging techniques have
now become essential tools, not only in
clinical diagnostics and therapy monitor-
ing, but also for pharmaceutical industries.
The reason is clear: bringing a new drug
to market is a time-consuming and expen-
sive undertaking, requiring approximately
15 years and an investment of US$ 800
900 million. One costly factor is caused by
the standard trial end-points of morbidity
and mortality. Using imaging, other mark-
ers of drug efficacy and toxicity may be de-
veloped, potentially offering faster and
more quantitative data. The advantages
would be earlier decisions on whether to
continue studies on a potential drug or to
discard it, and earlier starts of clinical
trials, thus reducing time to market. The
United States Food and Drug Administra-
tion (FDA) has also recognized this and re-
cently started an initiative to investigate
the use of imaging results as biomarkers
for biopharmaceutical development.
Traditionally, medical imaging provides
information on the anatomical level, and
diseases or effects of therapy are mainly
detected when structural abnormalities oc-
cur. Additional contrast between different
tissues may be gained through the use of
contrast agents such as paramagnetic na-
noparticles or radiolabeled molecules.
About 10 years ago, a trend started to-
wards the development of contrast agents
that carry a recognition element for target-
ing to a certain molecule. Using such
agents to visualize the concentration or ac-
tivity of a specific molecule in vivo is called
molecular imaging [1]. This technology
has very important advantages for the di-
agnosis of cancer and other diseases,
where subtle cellular changes occur well
before an anatomical abnormality such as
a tumor becomes detectable. Molecular
imaging may also aid therapy decisions if
it can be used to distinguish between
structures with a similar morphology but
different molecular malfunctions, such as
benign and malignant tumors. Further-
more, imaging on a molecular level can
provide information on therapy response
before an anatomical effect on diseased tis-
sue can be detected, enabling much faster
treatment optimization. Molecular imaging
not only has a high potential impact on
medical diagnostics but, in combination
with the advances in small animal imag-
ing equipment and animal models for dis-
eases, also on drug discovery and develop-
ment. In this chapter, we will discuss
these promising technologies and provide
application examples. We would like to
give a broad overview, and thus will by no
means be comprehensive. Therefore,
throughout the text we refer to outstand-
ing reviews that deal more extensively with
a certain subject. Before we dive into mo-
lecular imaging, we will first examine the
different imaging techniques available.
4.2
Imaging Modalities and Contrast Agents
Several techniques have been developed
that non-invasively provide images of the
body in vivo. As each technique detects
how another form of energy interacts with
tissue, each has its own characteristics in
terms of sensitivity, spatial resolution,
available contrast agents, ease of use and
costs, several of which are listed in Table
4.1 in Section 4.3.3. Together, the modali-
ties offer a wide spectrum of possibilities
and applications and can, in general, be
considered complementary.
4.2.1
X-ray Imaging
X-ray imaging is a transmission-based
technique, in which X-rays from a source
pass through the patient and are detected
on the opposite side. The contrast in the
image arises from the different attenuation
of X-rays in different tissues, and the
amount of absorption depends on the tis-
sue composition. For example, dense bone
matter will absorb many more X-rays than
soft tissues, such as muscle and fat. The
contrast also depends on the energy of the
X-rays in soft tissue, low-energy X-rays
result in better contrast. Thus, another im-
portant factor for the overall image quality
of X-ray images is the X-ray tube. This
tube consists of two electrodes, a negative-
ly charged cathode and a positively
charged anode. The heated cathode gener-
ates electrons, which are accelerated by ap-
plying a high voltage (20150 kV). The ac-
celerated electrons strike on the anode,
generating X-rays. The resultant beam con-
sists of X-rays of a broad range of ener-
gies, where the maximum energy depends
on the applied voltage. In planar X-ray, the
line integral of the spatially dependent at-
tenuation is measured and the resulting
intensity is displayed in a two-dimensional
image. However, it is difficult to interpret
overlapping layers of soft tissue and bone
structures on such a projection image. In
order to resolve such three-dimensional
(3D) structures, X-ray computed tomogra-
phy (CT) is used, which generates cross-
sectional, two-dimensional (2D) images of
the body. Due to the high spatial resolu-
tion of X-ray CT images, the technique is
widely applied in structural imaging, that
is it is used to depict morphology.
In CT, typically an X-ray tube and a de-
tector are rapidly rotated 3608 around the
patient (see Fig. 4.1). The acquisition time
of an image is determined by the rotation
time of the detectors and the X-ray tube,
which ranges from 0.3 to 1 second. Mod-
ern CT scanners use a number (1664) of
detector rows, which allows the simulta-
neous measurement of multiple slices.
Further reduction of the scan time can be
achieved by faster rotation times. There-
Fig. 4.1 Schematic set-up of X-ray computed to-
mography. An X-ray tube and a detector rotate
around the patient. A collimator in front of the
tube shapes the X-ray beam. Absorption of the
transmitted X-rays is measured on the other side
by a detector.
fore, the improvement of the mechanical
design of the gantry, which carries the X-
ray tube and the detector, is essential. The
final image is reconstructed from the mea-
surements by applying either a filtered
backprojection algorithm or more ad-
vanced iterative reconstruction techniques.
Each pixel contains the averaged attenua-
tion values within the corresponding voxel
(the smallest volume unit in the image).
This number is compared to the attenua-
tion value of water and displayed on a
scale of Hounsfield units (HU), named
after the inventor of CT, Sir Godfrey
Hounsfield. The scale assigns an attenua-
tion value of zero to water and regular at-
tenuation values range from 1000 to
3000 HU. The attenuation value of soft tis-
sue ranges between 40 and 100 HU.
The spatial resolution in X-ray CT de-
pends on the focal spot of the X-ray tube
and the size of the detector elements. The
spatial resolution of a clinical CT scanner
is less than 0.5 mm in the center of the
CT scanner. For pre-clinical imaging, sev-
eral micro-CT systems have been devel-
oped and are commercially available [2].
The spatial resolution of dedicated small-
animal scanners is much higher than for
clinical scanners, and is in the order of
20 lm. Fig. 4.2 shows high-resolution
images obtained with an animal CT scan-
ner. The major advantage of X-ray CT is
its ease of use (push-button-technology)
for acquiring large 3D datasets with struc-
tural information at a very high spatial re-
solution. The disadvantage of X-ray CT is
the use of ionizing radiation, which can
lead to cell death or to cancer due to ge-
netic mutations. For example, the effective
dose of a clinical CT scan of the abdomen
is about 10 mSv, which is about 400 times
higher than the dose of a chest projection
X-ray.
4.2.1.1 X-ray Contrast Agents
X-ray contrast agents are chemicals that
are introduced into the body to increase
the image contrast. They contain sub-
stances with a high atomic number that
increase the attenuation value in the re-
gions where they accumulate. A typical X-
ray agent used to image the gastrointesti-
nal tract is barium sulfate, a solution of
which the patient drinks hours prior to ex-
amination. For imaging of the colon, air is
also used, which has a high negative
Hounsfield value (1000). Iodine-contain-
ing contrast agents are widely used for X-
ray-based angiography, and are also ap-
Fig. 4.2 Dedicated animal X-ray computed tomo-
graphy system that allows high-resolution imaging
(up to 18 lm) of rodents. The different attenuation
of X-rays in various tissues and bones allows volume
rendering of the different structures (right image)
(Courtesy of ImTek, Inc., Knoxville, TN, USA).
plied in the detection of tumors. These
agents are injected into the bloodstream,
where they cause the attenuation of X-rays
to be higher than in the surrounding tis-
sue. The strength of the attenuation de-
pends on the iodine load of the agent.
However, a higher iodine load usually in-
creases the osmolarity of the solution,
which can change the cell volume and
thus can lead to adverse reactions. Re-
cently, a new generation of low- or iso-os-
motic agents has been introduced.
4.2.2
Magnetic Resonance Imaging
Magnetic resonance imaging (MRI) is a
non-ionizing imaging technique with su-
perior soft-tissue contrast, high spatial re-
solution and good temporal resolution.
MRI is capable of measuring a wide range
of endogenous contrast mechanisms that
include proton density, spin-lattice relaxa-
tion time (T
1
), spin-spin relaxation time
(T
2
), chemical shift, temperature and dif-
ferent types of motion, such as blood flow,
perfusion, or diffusion. MRI has become
the modality of choice in many pre-clinical
and clinical applications, because it can
provide structural and functional informa-
tion. With ongoing developments to im-
prove the image quality, acquisition speed
and quantitative accuracy, the range of ap-
plications for MRI continues to expand
rapidly.
Given the short length of this chapter,
the description of MRI must be superficial
(for a detailed description, see Refs. [3, 4]).
MRI is based on the nuclear magnetic res-
onance (NMR) effect that was observed by
Bloch and Purcell in 1946. Three basic re-
quirements must be satisfied to measure
an MR-signal. First, the nucleus of interest
must possess a non-zero magnetic mo-
ment. All nuclei that have an odd number
of protons and/or neutrons have this prop-
erty and behave as small magnets. Typical
nuclei are, for example, hydrogen (1-H),
phosphorus (31-P), carbon (13-C), sodium
(23-Na), or fluorine (19-F). In the absence
of an external magnetic field, these indi-
vidual magnetic moments are randomly
oriented and there is no net magnetiza-
tion. The second requirement is an exter-
nal static magnetic field B
0
. Typical field
strengths for clinical MRI are between
0.23 and 3 Tesla, whereas they are much
higher for pre-clinical systems (4.77 Tesla,
and even >10 Tesla). Due to the Zeemann
effect, magnetic moments align under spe-
cific angles along or opposed to the exter-
nal field B
0
, resulting in a precessional
movement of the magnetic moments. The
precessional frequency, also called Larmor
frequency, is given by f
0
=c B
0
, where c is
the gyromagnetic ratio, a constant for a
given nucleus. NMR of hydrogen is the
most important for clinical applications,
because hydrogen is highly present in bio-
logical tissues and its gyromagnetic ratio
is the largest of all nuclei. Since an align-
ment parallel to the field is the lower en-
ergy state, it is preferred and slightly more
nuclei will align along rather than opposed
to the field. As a result, the tissue will ex-
hibit a net magnetization, which is parallel
to the external magnetic field and is called
longitudinal magnetization. The amount
of the net magnetization depends on the
field strength and increases for higher
magnetic fields. The third requirement is
a time-varying magnetic field B
1
applied
perpendicularly to the static B
0
field and at
Larmor-frequency (i.e., at resonance condi-
tion). For this, an additional radiofre-
quency (RF-) coil produces a B
1
-pulse of a
certain amplitude and duration. Such a B
1
-
pulse flips the longitudinal magnetization
to an arbitrary angle (also called flip-an-
gle). The transverse component of the
flipped magnetization precesses around
the static B
0
field at Larmor frequency and
induces a time varying voltage signal in
the RF-coil (Fig. 4.3).
The detected transverse magnetization
does not remain forever, since two indepen-
dent relaxation processes take place. First,
the spin-lattice relaxation describes how fast
the longitudinal magnetization recovers
after applying the B
1
-pulse. The rate of the
recovery process is determined by the re-
laxation time T
1
. Second, the spin-spin re-
laxation describes how fast the transversal
magnetization loses its coherence and thus
decays. The rate of dephasing is determined
by the relaxation time T
2
. In addition to
spin-spin interactions, dephasing between
the coherently precessing magnetic mo-
ments can also be caused by B
0
-field inho-
mogeneities. As a result, an apparently
stronger relaxation process is visible, which
is called T
2
*
relaxation. This T
2
*
relaxation de-
scribes the decay of a time varying signal,
which is called the free induction decay
(FID).
Both the spin-lattice relaxation time T
1
and the spin-spin relaxation time T
2
vary
among different types of tissue, and T
1
is
always larger than T
2
. In addition, the T
1
relaxation time depends on the field
strength B
0
, whereas the T
2
time is inde-
pendent. The signal amplitude depends on
the timing of the experiment and the re-
laxation times T
1
and T
2
(*). It can also be
influenced by endogenous contrast mecha-
nisms such as diffusion or blood-oxygena-
tion.
In order to distinguish signals from dif-
ferent spatial locations, magnetic field gra-
dients are applied by using gradient coils.
The gradient coils create a linear variation
in the z-component of the static magnetic
field. Consequently, with the spatially vary-
ing field strength a spatially varying pre-
cessional frequency is connected. Usually,
for 3D encoding, not all gradients are ap-
plied simultaneously and the image forma-
tion process can be separated into three
phases: slice selection, phase encoding,
and frequency encoding. After performing
Fig. 4.3 Schematic set-up of an MRI system. The magnet
causes a strong homogeneous static magnetic field, B
0
. The
gradient coil creates a linear variation of the magnetic field in
all three dimensions. The RF-coil creates a time varying field
that is perpendicular to the static magnetic field B
0
.
a number of experiments with different
gradient values, an image can be recon-
structed by using a Fourier-transform of
the obtained signals. The spatial resolution
strongly depends on the amplitude of the
gradients and the acquisition bandwidth.
Typical values of the spatial resolution of
clinical scanners are in the order of 0.5
1 mm. However, high-resolution imaging
on clinical MR-scanners is possible using
dedicated RF-coils to increase the sensitiv-
ity. With this a spatial resolution of about
100 lm can be achieved (Fig. 4.4). A high-
er spatial resolution is possible in dedi-
cated animal MR-scanners that operate at
a higher magnetic field strength (e.g.,
7 Tesla) and which apply strong gradients.
4.2.2.1 MRI Contrast Agents
In some clinical situations, the intrinsic
contrast of the tissue is not sufficient to
distinguish pathological from healthy tis-
sue. Therefore, the use of contrast-enhanc-
ing agents has become an integral part of
MR imaging. There are two basic classes
of MRI contrast agents: paramagnetic and
superparamagnetic agents. Paramagnetic
agents primarily shorten the T
1
relaxation
time of the tissue in which they accumu-
late. A more detailed description of MRI
contrast agents can be found in Ref. [5].
Paramagnetic agents are based on metal
ions with one or more unpaired electrons.
These unpaired electrons result in a very
large magnetic moment that interacts with
the much smaller magnetic moments of
the nucleus. Molecular motions result in
random fluctuations of the dipolar mag-
netic interaction that reduces both the T
1
and the T
2
relaxation times. Gadolinium
(Gd
3+
) and manganese (Mn
2+
) are exam-
ples of paramagnetic ions that are used in
MR contrast agents. Since these metal
ions are highly toxic, they must be con-
tained in a chelate to prevent circulation of
free ions in the body. Most clinically used
agents base on gadolinium and differ only
in the chelating agents; for example, the
most commonly used clinical paramag-
netic contrast agent is gadolinium diethyl-
enetriaminepenta-acetic acid (Gd-DTPA;
tradename Magnevist
, Schering AG, Ber-

lin, Germany). The DTPA-chelate does not
bind to blood proteins, ensuring rapid dis-
tribution through the bloodstream and fast
clearance through the kidneys. Another
class of paramagnetic agents is specifically
designed to remain in the blood pool for a
longer period of time. These blood pool
agents either are of larger size, or they
bind reversibly to albumin in blood plas-
ma [6]. Recently, a new class of contrast
agents was proposed that are based on
chemical exchange-dependent saturation
transfer (CEST) and which can be used to
measure pH in vivo [7].
Superparamagnetic agents consist of
small magnetic particles. These usually
consist of a crystalline core comprising a
Fig. 4.4 Animal MR image with 100 lm resolution
obtained on a clinical 3 Tesla MR scanner using a
dedicated MR receive coil (Philips Research, Ham-
burg, Germany).
mixture of iron oxides (Fe
2
O
3
and Fe
3
O
4
)
coated in a polymer matrix. The particles
are divided into two classes according to
their overall size: if the diameter is larger
than 50 nm they are called superparamag-
netic iron oxide (SPIO) particles, but if the
diameter is smaller they are categorized as
ultra-small superparamagnetic iron oxide
(USPIO) particles. These contrast agents
function by causing local field inhomo-
geneities that result in different relaxation
regimes. Water molecules that diffuse
through the local field inhomogeneities
undergo T
2
and/or T
2
*
decay. In addition,
USPIO particles also have excellent T
1
-en-
hancing properties.
4.2.3
Ultrasound Imaging
Ultrasound imaging is a non-invasive, por-
table and relatively inexpensive imaging
modality, which is used extensively in the
clinic. An ultrasound transducer (also
called scanhead) sends short pulses of a
high-frequency sound wave (110 MHz)
into the body. At interfaces between two
types of tissue, the wave will be refracted
and part of the sound wave is reflected
back due to Snells law. How much is re-
flected depends on the densities of the re-
spective tissues, and thus the speed of the
sound wave within the different tissues. In
addition, parts of the sound wave are also
backscattered from small structures at tis-
sue boundaries or within the tissue. High-
frequency sound waves propagate well
through soft tissue and fluids, but they are
more or less stopped by air or bone. In
clinical practice, this limitation is referred
to as an acoustic window. The transducer
not only sends the wave into the body but
also receives part of the reflected and/or
backscattered wave, also named echo. In
clinical practice, ultrasound is used in a
wide variety of imaging situations includ-
ing imaging of the heart (echocardiogra-
phy), liver, kidney, ovaries, breast, periph-
eral vascular system, and even portions of
the brain.
Diagnostic ultrasonic images can be gen-
erated using a variety of clinical imaging
modes including A-lines, B-mode, and M-
mode. The nomenclature is inherited from
the world of radar. A-line mode is an older
imaging mode where a one-dimensional
amplitude line representing the propaga-
tion of sound along one line is shown. B-
mode scanning is the most familiar clinical
imaging mode and is represented by a 2D
image. M-mode is the practice of rapidly fir-
ing one line of sight through a moving or-
gan. This allows the tracking of motion of
a structure such as a cardiac valve or cardiac
wall with great time resolution.
Each line can be obtained in less than
100 ls. Thus, an image consisting of 100
lines can be obtained in less than 10 ms,
which means that real-time imaging is
possible. Therefore, in addition to imaging
morphology, ultrasound is also capable of
measuring the velocity of blood in circula-
tion using the Doppler effect. The move-
ment of red blood cells causes a shift in
the frequency of returning ultrasound
waves. The Doppler frequency shift is pro-
portional to the blood flow velocity and
thus allows quantification.
For the spatial resolution of an ultra-
sound system, three different dimensions
must be considered: the axial, the lateral,
and the elevation dimension. The axial re-
solution along the axis of the transducer is
defined as the closest separation of two
echoes that can be resolved and improves
at higher frequencies. However, the pene-
tration depth of ultrasound waves de-
creases with increasing frequencies. There-
fore, lower frequencies (13 MHz) are
used for studies of deep-lying structures,
while higher frequencies (510 MHz) are
used to image regions that are closer to
the body surface. Typical values of the ax-
ial resolution are 12 mm at 1 MHz and
0.3 mm at 5 MHz. The lateral and eleva-
tion resolution of the ultrasound beam are
determined by its thickness at the focal
plane. For a single-crystal transducer, the
lateral width of the ultrasound beam is de-
termined by the diameter of the transdu-
cer and for clinical scanheads typically
ranges from 0.5 to 2 mm at the focal
plane. Nowadays, almost all commercially
available transducers consist of an array of
small piezoelectric crystals. Each of these
array elements can be electronically con-
trolled for transmission and reception of
the ultrasound signal. In a phased array, a
large number of array elements are con-
trolled simultaneously to shape and to
steer the location of focal plane of the ul-
trasound beam. Due to this dynamic fo-
cusing approach, the lateral resolution can
be increased. The elevation dimension is
determined by the length of the crystal ele-
ments and is in the order of 23 mm for
clinical scanheads. However, with dedi-
cated animal scanheads and/or systems
operating at higher frequencies (up to
50 MHz) a much higher spatial resolution
(below 100 lm) can be obtained on small
animals. Fig. 4.5 shows an ultrasound im-
age of a mouse heart taken on a clinical
scanner using such a dedicated scanhead.
4.2.3.1 Ultrasound Contrast Agents
In several instances, the contrast in ultra-
sound is not high enough. For example, in
1015% of patients normal echocardiogra-
phy is not possible for anatomic or other
reasons. In addition, Doppler-based meth-
ods to measure blood flow sometimes fail,
due to masking of the signal by overlying
tissue. For these applications, microbubbles
can markedly enhance contrast. Microbub-
bles are gas-filled microspheres with a di-
ameter of several micrometers, which are
stabilized by a shell composed of, for exam-
ple, albumin or lipids. Several versions of
these diagnostic blood pool contrast agents
are available commercially, and approved
for better delineation of the chambers of
the heart [8]. Microbubbles can be used to
increase the echogenicity of blood via two
different mechanisms. The first mecha-
nism is resonance of microbubbles that ex-
pand and contract in an ultrasound field. At
the resonant frequency, strong signals are
generated at multiples of the transmitted
frequency, also called harmonics. These
can be very well detected with an ultrasound
technique called harmonic imaging
(Fig. 4.6), in which frequency filtering can
be used to receive the harmonics. The sec-
ond mechanism bases on differences in
the acoustic impedance and thus increases
the backscattering. In addition to resonance
phenomena, microbubbles also show an in-
creased scattering of the ultrasound wave
[9], which is stronger than that for red blood
cells. This property can be used to enhance
Fig. 4.5 Ultrasound image of a mouse heart using
a dedicated animal scanhead on a clinical ultra-
sound scanner. A microbubble contrast agent pro-
vides a high contrast between blood (white) and
myocardium. (Data courtesy K. Tiemann, Univer-
sity of Bonn, Germany).
Doppler imaging of blood flow. Besides mi-
crobubbles, a targeted perfluorocarbon
emulsion also showed an increase of back-
scattering upon accumulation and can thus
be considered an acoustic mirror [10].
4.2.4
Nuclear Imaging
Nuclear imaging is based on the detection
of gamma rays that are emitted by radio-
nuclides. For this, small amounts (typically
nanograms) of radionuclides are injected
into the organisms. In contrast to other
imaging modalities such as MRI, X-ray or
ultrasound, nuclear imaging does not pro-
vide morphology information but images
the spatial distribution of radionuclides in
the organism. This distribution depends
strongly on the biological behavior of a
radiopharmaceutical. Therefore, the devel-
opment and synthesis of radiopharmaceu-
ticals is key in nuclear imaging to obtain
physiological, metabolic, and molecular in-
formation. Two main nuclear imaging
techniques can be distinguished due to the
use of different types of radionuclides:
Single photon emitters decay under the
emission of gamma rays with energies
between 100 and 360 keV.
Positron emitters decay under the emis-
sion of positrons that result in a pair of
high-energy gamma rays (511 keV) after
annihilation with an electron.
The corresponding 3D imaging tech-
niques are called single photon emission
computed tomography (SPECT) and posi-
tron emission tomography (PET). As PET
is discussed in detail elsewhere in this
book (see Part V, Chapter 5), we will focus
here on SPECT.
Fig. 4.7 shows the basic principles and
instrumentation of SPECT. The injected
radiopharmaceutical has accumulated in a
suspicious region in the body. During de-
cay of the radionuclides, gamma rays are
emitted in all directions. Some of the gam-
ma rays are attenuated and scattered in
the body. In order to detect the gamma
rays, a gamma camera is rotated around
the body. The basic design of a gamma
camera was described by Hal Anger in
1953 [11], which mainly consists of three
parts: a collimator, a scintillation crystal,
and a number of photomultipliers.
The collimator selects only those gamma
rays that have a trajectory at an angle of
90
o
to the detector plane, and blocks all
others. The collimator is generally a lead
structure with a honeycomb array of holes,
where the lead walls (septa) are designed
to prevent penetration of gamma rays
from one hole to the other. The parallel
hole collimator is the most widely used
collimator. Other types of collimators can
be used to magnify the size of an object
on the image. A pinhole collimator is an
extreme form of a converging collimator
and allows magnifying small objects
placed very close to the pinhole [12]. The
disadvantage of collimators is the low effi-
ciency of gamma ray utilization, because
they absorb most of the emitted gamma
rays. Typically about one of 10000 emitted
Fig. 4.6 Schematic set-up of harmonic ultrasound
(US) imaging. Injected microbubbles resonate in
a transmitted US field and emit US waves at high-
er harmonics that can be detected by the scan-
head.
gamma rays is transmitted through a colli-
mator, resulting in reduction of the signal-
to-noise ratio.
The gamma rays that pass through the
collimator are converted into a detectable
signal. Usually, a single sodium iodide
crystal doted with thallium is used. When
a gamma ray strikes this scintillation crys-
tal, it loses energy through photoelectric
interactions. Consequently, light is
emitted, the intensity of which is propor-
tional to the energy of the gamma rays.
Overall, approximately 15% of the ab-
sorbed energy is converted into visible
light. Typically, about 100 photomultipliers
are closely coupled to the scintillation crys-
tal to convert the light signal into an elec-
trical signal. The position of the scintilla-
tion point is determined from the relative
signal outputs of the photomultipliers
using an Anger position network. In addi-
tion, the sum of the signals is proportional
to the energy of the absorbed gamma rays,
which can be used to differentiate between
non-scattered and scattered gamma rays.
This energy window selection process can
improve image quality, because scattered
gamma rays produce a high background
noise. Furthermore, windows at different
energies can also be used to discriminate
between gamma rays of different energies,
which are emitted from different radionu-
clides. Multi-energy windows thus allow
for simultaneous imaging of various tra-
cers.
In SPECT, one or more gamma cameras
are rotated around the patient. The camera
acquires a number of planar images from
different view angles in a stop-and-go
mode. Typically, 32 to 128 views are ac-
quired to reconstruct a 3D image of the
object using a filtered backprojection algo-
rithm. Usually, images with a numerical
resolution of 6464 and 128128 are re-
constructed. In order to improve the im-
age quality in SPECT, the tissue attenua-
tion must be corrected for. The spatially
dependent tissue attenuation can be deter-
mined from a pre-acquired CT-scan, which
provides a 3D attenuation map. Alterna-
tively, an approximation of an attenuation
correction can be applied by performing a
360
o
rotation SPECT scan.
In first instance, the spatial resolution of
SPECT is determined by the gamma cam-
era. The resolution of the scintillator-
Fig. 4.7 Schematic set-up of a SPECT system. One or more gamma cam-
eras slowly rotate around the patient. The gamma cameras detect gamma
ray photons that are emitted by a radiopharmaceutical injected into the
patient.
photomultiplier combination is in the or-
der of 3 mm, and depends on the thick-
ness of the scintillation crystal and the di-
ameter of the photomultiplier. However,
the practical resolution of gamma cameras
is less than 3 mm and is mainly limited
by the collimator. The spatial resolution of
a collimator depends on its geometry
(length and the distance of the lead strips)
and its distance to the gamma ray source.
The overall spatial resolution of clinical
SPECT ranges from less than 1 cm to
about 2 cm, depending on the collimator
type and its distance from the gamma ray
source. In dedicated animal systems, spe-
cial collimators are used, such as pinhole
collimators and their resolution can be be-
low 3 mm.
4.2.4.1 Nuclear Imaging Contrast Agents
Nuclear imaging can be considered as a
pure molecular imaging technique, since
it can directly detect the molecules in
which the radionuclides have been em-
bedded. These compounds are called
radiopharmaceuticals or radiotracers.
Typical radionuclides for SPECT imaging
are
99m
Tc,
111
In,
123
I,
201
Tl, and
67
Ga. Be-
cause of their toxicity, these isotopes are
usually contained in chelates. The chemi-
cal structure of the radiopharmaceuticals
determines their biodistribution and up-
take in the body. Under pathological condi-
tions, radiopharmaceuticals accumulate in
certain regions and/or particular tissues,
and thus can be used for early disease di-
agnosis. The major applications of SPECT
imaging are assessment of cardiac func-
tion, measurement of blood perfusion in
various organs (e.g., heart, brain or lung),
detection of tumors and measurement of
the renal function.
4.2.5
Optical Imaging
Optical imaging encompasses a large set of
imaging technologies that use light from
the ultraviolet to the infrared region to im-
age tissue characteristics. These techniques
rely on different contrast mechanisms such
as transmission, absorption, reflectance,
scattering, luminescence, and fluorescence.
These mechanisms provide information on
structure, physiology, biochemistry and mo-
lecular function. Optical imaging is a com-
mon tool for high-resolution imaging of
surface structures. Microscopes are for in-
stance used to characterize pathologies of
the skin, whereas endoscopes are used to
image structures inside the body. Naturally,
these techniques currently are primarily
limited to surface imaging or experimental
imaging in small animals, because the pen-
etration depth of light is very limited. How-
ever, light within a small spectral window of
the near infrared (NIR) region (600
900 nm) can penetrate more than 10 cm
into tissue due to the relatively low absorp-
tion rates at these wavelengths [13]. The
lower boundary in this window is given by
the high absorption rate of blood (hemoglo-
bin), whereas the absorption above 900 nm
rises due to the presence of water. For this
reason, the NIR part of the spectrum is typ-
ically selected for non-surface optical imag-
ing. The resolution of optical imaging is
limited by the scattering of light: light
photons propagating through tissue diffuse
and follow random paths. Both, absorption
and scattering are intrinsic contrast param-
eters of tissue that can be assessed by opti-
cal imaging. Measuring light absorption
provides functional information on tissue,
because oxyhemoglobin preferentially ab-
sorbs light at lower wavelengths than deox-
yhemoglobin. This difference offers a non-
invasive tool to quantify the vascularization
and/or oxygenation status of tissue. On the
other hand, scattering is associated with the
structural properties of tissue. In general,
two main optical imaging techniques can
be differentiated to assess the absorption
and scattering in tissue: transillumination
and diffuse optical tomography.
In 1929, Cutler developed a technique
called transillumination [14]: light was
shined on one side of a breast, and the ab-
sorption behavior was examined on the
other side. This approach is similar to pro-
jection X-ray imaging, but in transillumi-
nation the spatial resolution is signifi-
cantly reduced due to the scattering of
light photons. Because the absorption of
hemoglobin depends on its oxygenation
state, regions with increased vascularity
could be detected with this technique.
However, it did not provide sufficient spec-
ificity to distinguish between malignant
and benign lesions. During the past 20
years, transillumination has been im-
proved by employing advances in light
sources (e.g., pulsed lasers) and detection
techniques (e.g., charged coupled device
detectors, time-of-flight techniques). Im-
provements in sensitivity and specificity
for detection of breast tumors are cur-
rently under investigation.
During the past decade, diffuse optical to-
mography (DOT) was developed [15]. In this
technique, light is applied from different
angles, and scattered light is detected from
all directions. In contrast to X-ray CT, in op-
tical tomography proper modeling of the
scattering process is essential. Typically, a
numeric solution of the diffusion equation
is used to describe the propagation of light
photons in diffuse media and to predict
the measurements of the experimental set-
up. Due to the strong influence of scatter-
ing and since the reconstruction problem
is ill-posed, the spatial resolution of optical
tomography is rather poor and on the order
of 510 mm. In comparison with transillu-
mination, DOT allows better quantification
of absorption, scattering, or fluorescence
in three dimensions.
4.2.5.1 Optical Imaging Contrast Agents
Similar to other imaging modalities, the in-
trinsic contrast of optical imaging is not suf-
ficient for certain applications, and imaging
agents are necessary. In general, two differ-
ent principles can be differentiated: fluores-
cence and bioluminescence. For fluores-
cence applications, fluorophore-labeled con-
trast agents are administered. The fluoro-
phores are excited using light of an
appropriate wavelength, which is generated
either by a laser or by a white-light source
using filters blocking light above the fluoro-
phore absorption wavelengths. The emitted
photons are detected using a high-sensitiv-
ity CCD camera. Two main set-ups are used
[16]. In reflectance imaging, the light source
and the CCD camera are both on the same
side of the body. This provides relatively
good images from probes that are no more
than a few millimeters deep within the tis-
sue. However, quantification of the signal
is not possible, as it cannot be determined
whether for example a lower signal is due
to a deeper location or a lower concentration
of the probes. This disadvantage is not pres-
ent when capturing 3D images with tomo-
graphic set-ups such as DOT, which also
permit the investigation of deeper-lying tis-
sues, especially in animals. In these ar-
rangements, fibers are used to guide the ex-
citation light to different positions around
the animal and to direct the emitted
photons to the CCD camera. An algorithm
designed especially for the reconstruction
of fluorescence in media such as tissue im-
proves and simplifies fluorescence imaging.
The two types of set-up both allow multi-
plexing through the measurement of sev-
eral different fluorophores and thus differ-
ent targets in the same animal, either si-
multaneously or in fast sequence. In the
NIR range, indocyanine green (ICG) is a
widely used fluorescence agent, because it
is safe and has been approved by the FDA.
ICG is an intravascular agent that extrava-
sates through vessels with a high perme-
ability, such as those in tumors. Recently,
more hydrophilic ICG derivatives were
synthesized which showed a different bio-
distribution and provided a better contrast
between healthy and tumor tissue [17].
In bioluminescence applications, an in-
ternal signal source is used that is,
photons are generated from an enzymatic
reaction and no external light source is
needed. This means, that bioluminescence
imaging has practically no background.
The enzymes used are luciferases, and
their substrates are named luciferins. Luci-
ferase: luciferin pairs occur in many differ-
ent organisms that can glow or emit
flashes of visible light, such as fireflies,
bacteria, and many marine organisms. Lu-
ciferases and luciferins from different or-
ganisms are not necessarily structurally re-
lated. Each luciferase oxidizes its own spe-
cific substrate to form a product in an
electronically excited state, which emits a
photon upon decay. The emission spectra
are relatively broad [18]. The luciferase: lu-
ciferin pair that is used most for optical
imaging is that of the firefly, as a consider-
able part of its emission spectrum is above
600 nm and this light has better tissue-
penetration properties. Bioluminescence
is, in principle, detected with set-ups simi-
lar to those used in fluorescence imaging.
Since bioluminescence imaging requires
the stable expression of exogenous genes
or modified endogenous genes in the or-
ganism under investigation, it is used only
to investigate gene expression in trans-
genic or xenografted animals [19].
4.2.6
Multimodality Techniques
In general, the different imaging modali-
ties provide different information and thus
can be considered as being complementary
rather than competitive. Therefore, the
combination of techniques is of high inter-
est. This can be done by image processing
techniques (i.e., image registration) or by
using integrated systems. In particular, the
combination of structural imaging tech-
niques (e.g., X-ray CT) and functional
imaging (e.g., nuclear imaging) is of high
interest, because it allows the co-registra-
tion of anatomy and molecular informa-
tion. Nowadays, clinical PET-CT and
SPECT-CT scanner combinations are com-
mercially available, whereas other config-
urations such as PET-MR [20] and DOT-
MR [21] are undergoing tests in academic
research. In addition, micro-SPECT/CT
systems are commercially available for ani-
mal imaging. Fig. 4.8 shows a micro-
SPECT/CT (ImTek, Inc., Knoxville, TN,
USA) [22]. This contains two 1020 cm
detector heads for whole-animal imaging
[23], can accommodate different collima-
tors, and provides high-resolution images
of anatomy (20 lm) and function (2 mm).
In addition to the integration of differ-
ent modalities into one system, a common
table can be used to exchange the patient
or animal rapidly and reproducibly. This
approach can additionally be supported by
a position tracking tool, which has already
been used for animal imaging without the
use of anesthetics [24]. Based on the con-
cept of exchangeable tables, new scanner
combinations are expected to become
available in the near future.
4.3
Molecular Imaging
In the previous sections we have seen that
substantial information can be gained
from traditional imaging techniques.
However, these provide mainly anatomical
or functional information. Recently, an
evolution in the development of contrast
agents has resulted in an expansion of the
number of probes that allow to image criti-
cal molecules and their interactions within
the living body non-invasively. These tech-
niques combine a regular contrast-confer-
ring agent with a moiety that interacts spe-
cifically with a target molecule. Such moi-
eties include receptor ligands, enzyme
substrates and recognition elements such
as an antibody or aptamer. They may be
attached to the contrast agent via a linker
molecule, for example to reduce steric hin-
drance for the interaction of the recogni-
tion moiety with its target. In the next sec-
tions, we will discuss different types of
contrast agents, genetic technologies and
suitable modalities for molecular imaging.
4.3.1
Contrast Agents for Molecular Imaging
After administration, traditional contrast
agents will be distributed over and cleared
from the body in patterns that depend on
their physico-chemical properties. Charac-
teristics such as molecular weight, hydro-
phobicity and charge may cause a contrast
agent to accumulate preferentially in cer-
tain cells or tissues. This passive targeting
is used, for example, in the imaging of tu-
mors, as the increased permeability of tu-
mor vasculature allows macromolecules to
extravasate and build up to a larger degree
in tumor tissue [25]. Active targeting of
contrast agents will specifically increase
their density at the target location. How-
ever, passive targeting is still needed to
reach the site of interest. In general, both
passive and active targeting contribute to
the specificity of a contrast agent for mo-
lecular imaging.
Active targeting of contrast agents re-
quires that molecular targets are known.
With the advances in gene expression pro-
filing and proteomic analyses of pathologi-
cal human tissues, an increasing number
of potential disease markers have been
identified. Depending on their disease
specificity and their accessibility such as
expression on the cell surface and in the
vascular lumen these markers may be
used for molecular imaging.
Many potential disease markers and
imaging targets are located inside the cell
and cannot be reached by most types of
contrast agents. Recently, several different
peptides have been used for the intracellu-
Fig. 4.8 Combined animal CT and SPECT system,
which allows high-resolution SPECT imaging of
mice (spatial resolution <2 mm). The image on
the left shows the anatomy of a mouse, while the
image on the right shows a SPECT overlay on the
anatomy. (Data courtesy of ImTek, Inc., Knoxville,
TN, USA).
lar delivery of imaging agents (for reviews,
see Refs. [26, 27]). Generally, these pep-
tides have no specificity for a certain cell
type. Improving the specificity of delivery
systems will reduce the amount of contrast
agent that needs to be administered. Alter-
natively, more radionuclide-labeled small
molecules with improved cellular permea-
tion may be developed. When well-de-
signed contrast agents interact with a tar-
get, the resulting image of probe localiza-
tion and concentration should be directly
related to the localization and concentra-
tion or activity of the target. In general,
three main types of contrast agents for
molecular imaging can be distinguished:
indirect probes, which can be divided into
accumulatable and activatable probes, and
direct-binding probes [28].
Direct-binding or targeted probes bind
their targets stoichiometrically and thus
provide exact information on their localiza-
tion and concentration. In principle, any
protein can be targeted with labeled anti-
bodies or aptamers. The imaging of cell
surface-specific antigens with radiolabeled
antibodies has developed over the past 30
years. The expression of receptors can also
be monitored using labeled ligand analogs.
As peptide receptors can be massively
overexpressed in certain tumors, many
studies have focused on the development
of radiolabeled peptide derivatives for
imaging (for reviews, see Refs. [29, 30]). A
111
In-labeled analog of somatostatin (Oc-
treoscan
:
111
In-DTPA-[D-Phe1]-octreotide;
Mallinckrodt, Inc., St. Louis, MO, USA) is
one of the few FDA-approved peptides for
imaging, and is used for the diagnosis of
neuroendocrine cancer. More recently, an-
nexin-V has been labeled with
99m
Tc for
SPECT imaging of apoptosis [31, 32]. This
protein binds to the phospholipid phos-
phatidylserine, which is present in higher
concentrations in the outer leaflet of the
cell membrane of apoptotic cells. In gener-
al, the imaging of receptors that are patho-
logically overexpressed, such as the HER2/
neu receptor in breast cancer, with direct
binding probes will allow the monitoring
of global tumor burden as well as selection
of patients for receptur tangeted therapy
in a a find, fight, follow strategy (see Part
I, Chapter 5).
For many disease processes, an increase
in enzyme activation not enzyme con-
centration is an important marker. For
example, in gastrointestinal stromal tu-
mors, it is not the number but the kinase
activity of c-Kit receptors that is increased
[33]. This means that direct-binding probes
cannot distinguish between healthy and
diseased tissue. However, enzyme activity
can be visualized using indirect probes,
which do not bind their targets stoichio-
metrically but are changed upon interact-
ing with them. These agents have a high
potential for the imaging of very early ther-
apy effects.
Accumulatable indirect probes become
locally increased in concentration as a con-
sequence of interaction with their target.
The most well-known example is 18F-
fluorodeoxyglucose (FDG), which becomes
trapped within the cell after phosphoryla-
tion by the enzyme hexokinase [34]. Thus,
a higher signal intensity visualized with
PET indicates tissues with increased glu-
cose utilization, and this is widely used to
determine tumor malignancy, to detect
metastases, and to follow therapy effects.
In oncology in general, a major goal is the
development of contrast agents that high-
light the increased activity of critical ki-
nases [35]. Such agents should remain in-
side the cell upon phosphorylation and be
highly specific for the kinase under inves-
tigation. As cellular permeation is a prere-
quisite of accumulatable probes, only
small labels can be incorporated, and con-
sequentially PET and SPECT are used to
visualize this type of contrast agent.
Activatable indirect probes are injected
into the patient in a quenched state. The
conversion of a probe molecule by its tar-
get enzyme increases its signal intensity,
but has no effect on the concentration of
the probes. Activatable probes are rather
new and, until now, mainly fluorescent
probes have been applied for the optical
imaging of protease activity. For example,
Bremer et al. used non-immunogenic co-
polymers to which fluorophores are at-
tached via short peptides, which are sub-
strates for matrix metalloproteinase 2
(MMP-2), a tumor marker. Due to their
close proximity on the polymer, the fluoro-
phores are quenched. Cleavage of the pep-
tides releases the fluorophores, and their
fluorescence signal increases. Using these
probes, it was possible to visualize MMP-2
activity and its inhibition by the potential
drug prinomastat in xenografted mice
after 2 days of treatment [36]. Some princi-
ples of activatable probes for MRI have
also been published [37, 38].
A fundamental difference between di-
rect-binding and indirect probes is the in-
tensity of the overall background signal.
Direct-binding probes are visible through-
out the entire body and require a waiting
period until the probe is enriched at the
target site and the non-bound probe has
largely been cleared from the rest of the
body. In contrast, many indirect probes
can only be imaged after interaction with
their target. In addition, one target en-
zyme can convert many probe molecules.
This built-in amplification causes the back-
ground to be practically non-existent, even
to the point of being a disadvantage, as it
impedes the exact localization of the tar-
get-containing tissue.
To date, a variety of probes for a consid-
erable number of targets have been devel-
oped [39]. However, the numbers of possi-
ble applications, targets and probes are
daunting, and a specific imaging probe
needs to be developed for each molecular
target. Like drugs, contrast agents must be
safe and specific, and they must also pos-
sess the right balance between clearance,
biodegradability and stability to allow an
optimal time window for imaging. In addi-
tion, they should preferably provide signal
amplification to enable visualization of tar-
get molecules in physiological concentra-
tions.
Different probes offer different levels of
molecular information and have different
application ranges. For example, a contrast
agent targeted to a protein that is highly
specific for a certain disease will provide
very detailed information, but only for this
one disease. On the other side of the spec-
trum is an agent such as FDG, which can-
not reveal the biochemical reasons for a
high glucose metabolism, but can be used
in the detection of many different types of
tumor. The sensitivity and specificity of
each probe-target couple must be validated
for its intended application. This can be a
very time-consuming and costly process,
and it may pose problems similar to those
encountered in developing new drugs,
while the criteria for imaging probes are
often more stringent [40]. Therefore, new
developments that will make the most
headway into clinical practice will probably
be those with a broad application range.
This means that they should enable visua-
lization of processes common to many dis-
eases, such as apoptosis, angiogenesis,
and inflammation. Another approach is to
focus on platform technologies that can
easily be adapted to various applications.
Nevertheless, a general approval for such
technologies remains a major obstacle.
4.3.2
Non-invasive Reporter Gene Assays
Important progress in pre-clinical studies is
facilitated by so-called reporter gene assays
(for excellent reviews, see Refs. [1, 40, 41]).
For these assays, a measurable reporter
gene is linked to a gene under investigation
or only brought under control of its promo-
tor. Consequentially, when the gene of inter-
est is turned on, the reporter gene is tran-
scribed and translated into a protein, usual-
ly an enzyme. The presence of this reporter
gene product can then be assessed using a
molecular imaging contrast agent, usually
an indirect probe. Because a reporter gene
can be linked to virtually any gene, a repor-
ter gene assay is a general method to study
non-invasively the expression of genes,
eliminating the need to develop a specific
contrast agent for each target.
For reasons of specificity, reporter genes
can be chosen to be exogenous, stemming
from a completely different type of organ-
ism than the animal under investigation,
and the substrates for the enzymes they
encode are selected to be not, or to a
much lesser extent, convertible by endoge-
nous proteins. Typically used exogenous
enzymes are luciferases from the firefly or
the sea pansy, which can be detected using
bioluminescent imaging (see Section
4.2.5.1). Advantages of the firefly luciferase
system are the very broad dynamic range
and linearity of the reaction and the possi-
bility of real-time measurements because
of the high turnover rate of the enzyme. It
is thus well suited for the monitoring of
changes in gene expression on a relatively
short timescale. Another much-used repor-
ter gene system employs the thymidine
kinase from type 1 herpes simplex virus
(HSV1-TK). HSV1-TK activity can be as-
sessed in a manner similar to hexokinase,
namely by nuclear imaging of radiolabeled
substrates that become intracellularly en-
trapped upon phosphorylation. The sub-
strate label can be a positron emitter, for
PET imaging, or a gamma-ray emitter, for
SPECT. A few strategies were devised for
reporter gene assays that can be visualized
with MRI, such as EgadMe, a substrate for
the enzyme beta-galactosidase. This meth-
od was demonstrated in vivo in Xenopus
laevis embryos after injection of the sub-
strate but cannot yet be used in mice until
a version of EgadMe that can enter the cell
has been developed [37]. Several reporter
gene assays have been designed employing
modified endogenous enzymes that only
have a very narrow expression pattern un-
der natural circumstances. Usually, these
are receptors for which a radiolabeled li-
gand has already been developed. In case
reporter gene assays will be applied in hu-
mans in the future, the lower or absent
immunogenicity of endogenous enzymes
would also be of advantage.
Reporter gene assays can be used for
many different types of studies, such as
the regulation of expression of genes of in-
terest in xenografted and transgenic ani-
mals, as well as the tracking of migrating
cells and even the assessment of gene
therapy and the in vivo measurement of
proteinprotein interactions. Two examples
of such applications will be provided at the
end of this chapter. As the technology re-
quires the stable expression of exogenous
or modified endogenous genes in target
tissues, it will be limited to animal studies
in the near future.
4.3.3
Suitable Modalities for Molecular Imaging
Molecular imaging focuses on the visuali-
zation of molecules and molecular pro-
cesses. Thus, especially in the case of di-
rect-binding probes, the imaging tech-
niques that are used should be sensitive
enough to detect the molecule of interest
in its physiological concentration. Since X-
ray CT offers only millimolar sensitivity
(see Table 4.1), it is not possible to detect
sparse targets with this modality. However,
it provides 3D anatomical information
with a superior spatial resolution and is
used in combination with PET or SPECT
for a better localization of the radionuclide
signal.
MRI also offers a good spatial resolution
and superior soft tissue contrast, but it can
only detect contrast agents in micromolar
concentrations. Although MR imaging of
receptors is possible considering the phys-
ics of MRI, there are biological limitations,
such as delivery of the agent to the site in
high enough quantities, which make this
combination questionable [42]. In order to
sufficiently amplify the signal, it requires
a bulky reporter moiety, such as nanoparti-
cles [43], dendrimers [44], buckeyballs
[45], or polymers carrying a large number
of lanthanide molecules or an iron oxide
nanoparticle. Therefore, MRI contrast
agents seem to be more practical for the
visualization of intravascularly expressed
targets.
Ultrasound imaging requires contrast
agents that are even larger than those
needed for MRI, but then a single micro-
bubble can be visualized, which in princi-
ple can yield a very good sensitivity.
Furthermore, its spatial resolution lies be-
low 1 mm. In addition, ultrasound is a
rather cheap and accessible imaging mo-
dality, and several targeted contrast agents
have been developed.
Due to their high sensitivity, nuclear
imaging techniques are well-suited to vi-
sualize targets present at low concentra-
tions. PET and SPECT, with their picomo-
lar sensitivity, allow for the imaging of
most known targets using ligands that car-
ry only one label each. As the radionuclide
label is relatively small, the probes may
even permeate into cells. However, the
spatial resolution of nuclear techniques
lies in the order of a few to tens of milli-
Table 4.1 Properties of imaging modalities
Modality Sensitivity (concentration
of contrast agent)
Spatial resolution Acquisition time
X-ray-CT
Animal CT
Approx. 10
3
M <500 lm
50 lm
Sub-seconds
MRI
Animal MRI
Approx. 10
6
M 250 lm1 mm
<100 lm
Sub-minutes
Ultrasound
Animal US
Single micro-bubble 100 lm1 mm
(depends on
penetration depth)
40 lm
Sub-seconds
SPECT
Animal SPECT
Approx. 10
10
M 520 mm
13 mm
Minutes
PET
Animal PET
Approx. 10
12
M 410 mm
24 mm
Minutes
Optical (Fluorescence) Approx. 10
10
M 100 lm10 mm
(depends on
penetration depth)
Seconds
meters. In addition, the very low back-
ground in nuclear imaging often requires
the use of an additional modality such as
X-ray CT to provide structural information.
Furthermore, nuclear imaging techniques
(especially PET) require expensive equip-
ment for probe synthesis, such as a nearby
cyclotron to generate the short-lived iso-
topes, and rapid synthesis schemes to in-
corporate these isotopes into the final
imaging probe.
Optical molecular imaging has a sensi-
tivity similar to that of nuclear imaging,
and its spatial resolution can be very high
if surface structures are imaged. Optical
molecular imaging is developing rapidly in
the field of small-animal imaging for con-
trast agent and pharmaceutical R&D. The
technology needed for optical imaging is
relatively cheap and simple, and 3D optical
imaging has been demonstrated [46]. The
considerable attenuation of light by tissue
does not pose a major problem in mice.
However, translation of optical imaging to
the clinic is still limited to fluorescence
imaging applications that investigate acces-
sible body surfaces. This may change in
the future, as a penetration depth of
10 cm has been claimed for NIR fluoro-
phores [13]. Fluorescence imaging also of-
fers the advantage that no radionuclides
are needed, and thus the synthesis of the
probes is easier and cheaper. Biolumines-
cence imaging will probably remain exclu-
sively a small-animal imaging modality for
many years, because it requires the intro-
duction of exogenous or modified endoge-
nous genes into an organism.
4.4
Molecular Imaging for Drug Discovery
and Development
Recent advances in genomics and molecu-
lar biology have raised new hopes for the
prevention, treatment, and even cure of
serious illnesses. However, many of the in-
novations of basic science are not trans-
ferred quickly enough into more effective
and safe drugs. This is because the current
medical product development path is be-
coming increasingly challenging and
costly. Consequently, during the past few
years, the number of new drugs submitted
to the FDA has declined considerably,
while the costs of product development
have increased significantly. For example,
in the year 2000, half of all potential drugs
were discarded during clinical develop-
ment because they lacked in safety and ef-
ficacy. The FDA identified two main rea-
sons for this trend:
1. The profits from a decreasing number
of successful products need to subsidize
a growing number of expensive failures.
2. The path to market even for successful
candidates is long, costly, and inefficient,
due in large part to the current reliance
on cumbersome assessment methods
[47] (see Part VII, Chapter 4).
Fig. 4.9 depicts schematically the drug
development process, which typically takes
about 7 to 10 years to complete. Currently,
an FDA-initiative [47] is trying to improve
the predictability and efficiency along the
critical path from laboratory concept to
commercial product. In this context, mo-
lecular imaging promises many benefits,
since it allows the measurement of drug
absorption, distribution, and target bind-
ing. Even more importantly, molecular
imaging has the potential to provide bio-
markers. A biomarker is . . . a characteris-
tic that is objectively measured and evalu-
ated as an indicator of normal biological
processes, pathogenic processes, or phar-
macological responses to a therapeutic in-
tervention [48]. Several classes of biomar-
kers exist that can be used to determine
drug safety and efficacy in a pre-clinical
and clinical phase and thus can potentially
reduce the costs of drug development [49].
A biomarker is called a surrogate end-
point when it is used to predict therapy
effect without looking at patient well-
being, functioning or survival. Well-known
surrogate endpoints are the reduction of
blood pressure or blood cholesterol level
for the approval of drugs for cardiovascular
diseases and stroke. The non-invasive na-
ture of imaging technologies makes them
very attractive for the use in surrogate end-
points. In oncology, the imaging of tumor
size is a well-accepted marker of therapy
response. With the advent of new medi-
cines that do not necessarily reduce tumor
size, molecular imaging has a high poten-
tial to provide rapid measures of therapy
response. For example, recent progress in
cancer therapies has resulted in new
classes of drugs that can specifically target
and inhibit a molecule, exerting more tar-
geted cytostatic effects rather than overall
cytotoxic effects. Examples are tyrosine
kinase inhibitors such as Gleevec [50], an-
giogenic modulators [51] and inhibitors of
proteases such as MMPs [52]. The usual
measurements of drug plasma concentra-
tions to determine dosage are inappropri-
ate for such a drug, as these data do not
necessarily reflect its concentration at the
site of action, such as a tumor. Even more
important is that the standard clinical trial
endpoints of morbidity and mortality
and newer ones such as tumor size are
insufficient to evaluate cytostatic therapies.
In these cases, molecular imaging can
help to answer the fundamental questions
of drug discovery and development
namely, if and how those drugs work in
vivo. Furthermore, if molecular imaging
has been proven to show the effectiveness
of a specific drug in drug development,
the same technology could be applied in
the clinical setting to monitor early drug
response and to adapt the therapy to the
individual patient (personalized medi-
cine) (see Part I, Chapter 2).
Obtaining molecular information in vivo
can aid many steps in the drug discovery
and development chain. In the following
section, the role of molecular imaging will
be discussed for the different phases of
the drug development process.
4.4.1
Drug Discovery
The major task of the drug discovery
phase is the identification and screening
of new drug targets and lead compounds.
For target identification, microarrays can
be used that allow analysis of gene and
protein expression patterns specific for dis-
eased states. Recently, it was shown that
molecular imaging can provide a cell-
based, high-throughput method to screen
thousands of samples against known tar-
get molecules and cells [53]. When a target
is expressed, its functionality depends
heavily on a number of factors such as ex-
Fig. 4.9 Phases in drug development.
pression level, turnover rate, post-tran-
scriptional modifications and feedback reg-
ulations, which are affected by complex in-
tra- and intercellular processes. Therefore,
target expression and the effect of drugs
should ideally be studied in vivo, and con-
sequently molecular imaging has a high
potential in drug discovery. It can provide
methods to measure drugtarget interac-
tions in vivo and to determine whether a
drug affects the expression and/or func-
tion of a specific target. In particular, mo-
lecular imaging techniques can help to
identify a lead compound, by comparing
the efficacy of different preselected mole-
cules in small animals. For example, it can
be assessed whether drug candidates find
the target, and whether and how they in-
teract with it. One possibility to determine
the delivery and affinity of an unlabeled
drug candidate to its target is to use nucle-
ar imaging to determine how well it inhib-
its the specific binding of a well-character-
ized radiolabeled ligand. This precludes
the need to design and synthesize a func-
tional, radiolabeled analog of each poten-
tial drug. Alternatively, optical imaging
techniques can be used to study target in-
teraction in vivo. In addition, both SPECT
and optical imaging allow multiplexing
through the measurement of several differ-
ent probes and thus different drug candi-
dates in the same animal, either simulta-
neously or in fast sequence [21].
4.4.2
Pre-clinical Testing
Pre-clinical testing is used to study lead
compounds with respect to their biodistri-
bution (pharmacokinetics), dose, toxicity,
and efficacy in small animals. Usually,
time-consuming dissection of animals and
histological analysis of the tissue is per-
formed. In general, pre-clinical imaging al-
lows the collection of data for different
time points and doses for a single animal.
Thus, the number of animals per study
can be reduced, which is more ethical than
sacrificing groups of animals for each data
point. Furthermore, this procedure is more
cost-effective and saves much time, as the
dissection and histology procedures are
slow and subject to sampling errors. More-
over, because each animal serves as its
own control, the statistical relevance of a
study increases as inter-animal variations
become less important. Many technologi-
cal developments have been made during
the past 15 years for small animal imag-
ing. For example, dedicated animal imag-
ing equipment with a (much) higher reso-
lution than for clinical scanners has been
developed, and microsystems are now
being marketed for almost all modalities.
However, the findings based on structural
and functional imaging are often too un-
specific to replace histological techniques.
Therefore, molecular imaging techniques
have a great potential for the study of
drugtarget interactions as well as their
functional consequences in living animals.
In particular, nuclear imaging can be used
to image the biodistribution of drugs. No-
tably, quantitative PET imaging is appro-
priate for this, since drugs can be labeled
with
11
C or
18
F without changing the
chemical properties of the compound to
any degree [54, 55]. Another advantage of
nuclear imaging techniques is that the up-
take rate of the labeled drugs can be quan-
tified more or less directly from the imag-
ing data. As the time point of radionuclide
measurement markedly affects the relative
radiation intensity, maps of tracer kinetics
are used. These are based on pharmacoki-
netic models, which describe the transport
mechanisms of the tracer [56, 57].
In addition, molecular imaging can pro-
vide strategies to visualize the downstream
consequences of administration of a lead
compound on diseased and normal tis-
sues, and to determine whether it has the
desired disease-modifying effect [58], for
example by visualizing cell proliferation,
hypoxia, apoptosis, or (anti-)angiogenesis.
Being able to use such markers of therapy
success may drastically shorten the dura-
tion of pre-clinical studies. Most modali-
ties can be used for these studies, depend-
ing on target localization and other proper-
ties, and examples will be provided in Sec-
tion 4.4.5.
4.4.3
Clinical Trials
Molecular imaging is expected to have a
dramatic impact on clinical trials, using
strategies similar to those in pre-clinical
studies. As in animal systems, pharmaco-
kinetic data in patients can be obtained
using labeled drug analogs. In fact, be-
cause the high sensitivity of nuclear imag-
ing enables the detection of very low doses
of radiolabeled molecules, it allows infor-
mation to be gained on the biodistribution
of a drug far below toxic levels, and even
below therapeutic levels. This microdosing
concept means that these data can be gath-
ered at a very early stage in the drug devel-
opment process [59]. Another possible ap-
plication area of molecular imaging is the
use of imaging biomarkers to stage pa-
tients and select those who will benefit
from a drug (see Part I, Chapter 3). This
will help in defining more stratified and
relevant study groups.
Furthermore, molecular imaging meth-
ods that have been validated as a disease
biomarker in the pre-clinical phase may
also be applied to test drug efficacy in clin-
ical trials. Although clinical trials which
represent the most expensive phase in drug
development would benefit heavily from
molecular imaging techniques, there is cur-
rently a wide gap between their use in the
pre-clinical and clinical phases. There are
several reasons for this. First, only a few
specific contrast agents are available that
have been approved for patient studies.
The development and approval of targeted
contrast agents is the central challenge for
molecular imaging, and this is a costly
and time-consuming process (similar to
that for drugs). Hence, the number of avail-
able targeted agents and probes is unlikely
to increase in the near future. However,
new PETagents may be an exception to this
trend, if their approval process is changed
due to the microdosing concept [59]. An-
other factor that limits the transfer from
pre-clinical to clinical trials is the variation
between pre-clinical and clinical imaging
systems. Protocols that have been opti-
mized on pre-clinical instruments cannot
easily be transferred to clinical scanners.
One way of circumventing these problems
is to support pre-clinical imaging on clinical
scanners, and this is possible by using dedi-
cated animal handling, detectors, and soft-
ware. Nevertheless, a strong interdisciplin-
ary scientific effort is needed to move basic
discoveries tested on animals into the clinic
more efficiently (from bench to bed side).
A number of initiatives are currently ad-
dressing the importance of this transla-
tional research for drug development, in-
cluding the National Institutes of Health
(NIH) roadmap [60] and the European orga-
nization for the treatment of cancer
(EORTC) [61].
4.4.4
Clinical Applications
The main bottleneck of molecular imaging
for clinical applications is the availability
of approved targeted contrast agents. Cur-
rently, the greatest number of targeted
agents is available for SPECT, and addi-
tional targeted agents are undergoing the
approval process. In particular, numerous
antibodies against different molecular tar-
gets and labeled with different gamma-
emitting radionuclides have been devel-
oped over the past two decades. For exam-
ple, arcitumomab-99m-technetium (CEA-
Scan; Immunomedics, NJ, USA) is an
antibody Fab fragment which is labeled
with
99m
Tc and directed against carcinoem-
bryonic antigen (CEA). The targeted con-
trast agent has been approved for the de-
tection of colorectal cancer, and has also
been found to provide imaging of both
palpable and non-palpable breast lesions
that appeared suspicious on screening
mammograms. Currently, CEA-Scan is un-
dergoing Phase III clinical trials for breast
cancer. In addition to diagnostics, the radi-
olabeling of antibodies also has huge po-
tential for cancer therapy. Zevalin
(Bio-
gen Idec, Cambridge, MA, USA, and
Schering AG, Berlin, Germany) is used for
cancer radioimmunotherapy of non-Hodg-
kins lymphoma, and consists of the mono-
clonal antibody ibritumomab and the at-
tached chelator/linker tiuxetan. The chela-
tor can bind indium-111 as well as the
high-energy radioemitter yttrium-90. The
biodistribution of the pharmaceutical can
be visualized with SPECT using indium-
111, before injecting the yttrium-90-labeled
substance for therapy, and finally again
with indium-111 to monitor the progress
of treatment, using SPECT. This approach
can be considered either as a see and
treat approach or as a find, fight, follow
strategy. Gamma emitters are more suited
for this approach than positron emitters
due to their longer half-life.
Clinical applications can strongly benefit
from molecular imaging in early diagno-
sis, disease staging, therapy monitoring,
and follow-up studies. Molecular imaging
is also expected to be of major importance
in the assessment of drug response. Con-
sequently, the co-development of drugs
and imaging-based biomarkers is likely to
have a major impact on treatment
schemes in the future. In particular, the
measurement of drug response would al-
low the customization of treatment re-
gimes and dose optimization for certain
patient groups (personalized medicine)
(see Part I, Chapter 2). In addition, diag-
nosis and treatment could further be inte-
grated into a see and treat approach,
when imaging proves that a targeted probe
does indeed interact with specific disease
molecules, and the same targeted probe is
subsequently loaded with a drug. The
combination of such an approach with
therapy monitoring then results in a find,
fight, follow strategy.
4.4.5
Molecular Imaging Examples
As described above, molecular imaging
can be used in the different phases of
drug development as well as in clinical ap-
plications. One important application of
molecular imaging is the detection of re-
ceptors and their interactions with labeled
ligands in vivo. In clinical oncology, the so-
matostatin/somatostatin receptor system is
used most frequently for contrast-en-
hanced detection and radiotherapy of can-
cer [62]. Somatostatin receptors are overex-
pressed in many tumors, and are present
in especially high concentrations in gastro-
enteropancreatic neuroendocrine tumors
[29, 30]. Radiolabeled synthetic analogs of
somatostatin have been applied success-
fully for routine molecular imaging of
these tumors and their metastases for
more than 15 years, and analog conjugates
for radiation therapy have been under clin-
ical trial for several years [63]. Recently, op-
tical imaging of the somatostatin receptor
has been demonstrated in a mouse model
using a NIR fluorescent probe [64].
Fig. 4.10 illustrates results of this study.
The indotricarbocyanine-conjugate of oc-
treotate (a somatostatin analog) was specif-
ically accumulated at the site of the tumor,
and not taken up by surrounding fibro-
blasts, proving again the high specificity of
this ligandreceptor interaction. Due to
the active targeting of the agent, the tumor
could be visualized with high contrast at
only 1 hour after administration. Fluoro-
phore-conjugated peptides show great pro-
mise for early tumor detection using fluo-
rescence endoscopy, intraoperative imag-
ing, and optical mammography.
Molecular imaging has also been used
to measure general processes such as an-
giogenesis, inflammation, hypoxia. or ar-
teriosclerosis. As many diseases result in
abnormalities of these processes (also
named common denominators), drug-
based therapies aim at their modification.
In other cases such as the apoptosis of
tumor cells a change in this process is a
desired therapeutic effect. For these rea-
sons, quantitative measurement of the sta-
tus of these processes may represent po-
tential biomarkers of drug efficacy.
Angiogenesis occurs both in normal pro-
cesses (e.g., wound repair) and in disease
states, and is a common denominator in
cancer and cardiovascular diseases. Usual-
ly, indirect imaging methods are used to
characterize changes in angiogenesis, such
as measurements of vessel density, perfu-
sion, or vascular permeability. However,
angiogenesis can also be detected at an
early stage with targeted contrast agents,
using a variety of modalities [65]. Angio-
genesis is a complex process involving a
large number of molecules, such as
growth factors [e.g. vascular endothelial
growth factor (VEGF), transforming
growth factor, and fibroblast growth fac-
tor], MMPs, fibronectin, integrins, and en-
dostatin. These molecules are potential tar-
gets for imaging and therapy. For example,
a transgenic mouse model was recently de-
veloped to monitor the effect of potential
(anti-) angiogenic therapeutics in vivo
using bioluminescence imaging [66]. In
these mice, firefly luciferase is brought un-
der control of the promoter for murine
VEGF receptor-2 (VEGFR2), a gene that is
transcriptionally regulated during angio-
genesis. The model was validated using
skin wounding, which induced VEGFR2
Fig. 4.10 Optical detection of a tumor in a mouse
model using a fluorescently labeled somatostatin
analog. The mouse is imaged before administra-
tion of the contrast agent (A) and 6 h after intra-
venous injection (B). Adapted from Ref. [64] with
copyright permission from the Nature Publishing
Group.
and thus luciferase expression. Treatment
of the wounds with a glucocorticoid re-
duced VEGFR2 expression, which could be
followed kinetically using bioluminescence
imaging.
In contrast to many growth factors, ex-
tra-domain B-fibronectin (ED-B-FN) is a
highly disease-specific target, meaning that
it is present during angiogenesis in vessels
of neoplastic tissues, but not in mature
vessels [67]. Recently, NIR fluorescent fi-
bronectin analogs have been applied for
optical imaging of tumor angiogenesis in
mice [68]. In addition, the a
v
b
3
integrin
plays an important role during tumor-in-
duced angiogenesis, and a
125
I-labeled an-
tagonist analog was used in SPECT imag-
ing of angiogenesis [69]. Since the a
v
b
3
epitope is highly expressed on activated
neovascular endothelial cells, targeted ul-
trasound [70] and MR agents [71] can also
be used. However, in order to detect low
concentrations of such a target, the sensi-
tivity of MRI must be improved. One pos-
sibility is to use a paramagnetic liquid per-
fluorocarbon nanoparticle with a
v
b
3
antag-
onist molecules and ca. 10
5
Gd-containing
chelates on its surface. Due to the high
number of Gd-chelates on these particles,
the in vivo detection of a
v
b
3
receptors at
nanomolar to picomolar concentrations in
tumor neovasculature is possible [72]. Re-
sults of this study, in which early tumor-
induced angiogenesis was measured on a
clinical 1.5 Tesla MRI scanner are shown
in Fig. 4.11. Hereto, targeted nanoparticles
were injected systemically into rabbits car-
rying a Vx-2 tumor to fibronectin xeno-
graft. At 2 hours after injection, an in-
creased MR-signal could be seen on T
1
-
weighted images at the periphery of the
tumor. In addition, enhanced MR contrast
was observed within the walls of some ves-
sels close to the tumor.
Angiogenesis is also a critical feature in
the development of atherosclerotic plaques
which, after rupture, can lead to myocar-
dial infarction and stroke. Thus, the same
a
v
b
3
integrin-targeted paramagnetic nano-
particles could be applied to detect athero-
sclerosis in rabbits [73]. Fibrin is another
important target that is associated with
Fig. 4.11 T
1
-weighted MR images of angiogenesis
in a 3 mm Vx-2 tumor (rabbit model) obtained on
a 1.5 Tesla clinical MR scanner. Perfluorocarbon
nanoparticles studded with Gd-chelates and an
a
v
b
3
epitope antagonist are used to confer tar-
geted contrast. Overlays are shown of regions of
enhancement at 2 h post contrast (yellow pixels)
of (A) the tumor and (B) a vein adjacent to the
tumor. (Courtesy of P. Winter, G. Lanza, S. Wick-
line, Washington University, St. Louis, MO, USA)
vulnerable plaque and thrombi (see Part
II, Chapter 1). Therefore, anti-fibrin Fab
fragments were conjugated to the nanopar-
ticles, which were then capable of detect-
ing vulnerable plaques in vivo [74]. Thus,
the paramagnetic nanoparticles may be
considered to be a contrast agent platform.
Other studies with another fibrin-binding
MR-agent (EP-2104R; EPIX Pharmaceuti-
cals, Cambridge, MA, USA) and advances
in coronary MRI techniques demonstrated
the potential for direct imaging of coro-
nary thrombosis [75]. In addition to specif-
ically targeting molecules such as a
v
b
3
in-
tegrin or fibrin, the detection of atheroscle-
rotic plaques is also possible with gado-
fluorine-enhanced MRI [76] or with
passive targeting using ultra-small iron ox-
ide agents [77] that are non specifically
phagocytosed by macrophages present in
the plaques. These irone oxide particles
also allow macrophage tracking associated
with multiple sclerosis [78]. A more specif-
ic diagnosis and monitoring of treatment
response in multiple sclerosis is possible
when targeting the endothelial cell inflam-
matory marker intercellular adhesion mol-
ecule-1 (ICAM-1). Results of a quantitative
ultrasound study in experimental autoim-
mune encephalomyelitis (EAE) rats using
targeted ICAM-1 air-filled microparticles
(MPs) [79] are shown in Fig. 4.12. In this
study, targeted MPs were systemically in-
jected into anesthetized EAE-rats (two
groups, each n=4) and healthy control rats
(n=4). A few minutes after compound ad-
ministration (dose=510
8
MPs kg
1
body
weight) the brains of one EAG group were
examined ex vivo and of the other two
groups in vivo with quantitative ultrasound
Fig. 4.12 Results of a quantitative ultrasound
study in experimental autoimmune encephalomye-
litis (EAE) rats. Application of ICAM-1-targeted
microparticles results in a highly significant
(P=0.001) increase of acoustic counts in EAE-rats
in comparison to healthy rats. (Courtesy of P.
Hauff, Schering AG, Berlin, Germany)
imaging. This imaging technique is based
on the stimulated acoustic emission effect,
which occurs after the destruction of air-
filled MPs during Doppler ultrasound
imaging, thereby allowing the quantifica-
tion of MP-based acoustic counts [80]. The
figure shows that a high number of acous-
tic counts was obtained at similar levels in
both groups of EAE-rats due to the pres-
ence of targeted MPs, demonstrating a
high correlation between ex vivo and in
vivo examination. In contrast, in the
healthy control rats and treated EAE rats
only a few signals were obtained, mainly
reflecting constitutive ICAM-1 expression
at the bloodbrain barrier.
Programmed cell death (apoptosis) plays
a crucial role in the pathogenesis of auto-
immune and neurodegenerative disease,
cerebral and myocardial ischemia, and in
the therapeutic response of cancer. One of
the earliest events in apoptosis is the exter-
nalization of phosphatidylserine, a mem-
brane phospholipid. Annexin-V has a high
affinity for phosphatidylserine, and can
thus be used for molecular imaging of
apoptosis in vivo [81]. Nuclear imaging of
apoptosis has also been demonstrated in
patients with acute myocardial infarction
[82]. In all patients of this study, reperfu-
sion of myocardium was ensured by per-
cutaneous transluminal coronary angio-
plasty. SPECT imaging of annexin-V-la-
beled
99m
Tc was performed several hours
after the acute infarct. Fig. 4.13(A) shows
an increased uptake in specific regions of
the heart, which was validated by a perfu-
sion study performed several weeks later
(Fig. 4.13(B)).
Two fundamental processes which are
also flawed in oncogenesis and may pro-
vide therapeutic targets are signal trans-
duction and the regulation of gene tran-
scription. At certain steps in these pro-
cesses, protein dimerization is required,
which can be triggered by small mole-
cules. One naturally occurring example of
such a molecule is rapamycin, of which
Fig. 4.13 Apoptosis imaging of acute infarction in
a patient using transversal body slices. (A) The ar-
row shows increased Tc-99m-labeled annexin-V
uptake 22 h after reperfusion. (B) Perfusion scinti-
graphy with sestamibi at 68 weeks after dis-
charge shows an irreversible perfusion defect
(arrow), which coincides with the previously mea-
sured area of increased Tc-99m-labeled annexin-V
uptake. In both images, a high signal is obtained
from the liver (L) due to clearance of annexin V.
(Courtesy of L. Hofstra, Maastricht University,
Maastricht, The Netherlands.)
several analogs have completed Phase I
studies and have shown promising poten-
tial as anti-cancer drugs [35]. The efficacy
of these drugs can now be imaged in vivo
in mice using bioluminescence [83]. In
this procedure, a luciferase gene is split
and each half is linked to one of the pro-
teins, the dimerization of which is under
study. When a compound induces such di-
merization, the luciferase halves can also
bind each other and form a functional pro-
tein, and this results in light emission.
The same trick can also be performed
using split reporter genes, the expression
of which can be imaged using either PET
or fluorescence [84].
4.5
Concluding Remarks
Molecular imaging techniques promise
many advantages for the field of drug dis-
covery and development. They allow the
non-invasive procurement of whole-body ki-
netic data on biodistribution and toxicity in
a single animal, thus providing faster,
cheaper, more ethical and probably more
relevant information. Perhaps the greatest
impact from molecular imaging will be
seen in the measuring of surrogate end-
points: monitoring the target or a short-
term therapy effect rather than morbidity
and mortality will significantly reduce the
time required to determine therapeutic suc-
cess. The use of an imaging biomarker in
patients that has already been validated
and used in pre-clinical research allows
the direct linking of results from different
phases, and also facilitates pre-clinical to
clinical transfer. Furthermore, real-time
therapy monitoring and dose-optimization
by testing different doses in one patient will
help to reduce the time needed to establish
therapy efficacy, and also accelerate therapy
development. Finally, by enabling the selec-
tion of those patients who will benefit from
therapy, the size of a clinical trial can be
markedly reduced, while increasing the
chances of the success of a compound.
Overall, molecular imaging promises a dra-
matic reduction in the time to market and
thus cost for new therapies.
In order to profit fully from these new
advances in molecularly targeted drugs,
and to improve attrition rates, the drug
discovery and validation field will need to
adopt techniques with which molecular in-
formation can be gained in vivo. Molecular
imaging offers several such techniques.
We have seen that ingenious compounds
and strategies have been devised to obtain
in vivo molecular information. In addition,
imaging equipment with ever-improving
sensitivity and resolution is available, and
each of the modalities has shown its po-
tential value in molecular imaging applica-
tions for therapeutic R&D. Although mo-
lecular imaging presents huge prospective
benefits for drug discovery and develop-
ment, it is still in its infancy. Increased de-
velopment, validation and the approval of
targeted contrast agents, biomarkers and
clinical imaging endpoints are still re-
quired. As the field is highly interdisciplin-
ary, close collaboration between pharma-
ceutical companies, developers of contrast
agents and manufacturers of imaging
equipment will accelerate progress and
help to ensure that molecular imaging ful-
fils its promise in the development of
Acknowledgement
We are very grateful to Leo, the newborn
son of Joke Orsel, for his cooperation, al-
lowing us to finish this chapter in time.
References
1 R. Weissleder, U. Mahmood, Radiology 2001,
219, 316333.
2 E. L. Ritman, Annu. Rev. Biomed. Eng. 2004, 6,
185208.
3 E. M. Haacke, R. W. Brown, M. R. Thomson, et
al., Magnetic Resonance Imaging, Physical Prin-
ciples and Sequence Design, Wiley Liss, USA,
1999.
4 M. T. Vlaardingerbroek, J. A. den Boer, Mag-
netic Resonance Imaging, Theory and Practice,
3rd edn. Springer-Verlag, Germany, 2002.
5 A. E. Merbach, E. Toth, The Chemistry of Con-
trast Agents in Medical Magnetic Resonance
Imaging, John Wiley and Sons, USA, 2001.
6 R. B. Lauffer, D. J. Parmelee, S. U. Dunham,
Radiology 1998, 207, 529538.
7 K. M. Ward, A. H. Aletras, R. S. Balaban, J.
Magn. Reson. 2000, 143, 7987.
8 J. R. Lindner, Nat. Rev. Drug Discov. 2004, 3,
527532.
9 T. Fritzsch, M. Schartl, J. Siegert, Invest. Radi-
ol. 1988, 23 (Suppl 1), S302S305.
10 G. M. Lanza, K. D. Wallace, M. J. Scott, et al.,
Circulation 1996, 94, 33343340.
11 H. O. Anger, Radioisotope cameras. In: G. J.
Hine (Ed.), Instrumentation in nuclear medi-
cine, Vol. 1, pp. 485552, Academic Press,
USA, 1967.
12 F. J. Beekman, D. P. McElroy, F. Berger, et al.,
Eur. J. Nucl. Med. Mol. Imaging 2002, 29, 933
938.
13 W. F. Cheong, S. A. Prahl, A. J. Welch, IEEE
J. Quantum Electronics 1990, 26, 21662185.
14 M. Cutler, Surg. Gynecol. Obstet. 1929, 48,
721729.
15 S. R. Arridge, Inverse Problems 1999, 15, R41
R93.
16 V. Ntziachristos, C. Bremer, R. Weissleder,
Eur. Radiol. 2003, 13, 195208.
17 K. Licha, B. Riefke, V. Ntziachristos, et al.,
Photochem Photobiol 2000, 72, 392398.
18 J. W. Hastings, Gene 1996, 173, 511.
19 P. R. Contag, I. N. Olomu, D. K. Stevenson et
al., Nat. Med. 1998, 4, 245247.
20 P. K. Marsden, D. Strul, S. F. Keevil, et al.,
Br. J. Radiol. 2002, 75 Spec No, S53S59.
21 V. Ntziachristos, A. G. Yodh, M. Schnall, et al.,
2772.
22 J. Gregor, S. Gleason, S. Kennel, et al., Small-
Animal SPECT Workshop, Tucson, AZ, 2004.
23 A. G. Weisenberger, B. Kross, S. Majewski, et
al., Proc. IEEE Nuclear Science Symposium and
Medical Imaging Conference, Rome, Italy, Octo-
ber 2004.
24 S. S. Gleason, J. S. Goddard, M. J. Paulus, et
al., Proc. IEEE Nuclear Science Symposium and
Medical Imaging Conference, Rome, Italy, Octo-
ber 2004.
25 Y. Matsumura, H. Maeda, Cancer Res. 1986,
46, 63876392.
26 B. L. Franc, S. J. Mandl, Z. Siprashvili et al.,
Mol. Imaging 2003, 2, 313323.
27 M. Zhao, R. Weissleder, Med. Res. Rev. 2004,
24, 112.
28 H. R. Herschman, Science 2003, 302, 605608.
29 R. R. P. Warner, T. M. ODorisio, Semin. Nucl.
Med. 2002, 32, 7983.
30 J. C. Reubi, Endocr. Rev. 2003, 24, 389427.
31 F. G. Blankenberg, P. D. Katsikis, J. F. Tait, et
al. Proc. Natl. Acad. Sci. USA 1998, 95, 6349
6354.
32 L. Hofstra, I. H. Liem, E. A. Dumont, et al.,
Lancet 2000, 356, 209212.
33 R. Capdeville, E. Buchdunger, J. Zimmer-
mann, J. Matter, Nat. Rev. Drug Discov. 2002,
1, 493502.
34 M. Reivich, D. Kuhl, A. Wolf, et al., Acta Neu-
rol. Scand. Suppl. 1977, 64, 190191.
35 Y. Lu, H. Wang, G. B. Mills, Rev. Clin. Exp.
Hematol. 2003, 7, 205228.
36 C. Bremer, C. H. Tung, R. Weissleder, Nat.
Med. 2001, 7, 743748.
37 A. Y. Louie, M. M. Hber, E. T. Ahrens, et al.,
Nat. Biotechnol. 2000, 18, 321325.
38 J. M. Perez, L. Josephson, T. OLoughlin, et al.,
Nat. Biotechnol. 2002, 20, 816820.
39 M. Rudin, R. Weissleder, Nat. Rev. Drug Dis-
cov. 2003, 2, 123131.
40 R. G. Blasberg, J. G. Tjuvajev, J. Clin. Invest.
2003, 111, 16201629.
41 T. F. Massoud, S. S. Gambhir, Genes Dev. 2003,
17, 545589.
42 A. D. Nunn, K. E. Linder, M. F. Tweedle, Q. J.
Nucl. Med. 1997, 41, 155162.
43 G. M. Lanza, C. H. Lorenz, S. E. Fischer, et al.,
Acad. Radiol. 1998, 5 (Suppl 1), S173S176.
44 C. Wiener, M. W. Brechbiel, H. Brothers, et
al., Magn. Reson. Med. 1994, 31, 18.
45 W. H. Noon, Y. Kong, J. Ma, Proc. Natl. Acad.
Sci. USA 2002, 99, 64666470.
46 V. Ntziachristos, C. H. Tung, C. Bremer, et al.,
Nat. Med. 2002, 8, 757760.
47 Food and Drug Administration. Innovation or
Stagnation? Challenge and Opportunity on
the critical path to new medical products,
White Paper 2004.
48 R. Frank, R. Hargreaves, Nat. Rev. Drug Discov.
2003, 2, 566580.
49 J. F. Niblack. In: G. J. Downing (Ed.), Biomar-
kers and Surrogate Endpoints: Clinical Research
and Applications. Elsevier Science, 2000.
50 L. K. Shawver, D. Slamon, A. Ullrich, Cancer
Cell 2002, 1, 117123.
51 M. Cristofanillli, C. Charnsangajev, G. N. Hor-
tobagyi, Nature Rev. Drug Discov. 2002, 1, 415
426.
52 L. M. Coussens, B. Fingleton, L. M. Matrisian,
Science 2002, 295, 23872392.
53 D. Hogemann, V. Ntziachristos, L. Josephson,
et al., Bioconjug. Chem. 2002, 13, 116121.
54 A. J. Fischman, N. M. Alpert, R. H. Rubin,
Clin. Pharmacokinet. 2002, 41, 581602.
55 C. Halldin, B. Gulyas, L. Farde, Ernst Schering
Res. Found. Workshop 2004, 48, 95109.
56 K. C. Schmidt, F. E. Turkheimer, Q. J. Nucl.
Med. 2002, 46, 7085.
57 C. S. Patlak, R. G. Blasberg, J. Cereb. Blood
Flow Metab. 1985, 5, 584590.
58 F. Brady, S. K. Luthra, G. D. Brown, Curr.
Pharm. Des. 2001, 7, 18631892.
59 G. Lappin, R. C. Garner, Nat. Rev. Drug Discov.
2003, 2, 233240.
60 See http://nihroadmap.nih.gov/overview.asp.
61 A. Eggermont, H. Newell, Eur. J. Cancer 2001,
37, 1965.
62 C. Grotzinger, B. Wiedenmann, Ann. N. Y.
Acad. Sci. 2004, 1014, 258264.
63 A. Heppeler, S. Froidevaux, A. N. Eberle, Curr.
Med. Chem. 2000, 7, 971994.
64 A. Becker, C. Hessenius, K. Licha, et al., Nat.
Biotechnol. 2001, 19, 327331.
65 M. Schirner, A. Menrad, A. Stephens, Ann.
N. Y. Acad. Sci. 2004, 1014, 6775.
66 N. Zhang, Z. Fang, P. R. Contag, et al., Blood
2004, 103, 617626.
67 D. Neri, B. Carnemolla, A. Nissim, et al., Nat.
Biotechnol. 1997, 15, 12711275.
68 K. Licha, C. Perltz, P. Hauff, et al., Proc. Soc.
Molec. Imaging 2004, 264.
69 R. Haubner, H. J. Wester, U. Reuning, et al.,
J. Nucl. Med. 1999, 40, 10611071.
70 H. Leong-Poi, J. Christiansen, A. L. Klibanov,
et al., Circulation 2003, 107, 455460.
71 D. A. Sipkins, D. A. Cheresh, M. R. Kazemi, et
al., Nat. Med. 1998, 4, 623626.
72 P. M. Winter, S. D. Caruthers, A. Kassner, et
al., Cancer Res. 2003, 63. 58385843.
73 P. M. Winter, A. M. Morawski, S. D. Caruthers,
et al., Circulation 2003, 108, 22702274.
74 S. Flacke, S. Fischer, M. J. Scott, Circulation
2001, 104, 12801285.
75 R. M. Botnar, A. Buecker, A. J. Wiethoff, et al.,
Circulation 2004, 110, 14631466.
76 J. Barkhausen, W. Ebert, C. Heyer, et al., Cir-
culation 2003, 108, 605609.
77 S. G. Ruehm, C. Corot, P. Vogt, et al., Circula-
tion 2001, 103, 415422.
78 M. Rausch, P. Hiestand, D. Baumann, et al.,
Magn. Reson. Med. 2003, 50, 309314.
79 P. Hauff, M. Reinhardt, M. Murer, et al.,
Proc. Soc. Molec. Imaging 2004, 121.
80 M. Reinhardt, P. Hauff, A. Briel, et al., Invest.
Radiol. 2004, 39.
81 F. Blankenberg, C. Mari, H. W. Strauss, Q. J.
Nucl. Med. 2003, 47, 337348.
82 L. Hofstra, I. H. Liem, E. A. Dumont, et al.,
Lancet 2000, 356, 209212.
83 R. Paulmurugan, T. F. Massoud, J. Huang, et
al., Cancer Res. 2004, 64, 21132119
84 G. D. Luker, V. Sharma, C. M. Pica, et al., Proc.
Natl. Acad. Sci. USA 2002, 99, 69616966.
References 1241
Abstract
Molecular imaging with positron emission
tomography (PET) is now evolving as a
unique non-invasive method for studying
tumor and normal tissue biochemistry,
physiology, and pharmacology. In oncol-
ogy, a range of drugs can be radiolabeled
for pharmacokinetic studies including mi-
crodosing of human subjects prior to
Phase I trials. Gene delivery can be as-
sessed by incorporating a reporter gene
that is detectable by PET within the vector
of interest. This chapter also reviews pro-
gress made in the PET imaging field in
the design of pharmacodynamic markers
including assays of cell surface receptor
status, angiogenesis, apoptosis, prolifera-
tion, glucose metabolism, and hypoxia.
Such probes are potentially useful in pa-
tient management and for drug develop-
ment. PET is particularly attractive for the
development of cancer-targeted therapies
where assessment of plasma drug levels or
assays of the target protein in peripheral
blood cells are less specific than direct as-
sessment of the tumor or tumor material.
The recent introduction of dedicated
small-animal scanners has helped bridge
in vitro science with in vivo clinical studies
for the efficient development of modern
biopharmaceuticals.
Abbreviations
ATSM diacetyl-bis(N4-methylthiosemi-
carbazone)
BrdU bromodeoxy uridine
CDK cyclin-dependent kinase
CT computed tomography
D2R dopamine D2 receptor
DACA N-[2-(dimethylamino)ethyl]acri-
dine-4-carboxamide
DLT dose-limiting toxicity
DOTA 1,4,7,10-tetraazacyclododecane-
N,N',N'',N'''-tetraacetic acid
DPD dihydropyrimidine dehydrogenase
DTPA diethylenetriamine pentaacetic
acid
ECM extracellular matrix
EGFR epidermal growth factor receptor
ELISA enzyme-linked immunosorbent
assay
ER estrogen receptor
FDG fluorodeoxyglucose
FES 16a-[
18
F]fluoro-17b-estradiol
FESP fluoroethylspiperone
FHBG 9-(4-[
18
F]fluoro-3-hydroxy methyl
butyl) guanine
FIAU 2'-fluoro-2'deoxy-1-b-D-arabinofur-
anosyl-5-[
124
I]iodo-uracil
FLT fluorothymidine
1243
5
Design and Development of Probes for In vivo Molecular
and Functional Imaging of Cancer and Cancer Therapies
by Positron Emission Tomography (PET)
Eric O. Aboagye
ISBN: 3-527-31184-X
FMAU 2-fluoro-5-methyldeoxyuracil-b-
D-arabinofuranoside
FMISO fluoromisonidazole
FU fluorouracil
HDAC histone deacetylase
HIDAC high density avalanche chamber
HSP heat shock protein
HSV1 herpes simplex virus type I
HYNIC hydrazino nicotinamide
IUdR iododeoxyuridine
MAPK mitogen-activated protein kinase
MMP matrix metalloproteinase
MTD maximum tolerated dose
NIS nal symporter
PET-CT positron emission tomography
computed tomography
PK1 N-(2-hydroxypropyl)methacryl-
amide copolymer of doxorubicine
PS phosphatidylserine
PTAC Pharmakokinetic & Pharmacody-
namic Technology Advisory Com-
mittee
RECIST response evaluation criteria in
solid tumours
RGD arginine-glycine-aspartic acid
SIB N-succinimidyl-3-iodobenzoate
SPECT single photon emission computed
tomography
TK1 thymidine kinase Typ 1
TMP thymidine monophosphate
TNF tumour necrosis factor
TRAIL tumor necrosis factor-related
apoptosis-inducing ligand
5.1
What is Positron Emission Tomography?
Positron emission tomography (PET)
imaging is one type of nuclear imaging
that utilizes short-lived, positron-emitting
isotopes to allow visualization and quanti-
fication of biological processes or drug ki-
netics [1]. Positron-emitting isotopes such
as
15
O (t
1/2
=2 min),
11
C (t
1/2
=20 min),
18
F
(t
1/2
=109 min), and
124
I (t
1/2
=4.2 days)
can be incorporated into many compounds
of biological interest to produce radiotra-
cers such as [
18
F]fluorodeoxyglucose (FDG)
for PET studies. Some examples of posi-
tron-emitting isotopes are listed in Table
5.1.
A PET study begins with the injection or
inhalation of a radiotracer, followed by
scanning. When the radiotracer decays, it
emits a positron that travels a short dis-
tance and annihilates with an electron
(Fig. 5.1). Annihilation produces two 511-
keV photons, which propagate at approxi-
mately 180
o
apart. These photons can be
detected within a short time window called
the coincidence time window (~10 ns).
Many such events are summed to provide
the distribution of the radiotracer. Radio-
tracer transport, washout and retention
can be monitored by PET and, if cali-
brated, the PET images can yield quantita-
tive estimates of the amount of radiotracer
5 Design and Development of Probes for In vivo Molecular and Functional Imaging of Cancer 1244
Table 5.1 Positron-emitting radioisotopes
Radioisotope Decay
mode
[% b
+
]
Half-life
[min]
Decay product
Carbon-11 99.8 20.38 Boron-11
Nitrogen-13 100 9.96 Carbon-13
Oxygen-15 99.9 2.03 Nitrogen-15
Fluorine-18 96.9 109.8 Oxygen-18
Iron-52 57 496 Manganese-52
Cobalt-55 77 1050 Iron-55
Copper-62 98 9.7 Nickel-62
Bromine-75 75.5 98 Selenium-75
Bromine-76 57 966 Selenium-76
Technetium-
94m
72 52 Molybdenum-
94
Iodine-124 25 6048 Tellurium-124
Gallium-68 90 68.3 Germanium-
68
in specific parts of the body. Additional in-
formation of blood radioactivity levels or
radioactivity in reference regions enables
the calculation of exchange rate constants
of transport, receptor binding, retention,
and metabolic rate. Mathematical kinetic
modeling is often employed to enhance
data interpretation within a framework of
important kinetic behaviors, and to obtain
quantitative parameters of relevance and
universal comprehension. As indicated be-
low, PET is being used as a tool to under-
stand focal pathophysiology, as well as to
monitor response to drug treatment. Fig-
ure 5.1 also shows examples of clinical
and pre-clinical PET scanners. Commercial
Fig. 5.1 (a) Positron emission. A positron emitter
such as carbon-11 decays to boron-11 with the
emission of a positron and a neutrino. The posi-
tron travels a finite distance, colliding with elec-
trons to form a positronium. Annihilation of the
positronium leads to the emission of two gamma
photons at approximately 1808 apart. These
photons can be detected by coincident detectors
within a scanner. (b) Example of a clinical PET
scanner, Siemens HR+ (published with permission
from Dr. Terry Spinks, Hammersmith Imanet, Lon-
don, UK). (c) Example of a dedicated small-animal
PET scanner, the Oxford Positron Systems nano-
PET scanner, quad-HIDAC.
clinical scanners have been in use for
several years, and more recently the
emergence of commercial PET-computed
tomography (PET-CT) scanners implies
that true fusion of anatomical and func-
tional data can be achieved [1]. On the
other hand, the recent development of
commercial dedicated small-animal scan-
ners has helped to bridge in vitro science
with in vivo clinical studies [2].
Advancement in radiochemistry meth-
ods has made it possible to develop several
probes to study biology. The development
of PET radiopharmaceuticals for use in
cancer and cancer therapeutics studies is
presented in the next section.
5.2
Radiochemistry Considerations
Radiopharmaceuticals with very high spe-
cific activity (515 Ci mmol
1
) are used in
PET studies so as to retain the pharmaceu-
tical, biological, and biochemical properties
of the compound being studied. This sec-
tion reviews the common radiosynthetic
methods, the selection of a suitable radio-
tracer, and limitations in the preparation
of radiopharmaceuticals.
5.2.1
Radiosynthetic Methods
The labeling agents for PET radiopharma-
ceuticals are prepared from very simple
chemicals such as [
11
C]CO, [
11
C]CO
2
,
[
18
F]F
and [
18
F]F
2
(Fig. 5.2). Of the var-
ious possibilities available, [
11
C] methyla-
tion, nucleophilic substitution with [
18
F]F
and electrophilic fluorination with [

18
F]-la-
beled intermediates form the majority of
labeling approaches used for biological
compounds [3]. The labeling agent is re-
acted with a precursor that is often pre-
pared in a form such that a single-step (or
a few steps) radiolabeling reaction is per-
formed, followed by purification and for-
mulation for intravenous injection. In
some cases for example, the preparation
of 2-[
11
C]thymidine [4] single-step chem-
istry is not possible and a multi-step reac-
tion is performed.
5.2.2
Selection of a Suitable Radiotracer
What is a suitable radiotracer for PET
studies? The selection of any radiotracer
for a biological application depends on the
objectives of the study. Generic considera-
tions include: the type of biological target;
the affinity and specificity of the radiotra-
cer for the target; the physico-chemical
properties, pharmacokinetics and metabo-
lism of the radiotracer; and the ability to
synthesize the compound.
Small molecule-, peptide-, and antibody-
based probes all tend to be suitable for tar-
geting cell surface receptors. However,
only small molecule-based probes and
membrane-permeable peptides are suitable
for targeting intracellular and nuclear en-
zymes or receptors. Consideration should
be given to the existence of membrane car-
rier proteins such as nucleoside transport-
ers or expression of efflux pumps such as
the p-glycoprotein pump for poly-aromatic
compounds in tissues. Additional consid-
erations include barriers to radiotracer de-
livery (e.g., the bloodbrain barrier), the
concentration and location of the target,
and the heterogeneity of the target within
the tissue. The goal of most radiotracer
studies is to achieve a high target-to-back-
ground ratio. This means that the affinity
of the radiotracer for the receptor or en-
zyme system should be high, usually low
Fig. 5.2 Synthesis of labeling agents. (a) Small building blocks derived from [
11
C]CO
2
and [
11
C]CH
4
.
(b) Incorporation of [
18
F]F
into small building blocks.

nanomolar binding affinity (often denoted
as K
d
, the concentration of tracer that is
associated with half-maximal receptor
binding, half-B
max
); the concentration of
target should be high (>10 pmol mg
1
tis-
sue); and non-specific accumulation in the
tissue of interest should be low. For recep-
tor-based studies, a general guiding princi-
ple is to select radioligands with high
B
max
/K
d
of 4. Binding of a radioligand to
a receptor, phosphorylation of a radiotracer
by an intracellular enzyme or other mech-
anism that increases the charge of the
compound, and covalent interactions with
cellular macromolecules provide the specif-
ic signal. It is then desirable to have rapid
clearance of the non-specific signal. Clear-
ance is largely determined by the physico-
chemistry and metabolic fate of the com-
pound. If a radiotracer has rapid clearance,
then it is acceptable to radiolabel it with
any isotope. For compounds with long
half-lives (e.g., antibodies), long-lived iso-
topes such as
124
I are often used.
Non-specific accumulation of radiotra-
cers, on the other hand, is more difficult
to define for any particular radiotracer, and
depends on specific activity (the presence
of unlabeled, cold material increases
non-specificity), specificity for the target
versus other related targets, non-specific
transport, and physico-chemical properties
such as stereochemistry, inductive effect,
lipophilicity, and ionization potential. The
latter two characteristics can be altered
fairly easily in the design of radiotracers to
reduce non-specific binding. Whereas high
lipophilicity (log p >1.5 and <3) is desir-
able in neuroscience applications to enable
the radiotracer to traverse the bloodbrain
barrier, this is not necessarily the case in
some oncology and cardiology applica-
tions, and it could lead to non-specific ac-
cumulation. Lipophilicity should be con-
sidered together with the ionization rate
constant, pK
a
, of the compound, which
could be designed to enhance or reduce
membrane transport.
5.2.3
General Limitations in the Preparation
of Radiopharmaceuticals for PET Studies
Due to the short physical half-lives of most
positron-emitting radionuclides, radiotra-
cer synthesis should be fast enough to al-
low the target drug to be isolated, purified,
and formulated as a sterile, pyrogen-free,
isotonic solution within two to three half-
lives of the radionuclide. Consequently,
large amounts of radioactivity must be
handled, and the limitation of potential ex-
posure to personnel is an extremely impor-
tant consideration. For most compounds,
these issues require the design and syn-
thesis of precursors that can be radiola-
beled in a single step, and in turn this has
fostered the development of new methods
for rapid remote-controlled and robotics-
based chemistry. The presence of radio-
chemistry facilities and on-site expertise in
radiochemistry are also required. The abil-
ity to radiolabel a compound depends on
the availability of suitable functional
groups. For example, compounds with N-,
S- or O-methyl (or -ethyl) groups, and ly-
sine and tyrosine functionalities, can be
radiolabeled fairly easily. In some cases,
the multi-step chemistry required to pro-
duce a radiotracer precludes radiolabeling
and purification of molecules rapidly en-
ough to avoid substantial decay of radioac-
tivity. Constraints in the availability of suit-
able labeling reagents including precursors
can also limit the ability to synthesize a
radiotracer. The position of labeling should
be robust towards metabolic degradation,
which further limits the number of com-
pounds that can be radiolabeled.
5.3
Pharmacological Objectives in Oncology
Imaging Studies
In the past, most oncology therapeutics
were designed to target DNA directly or in-
directly via modulation of the enzymes in-
volved in its synthesis, coiling/uncoiling,
and segregation. In the past decade, our
knowledge of the processes that initiate
and drive tumorigenesis has increased,
and so has the development of drugs that
target these processes (Table 5.2). Such tu-
mor-targeted drugs, however, bring along
with them difficulties in drug develop-
ment. In most cases, the use of systemic
toxicity end-points (maximum tolerated
dose, MTD; dose-limiting toxicity, DLT) are
inappropriate. Furthermore, the drugs are
largely cytostatic and do not cause overt
changes in tumor size within several
weeks to months of treatment. Methods
which can show that the drug has reached
its target, has modulated its biological tar-
get (or cognate biochemical events), and
has caused a biological response are now
required to provide proof of concept. The
incorporation of such design into early-
phase trials makes the trial an extension
of the pre-clinical testing of the com-
pound. A summary of generic pharmaco-
logical objectives in the development of
novel targeted therapies is listed in Table
5.3. The end-points could be assessed di-
rectly on biopsy material from the patient.
In the absence of biopsy material, investi-
gators have used peripheral blood mono-
nuclear cells and buccal scrapings to evalu-
ate the mechanism of action. It should be
noted that the target may not be expressed
to the same extent in these tissues as in
Table 5.2 Paradigm shift in oncology from DNA-
targeting to targeting factors that initiate and drive
tumorigenesis
Classical targets of systemic chemotherapy
DNA: alkylators, platinators
- Incorporation of nucleotides into DNA:
antimetabolites
Topoisomerases: anthracyclines, camptothe-
cins
Microtubules: vinca alkaloids, taxanes
New targets for systemic therapy
Growth factors: erbB2, EGFR, VEGF,
VEGFR
Angiogenesis/vascular: VEGF, FGF, PDGF,
integrins
Signal transduction: protein kinases, cKit,
ras, MAPK, TRAIL
Cell cycle regulation: cyclins, CDKs, p53
Invasion and metastasis: matrix metallopro-
teinases
Multiple targets: Hsp90, HDAC, Cox-2
EGFR, epidermal growth factor receptor; VEGF, vas-
cular endothelial growth factor; VEGFR, VEGF re-
ceptor; MAPK, mitogen-activated protein kinase;
TRAIL, TNF-related apoptosis-inducing ligand;
CDKs, cyclin-dependent kinases; Hsp90, heat-shock
protein 90; HDAC, histone deacetylase; Cox-2, cyclo-
oxygenase type-2.
Table 5.3 Pharmacological objectives for testing tar-
geted therapies
Objectives Measurable end-points
Select patients express-
ing specific target
e.g., erbB2 status,
hypoxia
Ensure adequate/opti-
mal exposure in experi-
mental animal model
or patient
Pharmacokinetics in
plasma or tissues
Demonstrate target
modulation
e.g., kinase inhibition,
demethylation
Demonstrate induction
of desired biological
effect
e.g., inhibition of prolif-
eration, invasion, angio-
genesis, or induction of
differentiation, apopto-
sis
Resulting clinical
response
e.g., disease-free surviv-
al, cytostasis, tumor
shrinkage
tumors, and where the target is expressed,
target modulation could occur at a lower
dose compared to that in tumors [5]. Mo-
lecular imaging methods (notably PET) are
potential alternatives for clinical decision
making, including dose and schedule se-
lection for Phase II, and the selection of
patient subpopulations enriched for re-
sponse, as well as early detection of re-
sponse. The imaging methods are attrac-
tive because:
they are surgically non-invasive;
they can be repeated in the same patient
several times before and after treatment;
they allow heterogeneity within tumors
or between a primary tumor and meta-
stases to be determined; and
they provide quantitative information.
For a more detailed description of how to
use imaging and non-imaging methods to
develop novel targeted therapies, the read-
er should refer to the Cancer Research UK
Pharmacokinetic & Pharmacodynamic
Technologies Advisory Committee (PTAC)
guidance document at the following URL
(http://science.cancerresearchuk.org/reps/
pdfs/PTACguidelines.pdf). The use of PET
in studying the pharmacokinetics and
pharmacodynamics of oncology drugs will
be reviewed in the next section.
5.4
The Use of Radiolabeled Drugs
to Image Tumor and Normal Tissue
Pharmacokinetics
A number of drugs have been radiolabeled
to enable their pharmacokinetics in tumor
and normal tissues to be studied. In oncol-
ogy, the large majority of these drugs are
small aliphatic, aromatic or heterocyclic
molecules identical (isotopic) or similar
(non-isotopic; e.g., the replacement of a
hydrogen or hydroxyl with fluorine) in
structure to the compound of interest. For
biopharmaceuticals such as antibodies and
polymers, non-isotopic labeling methods
are often employed. Examples of these
radiopharmaceuticals will be given to illus-
trate the application of the technology in
oncology. In most cases, the key objective
of the study is to investigate whether the
pharmacokinetic properties predicted in si-
lico or seen in vitro are similar to those
seen in animals and humans; this should
enable the confirmation of drug-design ob-
jectives, and also enable dose versus tis-
sue-exposure relationships to be assessed.
5.4.1
Pharmacokinetics of Small Molecules
In addition to toxicity, Phase I trials often
involve assessment of the plasma pharma-
cokinetics of the drug to obtain parameters
such as the elimination half-life (Ke), sys-
temic clearance (sCL), area under the plas-
ma drug concentration versus time curve
(pAUC, the systemic exposure), and sys-
temic volume of distribution (sV
d
, the ratio
of the amount of drug to the plasma drug
concentration at steady state). The overall
(systemic) extravascular distribution of
drugs can be predicted from the sV
d
; for
example, if sV
d
is higher than the plasma
volume, then the drug shows extravascular
distribution. sV
d
does not indicate, how-
ever, the tissues to which the drug is dis-
tributing. The delivery, washout and reten-
tion of drugs in tumor and specific normal
tissues are easily assessable by PET imag-
ing of the radiolabeled compound in ani-
mal models and in patients. To this end, a
number of anti-cancer drugs including 5-
fluorouracil [6], cisplatin [7], temozolomide
[8], and N-[2-(dimethylamino)ethyl]acri-
dine-4-carboxamide (DACA) [9] have been
radiolabeled for pharmacokinetic studies
5.4 The Use of Radiolabeled Drugs to Image Tumor and Normal Tissue Pharmacokinetics 1251
Fig. 5.3 Radiolabeling of anti-cancer drugs for pharmacokinetic studies. DACA, N-[2-(dimethyl-
amino)]acridine-4-carboxamide.
11
CH
3
in humans. More recently, the taxanes
(docetaxel and paclitaxel), the oral fluoro-
pyrimidine prodrug, capecitabine, and the
epidermal growth factor receptor tyrosine
kinase inhibitor, Iressa, have been radiola-
beled, although clinical studies have not
been performed with these materials [10
13]. Some examples of radiolabeling reac-
tions are illustrated in Fig. 5.3.
The tissue distribution of a selection of
radiolabeled drugs in cancer patients is
shown in Fig. 5.4. From such studies, the
time-course of drug distribution can be de-
termined for any region of interest within
the field of view. PET pharmacokinetic
studies can be performed with the radio-
tracer alone (high specific activity), or
mixed with a suitable dose of the unla-
beled drug (low specific activity). Perform-
ing both studies in the same patient can
give an indication of saturation effects.
The studies can be carried out before
Phase I trials (pre-Phase I, also called mi-
cro-dosing) [9, 14], or as part of a Phase I
trial [15]. An example of a pre-Phase I
radiotracer study is that of [
11
C]DACA, a
DNA intercalating and topoisomerase I/II
inhibitor. This study was conducted at
one-thousandth of the Phase I starting
dose at one year before the Phase I trial,
and demonstrated that the drug distrib-
uted well to tumors [9]. Radioactivity local-
ized to tissues in the order: vertebra <
brain < tumor < kidney < lung < myocar-
dium < spleen < liver. The low peak con-
centrations and overall exposure in brain
and vertebra contrasted with the high dis-
tribution to brain in rodents, and sug-
gested that neurotoxicity and myelotoxicity
were less likely to be dose-limiting. On the
other hand, the high localization of radio-
activity in the myocardium (saturable at
Phase I doses of the drug) [9, 15] war-
ranted close monitoring of cardiovascular
effects.
High distribution of [methyl-
11
C]temozo-
lomide, a DNA methylating agent, to tu-
mors has also been demonstrated by PET.
In patients with brain tumors, delivery
and exposure of the radiotracer was found
Fig. 5.4 Distribution of radiolabeled drugs in hu-
mans monitored by PET. (a) A brain image of
[
11
C]temozolomide, showing high localization of
radioactivity in the tumor (glioma). (b) A thoracic
image of [
11
C]DACA scan, showing localization in
the myocardium and tumor (mesothelioma).
(c) An abdominal image of 5-[
18
F]fluorouracil,
showing a hyperintense normal liver and low up-
take in the liver metastases.
to be higher in the tumor compared to nor-
mal brain tissue [16], and was suggestive of
some selectivity for tumors. In another
study, Saleem et al. showed that radio-
labeled temozolomide (radiolabeled in the
N-methyl or carbonyl position) undergoes
ring opening selectively in tissues in com-
parison to plasma, but did not show selec-
tivity for tumor versus brain [8]. The reason
for the higher exposure of [methyl-
11
C]-te-
mozolomide seems to be a higher delivery
(K
1
) of the radiotracer to the tumor [16].
Studies with [methyl-
11
C]temozolomide
also illustrate one of the limitations of
PET studies of drug pharmacokinetics. Of-
ten, the decay constant of the radioisotope
(k) is much higher than the plasma Ke,
which means that only the initial delivery-
phase of the drug is accurately measured.
Among anticancer drugs, 5-[
18
F]fluor-
ouracil ([
18
F]FU) has been the most widely
studied using PET, it having been shown
that:
retention of the drug in tumors is low
[17];
the drug is catabolized by the liver to a
transiently trapped catabolite, [
18
F]fluoro-
b-alanine; a large proportion of the ad-
ministered radioactivity is, therefore, lo-
calized in the liver [6, 17];
eniluracil, a dihydropyrimidine dehydro-
genase (DPD) inhibitor, can inhibit he-
patic clearance of the drug and increase
drug exposure in tumors; these effects
occur in concert with an increase in
plasma uracil levels (a systemic measure
of DPD inhibition) [6, 17, 18];
alpha interferon increases tumor expo-
sure of 5-FU [19];
folinic acid, a modulator of the thymidy-
late synthase activity, has no effect on
[
18
F]FU pharmacokinetic [19]; and
intra-arterial administration gives rise to
a higher tumor exposure of the radiotra-
cer than the intravenous route [20].
These studies have demonstrated that PET
imaging of small molecules radiolabeled
with positron emitters can add value to
classical studies of new drugs. The PET
radiolabeling studies provide very impor-
tant tissue pharmacokinetics information.
It is worth noting, however, that the
achievable pharmacokinetic parameters
may be limited for three reasons:
The use of radiotracers (high specific ac-
tivity) at doses much below the point
where metabolism and protein binding
become important (K
m
) can lead to al-
tered systemic clearance than when the
studies are carried out at relevant Phase
I doses.
The half-life of
11
C is often shorter than
the half-life of the drug of interest; thus,
delivery and partitioning of the radiotra-
cer are more accurately estimated than
retention parameters. The use of
18
F al-
leviates the problem of a short half-life
in some cases.
The metabolism of the radiotracer can
complicate interpretation of the data.
Despite these difficulties, PET studies of
radiolabeled drugs have provided unique
pharmacokinetic information in patients.
5.4.2
Pharmacokinetics of Biopharmaceuticals
Although most radiolabeling and PET
pharmacokinetic studies performed to date
have utilized small-molecule drugs, there
is potential to apply the technologies to
macromolecular agents such as biophar-
maceuticals. Applications to peptides and
antibodies will be reviewed under pharma-
codynamic studies, as the majority of such
applications exist there. For polymer-based
drugs, including N-(2-hydroxypropyl)meth-
acrylamide copolymer doxorubicin (PK1; a
doxorubicin-polymer conjugate), either the
5.4 The Use of Radiolabeled Drugs to Image Tumor and Normal Tissue Pharmacokinetics 1253
active agent or the polymer can be radio-
labeled, depending on the clinical question
asked. Long-lived isotopes such as
124
I are
preferred for radiolabeling polymers as the
pharmacokinetics can be measured over
days, thereby allowing long-term preferen-
tial tumor localization (if any) to be as-
sessed. In this case, the radiochemistry
will require the synthesis of appropriate
precursors containing for instance tyrosine
or lysine residues to enable direct iodina-
tion (with Iodogen) or indirect iodination
(with iodinated aromatic labeling agents).
Previously, PK1 was radiolabeled with
131
I
for gamma-camera imaging [21].
5.4.3
Pharmacokinetics of Gene Delivery Systems
An increasing number of PET studies are
dedicated to detection of the efficiency of
gene delivery. To date, utility with viral,
liposome and stem cell delivery systems
have been investigated. PET and other
forms of molecular imaging modalities
allow the location, magnitude and time-
course of gene expression to be deter-
mined using a marker or reporter gene to-
gether with a reporter substrate a sub-
strate for the protein product of the report-
er gene (Fig. 5.5). Gene expression is,
thus, monitored indirectly by the detection
of the activity of the reporter gene.
Although reporter gene studies alone can
be performed for characterization of vector
systems, they are probably more useful
when studied together with the therapeutic
gene. This can be done by: 1) using a con-
struct (with an internal ribosomal entry
site) that allows both reporter and thera-
peutic gene to be transcribed as one
mRNA and translated into two proteins; 2)
using a bi-directional vector in which tran-
scription of both reporter and therapeutic
genes are initiated by a single event such
as doxycycline or tetracycline; and 3) using
two separate vectors (for a review, see Ref.
[22]). There are several reporter genere-
porter substrate pairs for use in gene ex-
pression studies, including: 1) herpes sim-
plex virus type 1-thymidine kinase gene
versus 2'-fluoro-2'-deoxy-1-b-D-arabinofura-
nosyl-5-[
124
I]iodo-uracil (FIAU) or 9-(4-[
18
F]-
fluoro-3-hydroxy methyl butyl) guanine
(FHBG); 2) dopamine type-2 receptor gene
versus [
18
F]fluoroethylspiperone (FESP);
and 3) sodium iodide symporter versus
[
124
I]NaI [22]. The latter system is particular-
ly useful for performing PET studies in in-
stitutions that do not have a cyclotron, since
the half-life of iodine is 4.2 days and so can
be transported to distant sites [23, 24]. More
recently, PET has been used to assess the
transcriptional activity of tissue/tumor-spe-
cific promoters, including the human telo-
merase RNA and protein promoters as a
prelude to using such promoters for suicide
gene therapy [25], and hypoxia response ele-
ments for reporting hypoxia-inducible fac-
tor 1 signal transduction [26]. As expected,
viral therapy constitutes the bulk of these
PET studies, although initial investigations
involving liposome delivery vectors and
stem cells are being pursued.
5.5
Pharmacodynamic Studies
As mentioned earlier, the early clinical de-
velopment of tumor-targeted anticancer
agents requires the use of non-traditional
methods such as molecular drug end-
points (Western blots, activity assays) in tu-
mor or surrogate tissue, and functional
imaging studies. A review of the literature
[27] showed that such methods were not
routinely incorporated into the study de-
sign for early trials of anticancer agents,
and rarely formed the primary basis for
dose selection. This may be due to the lack
of sufficiently validated functional imaging
end-points for clinical studies. In this sec-
tion, an overview is provided of new imag-
ing assays that have the potential to be
used in drug trials.
5.5.1
Assessment of Receptors, Antigens
and Extracellular Matrix Proteins
Receptor imaging in cancer is less well de-
veloped than it is in neuroscience, in part
due to the relative importance of post-re-
ceptor signal transduction mechanisms.
Fig. 5.5 PET imaging of gene expression. (a) Illus-
tration of the three major methods for assessing
gene expression in vivo: dopamine D
2
receptor
(D2R), HSV1-thymidine kinase (HSV1-tk), and NaI
symporter (NIS) methods. (b) [
124
I]-PET images
obtained from untreated and adenovirus-treated
mice. The high localization in the liver after virus
treatment is due to tropism of adenovirus for spe-
cific receptors expressed on liver cells. Uptake in
the thyroid and stomach is due to physiological
expression of NIS; the bladder signal is due to uri-
nary excretion.
Receptor imaging is, however, particularly
important for the early detection of cancer,
staging and selecting patients for specific
receptor-based therapies such as tamoxi-
fen, herceptin and Iressa, which target the
estrogen receptor (ER), c-erbB2 and epider-
mal growth factor receptor (EGFR), respec-
tively. There is therefore active research in
this area at present, albeit at an early
stage, and these investigations will be dis-
cussed from the viewpoint of the type of
radiotracer utilized, whether small mole-
cules, peptides, or antibodies.
5.5.1.1 Small-molecule Radiotracers
A number of small-molecule radiotracers
are currently under development for the
imaging of EGFR. Overexpression of this
transmembrane receptor has been found
in breast, lung, ovarian, colon and prostate
cancer, and is associated with poor prog-
nosis in these cancers [28, 29]. In most
specialist centers, the assay of receptor sta-
tus by immunohistochemistry forms part
of the patient work-up, before decisions
are made on therapeutic management.
Radiotracer methods are being developed
to provide quantitative and non-invasive al-
ternatives to biopsy-based methods. Most
of the radiotracers are quinazoline deriva-
tives, and similar to drug candidates for
the receptors [29, 30]. Initial attempts to
develop reversible
18
F- and
11
C-inhibitors
of the intracellular ATP binding site of
EGFR were unsuccessful, presumably be-
cause of competition for the radiotracers
with the high levels of intracellular ATP in
cells. A group at the Hebrew University,
Israel has now developed irreversible
EGFR inhibitors, including [
11
C]ML03
(Fig. 5.6), with slower tumor washout ki-
netics because of covalent association
through Michael addition, between the
double bond of the acryl-amide group at
the 6-position of the quinazoline ring and
Cys-773 at the receptors tyrosine kinase
domain [31]. It is hoped that these studies
will yield clinical candidates for PET imag-
ing of EGFR receptor status. Not all recep-
tors are extracellular in nature, however.
Nuclear receptors are also attractive candi-
dates for imaging, although for these the
radiotracers must traverse the plasma and
nuclear membrane before binding to the
receptor. The ER is an important nuclear
receptor in breast cancer, and a target for
drugs such as tamoxifen that bind to the
receptor or aromatase inhibitors such as
anastrozole that reduce circulating levels
of the natural ligand, estrogen. The most
widely used receptor ligand for ER is 16a-
[
18
F]fluoro-17-estradiol (FES). A good cor-
relation between FES-PET imaging and
biopsy-based measures of ER status has
been reported [32]. Furthermore, ER occu-
pancy as measured by FES-PET showed a
greater decrease after tamoxifen in pa-
tients who responded to hormonal therapy
than non-responders, thereby demonstrat-
ing promise of this radiotracer for predict-
ing patients likely to respond to estrogen-
based therapy [33, 34].
A number of PET radiotracers are being
developed for imaging extracellular matrix
(ECM) proteins. Of particular interest are
the matrix metalloproteinases (MMPs);
these are usually classified into five groups
according to their domain structure col-
lagenases, gelatinases, stromelysins, mem-
brane-type MMPs, and others and are a
family of zinc-dependent enzymes that de-
grade specific components of the ECM.
The active forms of MMPs (after proteolyt-
ic cleavage of the inactive zymogen) are
highly expressed during tumor growth and
invasion compared to their expression in
normal tissues. A suitable PET marker will
thus find utility in disease prognosis, as
well as for monitoring the pharmacody-
namics of MMP inhibitors [35]. Furumoto
and co-workers have developed
18
F probes
((2R)-2-[4-(6-[
18
F]fluorohex-1-ynyl)-benzene-
sulfonylamino]-3-methylbutyric acid and
its ester) for PET imaging of MMP type 2
(MMP-2) [36]. The key design features
(Fig. 5.6) include a carboxylic acid group,
which binds to active-site zinc ion, and a
linear side chain that interacts with the
channel-like S'-1 subsite of MMP [35]. The
Fig. 5.6 Examples of novel small molecule and
peptide radiotracers for in vivo pharmacodynamic
imaging. (a) [
11
C]ML03, an irreversible inhibitor of
EGFR tyrosine kinase and potential PET marker
for the receptor (b) [
18
F]SAV03, an MMP-2 inhibi-
tor developed for imaging the levels of this en-
zyme. (c) [
18
F]Galacto-RGD, an a
V
b
3
integrin re-
ceptor ligand developed for imaging the levels of
this receptor, which is highly expressed on angio-
genic vessels.
currently available radiotracers have a
number of limitations, and new-generation
probes are being developed to overcome
this. Significant first-pass metabolism has
been reported for the free carboxylic acid-
containing radiotracer; ester analogues of
these have reduced first-pass metabolism
and enhanced tumor localization [36]. De-
fluorination of the fluoroalkyl moiety leads
to high non-specific uptake in the bone
a limitation that the new candidate radio-
tracers will attempt to overcome. Some of
the new candidates are based on pan-
MMP-inhibitors. For example, Cheesman
et al. have reported promising biodistribu-
tion data with an [
111
In]-DTPA-linked mac-
rocyclic succinic acid hydroxamate conju-
gate [37].
5.5.1.2 Antibody Radiotracers
Antibodies are promising alternatives to
small molecules. Because of their size and
charge, they do not cross the plasma
membrane, and are therefore useful for
imaging the extracellular domains of re-
ceptors and antigens. Intact monoclonal
antibodies (IgG, ~150 kDa) have extremely
high selectivity for the target, but often
present with poor pharmacokinetics and
high immunogenicity. Interest in the use
of antibody fragments, derived from pro-
tein digestion or by recombinant methods
for radiolabeling, stems from the potential
gain in enhanced clearance of these mole-
cules (see Part V, Chapters 1, 2 and 6).
The c-erbB2 receptor has also been
traced using PET. Like EGFR, this receptor
is also accessed routinely in breast cancer
patients by immunohistochemistry for ex-
pression, or by fluorescence in situ hybridi-
zation for amplification, to inform patient
management. Smith-Jones et al. [38] radi-
olabeled a F(ab)'
2
fragment of the anti-c-
erbB2 antibody herceptin with
68
Ga via a
1,4,7,10-tetra-azacyclododecane-N,N',N'',N'''-
tetra-acetic acid (DOTA) linker (see Part I,
Chapter 5). The resultant radiotracer was
used to image c-erbB2-expressing mouse
BT-474 tumors; a reduction in tumor
localization of the radiotracer was seen
at 24 hours after treatment with geldana-
mycin, an hsp90 inhibitor that degrades
c-erbB2 [38]. Other investigators have
explored single-chain Fv fragments (scFv:
~2530 kDa) due to their potentially rapid
clearance from the circulation and re-
sultant high target-to-background ratio.
Sundaresan et al. [39] radiolabeled anti-
carcinoembryonic antigen (anti-CEA) scFv-
C
H
3 minibody and diabody with
124
I via
the standard Iodogen method (which tar-
gets the radiolabel predominantly to tyro-
sine residues on the antibody). This
allowed CEA-positive tumor xenografts
(LS174T) which were <3 mm in diameter
to be imaged in vivo by PET [39]. With the
rapid development of recombinant strate-
gies for preparing pharmacokinetically su-
perior scFvs coupled with new ways of
radiolabeling various antibodies, it is anti-
cipated that most extracellular receptors
could be imaged by PET in the very near fu-
ture. These probes could be used to provide
proof of the mechanism of action in vivo.
5.5.1.3 Peptide Radiotracers
Of the three classes of radiotracers used to
image receptors, this is probably the most
attractive. There are several reasons for this:
Peptides are less-immunogenic mole-
cules that show rapid distribution to the
target tissue and which, unlike antibod-
ies, have a rapid systemic clearance.
There is rapid development in the field
of peptide synthesis via solid phase and
from phage libraries.
There are new and attractive methods
for metabolic stabilization and coupling
of peptides to chelators or prosthetic
groups for indirect labeling, as well as
methods for direct radiolabeling with
radiohalogens to tyrosine residues.
Despite these advantages, there are a num-
ber of important considerations. First, the
metabolic degradation of peptide radiotra-
cers by endogenous peptidases or proteases
can make their clearance too rapid, leading
to low sensitivity for detection of target ac-
tivity. The incorporation of linkers such as
polyethylene glycol (PEG) can reduce the
clearance (see Part VI, Chapter 2). Second,
peptide labeling may lead to a loss of affin-
ity and, like antibodies, affinity of the la-
beled peptide needs to be assessed (e.g.,
with an enzyme-linked immunosorbent as-
say; ELISA). Today, there are many diverse
peptide radiopharmaceuticals under devel-
opment, including radiotracers for the in-
tegrin a
v
b
3
receptor expressed in neovascu-
larization or angiogenesis, and radiophar-
maceuticals for the somatostatin receptor
expressed in neuroendocrine tumors/carci-
noids, small cell lung cancer and lympho-
mas.
The integrin a
v
b
3
receptor has attracted
much interest in the field of PET, as it is
known to be involved in angiogenesis.
Molecules containing the tri-peptide se-
quence arginine-glycine-aspartate (RGD)
have been shown to bind selectivity to the
a
v
b
3
receptor [40]. High in vitro affinity
and in vivo tumor selectivity was demon-
strated with the first candidate radiotracers
{[
124
I]cyclo(-Arg-Gly-Asp-D-Phe-Tyr-) and
[
124
I]cyclo(-Arg-Gly-Asp-D-Tyr-Val-)}. At the
time, the fluorination of peptides was a
rather challenging task, and was per-
formed in multiple steps. The break-
through in this field came with the devel-
opment of [
18
F]synthons and peptide pre-
cursors that permitted chemoselective,
single-pot fluorination and simple purifica-
tion of the radiolabeled product. For exam-
ple, an
18
F-labeling methodology based on
the chemoselective oxide formation be-
tween an unprotected amino-functiona-
lized RGD peptide and an
18
F-labeled alde-
hyde or ketone has been reported [40, 41];
Glaser and co-workers [42] also reported
the use of [
18
F]fluorothiols and methane-
sulfonyl precursors for labeling a model
peptide that can potentially be used for
radiolabeling RGD peptides. In addition to
the use of cyclic and bicyclic peptides to
improve systemic stability and receptor af-
finity and improved peptide labeling, a
number of investigators have also attached
PEG and sugar amino acids to improve
the pharmacokinetics of candidate pep-
tides [41, 4345], the initial studies with
which in tumor models have provided en-
couraging results.
5.5.2
PET Monitoring of End-points of Tumor
Growth and Response to Treatment
Deranged proliferation and apoptosis are
fundamental to the development of cancer
[46]. Tumor growth is a term that de-
scribes the balance between cell division
and cell death. This is arguably the most
important biological end-point for deter-
mining the sensitivity of cancer cells to
drugs, and is often estimated in patients
using radiological methods (RECIST crite-
ria [47]), by an analysis of biopsy material
for visible mitosis, cell and cycle markers,
and by autoradiography (for a review, see
Ref. [48]). The limitations of the RECIST
criteria and biopsy-based methods have led
to an interest in alternative imaging meth-
ods for monitoring tumor growth. Thus, a
review of PET methods for glucose meta-
bolism, proliferation and apoptosis in
monitoring the end-points of tumor
growth and response to treatment will be
presented in the next section.
5.5.2.1 PET Imaging of Glucose Metabolism
Fluorodeoxyglucose is the most commonly
used radiotracer for imaging glucose utili-
zation. The uptake of FDG into tissues is
determined mainly by high glucose trans-
porter and hexokinase activities and low
glucose-6-phosphatase activity [4951]. The
model for FDG uptake is illustrated in
Fig. 5.7. After transport into cells, FDG
(like glucose) is phosphorylated by hexo-
kinase to FDG-phosphate which, unlike
glucose-6-phosphate, is not a substrate for
Fig. 5.7 Imaging of glucose metabolism with
FDG. (a) A model of FDG uptake. FDG is deliv-
ered and retained in tissues according to this
model: k
1
, k
2
, k
3
, and k
4
are rate constants for
clearance into the cell from blood, clearance out
of the cell, phosphorylation and dephosphoryla-
tion, respectively. (b) High-resolution FDG-PET
images obtained using the nano-PET technology
of a tumor-bearing mouse showing localization of
the radiotracer in myocardium, brain, tumor, and
bladder. Four-dimensional data (three-dimensional
spatial + time) were obtained; 0.3-mm orthogonal
slices are shown. (c) Brain image of a glioma pa-
tient, showing localization of the radiotracer in
the tumor and forebrain.
further glycolytic metabolism. Trapping of
the radiotracer is effected by the high
charge of the phosphorylated product and
its low rate of dephosphorylation in tumor
tissues [52]. Tissues that have high glu-
cose-6-phosphatase activity (e.g., liver) [51]
show a low retention of FDG. Non-tumor
tissues such as brain and myocardium
(Fig. 5.7), as well as inflammatory cells
[53] also take up FDG; consequently, care
should be taken in interpreting FDG data.
From the above account it is clear that
FDG does not directly measure tumor pro-
liferation. Rather, it is used as a surrogate
for cell viability, which is related indirectly
to cell number and proliferation. Other
than applications in the diagnosis and
staging of disease, FDG has found use in
monitoring the response to drug therapy.
Compared to cross-sectional imaging,
FDG-PET is highly reproducible [54], dif-
ferentiates between viable and fibrotic tis-
sue, and changes in radiotracer uptake oc-
cur early after treatment [55, 56]. For cyto-
toxic therapies, the decrease in FDG up-
take after therapy generally mimics the
reduction in tumor cell viability. A number
of response studies showing early reduc-
tion in FDG uptake after treatment have
been published, including the treatment of
brain tumors with temozolomide [57], of
breast cancer patients with combined che-
mo- and hormonal therapy [58], and of
non-Hodgkins lymphoma with combina-
tion chemotherapy [59]. More recently,
FDG-PET has been used to image the re-
sponse of gastrointestinal stromal tumors
to molecular therapeutics such as the c-Kit
and bcr-abl inhibitor, Imatinib mesylate
(Gleevec; Glivec; STI-571). In this setting,
FDG uptake decreased dramatically as
early as 24 hours after drug treatment
[60, 61], and the early changes correlated
with a clinical response at 13 months
[60]. Due to a number of limitations with
FDG however, other tracers that more
closely monitor cellular proliferation or
cell death are being evaluated for imaging
drug response.
5.5.2.2 PET Imaging of Cell Proliferation
A number of radiolabeled pyrimidine nu-
cleosides have been synthesized for imag-
ing proliferation. 2-[
11
C]Thymidine and its
analogues, 3'-deoxy-3'-[
18
F]fluorothymidine
(FLT), 2-fluoro-5-[
11
C]methyldeoxyuracil-b-
D-arabinofuranoside (FMAU), [
76
Br]bromo-
deoxy uridine (BrdU), and [
124
I]iododeoxy-
uridine (IUdR) (Fig. 5.8) have been evalu-
ated pre-clinically and clinically for assess-
ing proliferation (for a review, see Ref.
[48]). These radiotracers are transported
into the cell by diffusion, as well as by nu-
cleoside transporters, and are phosphory-
lated by a thymidine kinase type-1 (TK1)
enzyme to form the corresponding mono-
phosphate (salvage pathway for DNA syn-
thesis; Fig. 5.9). TK1 activity and the sub-
sequent rates of phosphorylation of the
monophosphate to di- and tri-phosphate
vary for the different nucleosides [62]. Of
importance to the clinical use of these
radiotracers is the fact that the analogues
with a halogen substitution in the sugar-
ring have suitable in vivo stability [48].
Although more clinical studies have been
performed with [
11
C]thymidine, FLT is
probably the most promising of all these
radiotracers for monitoring proliferation
due to its superior in vivo stability. Several
in vitro and in vivo (rodent and clinical)
studies have now been carried out demon-
strating that FLT-PET uptake measures the
TK-1-dependent phosphorylation [63, 64],
which correlates with cellular proliferation
as measured by S-phase fraction or Ki-67
immunohistochemistry [6567]. This tech-
nology is currently being used pre-clini-
cally in rodents to evaluate the efficacy of
anti-cancer agents [6870]. Prospective
clinical studies of drug response are on-
going, but results have not been published
for FLT. Such data do exist for [
11
C]thymi-
dine, however [71].
5.5.2.3 PET Imaging of Apoptosis
Programmed cell death, or apoptosis, is the
major mechanism of cell death following
anti-cancer drug therapy (see also video ani-
mation on supplement CD-ROM). Given its
importance, several investigators are study-
ing ways of imaging apoptosis in vivo.
Among various possibilities, the use of radi-
olabeled Annexin V is currently the most
promising. Annexin V is a 36-kDa cal-
cium-dependent protein that binds to phos-
phatidylserine (PS) with very high affinity.
When radiolabeled with positron emitters
Fig. 5.8 Radiolabeled pyrimidine nucleosides for imaging cellular proliferation.
or gamma emitters, the resultant radiophar-
maceuticals label apoptotic cells because, in
apoptotic cells (unlike viable intact cells) PS
translocates from the inner to the outer
plasma membrane, making it accessible
for binding. Annexin V has been indirectly
radiolabeled with
124
I to provide [
124
I]-SIB-
Annexin V for PET imaging studies [72,
73]. The most widely used radiopharmaceu-
tical, however, is [
99m
Tc]-HYNIC-annexin V
for single photon emission computed tomo-
graphy (SPECT) studies [7476]. Clinical
studies with this reagent have been per-
formed [77, 78]. Given that apoptosis is a dy-
namic process, the optimal timing of these
studies is crucial, and is likely to vary for dif-
ferent tumor types and for different thera-
peutic modalities. Indeed, this issue repre-
sents the biggest challenge for imaging
apoptosis in tumors.
5.5.3
Assessment of Tumor Hypoxia
The final section of this chapter deals with
an important predictor of therapeutic re-
sponse in oncology hypoxia. Imaging of
this physiological property has relevance
in selecting patients likely to benefit from
therapies aimed at modulating hypoxia.
The disorganized and inadequate vascula-
ture and blood flow in tumors often leads
to impaired oxygen delivery, and this leads
to the creation of areas of low oxygen ten-
sion (hypoxic regions). Hypoxia is a power-
ful trigger of gene expression, and thus
clonal selection of more aggressive pheno-
type, for example, diminished apoptotic
potential [79]. Hypoxia also predicts for lo-
cal tumor control by external beam radio-
therapy, and predicts general treatment
Fig. 5.9 Mode of action of radiolabeled thymidine
analogues used for imaging cellular proliferation.
The radiotracers are transported and phosphory-
lated by thymidine kinase (salvage pathway) to
the corresponding monophosphate that is subse-
quently phosphorylated to the triphosphate and
incorporated into DNA. For FLT, most of the label
remains as the monophosphate being a poor sub-
strate for deoxythymidine monophosphate kinase
(dTPM). The de-novo pathway for DNA synthesis
is also illustrated.
outcome, including metastatic potential
and survival following radio-/chemother-
apy and surgery in a number of human
cancers [8082]. Several radiotracers have
been developed for imaging hypoxia by
PET; these include 2-nitroimidazole-based
probes such as [
18
F]fluoromisonidazole
(FMISO) and [
18
F]fluoroetanidazole [83,
84], and copper bis-thiosemi-carbazones
probes such as [
60
Cu]ATSM [85]. In all
cases, proof that the technique measures
hypoxia involves comparisons with direct
measurements of pO
2
by oxygen elec-
trodes, assessment of uptake following
modulation of hypoxia, and radiation sen-
sitivity. Studies with FMISO have demon-
strated the existence of hypoxic regions in
a number of tumor types [83, 86]. With
the development of more modern radiotra-
cers, it is hoped that the measurement of
hypoxia can be performed efficiently to en-
able patient selection for hypoxia-targeted
therapy.
5.6
Conclusions
PET is a potentially powerful technology
for monitoring drug pharmacokinetics,
and for the prediction or assessment of re-
sponse to anti-cancer treatment. The tech-
nique allows quantitative measurements to
be made in animals and humans, in a
non-invasive manner. It is hoped that de-
velopments during the next few years will
provide the platform for incorporating this
technology into the discovery and develop-
ment of biopharmaceuticals and, in partic-
ular, for early drug trials in patients.
References
1 Valk, P. E., Bailey, D. L., Townsend, D. W., and
Maisey, M. N. Positron emission tomography:
basic science and clinical practice. Springer,
London 2003.
2 Aboagye, E. O. Positron emission tomography
of small animals in anticancer drug develop-
ment. Mol. Imaging Biol., 7: 5358, 2005.
3 Elsinger, P. H. Radiopharmaceutical chemistry
for positron emission tomography. Methods,
27: 208217, 2002.
4 Steel, C. J., Brady, F., Luthra, S. K., Brown, G.,
Khan, I., Poole, K. G., Sergis, A., Jones, T.,
and Price, P. M. An automated radiosynthesis
of 2-[
11
C]thymidine using anhydrous [
11
C]urea
derived from [
11
C]phosgene. Appl. Radiat.
Isot., 51: 377388, 1999.
5 Collins, J. M. Innovations in phase 1 trial de-
sign: where do we go next? Clin. Cancer Res.,
6: 38013802, 2000.
6 Saleem, A., Yap, J., Osman, S., Brady, F., Sut-
tle, B., Lucas, S. V., Jones, T., Price, P. M., and
Aboagye, E. O. Modulation of fluorouracil tis-
sue pharmacokinetics by eniluracil: in vivo
imaging of drug action. Lancet, 355: 2125
2131, 2000.
7 Ginos, J. Z., Cooper, A. J., Dhawan, V., Lai,
J. C., Strother, S. C., Alcock, N., and Rotten-
berg, D. A. [
13
N]cisplatin PET to assess phar-
macokinetics of intra-arterial versus intra-
venous chemotherapy for malignant brain tu-
mors. J. Nucl. Med., 28: 18441852, 1987.
8 Saleem, A., Brown, G. D., Brady, F., Aboagye,
E. O., Osman, S., Luthra, S. K., Ranicar,
A. S. O., Brock, C. S., Stevens, M. F. G., New-
lands, E., Jones, T., and Price, P. Metabolic ac-
tivation of temozolomide measured in vivo
using positron emission tomography. Cancer
Res., 63: 24092415, 2003.
9 Saleem, A., Harte, R. J., Matthews, J. C., Osman,
S., Brown, G. D., Bleehen, N., Connors, T.,
Jones, T., Price, P. M., and Aboagye, E. O.
Pharmacokinetic evaluation of N-[2-(Dimethyl-
amino)ethyl]acridine-4-carboxamide (DACA;
XR5000) in patients by positron emission
tomography. J. Clin. Oncol., 19: 14211429,
2001.
10 Fei, X., Wang, J. Q., Miller, K. D., Sledge,
G. W., Hutchins, G. D., and Zheng, Q. H. Syn-
thesis of [(
18
)F]Xeloda as a novel potential
PET radiotracer for imaging enzymes in can-
cers. Nucl. Med. Biol., 31: 10331041, 2004.
11 Kurdziel, K. A., Kiesewetter, D. O., Carson,
R. E., Eckelman, W. C., and Herscovitch, P.
Biodistribution, radiation dose estimates, and
in vivo Pgp modulation studies of
18
F-pacli-
taxel in nonhuman primates. J. Nucl. Med.,
44: 13301339, 2003.
12 De Jesus, O. T., Murali, D., Flores, L. G., Con-
verse, A. K., Dick, D. W., Oakes, T. R., Roberts,
A. D., and Nickles, R. J. Synthesis of [F-18]-
ZD1839 as a PET imaging agent for epider-
mal growth factor receptors. J. Label. Compds.
Radiopharm., 46: S1, 2003.
13 van Tilburg, E. W., Franssen, E. J. F., van der
Hoeven, J. J. M., Elshove, D., Lammertsma,
A. A., and Windhorst, A. D. Radiosynthesis of
[
11
C]docetaxel. J. Label. Compds. Radiopharm.,
46: S3, 2003.
14 Aboagye, E. O., Luthra, S. K., Brady, F., Poole,
K., Anderson, H., Jones, T., Boobis, A., Bur-
tles, S. S., and Price, P. Cancer Research UK
procedures in manufacture and toxicology of
radiotracers intended for pre-phase 1 positron
emission tomography studies in cancer pa-
tients. Br. J. Cancer, 86: 10521056, 2002.
15 Propper, D. J., de Bono, J., Saleem, A., Ellard,
S., Flanagan, E., Paul, J., Ganesan, T. S., Tal-
bot, D. C., Aboagye, E. O., Price, P., Harris,
A. L., and Twelves, C. Use of positron emis-
sion tomography in pharmacokinetic studies
to investigate therapeutic advantage in a phase
1 study of 120-hour intravenous infusion
XR5000. J. Clin. Oncol., 21: 203210, 2003.
16 Meikle, S. R., Matthews, J. C., Brock, C. S.,
Wells, P., Harte, R. J., Cunningham, V. J.,
Jones, T., and Price, P. Pharmacokinetic as-
sessment of novel anti-cancer drugs using
spectral analysis and positron emission
tomography: a feasibility study. Cancer
Chemother. Pharmacol., 42: 183193, 1998.
17 Aboagye, E. O., Saleem, A., Cunningham, V.,
Osman, S., and Price, P. Extraction of
5-fluorouracil by tumor and liver: a non-inva-
sive positron emission tomography study of
patients with gastrointestinal cancer. Cancer
Res., 61: 49374941, 2001.
18 Bading, J. R., Yoo, P. B., Fissekis, J. D., Alaud-
din, M. M., DArgenio, D. Z., and Conti, P. S.
Kinetic modeling of 5-fluorouracil anabolism
in colorectal adenocarcinoma: a positron emis-
sion tomography study in rats. Cancer Res.,
63: 36673674, 2003.
19 Harte, R. J., Matthews, J. C., OReilly, S. M.,
Tilsley, D. W., Osman, S., Brown, G., Luthra,
S. J., Brady, F., Jones, T., and Price, P. M. Tu-
mor, normal tissue, and plasma pharmacoki-
netic studies of fluorouracil biomodulation
with N-phosphonacetyl-L-aspartate, folinic
acid, and interferon alfa. J. Clin. Oncol., 17:
15801588, 1999.
20 Dimitrakopoulou-Strauss, A., Strauss, L. G.,
Schlag, P., Hohenberger, P., Irngartinger, G.,
Oberdorfer, F., Doll, J., and van Kaick, G. In-
travenous and intra-arterial oxygen-15-labeled
water and fluorine-18-labeled fluorouracil in
patients with liver metastases from colorectal
carcinoma. J. Nucl. Med., 39: 465473, 1998.
21 Vasey, P. A., Kaye, S. B., Morrison, R., Twelves,
C., Wilson, P., Duncan, R., Thomson, A. H.,
Murray, L. S., Hilditch, T. E., Murray, T., Bur-
tles, S., Fraier, D., Frigerio, E., and Cassidy, J.
Phase I clinical and pharmacokinetic study of
PK1 [N-(2-Hydroxypropyl)methacrylamide co-
polymer doxorubicin]: first member of a new
class of chemotherapeutic agents drug-poly-
mer conjugates. Clin. Cancer Res., 5: 8394,
1999.
22 Gambhir, S. S. Molecular imaging of cancer
with positron emission tomography. Nature
Med., 2: 683693, 2002.
23 Groot-Wassink, T., Aboagye, E. O., Glaser, M.,
Lemoine, N. R., and Vassaux, G. Adenovirus
biodistribution and noninvasive imaging of
gene expression in vivo by positron emission
tomography using human sodium/iodide
symporter as reporter gene. Hum. Gene Ther.,
13: 17231735, 2002.
24 Groot-Wassink, T., Aboagye, E. O., Wang, Y.,
Lemoine, N. R., Reader, A. J., and Vassaux, G.
Quantitative imaging of NaI symporter trans-
gene expression using positron emission to-
mography in the living animal. Mol. Ther., 9:
436442, 2004.
25 Groot-Wassink, T., Aboagye, E. O., Wang, Y.,
Lemoine, N. R., Keith, W. N., and Vassaux, G.
Non-invasive imaging of the transcriptional
activities of human telomerase promoter frag-
ments in mice. Cancer Res., 64: 49064911,
2004.
26 Serganova, I., Doubrovin, M., Vider, J., Pono-
marev, V., Soghomonyan, S., Beresten, T.,
Ageyeva, L., Serganov, A., Cai, S., Balatoni, J.,
Blasberg, R., and Gelovani, J. Molecular imag-
ing of temporal dynamics and spatial hetero-
geneity of hypoxia-inducible factor-1 signal
transduction activity in tumors in living mice.
Cancer Res., 64: 61016108, 2004.
References 1265
27 Parulekar, W. R. and Eisenhauer, E. A. Phase I
trial design for solid tumor studies of tar-
geted, non-cytotoxic agents: theory and prac-
tice. J. Natl. Cancer Inst., 96: 990997, 2004.
28 Herbst, R. S. Review of epidermal growth fac-
tor biology. Int. J. Radiat. Oncol. Biol. Phys.,
59: 2126, 2004.
29 Silchenmyer, W. J., Elliott, W. L., and Fry, D. W.
CI-1033, a pan-erb tyrosine kinase inhibitor.
Semin. Oncol., 28: 8085, 2001.
30 Gaul, M. D., Guo, Y., Affleck, K., Cockerill,
G. S., Gilmer, T. M., Griffin, R. J., Guntrip, S.,
Keith, B. R., Knight, W. B., Mullin, R. J., Mur-
ray, D. M., Rusnak, D. W., Smith, K., Tadepalli,
S., Wood, E. R., and Lackey, K. Discovery and
biological evaluation of potent dual erbB-2/
EGFR tyrosine kinase inhibitors: 6-thiazolyl-
quinazolines. Bioorg. Med. Chem. Lett., 13:
637640, 2003.
31 Mishani, E., Abourbeh, G., Rozen, Y., Jacob-
son, O., Laky, D., David, I. B., Levitzki, A., and
Shaul, M. Novel carbon-11 labeled dimethyl-
amino-but-2-enoic acid [4-(phenylamino)-qui-
nazoline-6-yl]-amides: potential PET bioprobes
for molecular imaging of EGFR-positive tu-
mors. Nucl. Med. Biol., 31: 469476, 2004.
32 Dehdashti, F., Mortimer, J. E., Siegel, B. A.,
Griffith, L. K., Bonasera, T. J., Fusselman, M. J.,
Detert, D. D., Cutler, P. D., Katzenellenbogen,
J. A., and Welch, M. J. Positron tomographic
assessment of estrogen receptors in breast
cancer: comparison with FDG-PET and in vi-
tro receptor assays. J. Nucl. Med., 36: 1766
1774, 1995.
33 Dehdashti, F., Flanagan, F. L., Mortimer, J. E.,
Katzenellenbogen, J. A., Welch, M. J., and
Siegel, B. A. Positron emission tomographic
assessment of metabolic flare to predict
response of metastatic breast cancer to anti-
estrogen therapy. Eur. J. Nucl. Med., 26: 51
56, 1999.
34 McGuire, A. H., Dehdashti, F., Siegel, B. A.,
Lyss, A. P., Brodack, J. W., Mathias, C. J., Min-
tun, M. A., Katzenellenbogen, J. A., and Welch,
M. J. Positron tomographic assessment of 16
alpha-[
18
F] fluoro-17 beta-estradiol uptake in
metastatic breast carcinoma. J. Nucl. Med., 32:
15261531, 1991.
35 Zucker, S., Cao, J., and Molloy, C. J. Role of
matrix metalloproteinases and plasminogen
activators in cancer invasion and metastasis:
therapeutic strategies. In: B. C. Baguley and
D. J. Kerr (Eds.), Anticancer drug develop-
ment, pp. 91122. Academic Press, San Die-
go, 2002.
36 Furumoto, S., Takashima, K., Kubota, K., Ido,
T., Iwata, R., and Fukuda, H. Tumor detection
using
11
F-labeled matrix metalloproteinase-2
inhibitor. Nucl. Med. Biol., 30: 119125, 2003.
37 Cheesman, E. H., Rajopadhye, M., Tran, Y. S.,
Liu, S., Ellars, C., Onthank, D., Silva, P., Yala-
manchili, P., Kavosi, M., Sachleben, R. A., Liu,
R. Q., Robinson, S. P., and Edwards, D. S.
Radiolabeled matrix metalloproteinase inhibi-
tors with high uptake in mouse tumor mod-
els. J. Label. Compds. Radiopharm., 46: S5,
2003.
38 Smith-Jones, P. M., Solit, D. B., Akhurst, T.,
Afroze, F., Rosen, N., and Larson, S. M. Imag-
ing the pharmacodynamics of Her2 degrada-
tion in response to Hsp90 inhibitors. Nat. Bio-
technol., 22: 701706, 2004.
39 Sundaresan, G., Yazaki, P. J., Shively, J. E.,
Finn, R. D., Larson, S. M., Raubitschek, A. A.,
Williams, L. E., Chatziioannou, A. F., Gambhir,
S. S., and Wu, A. M.
124
I-labeled engineered
anti-CEA minibodies and diabodies allow
high-contrast, antigen-specific small-animal
PET imaging of xenografts in athymic mice.
J. Nucl. Med., 44: 19621969, 2003.
40 Haubner, R., Wester, H.-J., Weber, W. A., and
Schwaiger, M. Radiotracer-based strategies to
image angiogenesis. Q. J. Nucl. Med., 47:
189199, 2003.
41 Poethko, T., Schottelius, M., Thumshirn, G.,
Hersel, U., Herz, M., Henriksen, G., Kessler,
H., Schwaiger, M., and Wester, H.-J. Two-step
methodology for high-yield routine radiohalo-
genation of peptides:
18
F-labeled RGD and oc-
treotide analogs. J. Nucl. Med., 45: 892902,
2004.
42 Glaser, M., Karlsen, H., Solbakken, M., Ar-
ukwe, J., Brady, F., Luthra, S. K., and Cuth-
bertson, A. 18F-Fluorothiols: a new approach
to label peptides chemoselectively as potential
tracers for positron emission tomography. Bio-
conj. Chem., 15: 14471453, 2004.
43 Chen, X., Liu, S., Hou, Y., Tohme, M., Park,
R., Bading, J. R., and Conti, P. S. MicroPET
imaging of breast cancer alphav-integrin ex-
pression with
64
Cu-labeled dimeric RGD pep-
tides. Mol. Imaging Biol., 6: 350359, 2004.
44 Chen, X., Hou, Y., Tohme, M., Park, R., Khan-
kaldyyn, V., Gonzales-Gomez, I., Bading, J. R.,
Laug, W. E., and Conti, P. S. Pegylated Arg-
Gly-Asp peptide:
64
Cu labeling and PET imag-
ing of brain tumor alphavbeta3-integrin ex-
pression. J. Nucl. Med., 45: 17761783, 2004.
45 Haubner, R., Kuhnast, B., Mang, C., Weber,
W. A., Kessler, H., Wester, H.-J., and Schwai-
ger, M. [
18
F]Galacto-RGD: synthesis, radiola-
beling, metabolic stability, and radiation dose
estimates. Bioconj. Chem., 15: 6169, 2004.
46 Hanahan, D. and Weinberg, R. A. The hall-
marks of cancer. Cell, 100: 5770, 2000.
47 Therasse, P., Arbuck, S. G., Eisenhauer, E. A.,
Wanders, J., Kaplan, R. S., Rubinstein, L., Ver-
weij, J., Van Glabbeke, M., van Oosterom,
A. T., Christian, M. C., and Gwyther, S. G. New
guidelines to evaluate the response to treat-
ment in solid tumors. European Organization
for Research and Treatment of Cancer, Na-
tional Cancer Institute of the United States,
National Cancer Institute of Canada. J. Natl.
Cancer Inst., 92: 205216, 2000.
48 Kenny, L. M., Aboagye, E. O., and Price, P. M.
Positron emission tomography imaging of cell
proliferation in oncology. Clin. Oncol., 16:
176185, 2004.
49 Weber, G. Enzymology of cancer cells (second
of two parts). N. Engl. J. Med., 296: 541551,
1977.
50 Weber, G. Enzymology of cancer cells (first of
two parts). N. Engl. J. Med., 296: 486492, 1977.
51 Nelson, C. A., Wang, J. Q., Leav, I., and Crane,
P. D. The interaction among glucose transport,
hexokinase, and glucose-6-phosphatase with
respect to 3H-2-deoxyglucose retention in
murine tumor models. Nucl. Med. Biol., 23:
533541, 1996.
52 Gallagher, B. M., Fowler, J. S., Gutterson, N. I.,
MacGregor, R. R., Wan, C. N., and Wolf, A. P.
Metabolic trapping as a principle of radiophar-
maceutical design: some factors responsible
for the biodistribution of [
18
F] 2-deoxy-2-
fluoro-D-glucose. J. Nucl. Med., 19: 1154
1161, 1978.
53 Kubota, R., Yamada, S., Kubota, K., Ishiwata,
K., Tamahashi, N., and Ido, T. Intratumoral
distribution of
18
F-fluorodeoxyglucose in vivo:
high accumulation in macrophages and gran-
ulation tissues studied by microautoradiogra-
phy. J. Nucl. Med., 33: 19721980, 1992.
54 Weber, W. A., Schwaiger, M., and Avril, N.
Quantitative assessment of tumor metabolism
using FDG-PET imaging. Nucl. Med. Biol.,
27: 683687, 2000.
55 Bourguet, P., Blanc-Vincent, M. P., Boneu, A.,
Bosquet, L., Chauffert, B., Corone, C., Cour-
bon, F., Devillers, A., Foehrenbach, H., Lum-
broso, J. D., Mazselin, P., Montravers, F., Mo-
retti, J. L., and Talbot, J. N. Summary of the
Standards, Options and Recommendations for
the use of positron emission tomography with
2-[18F]fluoro-2-deoxy-D-glucose (FDP-PET
scanning) in oncology (2002). Br. J. Cancer,
89: S8491, 2003.
56 Young, H., Baum, R., Cremerius, U., Herholz,
K., Hoekstra, O., Lammertsma, A. A., Pruim,
J., and Price, P. Measurement of clinical and
subclinical tumour response using [
18
F]-fluor-
odeoxyglucose and positron emission tomog-
raphy: Review and 1999 EORTC recommenda-
tions. Eur. J. Cancer, 35: 17731782, 1999.
57 Brock, C. S., Young, H., OReilly, S. M., Mat-
thews, J., Osman, S., Evans, H., Newlands,
E. S., and Price, P. Early evaluation of tumour
metabolic response using [
18
F]fluorodeoxyglu-
cose and positron emission tomography: a pi-
lot study following the phase II chemotherapy
schedule for temozolomide in recurrent high-
grade gliomas. Br. J. Cancer, 82: 608615,
2000.
58 Wahl, R. L., Zasadny, K., Helvie, M., Hutchins,
G. D., Weber, B., and Cody, R. Metabolic mon-
itoring of breast cancer chemohormonother-
apy using positron emission tomography: ini-
tial evaluation. J. Clin. Oncol., 11: 21012111,
1993.
59 Spaepen, K., Stroobants, S., Dupont, P., Van
Steenweghen, S., Thomas, J., Vandenberghe,
P., Vanuytsel, L., Bormans, G., Balzarini, J.,
De Wolf-Peeters, C., Mortelmans, L., and Ver-
hoef, G. Prognostic value of positron emission
tomography (PET) with fluorine-18fluorode-
oxyglucose ([
18
F]FDG) after first-line chemo-
therapy in non-Hodgkins lymphoma: is
[
18
F]FDG-PET a valid alternative to conven-
tional diagnostic methods? J. Clin. Oncol., 19:
414419, 2001.
60 van Oosterom, A. T., Judson, I., Verweij, J.,
Stroobants, S., Donato di Paola, E., Dimitri-
jevic, S., Martens, M., Webb, A., Sciot, R., Van
Glabbeke, M., Silberman, S., and Nielsen,
O. S. Safety and efficacy of imatinib (STI571)
in metastatic gastrointestinal stromal tu-
mours: a phase I study. Lancet, 358: 1421
1423, 2001.
61 Joensuu, H., Roberts, P. J., Sarlomo-Rikala,
M., Andersson, L. C., Tervahartiala, P., Tuve-
son, D., Silberman, S., Capdeville, R., Dimitri-
jevic, S., Druker, B., and Demetri, G. D. Effect
References 1267
of the tyrosine kinase inhibitor STI571 in a
patient with a metastatic gastrointestinal stro-
mal tumor. N. Engl. J. Med., 344: 10521056,
2001.
62 Grierson, J. R., Schwartz, J. L., Muzi, M.,
Jordan, R., and Krohn, K. A. Metabolism of 3'-
deoxy-3'[F-18]fluorothymidine in proliferating
A549 cells: validation for positron emission to-
mography. Nucl. Med. Biol., 31: 829837, 2004.
63 Barthel, H., Perumal, M., Latigo, J., He, Q.,
Brady, F., Luthra, S. K., Price, P., and Aboagye,
E. O. The uptake of 3'-deoxy-3'-[
18
F]Fluorothy-
midine into L5178Y tumors in vivo is depen-
dent on thymidine kinase 1 protein and ATP
levels. Eur. J. Nucl. Med. Mol. Imaging, 32:
257263, 2005.
64 Rasay, J. S., Grierson, J. R., Wiens, L. W., Kolb,
P. D., and Schwartz, J. L. Validation of FLT up-
take as a measure of thymidine kinase-1 activ-
ity in A549 carcinoma cells. J. Nucl. Med., 43:
12101217, 2002.
65 Buck, A. K., Halter, G., Schirrmeister, H., Kot-
zerke, J., Wurzinger, I., Glatting, G., Mattfeldt,
T., Neumaier, B., Reske, S. N., and Hetzel, M.
Imaging proliferation in lung tumors with
PET:
18
F-FLT versus
18
F-FDG. J. Nucl. Med.,
44: 14261431, 2003.
66 Buck, A. K., Schirrmeister, H., Hetzel, M., von
der Heide, M., Halter, G., Glatting, G., Matt-
feldt, T., Liewald, F., Reske, S. N., and Neu-
maier, B. 3-deoxy-3-[(
18
)F]fluorothymidine-posi-
tron emission tomography for non-invasive
assessment of proliferation in pulmonary
nodules. Cancer Res., 62: 33313334, 2002.
67 Toyohara, J., Waki, A., Takamatsu, S., Yone-
kura, Y., Magata, Y., and Fujibayashi, Y. Basis
of FLT as a cell proliferation marker: compara-
tive uptake studies with [
3
H]arabinothymidine,
and cell-analysis in 22 asynchronously grow-
ing tumor cell lines. Nucl. Med. Biol., 29:
281287, 2002.
68 Oyama, N., Ponde, D. E., Dence, C., Kim, J.,
Tai, Y. C., and Welch, M. J. Monitoring of ther-
apy in androgen-dependent prostate tumor
model by measuring tumor proliferation. J.
Nucl. Med., 45: 519525, 2004.
69 Sugiyama, M., Sakahara, H., Sato, K., Harada,
N., Fukumoto, D., Kakiuchi, T., Hirano, T.,
Kohno, E., and Tsukada, H. Evaluation of 3'-
deoxy-3'-
18
F-fluorothymidine for monitoring
tumor response to radiotherapy and photody-
namic therapy in mice. J. Nucl. Med., 45:
17541758, 2004.
70 Barthel, H., Cleij, M. C., Collingridge, D. R.,
Hutchinson, O. C., Osman, S., He, Q., Luthra,
S. K., Brady, F., Price, P. M., and Aboagye,
E. O. 3'-deoxy-3'-[
18
F]Fluorothymidine as a new
marker for monitoring tumor response to
anti-proliferative therapy in vivo with positron
emission tomography. Cancer Res., 3: 3791
3798, 2003.
71 Shields, A. F., Mankoff, D. A., Link, J. M., Gra-
ham, M. M., Eary, J. F., Kozawa, S. M., Zheng,
M., Lewellen, B., Lewellen, T. K., Grierson,
J. R., and Krohn, K. A. Carbon-11-thymidine
and FDG to measure therapy response.
J. Nucl. Med., 39: 17571762, 1998.
72 Collingridge, D. R., Glaser, M., Osman, S.,
Barthel, H., Hutchinson, O. C., Luthra, S. K.,
Brady, F., Bouchier-Hayes, L., Martin, S. J.,
Workman, P., Price, P., and Aboagye, E. O. In
vitro selectivity, in vivo biodistribution and tu-
mour uptake of annexin V radiolabelled with
a positron emitting radioisotope. Br. J. Cancer,
89: 13271333, 2003.
73 Glaser, M., Collingridge, D. R., Aboagye, E. O.,
Bouchier-Hayes, L., Hutchinson, O. C., Martin,
S., Price, P., Brady, F., and Luthra, S. K. Io-
dine-124 labelled annexin-V as a potential
radiotracer to study apoptosis using positron
emission tomography. Appl. Radiat. Isot., 58:
5562, 2003.
74 Blankenberg, F. G., Tait, J. F., and Strauss,
H. W. Apoptotic cell death: its implications for
imaging in the next millennium. Eur. J. Nucl.
Med., 27: 359367, 2000.
75 Blankenberg, F. G. Recent advances in the
imaging of programmed cell death. Curr.
Pharm. Des., 10: 14571467, 2004.
76 Blankenberg, F. G., Katsikis, P. D., Tait, J. F.,
Davis, R. E., Naumovski, L., Ohtsuki, K., Kopi-
woda, S., Abrams, M. J., Darkes, M., Robbins,
R. C., Maecker, H. T., and Strauss, H. W. In
vivo detection and imaging of phosphatidylse-
rine expression during programmed cell
death. Proc. Natl. Acad. Sci. USA, 95: 6349
6354, 1998.
77 van de Wiele, C., Lahorte, C., Vermeersch, H.,
Loose, D., Mervillie, K., Steinmetz, N. D., Van-
derheyden, J. L., Cuvelier, C. A., Slegers, G.,
and Dierck, R. A. Quantitative tumor apoptosis
imaging using technetium-99m-HYNIC an-
nexin V single photon emission computed
tomography. J. Clin. Oncol., 21: 34833487,
2003.
78 Haas, R. L., de Jong, D., Valdes Olmos, R. A.,
Hoefnagel, C. A., van den Heuvel I., Zerp,
S. F., Bartelink, H., and Verheij, M. In vivo
imaging of radiation-induced apoptosis in fol-
licular lymphoma patients. Int. J. Radiat. On-
col. Biol. Phys., 59: 782787, 2004.
79 Graeber, T. G., Osmanian, C., Jacks, T., Hous-
man, D. E., Koch, C. J., Lowe, S. W., and Giac-
cia, A. J. Hypoxia-mediated selection of cells
with diminished apoptotic potential in solid
tumours. Nature, 379: 8891, 1996.
80 Horsman, M. R. and Overgaard, J. Overcom-
ing tumour radiation resistance resulting
from acute hypoxia. Eur. J. Cancer, 28A: 2084
2085, 1992.
81 Hockel, M., Schlenger, K., Vorndran, B., Baus-
samann, E., Mitze, M., Knapstein, P. G., and
Vaupel, P. Intratumoral pO
2
predicts survival
in advanced cancer of the uterine cervix.
Radiother. Oncol., 26: 4550, 1993.
82 Brizel, D. M., Scully, S. P., Harrelson, J. M.,
Layfield, L. J., Bean, J. M., Prosnitz, L. R., and
Dewhirst, M. W. Tumor oxygenation predicts
for the likelihood of distant metastases in hu-
man soft tissue sarcoma. Cancer Res., 56:
941943, 1996.
83 Rasey, J. S., Koh, W. J., Evans, M. L., Peterson,
L. M., Lewellen, T. K., Graham, M. M., and
Krohn, K. A. Quantifying regional hypoxia in
human tumors with positron emission tomog-
raphy of [
18
F]fluoromisonidazole: a pretherapy
study of 37 patients. Int. J. Radiat. Oncol.
Biol. Phys., 36: 417428, 1996.
84 Barthel, H., Wilson, H., Collingridge, D. R.,
Brown, G., Osman, S., Luthra, S. K., Brady, F.,
Workman, P., Price, P. M., and Aboagye, E. O.
In vivo evaluation of [
18
F]fluoroetanidazole as
a new marker for imaging tumor hypoxia with
positron emission tomography. Br. J. Cancer,
90: 22322242, 2004.
85 Dehdashti, F., Mintum, M. A., Lewis, J. S.,
Bradley, J., Govindan, R., Laforest, R., Welch,
M. J., and Siegel, B. A. In vivo assessment of
tumor hypoxia in lung cancer with
60
Cu-
ATSM. Eur. J. Nucl. Med. Mol. Imaging, 30:
844850, 2003.
86 Koh, W. J., Bergman, K. S., Rasey, J. S., Peter-
son, L. M., Evans, M. L., Graham, M. M., Grier-
son, J. R., Lindsley, K. L., Lewellen, T. K.,
Krohn, K. A., et al. Evaluation of oxygenation
status during fractionated radiotherapy in
human non-small cell lung cancers using
[F-18]fluoromisonidazole positron emission
tomography. Int. J. Radiat. Oncol. Biol. Phys.,
33: 391398, 1995.
References 1269
Abstract
The targeted delivery of molecules to sites
of disease in vivo promises to open new
avenues for the imaging of pathologies,
and for the development of more selective
therapeutic agents. This chapter will re-
view progress made in the identification of
pathology-associated antigens and in the
development of binding molecules (anti-
bodies, peptides and small organic mole-
cules). Furthermore, we will present the
authors views on molecular strategies for
the conversion of binding molecules into
novel imaging or therapeutic biopharma-
ceutical agents.
Abbreviations
cytotoxicity
ADEPT Antibody-Directed Enzyme
Prodrug Therapy
aFGF acid fibroblast growth factor
APb42 amyloid-b-42 peptide
APP amyloid precursor protein
ARMD age-related macular
degeneration
BBB bloodbrain barrier
bFGF basic fibroblast growth factor
CLIO cross-linked iron oxide
DTPA diethylenetriamine penta-
acetate
ECAM endothelial cell adhesion
molecules
ECM extracellular matrix
ECs endothelial cells
EDB domain of the extracellular
matrix protein fibronectin
ESACHEL Encoded Self-Assembling
Chemical Libraries
FMT fluorescence-mediated
tomography
FR-b folate receptor b
ICAM intercellular adhesion molecule
Gd-DTPA gadolinium-DTPA
HIF-1 hypoxia-inducible factor
MMP matrix metalloproteinases
NIRF near-infrared fluorescence
PD-ECGF platelet-derived endothelial
cell growth factor
PET positron emission
tomography
PSMA prostate-specific membrane
antigen
SAP serum amyloid protein
SPECT single-photon emission
computed tomography
TGFa transforming growth factor a
TGFb transforming growth factor b
1271
6
Ligand-based Targeting of Disease:
From Antibodies to Small Organic (Synthetic) Ligands
ISBN: 3-527-31184-X
uPA urokinase-plasminogen
activator
VCAM vascular cell adhesion molecule
factor
6.1
Introduction
Chemotherapy that is, the administra-
tion of chemical compounds in order to
confer a therapeutic benefit to the patient
is often limited by the doses of drug
which can be reached, without observing
limiting toxicities. For example, in oncol-
ogy, many therapeutic strategies rely on
the expectation that anti-cancer drugs will
preferentially kill rapidly dividing tumor
cells, rather than normal cells. Since a
large proportion of tumor cells must be
killed in order to obtain and maintain a
complete remission, large doses of drugs
are typically used, with significant toxicity
towards proliferating non-malignant cells.
It is therefore not surprising that the
search for improved potency and selectivity
of therapeutic compounds is a common fea-
ture in most pharmaceutical development
activities. In principle, several strategies
could be considered in order to develop bet-
ter, more selective therapeutic agents. In
many cases, research is driven by the hope
to identify macromolecular targets, which
are not essential in normal physiology, but
the inhibition of which may revert the
pathological condition that one intends to
fight. While such prerequisites may be
met in certain therapeutic areas (e.g., the
use of antibiotics inhibiting microbial pro-
tein targets which do not have a counterpart
in the host), the discovery of selective tar-
gets remains a formidable challenge for
many relevant pathologies.
The selective delivery of bioactive com-
pounds to a site of disease (the magic bul-
lets first envisioned by Paul Ehrlich at the
end of the nineteenth century) appears to
be a general strategy for the development
of better, more selective therapeutic agents.
In most cases, the selective accumulation of
drugs at the site of disease will spare nor-
mal tissues and will increase the therapeu-
tic index of the drug that is, the relative ac-
tivity towards the diseased tissue, compared
to normal organs. In principle, targeted
strategies based on the selective delivery of
active compounds could be applicable both
in diseases in which cell growth has to be
limited (e.g., cancer) or promoted (i.e., tis-
sues regeneration after infarction).
The words targeting and targeted
therapy are often used for a variety of dif-
ferent pharmaceutical approaches, aimed
at achieving better in vivo selectivities. In
this chapter, however, we will concentrate
solely on those targeting strategies, which
rely on the ligand-based selective delivery
of bioactive agents to sites of disease.
Those readers interested in other targeting
strategies, which achieve a selective biodis-
tribution in vivo in the absence of a specif-
ic molecular recognition event and by
means of other physical principles (e.g.,
the enhanced permeability and retention
of polymers in tumors [1, 2]), are encour-
aged to consult other reviews which have
been written on this topic [35].
The ligand-based targeting of diseases is a
rational strategy for drug discovery. To some
extent, in fact, the performance of a targeted
drug can be predicted on the basis of how
selective is its localization on the target tis-
sue. Furthermore, a binding molecule cap-
able of disease targeting may be useful not
only for therapeutic applications, but also
for imaging purposes, after modification
with a suitable radionuclide or infrared
fluorophore. A number of parameters are
expected to influence the in vivo perfor-
mance of a targeting agent. Molecular
6 Ligand-based Targeting of Disease: From Antibodies to Small Organic (Synthetic) Ligands 1272
weight, binding affinity for the target, solu-
bility, valence are some of the parameters
which, in many cases, have been shown to
contribute to the overall performance of
the targeting process.
In this chapter, we will first review the
classes of molecules that are currently used
for targeting applications. We will then ana-
lyze classes of diseases which lend them-
selves to molecular targeting, and finally
will discuss how a ligand can be converted
into a diagnostic or therapeutic agent.
Different pathologies may require the li-
gand-based targeting of different antigens,
located on different structures. A sche-
matic representation of tissues and cellular
components which can be considered for
targeting applications is provided in
Fig. 6.1. Whilst in oncology most targeting
approaches focus on the tumor cells, other
structures such as altered vascular struc-
tures, modified extracellular matrix, infil-
trating leukocytes, areas of necrosis, pla-
ques and microbes can also be used as tar-
gets in oncology and in other diseases.
6.2
Ligands
6.2.1
Antibodies
At present, antibodies are the only general
class of affinity reagents which can be gen-
erated rapidly against virtually any biomole-
cular target. Monoclonal antibodies repre-
sent an ideal alternative to hyperimmune
sera for in vivo applications [6]. However, ro-
dent antibodies are immunogenic in hu-
mans. Early studies showed that human
monoclonal antibodies can be produced by
immortalizing B cells with EpsteinBarr
virus (EBV) [7, 8], or by fusing B cells with
an appropriate partner to produce hybrido-
mas [9, 10]. However, these methods have
6.2 Ligands 1273
Fig. 6.1 Schematic representation of tissue and cellular components, which can be considered
for targeting applications.
very low efficiency, and therefore alternative
strategies have been developed. These in-
clude: 1) humanization of murine monoclo-
nal antibodies through protein engineering
[11] (see also Part V, Chapters 1 and 2) selec-
tion of antibodies from phage-display li-
braries of human antibody fragments [12,
13] origin (see also Part V, Chapters 2 and
3) immunization of transgenic mice carry-
ing human immunoglobulin loci, followed
by production of monoclonal antibodies
using hybridoma technology [14].
Monoclonal antibodies exhibit a slow
elimination from the blood, and accumu-
late in the liver. For these reasons, rapidly
clearing antibody fragments are typically
preferred for imaging applications in nu-
clear medicine. By contrast, intact immu-
noglobulins continue to represent the anti-
body format of choice for many therapeu-
tic applications [15], which rely on the
antibodys ability to interfere with signal-
ing events, and to activate antibody-depen-
dent cellular cytotoxicity (ADCC) mecha-
nisms or complement.
The immunogenicity of rodent antibod-
ies continues to be a concern for repeated
administrations to humans, and the use of
chimeric, humanized or fully human anti-
bodies is generally preferred.
In our experience, antibody phage tech-
nology represents the most efficient avenue
for producing good-quality human mono-
clonal antibodies, whenever sufficient quan-
tities of pure antigen are available (12 mg).
The display of antibody fragments on the
surface of filamentous phage allows the fa-
cile construction of large (>10
9
antibodies)
libraries of human antibodies, from which
monoclonal antibodies can be isolated by
panning the phage library onto an immobi-
lized antigen [13, 16]. When required, anti-
body affinity can be matured using combi-
natorial mutagenesis of the antibody gene
and stringent selection strategies [17, 18].
Recently, ribosome display has been pro-
Fig. 6.2 Different antibody formats and
antibody fragments.
posed as a fully in vitro avenue for the isola-
tion and affinity maturation of human anti-
bodies [19].
Antibody phage technology directly
yields antibody fragments (typically in scFv
or Fab format). However, other antibody
formats (e.g., IgG) can easily be obtained
by transplanting the genes coding for the
variable antibody domains into suitable ex-
pression vectors (Fig. 6.2).
6.2.2
Peptides
A number of internalizing peptides, specific
to receptors which are overexpressed in tu-
mor cells, have been used for the imaging
of tumors and for the selective delivery of
therapeutic radionuclides to neoplastic le-
sions. The somatostatin analogue octreotide
[20], for example, has been approved in Eu-
rope or the USA for the imaging of tumors.
Several other agents are in development [21,
22], such as integrin binding peptides
(RGD-peptides) [23] and bombesin peptide
analogues [24]. Other areas in which natu-
rally occurring peptides (or peptides derived
from protein sequences) are used include
the ligand-based targeting of thrombotic
events, of microbial infections and of amy-
loidosis (see sections below).
In contrast to naturally occurring pep-
tides, high-affinity peptidic ligands to pro-
tein targets are often difficult to isolate.
Phage display libraries of linear and disul-
fide-constrained peptides are commercially
available, and have been used for the isola-
tion of binding specificities [25]. For exam-
ple, peptides specific to human lung tumor
cell lines have been selected from a phage
library [26]. Novel technologies for the isola-
tion of high-affinity binding peptides are
available [27], but the in vivo stability of lin-
ear peptides remains a cause for concern.
Peptide phage libraries have been used for
in vivo panning applications by the groups
of Pasqualini and Ruoslahti [28, 29], but
the real imaging and therapeutic potential
of these phage-derived peptides remains to
be investigated in advanced animal models,
as well as in the clinic.
6.2.3
Small Organic Molecules
In contrast to antibody technology, the isola-
tion of high-affinity small organic binders
to protein antigens can be a difficult task,
which often fails when the epitopes to be
recognized do not contain hydrophobic
pockets [30]. An increasing number of ex-
perimental evidences suggest that bidentate
ligands, recognizing adjacent but not-over-
lapping surfaces of the target protein, may
display high binding affinity and specificity,
as a result of the chelate effect [31]. Methods
for the identification of such bidentate li-
gands include SAR-by-NMR [32], dynamic
combinatorial chemistry [33], and tethering
approaches [34]. Our laboratory has recently
developed a novel technology (termed En-
coded Self-Assembling Chemical Libraries,
or ESACHEL), which allows the facile con-
struction of very large libraries of chemical
compounds by the DNA-mediated self-as-
sembly of smaller sub-libraries [35]. Each
pharmacophore in the library is covalently
coupled to an oligonucleotide, which med-
iates the self-assembly of the library and
provides the pharmacophore with a distinc-
tive identification DNA tag. Similar to anti-
body phage display libraries, ESACHEL li-
braries can be panned in solution, thus en-
riching bidentate ligands which display a
preferential binding to the target of interest.
After the capture of the desired binding
specificities on the protein target, the bind-
ing code associated with the selected phar-
macophores can be decoded by a number
of experimental techniques (e.g., hybridiza-
6.2 Ligands 1275
tion on DNA chips, by a modified PCR tech-
nique followed by sequencing). We have de-
scribed the isolation of ESACHEL-derived
bidentate molecules, with nanomolar affin-
ity to carbonic anhydrase [35].
6.3
Classes of Diseases
6.3.1
Cancer
Cancer chemotherapy can be successful in
certain specific indications, but suffers
otherwise from major drawbacks. The lack
of selectivity of anti-proliferative agents
may give rise to severe side effects, thus
limiting efficacy and facilitating the devel-
opment of acquired drug resistance. The
discovery of more selective anticancer
drugs, with better discrimination between
tumor and normal cells, is possibly the
most important goal of modern anticancer
research. The targeted delivery of bioactive
moieties (drugs, cytokines, procoagulant
factors, photosensitizers, radionuclides,
etc.) by means of binding molecules (re-
combinant antibodies, peptides, etc.) spe-
cific to tumor-associated markers can im-
prove the efficiency of tumor therapy and
limit non-specific toxicity.
6.3.1.1 Tumor-associated Markers
The selection of a suitable molecular target
is an essential step in the design of any li-
gand-based therapeutic. Tumor-associated
markers are usually proteins or carbohy-
drates that are abnormally expressed or
overexpressed in the tumor environment.
Fundamental prerequisites of an ideal tu-
mor-associated marker are specificity,
abundance and stability, together with
good accessibility for ligand molecules
transported by the blood stream. To date,
only few good-quality tumor-associated
markers are known (see also Part I, Chap-
ter 5). Most existing candidate markers are
also present in normal tissues, thus limit-
ing their usefulness for in vivo targeting
applications. Several methods, such as pro-
teomic and transcriptomic techniques, bio-
panning of phage display libraries and se-
rial analysis of gene expression, are now
available and may help to identify new tu-
mor-associated markers. However, the vali-
dation of the newly identified markers re-
quires the generation of specific monoclo-
nal antibodies, extensive immunohisto-
chemical analysis and biodistribution
experiments in tumor-bearing animals.
Tumor-associated markers (antigens) can
be grouped into two main categories ac-
cording to their localization in the tumor
tissue: 1) antigens on the surface of tumor
cells (tumor markers); or 2) stromal anti-
gens, which can be located either around
the tumor neovasculature or display a
more diffuse staining pattern correspond-
ing to the modified extracellular matrix
(ECM) of solid tumors.
6.3.1.2 Targeting Markers on Tumor Cells
In this section we will discuss the most ex-
tensively studied tumor markers and the
corresponding monoclonal antibodies that
have been approved by the FDA (US Food
and Drug Administration; www.fda.gov)
and in Europe for the imaging and thera-
py of cancer, or that are currently being
developed.
Tumor-associated glycoprotein 72 (TAG-
72) This tumor marker [36] was identi-
fied by means of a murine monoclonal
antibody, B72.3, raised against human,
metastatic mammary carcinoma cells [37].
The expression pattern of TAG-72 was ex-
tensively analyzed in a number of different
tumors, such as ovarian carcinoma, pan-
creatic adenocarcinoma, and colorectal
adenocarcinoma [3840]. Immunohisto-
chemical studies showed that TAG-72 is
expressed in more than 80% of colorectal
carcinomas, but is rarely expressed in nor-
mal epithelium and benign diseases. TAG-
72 can also be found in the body fluids of
patients with adenocarcinomas, and its di-
rect measurement can be used in conjunc-
tion with immunocytochemical analysis to
help in discriminating benign from malig-
nant effusions [38]. The murine monoclo-
nal antibody satumomab pentedite in-
dium-111 conjugate (OncoScint), specific
to TAG-72, was the first monoclonal anti-
body approved by the FDA for tumor
imaging (colorectal and ovarian cancer).
Carcinoembryonic antigen (CEA) This was
first described in 1965 by Freedman and
Gold [41]. CEA, which is a highly glycosy-
lated membrane protein, has a restricted
expression in normal tissues and is ex-
pressed at high levels in positive tumors
(colon carcinoma). CEA became one of the
most widely used tumor markers world-
wide. Its main application is mostly in gas-
trointestinal cancer, especially in colorectal
malignancy [42]. CEA-Scan, a murine
monoclonal antibody fragment (Fab)
linked to technetium-99m, was approved
in both Europe and the USA in 1996 for
the detection of recurrent/metastatic colo-
rectal cancer. Further anti-CEA monoclonal
antibody fragments were developed by the
group of Begent. The scFv fragment MFE-
23 [43], which shows high affinity to CEA,
was tested in biodistribution experiments
[44] and was genetically fused to several
bioactive molecules, such has TNF-a [45]
and carboxypeptidase G2 (CPG2) [46] used
for Antibody-Directed Enzyme Prodrug
Therapy (ADEPT).
Prostate-specific membrane antigen
(PSMA) This is a type 2 membrane pro-
tein that represents an attractive target for
cancer imaging and immunotherapy by
virtue of its abundant and restricted ex-
pression on the surface of prostate carcino-
mas, and on the neovasculature of most
other solid tumors. PSMA was originally
discovered in the androgen-dependent
LNCaP human prostatic adenocarcinoma
cell line [47]. ProstaScint, approved by the
FDA in 1996, is a murine monoclonal
antibody imaging agent linked to indium-
111 directed against PSMA [48,49] used
for the detection, staging and follow-up of
prostate adenocarcinoma.
HER2/neu oncogene This marker belongs
to a family of human epidermal growth
factor receptors (EGFRs) involved in the
transmission of signals controlling normal
cell growth and differentiation [50, 51].
HER2/neu is known to be overexpressed
in many different types of human cancers,
including breast, ovarian, lung, gastric,
and oral cancers [52] (see also Part I,
Chapter 5). The presence on their surface
of high amounts of HER2 enhances the
responsiveness to growth factors and ma-
lignant growth of breast tumor cells. The
humanized monoclonal antibody Hercep-
tin, specific for the protein product of
HER2 (p185HER2) [53, 54], was approved
by the FDA in 1998 and in Europe in 2000
for the treatment of metastatic breast can-
cer. Binding of Herceptin to the extracellu-
lar domain of the receptor results in
down-regulation of HER2 by inducing re-
ceptor internalization, inhibition of cell-cy-
cle progression and antibody-dependent
cellular cytotoxicity by inducing an im-
mune response [55]. Moreover, Herceptin
is able to block cleavage of HER2, which
would generate a membrane-bound trun-
cated receptor that is constitutively active.
Two other HER2-specific monoclonal anti-
bodies are currently in clinical trials,
namely 2C4 the activity of which (unlike
that of Herceptin) is not dependent on
HER2 amplification (Genentech) [56]; and
Osidem, a bispecific antibody that was de-
veloped to target cytotoxic effector cells ex-
pressing Fc gamma receptor type I (Fc
gammaRI, CD64) to HER2/neu-overex-
pressing tumor cells (Medarex) [57].
CD20 This is a signature B-cell differen-
tiation antigen. The function of CD20 is
unknown, although it is thought to be in-
volved in B-cell activation, regulation of
cellular growth, and transmembrane cal-
cium flux [58]. There are two main classes
of antibodies directed against the CD20
antigen that have been developed for ther-
apeutic intent: unconjugated and radiola-
beled antibodies. Rituxan, an unconjugated
chimeric monoclonal antibody, was the
first monoclonal antibody approved by the
FDA in 1997 for the therapy of cancer,
more precisely for the therapy of non-
Hodgkin lymphoma (NHL). Zevalin and
Bexxar are radiolabeled murine monoclo-
nal antibodies that were approved for the
therapy of NHL in 2002 and 2003, respec-
tively [59, 60].
EpCAM This is a 40-kDa epithelial trans-
membrane glycoprotein expressed on the
basolateral surface of simple, pseudostrati-
fied, and transitional epithelia. EpCam
mediates epithelium-specific, Ca
2+
-inde-
pendent homotypic cellcell adhesions. In
vivo expression of EpCam is related to in-
creased proliferation of epithelial cells, and
correlates negatively with cell differentia-
tion [61]. EpCam was found to be strongly
expressed in carcinomas of various origins,
including colon and rectum [62], prostate
[63], liver [64], esophagus , lung, head and
neck, pancreas, and breast [61]. Chimeric
and humanized antibodies have been gen-
erated, such as the chimeric antibody 17-
1A (edrecolomab; Panorex, Glaxo Well-
come GmbH). Edrecolomab immunother-
apy decreased the frequency of distant me-
tastasis in patients with colorectal cancer
and eliminated disseminated breast cancer
tumor cells in the bone marrow [65]. A re-
cently published study [66] could not show
a significant effect of edrecolomab in stage
III colon cancer therapy.
The instability and plasticity of tumor
genomes represents a major drawback of
the targeting approaches based on ligands
(e.g., monoclonal antibodies) specific to
antigens on the tumor cell membrane.
Events such as partial or complete deletion
of chromosomes, amplification of genes,
translocations or rearrangements of chro-
mosomes, and simple mutations ensure
efficient selection and overgrowth of drug-
resistant tumor cell during and after thera-
py. Furthermore, the accessibility of mark-
ers on tumor cells is not optimal for
agents coming from the blood stream, as a
high interstitial pressure and irregular tu-
mor vasculature may hinder the antibody
extravasation and tissue penetration. It is
likely that the absolute amount of tumor-
associated antigen in the neoplastic lesion
influences the performance of ligand-
based tumor targeting approaches.
6.3.2
Angiogenesis-related Diseases
6.3.2.1 Angiogenesis and Tumor
Angiogenesis
Angiogenesis is the process through which
new blood vessels form from pre-existing
ones. It occurs primarily during embryo-
genesis as an essential process for the de-
velopment of the vascular network of ar-
teries, veins, arterioles, venules and capil-
lary blood vessels that nourish and protect
the bodys tissues [67]. Once the vascular
network is in place in the adult, the endo-
thelial cells (ECs) lining the blood vessels
are quiescent, and angiogenesis is nor-
mally triggered only locally and transiently
during some physiological processes such
as the female reproductive cycle, hair
growth, wound healing and inflammation
[67]. Angiogenesis is a tightly controlled,
multistep process in which pro-angiogenic
and anti-angiogenic factors are in equilib-
rium to neutralize one another. Imbalance
of this equilibrium, either by the up-regu-
lation of pro-angiogenic or down-regula-
tion of anti-angiogenic mediators, induces
angiogenesis. Angiogenesis is an impor-
tant feature of a range of different patho-
logical conditions, cancer being one of the
most prominent examples [68]. The
growth of new capillaries is often triggered
in conditions of cellular proliferation, isch-
emia or chronic inflammation, where an
increase in blood supply may compensate
for hypoxia and insufficient delivery of nu-
trient to the tissue [69, 70]. Unlike the sit-
uation in physiological conditions, blood
vessels grow unabated in cancer and other
pathologies, and tumor angiogenesis sus-
tains the progression of the disease.
During angiogenesis, endothelial cells
detach from the pre-existing destabilized
vessel, migrate into the perivascular space,
and proliferate to finally mature and form
new vascular structures. A number of
growth factors, proteases, adhesion mole-
cules and other angiogenic mediators
which enable endothelial cell migration or
proliferation regulate this process. Vascular
endothelial growth factor (VEGF) is con-
sidered to be one of the most important
growth factors in angiogenesis [71]. It in-
creases the permeability of existing blood
vessels and acts as endothelial cell survival
factor, as well as being a potent endothelial
cell mitogen. The neutralizing humanized
monoclonal anti-VEGF antibody Avastin
has recently been approved for the treat-
ment of colorectal cancer [72], but showed
no survival benefit in patients with breast
cancer [73].
Most of the current knowledge about an-
giogenesis stems from investigations on
tumoral angiogenesis. A large number of
molecules involved in angiogenesis have
been first identified in tumors, and later
confirmed in other pathological conditions.
Many tumors in humans persist in situ
without being accompanied by angiogen-
esis [74, 75]. At that stage they tend to be
clinically undetectable and are rarely larger
than 12 mm in diameter, because diffu-
sion of oxygen and nutrients limit their
size. The high rate of proliferation in these
tumors is compensated by abundant inter-
nal apoptosis as a consequence of insuffi-
cient blood supply.
As the tumor adopts an angiogenic phe-
notype, the balance between pro- and anti-
angiogenic factors is upset and angiogen-
esis is triggered. The tumor mass is al-
lowed to overtake the apoptotic rate, and
consequently expands. This process is re-
ferred to as angiogenic switch [68, 76].
Not only is angiogenesis required for tu-
mors to grow beyond a certain size, but it
also enables tumor cells to migrate into
surrounding tissue and to colonize distant
sites, forming metastases. Metastases
again can only grow to threatening size if
the metastatic cells are able to trigger an-
giogenesis [68].
Although the mechanisms eliciting the
angiogenic switch are not entirely under-
stood to date, it is believed that besides tu-
mor-suppressor mutation and oncogene
activation, hypoxia plays a pivotal role [77].
There are at least two hypoxia-dependent
regulatory mechanisms which lead to
VEGF expression. The first mechanism re-
lies on the transcription factor hypoxia-in-
ducible factor (HIF-1) which controls
VEGF transcription [78]. The alpha sub-
unit of HIF-1, HIF-1a, is degraded under
normoxic conditions and stabilized under
hypoxia [69, 7981]. Second, VEGF mRNA
becomes stabilized under hypoxic condi-
tions [82]. VEGF concentrations stimulate
the proliferation of endothelial cells, which
in turn produce many unspecific angio-
genic stimulators, including basic fibro-
blast growth factor (bFGF), acid fibroblast
growth factor (aFGF), transforming growth
factor a and b (TGFa and TGFb) or plate-
let-derived endothelial cell growth factor
(PD-ECGF). Additionally, tumor cells pro-
duce proteases, among which are matrix
metalloproteinases (MMP) and serine pro-
teases like urokinase-plasminogen activator
(uPA) or tissue plasminogen activator
(tPA). Endothelial cells display cell adhe-
sion molecules such as integrins a
v
b
3
and
a
v
b
5
which mediate interaction with the
ECM. Laminin, type IV collagen and tenas-
cin are synthesized to constitute the new
basement membrane.
Reduced oxygen tension promotes an-
giogenesis not only by stimulating the pro-
duction of inducers, but also by reducing
the production of inhibitors.
Thrombospondin-1 was the first angio-
static protein for which anoxia-triggered
down-regulation during tumorigenesis was
demonstrated [83]. A number of endoge-
nous angiogenesis inhibitors have since
been identified.
The tumor vessels may be distinguished
from their normal counterparts: architectu-
rally, they are irregularly shaped, dilated,
tortuous and even contain dead ends [84].
Extensive fenestration, an abnormal base-
ment membrane and unusual wide gaps
between adjacent endothelial cells make
them leaky [8587].
The treatment of cancer with an anti-an-
giogenic approach was first proposed more
than two decades ago [74]. Accordingly,
various anti-angiogenic strategies have
been investigated preclinically. This exten-
sive research has culminated in the recent
approval of Bevacizumab (Avastin, Genen-
tech) as first-line treatment of metastatic
colon carcinoma [72, 88]. Furthermore,
several lines of evidence suggest that (at
least in part) the action of chemotherapeu-
tic agents against solid tumors may be re-
lated to the preferential killing of the tu-
mor endothelium, rather than the endo-
thelium of normal tissues [89].
6.3.2.2 Non-tumor Angiogenesis
A number of non-cancer disorders are
strongly associated with the overexuberant
proliferation of new blood vessels, and
may benefit from anti-angiogenesis treat-
ments. It is generally accepted that several
potentially blinding ocular disorders (e.g.,
the exsudative form of age-related macular
degeneration (ARMD), diabetic retinopa-
thy, retinopathy of prematurity, rubeosis ir-
idis, etc.) and chronic inflammatory condi-
tions (e.g., rheumatoid arthritis, psoriasis)
fall into the category of angiogenesis-re-
lated diseases [68, 70]. Consequently, the
identification of markers for angiogenesis,
and the validation of high-affinity ligands
to such markers, is expected to lead to
benefits both for the diagnosis and the
therapy of these diseases. In a number of
cases, anti-angiogenic treatments based on
vascular targeting approaches (e.g., the use
of laser irradiation and Visudine in
ARMD) are already widely diffused in the
clinical practice, even if their targeting
component is limited. We expect to see
more targeting approaches in the near fu-
ture, for the treatment of non-cancer an-
giogenesis-related diseases.
6.3.2.3 Markers of Angiogenesis
and Stromal Antigens
Several antigens have been proposed as
putative markers of angiogenesis, but only
a few have been extensively characterized
by immunohistochemistry, by in vivo bio-
distribution analysis and by scintigraphic
procedures in patients with cancer or other
diseases.
The antigens which have been character-
ized more extensively are possibly the
EDB domain of fibronectin [90, 91], the
large isoforms of tenascin-C [92, 93],
PSMA and the a
v
b
3
integrin. Indeed, for
all these antigens, extensive immunohisto-
chemical studies have been reported in the
literature, and monoclonal antibodies are
currently undergoing clinical trials. We
have described these antigens in detail in
other reviews, and refer the interested
readers to those articles [9496]. Other
markers of angiogenesis, which have dis-
played promising results but which are at
later stages of development, include Endo-
glin (CD105) [97, 98], VEGF and VEGF-re-
ceptor complex [99, 100], CD44 [101],
phosphatidyl serine phospholipids [102],
magic roundabout (ROBO-4) [103], Amino-
peptidase N [104] and Annexin A1 [100].
It may be worthwhile mentioning that
recent technological breakthroughs may fa-
cilitate the discovery of markers of angio-
genesis. For example, the group of Kinzler
and Vogelstein have reported a transcrip-
tomic analysis of endothelial cells purified
from colorectal cancer and from normal
tissues [105]. Recent experimental
approaches based on terminal perfusion of
tumor-bearing rodents have allowed, for
the first time, a direct proteomic analysis
of accessible antigens in vascular struc-
tures. The group of Schnitzer has reported
the use of terminal perfusion protocols
with silica beads for the identification of
tumor endothelial markers. In their work,
Annexin A1 emerged as a promising anti-
gen for the radiolabeled antibody-based
imaging and therapy of cancer [100]. In
our laboratory, we use terminal perfusion
protocols featuring active esters of biotin
for the selective chemical labeling of acces-
sible proteins in vascular structures. Bioti-
nylated proteins are then purified from dif-
ferent organs (collected separately) and are
submitted to a comparative proteomic
analysis [106].
6.3.3
Cardiovascular Diseases
Cardiovascular diseases are currently the
leading cause of death and illness in devel-
oped countries.
6.3.3.1 Atherosclerosis
Complications of atherosclerosis are the
leading cause of morbidity and mortality
in developed countries. It can be consid-
ered as a chronic inflammation resulting
from interaction between modified lipo-
proteins, monocyte-derived macrophages,
T cells, and the normal cellular elements
of the arterial wall [107109]. The earliest
lesion is a pure inflammatory lesion con-
sisting of monocyte-derived macrophages
and T cells. The presence of monocytes in
every phase of atherosclerosis, and of hy-
drolytic enzymes secreted by these and
other cells, play a central role in different
stages of the disease, particularly in the re-
sorption of the fibrous cap leading to
plaque rupture.
None of the current imaging techniques
(typically monitoring luminal diameter,
volume, and thickness of the plaque, etc.)
is capable of characterizing biological
plaque activity to identify high-risk pa-
tients. Therefore, a considerable research
effort concentrated in the development of
techniques that allows the high-resolution
detection of high-risk (or active) athero-
sclerotic lesions.
Imaging of protease activity Proteolytic en-
zymes are produced at sites of atherosclerot-
ic lesions by biologically active macro-
phages and endothelial cells. These en-
zymes, such as cathepsin B [110, 111] and
MMPs [108, 112], seem to be involved in
the degradation of the fibrous cap that can
lead to the rupture of the atherosclerotic
plaque. Recently, Chen et al. [113] described
the imaging of cathepsin B activity in vivo
by fluorescence-mediated tomography
(FMT). This was achieved by using an auto-
quenched cathepsin B sensitive near-infra-
red fluorescence (NIRF) probe [114] that
can generate a strong NIRF signal after en-
zyme activation (e.g., by cathepsin B activ-
ity). These activatable probes consist of
NIR fluorochromes linked to a delivery ve-
hicle via specific peptide sequences that
serve as a substrate for the protease of inter-
est [115]. The authors hypothesized that
imaging of cathepsin B activity in athero-
sclerotic plaques may serve as a new meth-
od to measure plaque inflammation and
vulnerability. Since MMP-2 has been sug-
gested to be a specific mediator of fibrous
cap destabilization [112], studies are cur-
rently being performed to image MMP-2 ac-
tivity in vivo in an atherosclerosis animal
model [109]. To this purpose, the MMP-2
peptide substrate was used to create an
autoquenced NIRF probe. The feasibility
of such an approach was proven by Bremer
et al. in 2001, who demonstrated the in vivo
imaging of MMP-2 expression using a
MMP-2 sensitive NIRF probe [116].
Imaging of activated macrophages Given
their ubiquitous presence in every stage of
the atherosclerosis disease, activated
macrophages are being recognized as an
important target for atherosclerosis treat-
ment and imaging. Detection of activated
macrophages in atherosclerotic lesions was
achieved by high-resolution MRI using
superparamagnetic iron oxide nanoparti-
cles [117]. Macrophages appear to phagocy-
tose nanoparticles, and the resulting iron
oxide accumulation generates strong T2
relaxation and MRI contrast.
Imaging of activated endothelial cells En-
dothelial cell adhesion molecules (ECAM)
are expressed at high levels on the plaque
surface, angiogenic vessels within the
plaque, and adventitial vessels, with low
expression levels in normal vessels [118].
ICAM-1, VCAM-1, P-selectin and integ-
rins, such as a
v
b
3
, have been associated
with advanced atherosclerosis. Several ul-
trasound contrast agents targeted to
ECAMs were developed in the recent past.
Acoustically active liposomes, conjugated
with monoclonal antibodies specific to
ICAM-1, were shown preferentially to ac-
cumulate in the endothelium overlying
atherosclerotic lesions [119]. Lipid micro-
bubbles conjugated to antibodies against
P-selectin could be used to image early in-
flammatory responses [120]. Angiogenesis-
targeted microbubbles were created by
conjugating antibodies or peptides binding
to a
v
-integrins [121]. Recently, Joseph et al.
described the creation of an air-filled mi-
croparticle conjugated to a L19 antibody
derivative, specific to the EDB domain of
fibronectin [122]. These microparticles are
aimed at the ultrasound in vivo imaging of
angiogenesis, but as yet no in vivo results
have been reported.
6.3.3.2 Thrombosis
Thrombosis that is, the formation of a
solid mass of blood products in a vessel
is the pathological hallmark of a number
of cardiovascular diseases (myocardial in-
farction, stroke, etc.). The imaging of mol-
ecules important for thrombogenesis could
provide a highly specific diagnostic throm-
bosis imaging method.
Imaging of platelet deposits Activated
platelets are usually found on the surface
of the thrombus. Ligands binding to recep-
tors found exclusively on the surface of acti-
vated platelets represent a highly specific
approach to detect platelet deposits. The
a
IIb
b
3
integrin on platelets is the most com-
monly targeted receptor for the detection of
platelet deposits. Several different peptide
ligands to a
IIb
b
3
have been developed and
tested for the imaging of thrombosis [123].
A linear peptide based on the amino acid se-
quence of the binding domain of a mono-
clonal antibody directed against a
IIb
b
3
,
PAC-1 [124] was synthesized and used as a
base to develop peptides for imaging throm-
bi [125]. A second possible approach to im-
age thrombi by means of peptides is to use
cyclic peptides based on the simplest known
integrin binding sequence, RGD. The cyclic
peptide P280 (Apcitide) [126] is an approved
thrombus-imaging radiopharmaceutical.
The third class of thrombus-imaging pep-
tides are natural polypeptides (disintegrins),
with high affinity for the receptor. Disinte-
grins are usually composed of 4884 amino
acids, and are rich in cysteine residues. The
formation of disulfide bridges confer to the
polypeptides a defined structure where the
RGD motif is exposed at the tip of a flexible
loop. Radioactively labeled Bitistatin [127]
produced images of intense uptake at the
thrombus site which corresponded to the
true dimension of the lesion.
Imaging of enzyme activities Thrombin, a
serine protease, plays an important role in
thrombogenesis, cleaving fibrinogen to
form fibrin monomers, which subse-
quently polymerize to form fibrin, the scaf-
folding of thrombus [115] (see also Part II,
Chapter 1). Jaffer et al. synthesized a NIRF
probe that consisted of a human throm-
bin-cleavable peptide that contained an N-
terminal NIR fluorochrome. The probe
successfully detected thrombi in animal
models [115]. Activated factor XIII is a tis-
sue transglutaminase that cross-links fi-
brin chains and plasmin inhibitors to form
mechanically and proteolytically stable
thrombi [109]. Factor XIII activity in
thrombi has been successfully imaged by
both NIRF [115, 128] and MRI [129]. For
MRI, the F13-CLIO agent consisted of a
dextran-coated caged iron-oxide particle
(CLIO) conjugated to an a2-antiplasmin
peptide that can be cross-linked by factor
XIII [129].
Imaging of fibrin As mentioned in the
previous section, fibrin represents the scaf-
folding of the thrombi. It is a favorable
thrombosis molecular imaging target, be-
cause it is usually present in all types of
thrombi and its plasma concentration is
low, thus minimizing the background sig-
nal [109]. Several fibrin-targeted molecular
imaging agents have been developed for
nuclear imaging [123], ultrasound imaging
[119, 130] and for high-resolution MRI
[131, 132]. Flacke et al. developed a novel
fibrin-specific MR contrast agent consist-
ing of a lipid-encapsulated liquid per-
fluorocarbon nanoparticle coated with an
anti-fibrin Fab antibody fragment, which
can carry high gadolinium-DTPA (Gd-
DTPA) payloads for high detection sensi-
tivity [131]. The authors demonstrated the
selective accumulation of the nanoparticles
in microthrombi overlying the atheroscle-
rotic intima. Moreover, a radiolabeled
12 kDa fragment of fibronectin (Fibrin-
binding domain) was shown by Taillefer
selectively to bind to thrombi in vivo [133].
6.3.3.3 Heart Failure
Although heart failure is one of the most
common human diseases in developed
countries, a limited understanding of the
underlying mechanisms leads to a lack of
effective treatments. The origins of the dis-
ease can be very diverse, ranging from hy-
pertension to viral infections to coronary
occlusion, but end-stage heart failure
shares many common pathologic features,
such as loss in myocyte viability, intersti-
tial remodeling, changes in gene expres-
sion and contractile dysfunction [134].
Imaging of myocardial apoptosis As men-
tioned above, cardiomyocyte apoptosis is a
pathologic feature of heart failure. A char-
acteristic feature of apoptosis is the exter-
nalization of phosphatidylserine phospholi-
pids [135] a lipid that, under normal con-
ditions, is present in only the inner layer
of the cell membrane. Radiolabeled annex-
in V, an intracellular phospholipid-binding
protein, was used clinically in patients
with acute myocardial infarction for the
non-invasive imaging of apoptosis [136].
Moreover, the use of NIR fluorochrome-
tagged annexin V has also been described
[137].
6.3.4
Inflammation
6.3.4.1 Rheumatoid Arthritis
Rheumatoid arthritis is an autoimmune
disease that affects multiple synovial joints
and involves inflammation of the synovial
membrane, often resulting in a loss of
function due to the erosion of bone and
cartilage [138].
Targeting activated macrophages Activated
macrophages are known to constitute the
key effector cells in rheumatoid arthritis
[139]. A clear correlation between the levels
of macrophage activity, joint inflammation,
articular pain and bone erosion was de-
fined. This correlation can be explained by
the fact that activated macrophages secrete
potent mediators of inflammation and tis-
sue destruction. Moreover, macrophages
participate in antigen presentation and
therefore contribute to the activation and
proliferation of antigen-specific T cells
[140]. One possibility of reducing the de-
structive effects of rheumatoid arthritis
might be the elimination of the cell popula-
tion that is mainly responsible for the in-
flammation that is, the activated macro-
phages. The folate receptor b (FR-b), a gly-
cosylphosphatidylinositol-anchored protein
that binds folic acid with high affinity, was
shown to be expressed on monocytic and
myelocytic lineages of hematopoietic cells
in a functional inactive form unable to bind
folic acid [141]. Interestingly, it was recently
shown that activated synovial macrophages
possess a functionally active FR-b [142].
EC20, a folate-conjugated radiopharmaceu-
tical complex with
99m
Tc was shown to accu-
mulate in arthritic extremities of diseased
rat, but not in the extremities of healthy an-
imals [143]. Fletcher et al. showed that a fo-
late-targeted immunotherapy reduced the
symptoms of rheumatoid arthritis in a sim-
ilar way as methotrexate [144]. Animals that
were previously immunized with haptens
were injected with folatehapten conju-
gates; in this way, the hapten-decorated ac-
tive macrophages were killed by the mecha-
nism of antibody-dependent cell cytotoxici-
ty. A similar approach the elimination of
active macrophage was undertaken by
van Roon et al., who demonstrated the in vi-
tro selective killing of activated macro-
phages, isolated from rheumatoid arthritis
patients, by means of an Fcb Receptor I-di-
rected immunotoxin [145].
It is worth mentioning here that Wun-
der et al. [146] were able to image inflam-
mation in arthritic joints by means of a ca-
thepsin B-specific autoquenced NIRF
probe. Cathepsin B activated NIRF probes
were shown to serve as reporters for the
imaging of treatment response to anti-
rheumatic drugs (e.g., methotrexate).
6.3.4.2 Other Inflammatory Diseases
The targeting of molecular markers asso-
ciated with inflammation could, in princi-
ple, be useful for the imaging and therapy
of several other inflammatory diseases. Be-
sides rheumatoid arthritis, these include
inflammatory bowel diseases (Crohns dis-
ease and ulcerative colitis), psoriasis,
atherosclerosis, and diseases of the central
nervous system (Alzheimers disease, mul-
tiple sclerosis, etc.).
The treatment of diseases such as psori-
asis and inflammatory bowel could in par-
ticular take advantage of these targeting
strategies, as angiogenesis appears to play
an important role in the pathology of these
conditions [70].
Psoriasis, a chronic inflammatory skin
disease that affects approximately 13% of
the western population [109], is character-
ized by hyperproliferation of keratinocytes,
infiltration of inflammatory cells, and in-
creased cytokine levels. Psoriasis is accom-
panied by an expansion of the superficial
dermal microvasculature and elongation of
capillary loops passing into dermal papil-
lae and the papillary tip [147].
Inflammatory bowel diseases (IBD) are
chronic inflammatory conditions which af-
fect the gastrointestinal tract. They are char-
acterized by a localized or diffuse granulo-
matous inflammatory process, accompa-
nied by systemic manifestations. As ulcera-
tion and regeneration of the intestinal
epithelium occurs during the course of the
disease, angiogenesis is undoubtedly an in-
tegral part of the IBD pathology [148].
It is likely that the ligand-based delivery
of anti-inflammatory drugs or anti-inflam-
matory cytokines will improve the efficacy
of therapies of psoriasis and inflammatory
bowel diseases.
Today, several animal models for psoria-
sis and inflammatory bowel diseases are
available [149151], and these should allow
investigations to be made of the benefits
of targeted therapy approaches in these
diseases.
6.3.5
Infection
Infections, which result from the invasion
of microorganisms, are usually diagnosed
on the basis of clinical history, physical ex-
amination, laboratory tests and the identi-
fication of pathogens in body fluids and
biopsy samples. The discrimination be-
tween infection and inflammation at an
early stage of the disease is considered to
be critical for a favorable outcome. Nuclear
medicine could contribute to the non-inva-
sive detection of infections, provided that
specific tracers are available which can dis-
criminate between infections and sterile
inflammations. To this purpose, two differ-
ent groups of tracers are currently under
development. The first group consists of
radiolabeled antibiotics and antifungal
compounds, while the second group con-
sists of radiolabeled peptides derived from
antimicrobial peptides/proteins.
Vinjamuri et al. showed that radiola-
beled ciprofloxacin (Infecton), a fluoroqui-
nolone antimicrobial agent that binds to
the DNA gyrase in all dividing bacteria,
could successfully discriminate between
bacterial infections and sterile inflamma-
tions [152]. A further example of this class
of tracers is
99m
Tc-labeled fluconazole. Lu-
petti et al. showed that
99m
Tc-fluconazole
is a very good marker for Candida albicans;
in fact, this tracer can detect C. albicans in-
fections but not bacterial infections in ani-
mal models [153]. The accumulation of
99m
Tc-fluconazole also correlated with the
number of viable C. albicans microorgan-
isms present at the infection site, making
this agent suitable for monitoring therapy.
Antimicrobial peptides belong to the sec-
ond group of tracers that are also under
development. Antimicrobial peptides play
a critical role in the defense system of
multicellular organisms against bacteria,
fungi, and viruses. They are produced by
macrophages, epithelial and endothelial
cells of all organisms, and their mecha-
nism of action is based on the interaction
between cationic residues of the peptides
and the negatively charged bacterial sur-
face. This feature determines the specifici-
ty of the antimicrobial peptides to bacteria,
since in mammalian cells negatively
charged lipids face the cytoplasm. Exam-
ples of this class of tracers are the ubiqui-
cidin-derived peptides (UBI-peptides). Ubi-
quicidin is a 6.7-kDa linear peptide, which
is a natural mammalian antimicrobial
agent [154]. Different
99m
Tc-labeled UBI-
derived peptides were tested for their abil-
ity to selectively detect infections in vivo
[155].
99m
Tc-labeled UBI 29-41 was able to
detect bacterial infections in mice, with
uptake at the infection site correlating with
bacterial density. Interestingly,
99m
Tc-la-
beled UBI 29-41 was shown not to elicit
any immune response in tested animals;
this was in contrast to defensin-derived
peptides (another class of antimicrobial
peptides used for the imaging of infec-
tions), which have been reported to induce
a potent immune response [156].
6.3.6
Amyloidosis
Amyloidosis is a condition which is char-
acterized by the extracellular deposition of
abnormal fibrillar proteins (amyloid), in
amounts sufficient to impair correct organ
function. Typical examples include Alzhei-
mers disease and prion disease.
6.3.6.1 Alzheimers Disease
Alzheimers disease (AD) is a complex
neurodegenerative dementing illness. Ex-
tracellular amyloid plaques consisting pre-
dominantly of the amyloid-b-42 peptide
(APb42), a proteolytic derivative of the
large transmembrane protein amyloid pre-
cursor protein (APP), are one of the patho-
logical hallmarks of Alzheimers disease
(see also Part VIII, Chapter 4). Progress in
an understanding of the mechanisms lead-
ing to the formation of extracellular amy-
loid plaques led to the development of
new classes of drugs for the therapy of
AD. b-Secretase is one of the proteases re-
sponsible for the proteolytic processing of
APP that leads to the formation of APb42,
a peptide that is prone to aggregation. Sev-
eral b-secretase inhibitors are currently
being developed by companies such as Pfi-
zer and Elan. In this section, we will focus
on the emerging studies on imaging of
amyloid plaques. Most efforts at in vivo
neuroimaging of amyloid plaques have
concentrated on developing radioactive li-
gands that can be detected by positron
emission tomography (PET) or single-
photon emission computed tomography
(SPECT). In order to be used as an amy-
loid plaque-imaging agent, molecules
must be able to cross the bloodbrain bar-
rier (BBB), and should therefore have a
molecular weight of 400600 kDa and
form as few hydrogen bonds with water as
possible (see also Part VIII, Chapter 4).
Congo Red (CR) and Thioflavin T are dyes
that bind to amyloid plaques in brain sec-
tions but do not cross the BBB [157]. Sev-
eral radiolabeled CR derivatives, such as
Chrysamine G [158] and X-34 [159], have
been created, but none of these gave satis-
factory results in vivo, mainly due to their
poor brain uptake. IBOX, a Thioflavin T
derivative, was shown to cross the BBB,
but no published in vivo studies are avail-
able [160]. Several peptide-based com-
pounds have also been designed. 10H3 is
a monoclonal antibody that binds specifi-
cally to Ab1-28 [161], and radioactive la-
beled fragments of this antibody have been
used for SPECT analysis. These studies
failed due to unspecific binding and the
inability of the peptide to cross the BBB.
Further examples of peptide-based imag-
ing agents are the serum amyloid protein
(SAP) and b-amyloid peptide. SAP is a
non-fibrillar glycoprotein, produced in the
liver, which was shown to be effective in
imaging systemic amyloidosis [162].
Although SAP is known to be able to cross
the BBB, in vivo studies have not shown
any difference between healthy and AD pa-
tients [163, 164]. As mentioned above, b-
amyloid peptide can also be used as a
plaque-imaging agent. Radiolabeled b-amy-
loid peptide 1-40 (Ab1-40) was shown to
bind with high affinity to amyloid plaques
in brain sections, though in vivo studies
have demonstrated only a limited brain
uptake of Ab1-40 [165].
To solve this problem, in vivo studies
have been performed in rats using the
Ab1-40 conjugated with a monoclonal anti-
body specific to the rat transferrin recep-
tor. The transferrin receptor allows recep-
tor-mediated transcytosis through the rats
BBB. Conjugation of the peptide with this
antibody increased its brain uptake and de-
creased its plasma clearance [166]. To our
knowledge, in vivo human studies have
not yet been reported.
Highly lipophilic fluorescent molecules
represent the last class of compounds that
will be discussed in this section. The com-
pound
18
FDDNP [167, 168], which binds
with high affinity to Ab1-40 fibrils in vivo
was shown readily to cross the BBB due to
its high lipophilicity. Human PET studies
were performed, and
18
FDDNP was shown
to localize in regions known to develop
plaques in Alzheimers disease [167].
6.3.6.2 Prion Diseases
Prion diseases in human and animals are
fatal neurodegenerative diseases. The pro-
tein-only hypothesis proposes that the
prion is a conformational isoform of the
normal host prion protein PrP
C
, which is
found predominantly on the outer surface
of neurons [169]. The abnormal PrP
Sc
iso-
form, which is protease-resistant, is
thought to cause the conversion of PrP
C
into a likeness of itself. PrP
Sc
has the ten-
dency to aggregate and to accumulate
mainly in the brain. The non-invasive di-
agnosis of prion diseases in humans is
challenging due to the lack of specific and
sensitive probes. However, during the past
few years several PET ligands have been
developed to image amyloid plaques in
Alzheimers disease (see Section 6.3.6.1).
A CR derivative, methoxy-X04 [170], was
tested for the ability of binding to prion
deposits in 87V-infected mice. This com-
pound was shown to localize at sites of
prion deposition in both symptomatic and
pre-symptomatic animals. The results of
these studies suggest that similar com-
pounds could be developed into useful
PET imaging agents to improve the diag-
nosis of prion diseases in humans.
6.4
From a Ligand to a Product
6.4.1
Imaging
The identification of a pathology-associated
antigen and the isolation of a specific
binder (antibody, peptide, small organic
molecule) does not automatically yield a
novel imaging agent. Indeed, a number of
conditions must be fulfilled in order to ob-
tain a successful imaging result. In part,
these conditions depend on the experi-
mental modality chosen for the in vivo
imaging application.
A general approach for converting a li-
gand into an imaging agent consists of
radioactive labeling with a suitable radio-
nuclide. Typically, short-lived gamma emit-
ters must be used for gamma-camera-
based imaging techniques (e.g.,
99m
Tc,
123
I,
111
In) in order to minimize exposure
of the patient to radiation.
Within the field of nuclear medicine,
most developments are expected to origi-
nate in the field of PET. This can be
thought of as a camera which can take pic-
tures of a subject of interest, and requires
an exposure time of a few seconds to sev-
eral minutes. The camera does not image
visible light, but rather high-energy gam-
ma-rays that are emitted from inside the
subject [171]. Natural biological molecules
can be labeled with an isotope which pro-
duces two gamma rays by emitting a posi-
tron from its nucleus. Frequently used po-
sitron-emitting isotopes include
15
O,
13
N,
11
C and
18
F; the latter is often used as a
substitute for hydrogen in the molecule of
interest.
124
I can conveniently be used to
radiolabel proteins. Labeled tracers are in-
troduced into the subject, after which PET
imaging is used to follow their distribution
and tissue concentrations. PET is at least
tenfold more sensitive than SPECT, and
positron-emitting isotopes can readily be
substituted for naturally occurring atoms,
producing less perturbation to the bio-
chemical behaviour of the radiolabeled par-
ent molecule. PET imaging might be used
to couple high sensitivity with the possibil-
ity of obtaining very good spatial resolu-
tion and quantitative pharmacokinetic re-
sults. Recent advances in animal micro-
PET will facilitate developments in this
field [172, 173].
Ideal targeting agents for imaging appli-
cations will show a rapid and selective ac-
cumulation at the site of disease, and a
rapid blood clearance. A number of anti-
body formats have been tested in animals
and patients [174], and suggested that a
mini-antibody format might be a good
compromise between efficient localization
at the site of disease and an acceptably
rapid blood clearance [174, 175]. The small
single-chain Fv fragments may yield better
target : blood ratios at early time points,
but their relatively low dose on the target
may compromise immunoscintigraphic de-
tection at acceptable doses of radioactivity.
A second approach to the ligand-based
molecular imaging of markers of pathol-
ogy relies on covalent modification of the
ligand with NIR fluorophores. Light pene-
tration of tissues is maximal at 800 nm,
with a 10% transmittance through 1 cm of
tissue [176]. A number of reactive ester de-
rivatives of NIR fluorophores which absorb
in this wavelength range are available
(e.g., Cy7, Alexa750) [177, 178]. While
imaging methodologies based on epi-illu-
mination are limited to the detection of
superficial lesions [179], and may be more
appropriate for endoscopic applications
[180], novel methodologies based on dif-
fuse optical tomography may allow the de-
tection of deeper lesions [181]. Compared
to very sensitive radioactive imaging tech-
niques, fluorescence-based detection meth-
ods may require the administration of a
1001000 nanomoles of fluorophore deri-
vative in order to obtain a sufficient signal
on the site of interest.
As mentioned previously, the targeted
delivery of suitable contrast agents for ul-
trasound or MRI imaging methodologies
would be desirable for many applications.
It remains to be seen whether microbub-
bles, magnetic nanoparticles or other ac-
tive compounds can cross barriers in vivo
(e.g., the endothelium) and can be deliv-
ered in sufficient amounts at sites of dis-
ease.
6.4.2
Therapy
Similar to imaging applications, most dis-
ease-targeting ligands must be modified in
order to convert them into therapeutic
agents.
In the special case of ligands capable of
internalization (e.g., in tumor cells), a
number of strategies may be considered,
including modification with short-range-
acting radionuclides, cytotoxic agents, and
toxins [182185]. However, in general,
most ligands to sites of diseases will not
be internalized. In oncology, the use of in-
tact immunoglobulins which recognize
antigens on tumor cells has led to the ap-
proval of several therapeutic antibodies in
Europe and the USA. However, until now
most of the anti-cancer antibodies ap-
proved have a signaling component which
may contribute to their potency.
Small antibody fragments can be con-
verted into therapeutic agents by modifica-
tion with suitable radionuclides, drugs
(using cleavable linkers or liposomal prep-
arations), enzymes catalyzing prodrug con-
version, cytokines, pro-coagulant factors,
photosensitizers, and other classes of
bioactive molecules (Fig. 6.3). Most of
these experimental strategies have been
tested in our laboratory using the L19 anti-
body, which is specific to the EDB domain
of fibronectin, a marker of angiogenesis
[186]. The relative advantages and disad-
vantages of the different methodologies
are still, to a large extent, a matter of spec-
ulation, and this situation will only be re-
solved by an analysis of the outcome of
clinical trials with different antibody deri-
vatives.
The fact that both pro-inflammatory cy-
tokines and anti-inflammatory cytokines
can be fused to antibody fragments sug-
gests a possible use of these fusion pro-
teins not only for augmenting an immune
response at suitable sites (e.g., against the
tumor), but also to attenuate an immune
response at sites of autoimmune diseases
and in chronically inflamed areas.
6.5
Concluding Remarks
More than a hundred years after Paul Ehr-
lichs vision of magic bullets, the targeted
delivery of ligands to sites of diseases has
been experimentally demonstrated for sev-
eral pathological conditions. In some ther-
apeutic areas (e.g., cancer), a large body of
information is available, and the first con-
clusions can be drawn about the most pro-
mising strategies and about the parame-
ters which are likely to influence targeting
performance. For other diseases, ligand-
based targeting approaches are still in
their infancy. During the past decade, tech-
nological revolutions have greatly facili-
tated the isolation of specific binding mol-
ecules and the identification of disease-as-
sociated antigens. We believe that the next
few years will continue to witness the
comparative analysis of targeting agents in
animal models of pathology, and transla-
tion of the most promising candidates
from the bench, via the clinic, into the bio-
pharmaceutical market.
References
1 Maeda, H., Fang, J., Inutsuka, T., and Kitamo-
to, Y. (2003) Vascular permeability enhance-
ment in solid tumor: various factors, mecha-
nisms involved and its implications, Int Im-
munopharmacol 3, 319328.
2 Kopecek, J., Kopeckova, P., Minko, T., Lu, Z.
R., and Peterson, C. M. (2001) Water soluble
polymers in tumor targeted delivery, J Control
Release 74, 147158.
3 Duncan, R. (1999) Polymer conjugates for tu-
mour targeting and intracytoplasmic delivery.
The EPR effect as a common gateway, Pharm
Sci Technol Today 2(11), 441449.
4 Duncan, R. (1992) Drugpolymer conjugates:
potential for improved chemotherapy, Antican-
cer Drugs 3, 175210.
5 Maeda, H. (2001) The enhanced permeability
and retention (EPR) effect in tumor vascula-
ture: the key role of tumor-selective macromo-
lecular drug targeting, Adv Enzyme Regul 41,
189207.
6 Kohler, G., and Milstein, C. (1975) Continu-
ous cultures of fused cells secreting antibody
of predefined specificity, Nature 256, 495497.
7 Steinitz, M., Klein, G., Koskimies, S., and Ma-
kel, O. (1977) EB virus-induced B lymphocyte
cell lines producing specific antibody, Nature
269, 420422.
8 Kozbor, D., and Roder, J. C. (1981) Require-
ments for the establishment of high-titered
human monoclonal antibodies against tetanus
toxoid using the Epstein-Barr virus technique,
J Immunol 127, 12751280.
9 Kozbor, D., Roder, J. C., Chang, T. H., Step-
lewski, Z., and Koprowski, H. (1982) Human
Fig. 6.3 Targeting moieties (ligands) can be converted into
imaging or therapeutic agents by modification with suitable
radionuclides, fluorophores, drugs, enzymes, cytokines or
other bioactive molecules.
anti-tetanus toxoid monoclonal antibody se-
creted by EBV-transformed human B cells
fused with murine myeloma, Hybridoma 1,
323328.
10 Karpas, A., Dremucheva, A., and Czepulkows-
ki, B. H. (2001) A human myeloma cell line
suitable for the generation of human mono-
clonal antibodies, Proc Natl Acad Sci USA 98,
17991804.
11 Jones, P. T., Dear, P. H., Foote, J., Neuberger,
M. S., and Winter, G. (1986) Replacing the
complementarity-determining regions in a hu-
man antibody with those from a mouse, Na-
ture 321, 522525.
12 McCafferty, J., Griffiths, A. D., Winter, G., and
Chiswell, D. J. (1990) Phage antibodies: fila-
mentous phage displaying antibody variable
domains, Nature 348, 552554.
13 Viti, F., Nilsson, F., Demartis, S., Huber, A.,
and Neri, D. (2000) Design and use of phage
display libraries for the selection of antibodies
and enzymes, Methods Enzymol 326, 480
505.
14 Green, L. L. (1999) Antibody engineering via
genetic engineering of the mouse: Xeno-
Mouse strains are a vehicle for the facile gen-
eration of therapeutic human monoclonal an-
tibodies, J Immunol Methods 231, 1123.
15 Walsh, G. (2003) Biopharmaceutical bench-
marks 2003, Nat Biotechnol 21, 865870.
16 Winter, G., Griffiths, A. D., Hawkins, R. E.,
and Hoogenboom, H. R. (1994) Making anti-
bodies by phage display technology, Annu Rev
Immunol 12, 433455.
17 Schier, R., Marks, J. D., Wolf, E. J., Apell, G.,
Wong, C., McCartney, J. E., Bookman, M. A.,
Huston, J. S., Houston, L. L., Weiner, L. M.,
and et al. (1995) In vitro and in vivo character-
ization of a human anti-c-erbB-2 single-chain
Fv isolated from a filamentous phage antibody
library, Immunotechnology 1, 7381.
18 Pini, A., Viti, F., Santucci, A., Carnemolla, B.,
Zardi, L., Neri, P., and Neri, D. (1998) Design
and use of a phage display library. Human an-
tibodies with subnanomolar affinity against a
marker of angiogenesis eluted from a two-di-
mensional gel, J Biol Chem 273, 2176921776.
19 Schaffitzel, C., Hanes, J., Jermutus, L., and
Pluckthun, A. (1999) Ribosome display: an in
vitro method for selection and evolution of an-
tibodies from libraries, J Immunol Methods
231, 119135.
20 Kowalski, J., Henze, M., Schuhmacher, J.,
Macke, H. R., Hofmann, M., and Haberkorn,
U. (2003) Evaluation of positron emission to-
mography imaging using [
68
Ga]-DOTA-D
Phe(1)-Tyr(3)-Octreotide in comparison to
[
111
In]-DTPAOC SPECT. First results in pa-
tients with neuroendocrine tumors, Mol Imag-
ing Biol 5, 4248.
21 Ginj, M., and Maecke, H. R. (2004) Radiome-
tallo-labeled peptides in tumor diagnosis and
therapy, Met Ions Biol Syst 42, 109142.
22 Behr, T. M., Gotthardt, M., Barth, A., and
Behe, M. (2001) Imaging tumors with pep-
tide-based radioligands, Q J Nucl Med 45,
189200.
23 Chen, X., Park, R., Hou, Y., Khankaldyyan, V.,
Gonzales-Gomez, I., Tohme, M., Bading, J. R.,
Laug, W. E., and Conti, P. S. (2004) MicroPET
imaging of brain tumor angiogenesis with
18
F-labeled PEGylated RGD peptide, Eur J
Nucl Med Mol Imaging 31, 10811089.
24 Okarvi, S. M., and al-Jammaz, I. (2003) Syn-
thesis, radiolabelling and biological character-
istics of a bombesin peptide analog as a tu-
mor imaging agent, Anticancer Res 23, 2745
2750.
25 Felici, F., Luzzago, A., Monaci, P., Nicosia, A.,
Sollazzo, M., and Traboni, C. (1995) Peptide
and protein display on the surface of filamen-
tous bacteriophage, Biotechnol Annu Rev 1,
149183.
26 Oyama, T., Sykes, K. F., Samli, K. N., Minna,
J. D., Johnston, S. A., and Brown, K. C. (2003)
Isolation of lung tumor specific peptides from
a random peptide library: generation of diag-
nostic and cell-targeting reagents, Cancer Lett
202, 219230.
27 Collins, J., Horn, N., Wadenback, J., and Szar-
denings, M. (2001) Cosmix-plexing: a novel re-
combinatorial approach for evolutionary selec-
tion from combinatorial libraries, J Biotechnol
74, 317338.
28 Pasqualini, R., and Ruoslahti, E. (1996) Organ
targeting in vivo using phage display peptide
libraries, Nature 380, 364366.
29 Trepel, M., Arap, W., and Pasqualini, R. (2002)
In vivo phage display and vascular heterogene-
ity: implications for targeted medicine, Curr
Opin Chem Biol 6, 399404.
30 Arkin, M. R., and Wells, J. A. (2004) Small-
molecule inhibitors of proteinprotein interac-
tions: progressing towards the dream, Nat Rev
Drug Discov 3, 301317.
References 1291
31 Mammen, M., C. S.-K., Whitesides G. M.
(1998) Polyvalent Interactions in Biological
Systems: Implications for Design and Use of
Multivalent Ligands and Inhibitors. Ange-
wandte Chemie International Edition, 2754
2794.
32 Shuker, S. B., Hajduk, P. J., Meadows, R. P.,
and Fesik, S. W. (1996) Discovering high-affin-
ity ligands for proteins: SAR by NMR, Science
274, 15311534.
33 Ramstrom, O., and Lehn, J. M. (2002) Drug
discovery by dynamic combinatorial libraries,
Nat Rev Drug Discov 1, 2636.
34 Erlanson, D. A., Lam, J. W., Wiesmann, C.,
Luong, T. N., Simmons, R. L., DeLano, W. L.,
Choong, I. C., Burdett, M. T., Flanagan, W. M.,
Lee, D., Gordon, E. M., and OBrien, T. (2003)
In situ assembly of enzyme inhibitors using
extended tethering, Nat Biotechnol 21, 308
314.
35 Melkko, S., Scheuermann, J., Dumelin, C. E.,
and Neri, D. (2004) Encoded self-assembling
chemical libraries, Nat Biotechnol 22, 568
574.
36 Johnson, V. G., Schlom, J., Paterson, A. J.,
Bennett, J., Magnani, J. L., and Colcher, D.
(1986) Analysis of a human tumor-associated
glycoprotein (TAG-72) identified by monoclo-
nal antibody B72.3, Cancer Res 46, 850857.
37 Nuti, M., Teramoto, Y. A., Mariani-Costantini,
R., Hand, P. H., Colcher, D., and Schlom, J.
(1982) A monoclonal antibody (B72.3) defines
patterns of distribution of a novel tumor-asso-
ciated antigen in human mammary carcinoma
cell populations, Int J Cancer 29, 539545.
38 Guadagni, F., Roselli, M., Cosimelli, M., Fer-
roni, P., Spila, A., Cavaliere, F., Arcuri, R.,
Carlini, S., Mariotti, S., Gandolfo, G. M., Cas-
ciani, C. U., Greiner, J. W., and Schlom, J.
(1996) TAG-72 expression and its role in the
biological evaluation of human colorectal can-
cer, Anticancer Res 16, 21412148.
39 Guadagni, F., Roselli, M., Cosimelli, M., Man-
nella, E., Tedesco, M., Cavaliere, F., Grassi, A.,
Abbolito, M. R., Greiner, J. W., and Schlom, J.
(1993) TAG-72 (CA 72-4 assay) as a comple-
mentary serum tumor antigen to carcinoem-
bryonic antigen in monitoring patients with
colorectal cancer, Cancer 72, 20982106.
40 Chieng, D. C., Rodriguez-Burford, C., Talley,
L. I., Sviglin, H., Stockard, C. R., Kleinberg,
M. J., Barnes, M. N., Partridge, E. E., Khazaeli,
M. B., and Grizzle, W. E. (2003) Expression of
CEA, Tag-72, and Lewis-Y antigen in primary
and metastatic lesions of ovarian carcinoma,
Hum Pathol 34, 10161021.
41 Gold, P., and Freedman, S. O. (1965) Demon-
stration of tumor-specific antigens in human
colonic carcinomata by immunological toler-
ance and absorption techniques, J Exp Med
121, 439462.
42 Duffy, M. J. (2001) Carcinoembryonic antigen
as a marker for colorectal cancer: is it clini-
cally useful?, Clin Chem 47, 624630.
43 Verhaar, M. J., Chester, K. A., Keep, P. A., Rob-
son, L., Pedley, R. B., Boden, J. A., Hawkins,
R. E., and Begent, R. H. (1995) A single chain
Fv derived from a filamentous phage library
has distinct tumor targeting advantages over
one derived from a hybridoma, Int J Cancer
61, 497501.
44 Verhaar, M. J., Keep, P. A., Hawkins, R. E.,
Robson, L., Casey, J. L., Pedley, B., Boden,
J. A., Begent, R. H., and Chester, K. A. (1996)
Technetium-99m radiolabeling using a phage-
derived single-chain Fv with a C-terminal cys-
teine, J Nucl Med 37, 868872.
45 Cooke, S. P., Pedley, R. B., Boden, R., Begent,
R. H., and Chester, K. A. (2002) In vivo tumor
delivery of a recombinant single chain Fv: tu-
mor necrosis factor-alpha fusion [correction of
factor: a fusion] protein, Bioconjug Chem 13,
715.
46 Francis, R. J., Mather, S. J., Chester, K., Shar-
ma, S. K., Bhatia, J., Pedley, R. B., Waibel, R.,
Green, A. J., and Begent, R. H. (2004) Radiola-
belling of glycosylated MFE-23: CPG2 fusion
protein (MFECP1) with
(99m)
Tc for quantita-
tion of tumour antibody-enzyme localisation
in antibody-directed enzyme pro-drug therapy
(ADEPT), Eur J Nucl Med Mol Imaging 31,
10901096.
47 Horoszewicz, J. S., Kawinski, E., and Murphy,
G. P. (1987) Monoclonal antibodies to a new
antigenic marker in epithelial prostatic cells
and serum of prostatic cancer patients, Anti-
cancer Res 7, 927935.
48 Gong, M. C., Chang, S. S., Sadelain, M., Ban-
der, N. H., and Heston, W. D. (1999) Prostate-
specific membrane antigen (PSMA)-specific
monoclonal antibodies in the treatment of
prostate and other cancers, Cancer Metastasis
Rev 18, 483490.
49 Freeman, L. M., Krynyckyi, B. R., Li, Y., Koru-
pulu, G., Saleemi, K., Haseman, M. K., and
Kahn, D. (2002) The role of
(111)
In Capromab
Pendetide (Prosta-ScintR) immunoscintigra-
phy in the management of prostate cancer,
Q J Nucl Med 46, 131137.
50 Rubin, I., and Yarden, Y. (2001) The basic
biology of HER2, Ann Oncol 12 Suppl 1, S3
8.
51 Menard, S., Pupa, S. M., Campiglio, M., and
Tagliabue, E. (2003) Biologic and therapeutic
role of HER2 in cancer, Oncogene 22, 6570
6578.
52 Hung, M. C., and Lau, Y. K. (1999) Basic
science of HER-2/neu: a review, Semin Oncol
26, 5159.
53 Carter, P., Presta, L., Gorman, C. M., Ridgway,
J. B., Henner, D., Wong, W. L., Rowland, A. M.,
Kotts, C., Carver, M. E., and Shepard, H. M.
(1992) Humanization of an anti-p185HER2
antibody for human cancer therapy, Proc Natl
Acad Sci USA 89, 42854289.
54 Park, J. W., Stagg, R., Lewis, G. D., Carter, P.,
Maneval, D., Slamon, D. J., Jaffe, H., and She-
pard, H. M. (1992) Anti-p185HER2 monoclo-
nal antibodies: biological properties and po-
tential for immunotherapy, Cancer Treat Res
61, 193211.
55 Fischer, O. M., Streit, S., Hart, S., and Ullrich,
A. (2003) Beyond Herceptin and Gleevec, Curr
Opin Chem Biol 7, 490495.
56 Mass, R. D. (2004) The HER receptor family: a
rich target for therapeutic development, Int J
Radiat Oncol Biol Phys 58, 932940.
57 Keler, T., Graziano, R. F., Mandal, A., Wallace,
P. K., Fisher, J., Guyre, P. M., Fanger, M. W.,
and Deo, Y. M. (1997) Bispecific antibody-de-
pendent cellular cytotoxicity of HER2/neu-
overexpressing tumor cells by Fc gamma re-
ceptor type I-expressing effector cells, Cancer
Res 57, 40084014.
58 Riley, J. K., and Sliwkowski, M. X. (2000)
CD20: a gene in search of a function, Semin
Oncol 27, 1724.
59 Krasner, C., and Joyce, R. M. (2001) Zevalin:
90
yttrium labeled anti-CD20 (ibritumomab
tiuxetan), a new treatment for non-Hodgkins
lymphoma, Curr Pharm Biotechnol 2, 341
349.
60 Vose, J. M. (2004) Bexxar: novel radioimmuno-
therapy for the treatment of low-grade and
transformed low-grade non-Hodgkins lympho-
ma, Oncologist 9, 160172.
61 Went, P. T., Lugli, A., Meier, S., Bundi, M.,
Mirlacher, M., Sauter, G., and Dirnhofer, S.
(2004) Frequent EpCam protein expression in
human carcinomas, Hum Pathol 35, 122128.
62 Herlyn, M., Steplewski, Z., Herlyn, D., and
Koprowski, H. (1979) Colorectal carcinoma-
specific antigen: detection by means of mono-
clonal antibodies, Proc Natl Acad Sci USA 76,
14381442.
63 Poczatek, R. B., Myers, R. B., Manne, U., Oel-
schlager, D. K., Weiss, H. L., Bostwick, D. G.,
and Grizzle, W. E. (1999) Ep-Cam levels in
prostatic adenocarcinoma and prostatic intra-
epithelial neoplasia, J Urol 162, 14621466.
64 Ruck, P., Wichert, G., Handgretinger, R., and
Kaiserling, E. (2000) Ep-CAM in malignant
liver tumours, J Pathol 191, 102103.
65 Braun, S., Hepp, F., Kentenich, C. R., Janni,
W., Pantel, K., Riethmuller, G., Willgeroth, F.,
and Sommer, H. L. (1999) Monoclonal anti-
body therapy with edrecolomab in breast can-
cer patients: monitoring of elimination of dis-
seminated cytokeratin-positive tumor cells in
bone marrow, Clin Cancer Res 5, 39994004.
66 Punt, C. J., Nagy, A., Douillard, J. Y., Figer, A.,
Skovsgaard, T., Monson, J., Barone, C., Fount-
zilas, G., Riess, H., Moylan, E., Jones, D.,
Dethling, J., Colman, J., Coward, L., and Mac-
Gregor, S. (2002) Edrecolomab alone or in
combination with fluorouracil and folinic acid
in the adjuvant treatment of stage III colon
cancer: a randomised study, Lancet 360, 671
677.
67 Bischoff, J. (1995) Approaches to studying cell
adhesion molecules in angiogenesis, Trends
Cell Biol 5, 6974.
68 Folkman, J. (1995) Angiogenesis in cancer,
vascular, rheumatoid and other disease, Nat
Med 1, 2731.
69 Pugh, C. W., and Ratcliffe, P. J. (2003) Regula-
tion of angiogenesis by hypoxia: role of the
HIF system, Nat Med 9, 677684.
70 Carmeliet, P. (2003) Angiogenesis in health
and disease, Nat Med 9, 653660.
71 Gerber, H. P., and Ferrara, N. (2003) The role
of VEGF in normal and neoplastic hematopoi-
esis, J Mol Med 81, 2031.
72 Hurwitz, H., Fehrenbacher, L., Novotny, W.,
Cartwright, T., Hainsworth, J., Heim, W., Ber-
lin, J., Baron, A., Griffing, S., Holmgren, E.,
Ferrara, N., Fyfe, G., Rogers, B., Ross, R., and
Kabbinavar, F. (2004) Bevacizumab plus irino-
tecan, fluorouracil, and leucovorin for meta-
static colorectal cancer, N Engl J Med 350,
23352342.
References 1293
73 Cobleigh, M. A., Langmuir, V. K., Sledge,
G. W., Miller, K. D., Haney, L., Novotny, W. F.,
Reimann, J. D., and Vassel, A. (2003) A phase
I/II dose-escalation trial of bevacizumab in
previously treated metastatic breast cancer,
Semin Oncol 30, 117124.
74 Folkman, J. (1971) Tumor angiogenesis: thera-
peutic implications, N Engl J Med 285, 1182
1186.
75 Folkman, J. (1972) Anti-angiogenesis: new
concept for therapy of solid tumors, Ann Surg
175, 409416.
76 Hanahan, D. (1998) A flanking attack on can-
cer, Nat Med 4, 1314.
77 Bergers, G., and Benjamin, L. E. (2003) Tu-
morigenesis and the angiogenic switch, Nat
Rev Cancer 3, 401410.
78 Forsythe, J. A., Jiang, B. H., Iyer, N. V., Agani,
F., Leung, S. W., Koos, R. D., and Semenza,
G. L. (1996) Activation of vascular endothelial
growth factor gene transcription by hypoxia-
inducible factor 1, Mol Cell Biol 16, 4604
4613.
79 Ivan, M., Haberberger, T., Gervasi, D. C., Mi-
chelson, K. S., Gunzler, V., Kondo, K., Yang,
H., Sorokina, I., Conaway, R. C., Conaway, J.
W., and Kaelin, W. G., Jr. (2002) Biochemical
purification and pharmacological inhibition of
a mammalian prolyl hydroxylase acting on
hypoxia-inducible factor, Proc Natl Acad Sci
USA 99, 1345913464.
80 Jaakkola, P., Mole, D. R., Tian, Y. M., Wilson,
M. I., Gielbert, J., Gaskell, S. J., Kriegsheim,
A., Hebestreit, H. F., Mukherji, M., Schofield,
C. J., Maxwell, P. H., Pugh, C. W., and Rat-
cliffe, P. J. (2001) Targeting of HIF-alpha to
the von Hippel-Lindau ubiquitylation complex
by O
2
-regulated prolyl hydroxylation, Science
292, 468472.
81 Ivan, M., Kondo, K., Yang, H., Kim, W., Va-
liando, J., Ohh, M., Salic, A., Asara, J. M.,
Lane, W. S., and Kaelin, W. G., Jr. (2001) HIF-
alpha targeted for VHL-mediated destruction
by proline hydroxylation: implications for O
2
sensing, Science 292, 464468.
82 Dibbens, J. A., Miller, D. L., Damert, A., Risau,
W., Vadas, M. A., and Goodall, G. J. (1999)
Hypoxic regulation of vascular endothelial
growth factor mRNA stability requires the co-
operation of multiple RNA elements, Mol Biol
Cell 10, 907919.
83 Tenan, M., Fulci, G., Albertoni, M., Diserens,
A. C., Hamou, M. F., El Atifi-Borel, M., Feige,
J. J., Pepper, M. S., and Van Meir, E. G. (2000)
Thrombospondin-1 is downregulated by anox-
ia and suppresses tumorigenicity of human
glioblastoma cells, J Exp Med 191, 17891798.
84 Konerding, M. A., Fait, E., and Gaumann, A.
(2001) 3D microvascular architecture of pre-
cancerous lesions and invasive carcinomas of
the colon, Br J Cancer 84, 13541362.
85 Roberts, W. G., and Palade, G. E. (1997) Neo-
vasculature induced by vascular endothelial
growth factor is fenestrated, Cancer Res 57,
765772.
86 Jain, R. K. (1987) Transport of molecules
across tumor vasculature, Cancer Metastasis
Rev 6, 559593.
87 Hashizume, H., Baluk, P., Morikawa, S.,
McLean, J. W., Thurston, G., Roberge, S., Jain,
R. K., and McDonald, D. M. (2000) Openings
between defective endothelial cells explain tu-
mor vessel leakiness, Am J Pathol 156, 1363
1380.
88 Ferrara, N., Hillan, K. J., Gerber, H. P., and
Novotny, W. (2004) Discovery and develop-
ment of bevacizumab, an anti-VEGF antibody
for treating cancer, Nat Rev Drug Discov 3,
391400.
89 Kerbel, R. S. (1991) Inhibition of tumor angio-
genesis as a strategy to circumvent acquired
resistance to anti-cancer therapeutic agents,
BioEssays 13, 3136.
90 Sekiguchi, K., Siri, A., Zardi, L., and Hako-
mori, S. (1985) Differences in domain struc-
ture between human fibronectins isolated
from plasma and from culture supernatants
of normal and transformed fibroblasts. Stud-
ies with domain-specific antibodies, J Biol
Chem 260, 51055114.
91 Castellani, P., Viale, G., Dorcaratto, A., Nicolo,
G., Kaczmarek, J., Querze, G., and Zardi, L.
(1994) The fibronectin isoform containing the
ED-B oncofetal domain: a marker of angiogen-
esis, Int J Cancer 59, 612618.
92 Carnemolla, B., Borsi, L., Bannikov, G., Troya-
novsky, S., and Zardi, L. (1992) Comparison
of human tenascin expression in normal, si-
mian-virus-40-transformed and tumor-derived
cell lines, Eur J Biochem 205, 561567.
93 Borsi, L., Carnemolla, B., Nicolo, G., Spina,
B., Tanara, G., and Zardi, L. (1992) Expression
of different tenascin isoforms in normal, hy-
perplastic and neoplastic human breast tis-
sues, Int J Cancer 52, 688692.
94 Brack, S. S., Dinkelborg, L. M., and Neri, D.
(2004) Molecular targeting of angiogenesis
for imaging and therapy, Eur J Nucl Med
Mol Imaging 31, 13271341.
95 Ebbinghaus, C., Scheuermann, J., Neri, D.,
and Elia, G. (2004) Diagnostic and therapeu-
tic applications of recombinant antibodies:
targeting the extra-domain B of fibronectin,
a marker of tumor angiogenesis, Curr
Pharm Des 10, 15371549.
96 Alessi, P., Ebbinghaus, C., and Neri, D.
(2004) Molecular targeting of angiogenesis,
Biochim Biophys Acta 1654, 3949.
97 Wang, J. M., Kumar, S., Pye, D., van Agtho-
ven, A. J., Krupinski, J., and Hunter, R. D.
(1993) A monoclonal antibody detects het-
erogeneity in vascular endothelium of tu-
mours and normal tissues, Int J Cancer 54,
363370.
98 Burrows, F. J., Derbyshire, E. J., Tazzari,
P. L., Amlot, P., Gazdar, A. F., King, S. W.,
Letarte, M., Vitetta, E. S., and Thorpe, P. E.
(1995) Up-regulation of endoglin on vascular
endothelial cells in human solid tumors: im-
plications for diagnosis and therapy, Clin
Cancer Res 1, 16231634.
99 Brekken, R. A., Huang, X., King, S. W., and
Thorpe, P. E. (1998) Vascular endothelial
growth factor as a marker of tumor endothe-
lium, Cancer Res 58, 19521959.
100 Oh, P., Li, Y., Yu, J., Durr, E., Krasinska,
K. M., Carver, L. A., Testa, J. E., and Schnit-
zer, J. E. (2004) Subtractive proteomic map-
ping of the endothelial surface in lung and
solid tumours for tissue-specific therapy, Na-
ture 429, 629635.
101 Taniguchi, K., Harada, N., Ohizumi, I., Tsut-
sumi, Y., Nakagawa, S., Kaiho, S., and Mayu-
mi, T. (2000) Recognition of human acti-
vated CD44 by tumor vasculature-targeted
antibody, Biochem Biophys Res Commun
269, 671675.
102 Ran, S., Downes, A., and Thorpe, P. E.
(2002) Increased exposure of anionic phos-
pholipids on the surface of tumor blood ves-
sels, Cancer Res 62, 61326140.
103 Huminiecki, L., Gorn, M., Suchting, S.,
Poulsom, R., and Bicknell, R. (2002) Magic
roundabout is a new member of the round-
about receptor family that is endothelial spe-
cific and expressed at sites of active angio-
genesis, Genomics 79, 547552.
104 Pasqualini, R., Koivunen, E., Kain, R., Lah-
denranta, J., Sakamoto, M., Stryhn, A., Ash-
mun, R. A., Shapiro, L. H., Arap, W., and
Ruoslahti, E. (2000) Aminopeptidase N is a
receptor for tumor-homing peptides and a
target for inhibiting angiogenesis, Cancer
Res 60, 722727.
105 St Croix, B., Rago, C., Velculescu, V., Traver-
so, G., Romans, K. E., Montgomery, E., Lal,
A., Riggins, G. J., Lengauer, C., Vogelstein,
B., and Kinzler, K. W. (2000) Genes ex-
pressed in human tumor endothelium,
Science 289, 11971202.
106 Rybak, J. N., Scheurer, S. B., Neri, D., and
Elia, G. (2004) Purification of biotinylated
proteins on streptavidin resin: a protocol for
quantitative elution, Proteomics 4, 2296
2299.
107 Glass, C. K., and Witztum, J. L. (2001)
Atherosclerosis. The road ahead, Cell 104,
503516.
108 Libby, P. (2002) Inflammation in athero-
sclerosis, Nature 420, 868874.
109 Jaffer, F. A., and Weissleder, R. (2004) Seeing
within: molecular imaging of the cardiovas-
cular system, Circ Res 94, 433445.
110 Reddy, V. Y., Zhang, Q. Y., and Weiss, S. J.
(1995) Pericellular mobilization of the tis-
sue-destructive cysteine proteinases, cathep-
sins B, L, and S, by human monocyte-de-
rived macrophages, Proc Natl Acad Sci USA
92, 38493853.
111 Leake, D. S., and Peters, T. J. (1981) Proteo-
lytic degradation of low density lipoproteins
by arterial smooth muscle cells: the role of
individual cathepsins, Biochim Biophys Acta
664, 108116.
112 Galis, Z. S., and Khatri, J. J. (2002) Matrix
metalloproteinases in vascular remodeling
and atherogenesis: the good, the bad, and
the ugly, Circ Res 90, 251262.
113 Chen, J., Tung, C. H., Mahmood, U., Ntzia-
christos, V., Gyurko, R., Fishman, M. C.,
Huang, P. L., and Weissleder, R. (2002) In
vivo imaging of proteolytic activity in athero-
sclerosis, Circulation 105, 27662771.
114 Weissleder, R., Tung, C. H., Mahmood, U.,
and Bogdanov, A., Jr. (1999) In vivo imaging
of tumors with protease-activated near-infra-
red fluorescent probes, Nat Biotechnol 17,
375378.
115 Jaffer, F. A., Tung, C. H., Gerszten, R. E., and
Weissleder, R. (2002) In vivo imaging of
References 1295
thrombin activity in experimental thrombi
with thrombin-sensitive near-infrared molec-
ular probe, Arterioscler Thromb Vasc Biol
22, 19291935.
116 Bremer, C., Bredow, S., Mahmood, U.,
Weissleder, R., and Tung, C. H. (2001) Opti-
cal imaging of matrix metalloproteinase-2
activity in tumors: feasibility study in a
mouse model, Radiology 221, 523529.
117 Kooi, M. E., Cappendijk, V. C., Cleutjens,
K. B., Kessels, A. G., Kitslaar, P. J., Borgers,
M., Frederik, P. M., Daemen, M. J., and van
Engelshoven, J. M. (2003) Accumulation of
ultrasmall superparamagnetic particles of
iron oxide in human atherosclerotic plaques
can be detected by in vivo magnetic reso-
nance imaging, Circulation 107, 24532458.
118 Lindner, J. R. (2002) Detection of inflamed
plaques with contrast ultrasound, Am J Car-
diol 90, 32L35L.
119 Demos, S. M., Alkan-Onyuksel, H., Kane,
B. J., Ramani, K., Nagaraj, A., Greene, R.,
Klegerman, M., and McPherson, D. D.
(1999) In vivo targeting of acoustically reflec-
tive liposomes for intravascular and trans-
vascular ultrasonic enhancement, J Am Coll
Cardiol 33, 867875.
120 Lindner, J. R., Song, J., Christiansen, J., Kli-
banov, A. L., Xu, F., and Ley, K. (2001) Ultra-
sound assessment of inflammation and re-
nal tissue injury with microbubbles targeted
to P-selectin, Circulation 104, 21072112.
121 Leong-Poi, H., Christiansen, J., Klibanov, A.
L., Kaul, S., and Lindner, J. R. (2003) Nonin-
vasive assessment of angiogenesis by ultra-
sound and microbubbles targeted to al-
pha(v)-integrins, Circulation 107, 455460.
122 Joseph, S., Olbrich, C., Kirsch, J., Hasbach,
M., Briel, A., and Schirner, M. (2004) A real-
time in vitro assay for studying functional
characteristics of target-specific ultrasound
contrast agents, Pharm Res 21, 920926.
123 Knight, L. C. (2001) Radiolabeled peptide li-
gands for imaging thrombi and emboli,
Nucl Med Biol 28, 515526.
124 Shattil, S. J., Hoxie, J. A., Cunningham, M.,
and Brass, L. F. (1985) Changes in the plate-
let membrane glycoprotein IIb.IIIa complex
during platelet activation, J Biol Chem 260,
1110711114.
125 Taub, R., Gould, R. J., Garsky, V. M., Cicca-
rone, T. M., Hoxie, J., Friedman, P. A., and
Shattil, S. J. (1989) A monoclonal antibody
against the platelet fibrinogen receptor con-
tains a sequence that mimics a receptor
recognition domain in fibrinogen, J Biol
Chem 264, 259265.
126 Lister-James, J., Knight, L. C., Maurer, A. H.,
Bush, L. R., Moyer, B. R., and Dean, R. T.
(1996) Thrombus imaging with a techne-
tium-99m-labeled activated platelet receptor-
binding peptide, J Nucl Med 37, 775781.
127 Shebuski, R. J., Ramjit, D. R., Bencen, G. H.,
and Polokoff, M. A. (1989) Characterization
and platelet inhibitory activity of bitistatin, a
potent arginine-glycine-aspartic acid-contain-
ing peptide from the venom of the viper Bi-
tis arietans, J Biol Chem 264, 2155021556.
128 Tung, C. H., Ho, N. H., Zeng, Q., Tang, Y.,
Jaffer, F. A., Reed, G. L., and Weissleder, R.
(2003) Novel factor XIII probes for blood co-
agulation imaging, Chembiochem 4, 897899.
129 Jaffer, F. A., Tung, C. H., Houng, A. K.,
OLoughin, T., Reed, G. L., Weissleder, R.
(2002) MRI of blood coagulation factor XIII
activity using a novel peptide derivatized
caged iron oxide nanoparticle (F13-CLIO),
Mol Imaging, 217218.
130 Lanza, G. M., Wallace, K. D., Scott, M. J., Ca-
cheris, W. P., Abendschein, D. R., Christy,
D. H., Sharkey, A. M., Miller, J. G., Gaffney,
P. J., and Wickline, S. A. (1996) A novel site-
targeted ultrasonic contrast agent with broad
biomedical application, Circulation 94,
33343340.
131 Flacke, S., Fischer, S., Scott, M. J., Fuhrhop,
R. J., Allen, J. S., McLean, M., Winter, P., Si-
card, G. A., Gaffney, P. J., Wickline, S. A.,
and Lanza, G. M. (2001) Novel MRI contrast
agent for molecular imaging of fibrin: impli-
cations for detecting vulnerable plaques, Cir-
culation 104, 12801285.
132 Barrett, J. A., Kolodziej, A. F., Caravan, P. D.,
Nair, S., Looby, R., Witte, S., Costello, C. R.,
Meslia, M. A., Drezwecki, L., Cesna, C.,
Pratt, C., McMurry, T. J., Lauffer, R. B., Yucel,
E. K., Zhao, L., Weisskoff, R. M., Carpenter,
A. P., Graham, P. B. (2002) EP-1873, a gado-
linium (Gd) labeled fibrin specific agent that
rapidly detects arterial and venous thrombi
with MRI, Circulation, II-120.
133 Taillefer, R. (2001) Radiolabeled peptides in
the detection of deep venous thrombosis,
Semin Nucl Med 31, 102123.
134 Petrich, B. G., and Wang, Y. (2004) Stress-ac-
tivated MAP kinases in cardiac remodeling
and heart failure; new insights from trans-
genic studies, Trends Cardiovasc Med 14,
5055.
135 Martin, S. J., Reutelingsperger, C. P., McGa-
hon, A. J., Rader, J. A., van Schie, R. C., La-
Face, D. M., and Green, D. R. (1995) Early re-
distribution of plasma membrane phosphati-
dylserine is a general feature of apoptosis re-
gardless of the initiating stimulus:
inhibition by overexpression of Bcl-2 and
Abl, J Exp Med 182, 15451556.
136 Hofstra, L., Liem, I. H., Dumont, E. A.,
Boersma, H. H., van Heerde, W. L., Doeven-
dans, P. A., De Muinck, E., Wellens, H. J.,
Kemerink, G. J., Reutelingsperger, C. P., and
Heidendal, G. A. (2000) Visualisation of cell
death in vivo in patients with acute myocar-
dial infarction, Lancet 356, 209212.
137 Schellenberger, E. A., Bogdanov, A., Jr., Pet-
rovsky, A., Ntziachristos, V., Weissleder, R.,
and Josephson, L. (2003) Optical imaging of
apoptosis as a biomarker of tumor response
to chemotherapy, Neoplasia 5, 187192.
138 Firestein, G. S. (2003) Evolving concepts of
rheumatoid arthritis, Nature 423, 356361.
139 Kinne, R. W., Brauer, R., Stuhlmuller, B., Pa-
lombo-Kinne, E., and Burmester, G. R.
(2000) Macrophages in rheumatoid arthritis,
Arthritis Res 2, 189202.
140 Paulos, C. M., Turk, M. J., Breur, G. J., and
Low, P. S. (2004) Folate receptor-mediated
targeting of therapeutic and imaging agents
to activated macrophages in rheumatoid ar-
thritis, Adv Drug Deliv Rev 56, 12051217.
141 Reddy, J. A., Haneline, L. S., Srour, E. F.,
Antony, A. C., Clapp, D. W., and Low, P. S.
(1999) Expression and functional character-
ization of the beta-isoform of the folate re-
ceptor on CD34(+) cells, Blood 93, 3940
3948.
142 Nakashima-Matsushita, N., Homma, T., Yu,
S., Matsuda, T., Sunahara, N., Nakamura, T.,
Tsukano, M., Ratnam, M., and Matsuyama,
T. (1999) Selective expression of folate recep-
tor beta and its possible role in methotrexate
transport in synovial macrophages from pa-
tients with rheumatoid arthritis, Arthritis
Rheum 42, 16091616.
143 Turk, M. J., Breur, G. J., Widmer, W. R., Pau-
los, C. M., Xu, L. C., Grote, L. A., and Low,
P. S. (2002) Folate-targeted imaging of acti-
vated macrophages in rats with adjuvant-in-
duced arthritis, Arthritis Rheum 46, 1947
1955.
144 Fletcher, D. S., Widmer, W. R., Luell, S.,
Christen, A., Orevillo, C., Shah, S., and Vis-
co, D. (1998) Therapeutic administration of
a selective inhibitor of nitric oxide synthase
does not ameliorate the chronic inflamma-
tion and tissue damage associated with adju-
vant-induced arthritis in rats, J Pharmacol
Exp Ther 284, 714721.
145 van Roon, J. A., van Vuuren, A. J., Wijngaar-
den, S., Jacobs, K. M., Bijlsma, J. W., Lafeber,
F. P., Thepen, T., and van de Winkel, J. G.
(2003) Selective elimination of synovial in-
flammatory macrophages in rheumatoid ar-
thritis by an Fcgamma receptor I-directed
immunotoxin, Arthritis Rheum 48, 1229
1238.
146 Wunder, A., Tung, C. H., Muller-Ladner, U.,
Weissleder, R., and Mahmood, U. (2004) In
vivo imaging of protease activity in arthritis:
a novel approach for monitoring treatment
response, Arthritis Rheum 50, 24592465.
147 Creamer, D., Sullivan, D., Bicknell, R., and
Barker, J. (2002) Angiogenesis in psoriasis,
Angiogenesis 5, 231236.
148 Kapsoritakis, A., Sfiridaki, A., Maltezos, E.,
Simopoulos, K., Giatromanolaki, A., Sivridis,
E., and Koukourakis, M. I. (2003) Vascular
endothelial growth factor in inflammatory
bowel disease, Int J Colorectal Dis 18, 418
422.
149 Xia, Y. P., Li, B., Hylton, D., Detmar, M.,
Yancopoulos, G. D., and Rudge, J. S. (2003)
Transgenic delivery of VEGF to mouse skin
leads to an inflammatory condition resem-
bling human psoriasis, Blood 102, 161168.
150 Boyman, O., Hefti, H. P., Conrad, C., Nicko-
loff, B. J., Suter, M., and Nestle, F. O. (2004)
Spontaneous development of psoriasis in a
new animal model shows an essential role
for resident T cells and tumor necrosis fac-
tor-alpha, J Exp Med 199, 731736.
151 Hoffmann, J. C., Pawlowski, N. N., Kuhl,
A. A., Hohne, W., and Zeitz, M. (2002) Ani-
mal models of inflammatory bowel disease:
an overview, Pathobiology 70, 121130.
152 Vinjamuri, S., Hall, A. V., Solanki, K. K., Bo-
manji, J., Siraj, Q., OShaughnessy, E., Das,
S. S., and Britton, K. E. (1996) Comparison
of
99m
Tc infecton imaging with radiolabelled
white-cell imaging in the evaluation of bacte-
rial infection, Lancet 347, 233235.
References 1297
153 Lupetti, A., Welling, M. M., Mazzi, U., Nib-
bering, P. H., and Pauwels, E. K. (2002) Tech-
netium-99m labelled fluconazole and antimi-
crobial peptides for imaging of Candida albi-
cans and Aspergillus fumigatus infections, Eur
J Nucl Med Mol Imaging 29, 674679.
154 Knight, L. C. (2003) Non-oncologic applica-
tions of radiolabeled peptides in nuclear
medicine, Q J Nucl Med 47, 279291.
155 Welling, M. M., Paulusma-Annema, A., Bal-
ter, H. S., Pauwels, E. K., and Nibbering,
P. H. (2000) Technetium-99m labelled anti-
microbial peptides discriminate between
bacterial infections and sterile inflamma-
tions, Eur J Nucl Med 27, 292301.
156 Tani, K., Murphy, W. J., Chertov, O., Salcedo,
R., Koh, C. Y., Utsunomiya, I., Funakoshi, S.,
Asai, O., Herrmann, S. H., Wang, J. M.,
Kwak, L. W., and Oppenheim, J. J. (2000) De-
fensins act as potent adjuvants that promote
cellular and humoral immune responses in
mice to a lymphoma idiotype and carrier
antigens, Int Immunol 12, 691700.
157 Sair, H. I., Doraiswamy, P. M., and Petrella,
J. R. (2004) In vivo amyloid imaging in Alz-
heimers disease, Neuroradiology 46, 93
104.
158 Klunk, W. E., Debnath, M. L., and Pettegrew,
J. W. (1995) Chrysamine-G binding to Alz-
heimer and control brain: autopsy study of a
new amyloid probe, Neurobiol Aging 16,
541548.
159 Styren, S. D., Hamilton, R. L., Styren, G. C.,
and Klunk, W. E. (2000) X-34, a fluorescent
derivative of Congo red: a novel histochem-
ical stain for Alzheimers disease pathology,
J Histochem Cytochem 48, 12231232.
160 Zhuang, Z. P., Kung, M. P., Hou, C., Plossl,
K., Skovronsky, D., Gur, T. L., Trojanowski,
J. Q., Lee, V. M., and Kung, H. F. (2001)
IBOX(2-(4'-dimethylaminophenyl)-6-iodo-
benzoxazole): a ligand for imaging amyloid
plaques in the brain, Nucl Med Biol 28,
887894.
161 Friedland, R. P., Majocha, R. E., Reno, J. M.,
Lyle, L. R., and Marotta, C. A. (1994) Devel-
opment of an anti-A beta monoclonal anti-
body for in vivo imaging of amyloid angiopa-
thy in Alzheimers disease, Mol Neurobiol 9,
107113.
162 Lovat, L. B., OBrien, A. A., Armstrong, S. F.,
Madhoo, S., Bulpitt, C. J., Rossor, M. N., Pe-
pys, M. B., and Hawkins, P. N. (1998) Scinti-
graphy with
123
I-serum amyloid P compo-
nent in Alzheimer disease, Alzheimer Dis
Assoc Disord 12, 208210.
163 Hawkins, P. N., Rossor, M. N., Gallimore,
J. R., Miller, B., Moore, E. G., and Pepys,
M. B. (1994) Concentration of serum amy-
loid P component in the CSF as a possible
marker of cerebral amyloid deposits in Alz-
heimers disease, Biochem Biophys Res
Commun 201, 722726.
164 Hawkins, P. N., Tyrell, P., Jones, T., et al.
(1991) Metabolic and scintigraphic studies
with radiolabeled serum amyloid P compo-
nent in amyloidosis: applications to cerebral
deposits and Alzheimer disease with posi-
tron emission tomography, Bull Clin Neu-
rosci, 178190.
165 Maggio, J. E., Stimson, E. R., Ghilardi, J. R.,
Allen, C. J., Dahl, C. E., Whitcomb, D. C.,
Vigna, S. R., Vinters, H. V., Labenski, M. E.,
and Mantyh, P. W. (1992) Reversible in vitro
growth of Alzheimer disease beta-amyloid
plaques by deposition of labeled amyloid
peptide, Proc Natl Acad Sci USA 89, 5462
5466.
166 Saito, Y., Buciak, J., Yang, J., and Pardridge,
W. M. (1995) Vector-mediated delivery of
125
I-
labeled beta-amyloid peptide A beta 1-40
through the blood-brain barrier and binding
to Alzheimer disease amyloid of the A beta
1-40/vector complex, Proc Natl Acad Sci
USA 92, 1022710231.
167 Agdeppa, E. D., Kepe, V., Liu, J., Flores-
Torres, S., Satyamurthy, N., Petric, A., Cole,
G. M., Small, G. W., Huang, S. C., and Bar-
rio, J. R. (2001) Binding characteristics of
radiofluorinated 6-dialkylamino-2-naphthyl-
ethylidene derivatives as positron emission
tomography imaging probes for beta-amy-
loid plaques in Alzheimers disease, J Neu-
rosci 21, RC189.
168 Barrio, J. R., Huang, S. C., Cole, G., et al.
(1999) PET imaging of tangles and plaques
in Alzheimers disease with a highly hydro-
phobic probe, J Label Comp Pharm, S194
195.
169 Weissmann, C., Enari, M., Klohn, P. C., Ros-
si, D., and Flechsig, E. (2002) Transmission
of prions, Proc Natl Acad Sci USA 99 Suppl
4, 1637816383.
170 Klunk, W. E., Bacskai, B. J., Mathis, C. A.,
Kajdasz, S. T., McLellan, M. E., Frosch, M. P.,
Debnath, M. L., Holt, D. P., Wang, Y., and
Hyman, B. T. (2002) Imaging Abeta plaques
in living transgenic mice with multiphoton
microscopy and methoxy-X04, a systemically
administered Congo red derivative, J Neuro-
pathol Exp Neurol 61, 797805.
171 Gambhir, S. S. (2002) Molecular imaging of
cancer with positron emission tomography,
Nat Rev Cancer 2, 683693.
172 Ugur, O., Kothari, P. J., Finn, R. D., Zanzoni-
co, P., Ruan, S., Guenther, I., Maecke, H. R.,
and Larson, S. M. (2002) Ga-66 labeled so-
matostatin analogue DOTA-DPhe1-Tyr3-oc-
treotide as a potential agent for positron
emission tomography imaging and receptor
mediated internal radiotherapy of somatosta-
tin receptor positive tumors, Nucl Med Biol
29, 147157.
173 Burt, B. M., Humm, J. L., Kooby, D. A.,
Squire, O. D., Mastorides, S., Larson, S. M.,
and Fong, Y. (2001) Using positron emission
tomography with [
(18)
F]FDG to predict tumor
behavior in experimental colorectal cancer,
Neoplasia 3, 189195.
174 Wu, A. M., Yazaki, P. J., Tsai, S., Nguyen, K.,
Anderson, A. L., McCarthy, D. W., Welch,
M. J., Shively, J. E., Williams, L. E., Raubit-
schek, A. A., Wong, J. Y., Toyokuni, T.,
Phelps, M. E., and Gambhir, S. S. (2000)
High-resolution microPET imaging of carci-
noembryonic antigen-positive xenografts by
using a copper-64-labeled engineered anti-
body fragment, Proc Natl Acad Sci USA 97,
84958500.
175 Borsi, L., Balza, E., Bestagno, M., Castellani,
P., Carnemolla, B., Biro, A., Leprini, A., Se-
pulveda, J., Burrone, O., Neri, D., and Zardi,
L. (2002) Selective targeting of tumoral vas-
culature: comparison of different formats of
an antibody (L19) to the ED-B domain of fi-
bronectin, Int J Cancer 102, 7585.
176 Wan, S., Parrish, J. A., Anderson, R. R., and
Madden, M. (1981) Transmittance of nonion-
izing radiation in human tissues, Photo-
chem Photobiol 34, 679681.
177 Licha, K., Riefke, B., Ebert, B., and Grotzin-
ger, C. (2002) Cyanine dyes as contrast
agents in biomedical optical imaging, Acad
Radiol 9 Suppl 2, S320322.
178 Licha, K., Riefke, B., Ntziachristos, V.,
Becker, A., Chance, B., and Semmler, W.
(2000) Hydrophilic cyanine dyes as contrast
agents for near-infrared tumor imaging: syn-
thesis, photophysical properties and spectro-
scopic in vivo characterization, Photochem
Photobiol 72, 392398.
179 Birchler, M., Neri, G., Tarli, L., Halin, C.,
Viti, F., and Neri, D. (1999) Infrared photo-
detection for the in vivo localisation of
phage-derived antibodies directed against an-
giogenic markers, J Immunol Methods 231,
239248.
180 Funovics, M. A., Alencar, H., Su, H. S., Kha-
zaie, K., Weissleder, R., and Mahmood, U.
(2003) Miniaturized multichannel near infra-
red endoscope for mouse imaging, Mol
Imaging 2, 350357.
181 Ntziachristos, V., Yodh, A. G., Schnall, M.,
and Chance, B. (2000) Concurrent MRI and
diffuse optical tomography of breast after in-
docyanine green enhancement, Proc Natl
Acad Sci USA 97, 27672772.
182 Payne, G. (2003) Progress in immunoconju-
gate cancer therapeutics, Cancer Cell 3, 207
212.
183 Michel, R. B., Brechbiel, M. W., and Mattes,
M. J. (2003) A comparison of 4 radionuclides
conjugated to antibodies for single-cell kill,
J Nucl Med 44, 632640.
184 Mattes, M. J. (2002) Radionuclide-antibody
conjugates for single-cell cytotoxicity, Cancer
94, 12151223.
185 Pastan, I., Beers, R., and Bera, T. K. (2004)
Recombinant immunotoxins in the treat-
ment of cancer, Methods Mol Biol 248, 503
518.
186 Neri, D. (2004) Tumor targeting, CHIMIA
58(10), 723726.
References 1299
Abstract
One major challenge facing the pharma-
ceutical industry today is to develop con-
trast-enhancing agents for molecular imag-
ing. Classic contrast agents primarily
document the anatomy. For pathophysiolo-
gical examinations using differential diag-
nostic techniques, i.e., characterizing the
development of a disease, they are only
suitable to a limited degree. Molecular
imaging selectively tracks down molecules
and cell structures to be able to establish
proof of disease at a very early stage and
then to make decisions about highly indi-
vidual treatment. The next straightforward
vision of medical imaging quite clearly lies
in the concept Find, Fight and Follow!.
In radiopharmaceuticals we are already
pursuing the approach of a triad consist-
ing of early diagnosis, therapy and therapy
control. Utilizing the nanotechnological
concepts of colloid and interface science,
imaging on a molecular level can also be
achieved via diagnostic ultrasound using
tiny gas-filled polymer particles coupled to
target-specific ligands. Additionally, nano-
sized polymeric drug carriers for targeting
and controlled release have been exten-
sively studied in the past. Here, a nanopar-
ticle or capsule acts like a container for a
pharmacologically active agent. Passive
and active targeting can be attained by
carefully chosen size and surface modifica-
tion of the carrier. Drug release can be
controlled via desorption of surface-bound
drugs, diffusion through the particle ma-
trix or the capsule wall, or matrix erosion.
Moreover, smart release can be achieved
by using smart polymers (pH or tempera-
ture sensitive) or, more interestingly, by
applying an external stress to the drug car-
rier. If the drug carrier is appropriately de-
signed, release can be induced by diagnos-
tic ultrasound. Building a bridge between
therapy and diagnosis opens the field of
theranostics. With a Find, Fight and
Follow! strategy, the tissue of interest can
first be imaged via target-specific ultra-
sound contrast particles. In a second step,
the same particles, now filled with a phar-
macologically active agent, can be used for
therapy. Finally, monitoring of treatment
effects is possible by sequential imaging.
This early approach demonstrates the suc-
cess of a resolute implementation of nano-
biotechnological concepts in a medical ap-
plication, and will be presented with re-
spect to polymer nanoparticle and micro-
capsule formation, the control of colloidal
structure, surface modification, antibody-
coupling strategies, and the resulting in vi-
tro properties as an ultrasound theranos-
tic. In vivo results will be addressed with
1301
7
Ultrasound Theranostics: Antibody-based Microbubble Conjugates
as Targeted In vivo Contrast Agents
and Advanced Drug Delivery Systems
ISBN: 3-527-31184-X
special emphasis on antibody-based target-
ing and gene delivery. Investigations with
different drugs and targeting sites demon-
strate that, in general, this approach can
serve as a biopharmaceutical platform
technology.
Abbreviations
3-D three-dimensional
CFT critical flooding temperature
CT computerized tomography
EAE experimental autoimmune ence-
phalomyelitis
ED-B extradomain B
EDC 1-ethyl-3-(3-dimethylaminopropyl)
carbodiimide
FN fibronectin
ICAM intercellular adhesion molecule
LOC loss of correlation
MS multiple sclerosis
pDNA plasmid DNA
SAE stimulated acoustic emission
SPAQ sensitive particle acoustic quantifi-
cation
SPECT single photon emission computer-
ized tomography
USCA ultrasound contrast agent
VCAM vascular cell adhesion molecule
7.1
Motivation: Find, Fight and Follow!
Dye see him? cried Ahab after allow-
ing a little space for the light to spread.
See nothing, sir.
Turn up all hands and make sail! he
travels faster than I thought for; the
top-gallant sails! aye, they should have
been kept on her all night. But no mat-
ter tis but resting for the rush. [ . . . ]
There she blows she blows! she
blows! right ahead! was now the
mast-head cry.
Aye, breach your last to the sun, Moby
Dick! cried Ahab, thy hour and thy
harpoon are at hand! Down! down all
of ye, but one man at the fore. The
boats! stand by! [ . . . ] [1].
Find, Fight and Follow!. It is not only
Captain Ahab that relied on that concept
to hunt whales with small harpoons by
hand in former times; the same strategy
was also used by the Indians in northern
America for hunting buffalo bows and
arrows were not efficient enough to kill a
strong animal like this immediately. In ad-
dition, the weapons of people in the stone
age were obviously not suitable for knock-
ing down a mammoth at first strike. To-
day, this is no longer a problem people
have developed specifically designed weap-
ons to kill buffalos and whales (and even
mammoths!) at the first shot.
However, what about our weapons to
fight against cancer? Cancer is one of the
most important killers today. Thousands
of humans die every minute and no thera-
py can rescue them from their destiny
although researchers continue to make
daily progress. We believe that we can
learn from the hunters in former times
Find, Fight and Follow!. With radiophar-
maceuticals, we are already pursuing the
approach of a triad consisting of early di-
agnosis, therapy and therapy control. We
have built a bridge between diagnosis and
therapy by coupling identical carriers with
different active substances diagnostic or
therapeutic.
For example, the radionuclide indium-
111 [
111
In] (or Technetium-99m [
99m
Tc] or
Yttrium-86 [
86
Y] as a positron emitter) can
be linked via a chelator to a cancer-specific
targeting moiety (a peptide or an antibody,
7 Ultrasound Theranostics 1302
etc.) (see also Part V, Chapter 2). Presuming
superior specificity of the conjugate the con-
trast agent (or, even better, the tracer) will
accumulate in the target tissue after admin-
istering to the patient. This leads to an in-
creased signal-to-background ratio in the
target and the position of a tumor can be
imaged with a c-camera [single photon
emission computerized tomography
(SPECT) or positron emission tomography
(PET), respectively, or even better a fusion
image with modern PET/CT settings] (see
also Part V, Chapters 46).
Furthermore, because the imaging agent
works by depositing radiation energy pro-
portionally to its concentration within the
tumor, imaging the target provides dosi-
metry about dose delivery and is therefore
important for planning the treatment.
After finding, fighting against the tu-
mor can be realized by replacing the
weak imaging radionuclide with a
strong therapeutic radionuclide like yt-
trium-90 [
90
Y], which shows chemically
similar behavior with regard to coupling to
the targeting moiety, but higher radiation
energy [2, 3] (see also Part II, Chapter 5).
Based on the dosimetry data, the therapeu-
tic dose can be calculated and targeted lo-
coregional radiotherapy can be started.
Finally, monitoring of treatment re-
sponse is possible by sequential imaging
using the same imaging agent. In sensi-
tive tumors, there should be a marked re-
duction of contrast uptake [4, 5]. In the
follow phase of the triad, these reporters
of efficacy determine the endpoint of ther-
apy or indicate further treatment. Differen-
tiation of responder from nonresponder is
a first step towards individualized medi-
cine (see also Part I, Chapter 2).
As genomics is now coming to the point
of indicating a predisposition for disease
and is also able to identify markers of ag-
gressive diseases, the combination with
modern molecular imaging techniques
provides exciting opportunities to exploit
the strategy described above.
The Find, Fight and Follow! concept
builds a bridge between therapy and diag-
nosis, and opens the field of theranostics
where genomics meets molecular imag-
ing and advanced drug delivery. Several lo-
gistical challenges, e.g., short half-life of
radiopharmaceuticals combined with the
production in-time and on-site, and,
furthermore, the radiation exposure to the
patient (especially in the follow-up phase),
make it important to evaluate alternative
approaches to theranostics.
In summary, a promising candidate
needs:
A high-resolution imaging modality
which is furthermore able to quantify
the amount of contrast conjugates in the
target tissue in order to indicate
responders and confirm therapeutic ef-
fects.
Excellent sensitivity to be able to image
small quantities of arresting contrast
conjugates.
High specificity and affinity to the target
to yield a significant signal-to-back-
ground ratio shortly after administra-
tion.
The possibility to deliver a drug and its
localized release.
Flexibility to act as a platform technol-
ogy to couple different targeting mole-
cules and deliver all kinds of relevant
drug substances.
This chapter attempts to identify whether
ultrasound in combination with smart ul-
trasound contrast agents (USCAs) can be a
powerful technique in applying the Find,
Fight and Follow! concept in future clini-
cal practice.
7.2
Ultrasound: Hear the Symptoms
Ultrasound is a versatile imaging tech-
nique that provides real-time diagnostic in-
formation on the morphology of internal
organs with high spatial resolution and
sufficient penetration depth. Imaging is
based on transmitting and receiving sound
waves with frequency ranges from 1 to 50
MHz. The resolution is directly related to
the frequency used, e.g., the higher the
frequency, the better the resolution. On
the other hand, high-frequency waves are
more attenuated than low-frequency
waves, which limits the penetration depth.
Consequently, the frequency used for
imaging is a trade-off between optimal re-
solution and desired penetration depth.
Typical frequencies of 27.5 MHz are used
to image large organs inside the body
(with a submillimeter range resolution
and penetration to about 30 cm) [6].
The acoustic energy used in diagnostic
ultrasound (intensity below 0.1 W cm
2
) is
very low compared to ultrasound for thera-
peutic applications [7] (usually much great-
er than 1 W cm
2
for clot lysis or transder-
mal drug delivery, which will not be ad-
dressed in this chapter). Noncontrast diag-
nostic ultrasound has been established as
an important, rapid and cost-effective
means in almost all medical fields. One
major advantage of ultrasound is the fact
that patients are not exposed to harmful
radiation. There are no limits with regard
to examination time and frequency, and
even investigations of embryos in utero
are now routine [an impressive video
showing the three-dimensional (3-D) scan-
ning of an embryo is shown on the sup-
plementary CD-ROM]. Additionally, ultra-
sound energy can be focused on a small
volume (down to cubic millimeters), which
makes it a perfect tool to trigger actions
inside the body (e.g., drug release from a
smart contrast agent).
To understand the following principles,
only two of the more than 10 ultrasound-
based imaging modalities are important.
The so-called B-mode utilizes the specific
ultrasound backscattering properties of dif-
ferent types of tissue and blood. Recon-
struction of the backscattering signals
leads to an image of the morphology. The
second so-called Doppler mode detects and
measures the blood flow by a frequency
shift (Doppler shift) between transmitted
and received ultrasound signals. Modern
3-D image processing enables a fused 3-D
image of anatomical structure (gray scale)
and functional blood flow (color coded) at
the same time [8].
Additionally, many different color Dop-
pler mode variations are implemented in
high-end imaging devices (power Doppler,
harmonic Doppler, harmonic power Dop-
pler, etc.) and in principle these variations
are suitable to highlight contrast agent-
specific signals in a color-coded manner
fused with a 3-D gray-scale image of the
morphology.
In summary:
Ultrasound imaging is safe (no harmful
radiation used).
The B-mode images the morphology at
submillimeter resolution with sufficient
penetration depth.
Different kinds of Doppler modes are
useful to highlight contrast-specific sig-
nals.
Fused 3-D images are state of the art.
Certain ultrasound pulses with focused
energy can induce drug release from
smart contrast agents.
7.3
Ultrasound Contrast: Tiny Bubbles
The diagnostic capabilities have not only
been improved by technological advances
in the ultrasound systems, but also by the
introduction of USCAs. Since 1980, exten-
sive research has been performed in order
to make contrast ultrasound an established
diagnostic technique [9]. At the beginning
of this period, investigators were forced to
use home-made contrast agents and it was
quickly realized that small bubbles in-
creased the contrast dramatically.
This is due to the fact that the contrast
factor in the underlying Rayleigh equation
[10] depends on the density difference and
compressibility difference of a dispersed
matter in the surrounding media. In par-
ticular, if a liquid is dispersed in another
liquid, the value of the contrast factor is
about 0 (i.e., there is almost no density dif-
ference and, thus, no compressibility dif-
ference between liquids) and a solid dis-
persed in a liquid increases the value just
slightly to about 12. However, in the case
of air in water (or blood) this contrast fac-
tor raises to 10
14
, which makes dispersed
air bubbles a perfect contrast agent for ul-
trasound.
Another important parameter is the size
of the bubbles. On the one hand, the scat-
tering cross-section increases proportion-
ally with the diameter of the bubble to the
power of 6 (Rayleigh scattering). On the
other hand, the bubbles must be small en-
ough to pass through the smallest capillar-
ies of the pulmonary system in order to
prevent an embolus in the blood vessel.
Considering safety and efficacy, a size
range of 17 lm is most preferable for a
gas-bubble-based contrast agent.
The first generation of commercialized
USCAs comprises free air bubbles or sur-
factant-stabilized air bubbles (Echovist or
Levovist). In order to increase their life-
time in blood, the second-generation US-
CAs (e.g., Echogen, Imagent, SonoVue or
Optison) were made out of gas with low
water solubility (like perfluorcarbon gases
or sulfurhexafluoride) and a surface stabili-
zer to prevent aggregation.
A scientific breakthrough in order to de-
sign USCAs on demand can be seen in
the third generation (Myomap, Quantison,
BiSphere and Sonavist). Compared to the
more or less free bubbles of the first and
second generations, the novel type of US-
CAs consist of encapsulated microbubbles
with a shell formed by a biopolymer (like
human albumin) and/or a biocompatible
synthetic polymer (like copolymers of poly-
lactide and polyglycolide or derivatives of
polycyanoacrylate). In addition to the pro-
longation of the lifetime in the blood
stream, these polymer-stabilized microbub-
bles can be manufactured to fulfill certain
needs, and to interact with diagnostic ultra-
sound in a defined and optimal manner.
The acoustic properties of encapsulated
gas bubbles were intensively discussed by
de Jong et al. [11, 12], Hoff [13] and Frink-
ing and de Jong [14] in the 1990s. Based
on their theoretical considerations, it can
be shown that after choosing the nature of
the gas and the type of polymer, the behav-
ior of a gas-filled microsphere in an ultra-
sound field depends only on the geometry
of the capsule; in particular, the size and
shell thickness. This means that scattering
and resonance properties can be tuned by
the geometry of the USCA.
Moreover, based on stability relations in
architecture it can be proven that the criti-
cal fragility of a hollow spherical body de-
pends on the modulus of elasticity of the
shell material (a physical constant deter-
mined by the nature and molecular weight
of the polymer used) and the quotient of
shell thickness over size. This means that
we can fix the pressure threshold of a gas-
filled microsphere by the geometry so that
the sphere collapses above a certain, criti-
cal value of the applied acoustical pres-
sure. This is extremely important due to
the fact that the USCA should resist the
blood pressure and a certain level of addi-
tional acoustic pressure that is just needed
for ultrasound imaging. However, at high-
er acoustic pressure the shell of the gas-
filled microcapsule should be destroyed.
It is now straightforward (see also Part
V, Chapter 6) to fill the capsule not only
with air, but a drug substance too, and to
use an ultrasound pulse to trigger the re-
lease from outside the body at a defined
place and time by bursting the bubble.
Utilizing nanotechnological concepts of
polymer, colloid and interface science we
have established a novel process to manu-
facture gas-filled microcapsules. On de-
mand, gas filling, elasticity, shell thickness
and overall size can be tailored indepen-
dently.
Using the so-called two-step process
[15, 16], polymer nanoparticles are first
synthesized via emulsion polymerization.
The size of the resulting nanoparticles can
be tuned by a simple process parameter
and covers a range of about 30400 nm.
In a second step these nanoparticles are
used to coat microbubbles in a controlled
bubble formation process. The nanoparti-
cles migrate to the surface of the bubbles
(this is related to the interface activity of
hydrophobic nanoparticles in general) and
build a monolayer around the bubbles.
Consequently, the size of the nanoparticles
determines the shell thickness of the final
microcapsules. Additionally, a carefully
chosen nanoparticle concentration regime
results in a certain microcapsule size dis-
tribution. In principle, particle sizes in the
range of 0.510 lm can be adjusted and
the microcapsule size distributions are ex-
tremely narrowly distributed. Last, but not
least, the molecular weight of the polymer,
which can be controlled in the first phase
of the two-step process, determines the
elasticity (within certain limits) of the
shell-forming polymeric material without
changing the chemical constitution.
An image of a gas-filled polymer-stabi-
lized microcapsule obtained by electron
microscopy is depicted in Fig. 7.1. The
overall spherical shape and the substruc-
ture of nanoparticles are visible. The un-
derlying two-step process is schemati-
cally described in Fig. 7.2. To highlight the
flexibility of the manufacturing process,
the tunable quality aspects like resonance
behavior and pressure stability are summa-
rized in Figs. 7.3 and 7.4, respectively.
In summary, the innovative two-step
process is an enabling technology to man-
ufacture gas-filled microparticles with de-
fined colloidal architecture to fulfill all the
Fig. 7.1 An image of one selected gas-filled poly-
mer-stabilized microcapsule obtained by electron
microscopy. The magic bullet is spherical with a
diameter of about 1.7 lm and the substructure
made of nanoparticles is clearly visible. In order
to indicate that the particles are in fact gas-filled,
an imperfect capsule with a bump has been care-
fully selected.
physical requirements on an ultrasound
mode-specific contrast agent (and ad-
vanced drug delivery system). Considering
the strong sensitivity of ultrasound to
small variations of the structure of USCAs,
even multiplexing of microcapsules due to
certain resonance behavior or critical fra-
gility is possible.
Fig. 7.2 The principles of the two-step process
suitable to control several parameters of an USCA
independently. In the first step, polymer nanopar-
ticles are synthesized via emulsion polymeriza-
tion. The size of the resulting nanoparticles can
be tuned by a simple process parameter and cov-
ers a range of about 30400 nm. In the second
step, these nanoparticles are used to coat micro-
bubbles in a controlled bubble formation process.
Consequently, the size of the nanoparticles deter-
mines the shell thickness of the final microcap-
sules. Additionally, a carefully chosen nanoparticle
concentration regime yields a certain microcap-
sule size distribution. Particle sizes in the range
of 0.510 lm can be adjusted and the size distri-
butions are extremely narrowly distributed. The
molecular weight of the polymer, which deter-
mines the elasticity (within certain limits) of the
shell-forming polymeric material, can be con-
trolled in the first phase of the two-step process.
Fig. 7.3 The resonance behavior of three different
microcapsule formulations obtained by ultrasonic
spectroscopy. The ultrasound attenuation (y-axis,
given relative to the maximum) is measured over
a driving frequency from 1 to 20 MHz (x-axis).
Each capsule population has the same mean di-
ameter of about 5 lm, but the shell thickness dif-
fers significantly (45, 105 and 236 nm). Tailoring
the geometry of the USCA tunes the resonance
properties. For example, increasing the
shell thickness at constant size of the gas-filled
microcapsules yields an increase in resonance fre-
quency (about 2, 5 and 9 MHz for the given ex-
amples). The sharply defined attenuation spectra
indicate that the size distribution of the capsules
is narrowly distributed. Using a mixture of bub-
bles with certain acoustic properties, combining
them in one contrast agent could be the basis for
ultrasound multiplexing (as in the case of optical
spectroscopy, the basis for optical imaging).
7.4
The Perfect Modality: Sensitive Particle
Acoustic Quantification (SPAQ)
Depending on the sound pressure of an
acoustic field, microbubbles can behave
quite differently. Using low sound pres-
sures, they oscillate and behave as linear
scatterers improving the signal-to-noise ra-
tio. Above a certain amplitude, gas-filled mi-
crocapsules disintegrate rapidly, thereby
emitting a strong nonlinear signal, which
is misinterpreted as a quick movement
and mapped as a characteristic random col-
or pattern in the color Doppler mode of the
ultrasound device. This characteristic bub-
ble signature firstly described by Reinhardt
et al. 1992 [17] is called stimulated acoustic
emission (SAE) or loss of correlation
(LOC) [1820], as depicted in Fig. 7.5. (Actu-
ally, a disintegrating bubble sends no acous-
tic signal, but the rapid shrinking process
caused by the dissolution of gas changes
the backscattering situation in such a way
that the algorithms of common Doppler
modes process an apparent movement).
However, the SAE effect enables the ex-
aminer to detect stationary microbubbles
[21, 22] a property that has generated a
large field of interest; in particular, be-
cause the amplitude of SAE signals is even
strong enough to detect individual micro-
bubbles within a given tissue. Thus, the
SAE effect is an extremely interesting tool
and could be the key to highly sensitive
molecular imaging.
Reliable quantification of microbubbles
within the tissue, even in high concentra-
tions, is needed to enable the efficient use
of molecular imaging for temporal and
spatial evaluation of molecular targets. A
major constraint in the quantification of
SAE signals is that SAE signals are dis-
played at millimeter-size in the image, be-
Fig. 7.4 To highlight a second tunable quality as-
pect of specially designed USCAs, the pressure
stability against a geometry factor (inner diameter
over the outer diameter) of capsules is plotted.
This factor has been selected in order to compare
the properties of small USCAs with commercially
available glass capsules, which are usually larger
(100700 lm). The fact that both fitted lines are
parallel proves that polymer and glass capsules
follow the same physical principles. Just the mod-
ulus of elasticity of glass is about
100 times higher than in the case of the polymeric
material used to manufacture the USCA. Compu-
ter-aided simulations have been performed to vali-
date the findings. It is easy to understand that
gas-filled polymer capsules with a quotient (inner
to outer diameter) larger than about 0.94 cannot
resist the blood pressure. Furthermore, the critical
fragility of capsules can be again exploited for
multiplexing and, most importantly, for drug deliv-
ery as well.
cause the device cannot locate the signal
within the resolution cell (voxel). Due to
the fact that a single SAE signal fills 1 vox-
el, more than a single microparticle within
1 voxel cannot be discriminated and any
amount of contrast agent within 1 voxel
would yield the same result in the image.
To overcome this inherent problem of
SAE signal quantification, we developed a
new quantification method called SPAQ,
which is based on a defined overlap be-
tween consecutive ultrasound images. The
SPAQ technology is schematically de-
scribed in Fig. 7.6, and has been validated
in agarose phantoms and ex-vivo in animal
tissue [23, 24].
The experiments show that an accurate
quantification of static microbubbles from
a single bubble event up to 300 000 bub-
bles mL
1
and 3-D mapping of large vol-
umes is possible with a resolution in a
scan direction of down to 10 lm.
The great advantage of gas-filled micro-
capsules as reporters for molecular imaging
is their outstanding detection sensitivity.
With SPAQ, it will now become possible
to quantify such targeted microbubbles with
maximum precision in an automatic scan
process that functions independently of
the examiner and is highly reproducible.
7.5
Targeting and Molecular Imaging:
The Sound of an Antibody
After discovering the suitable quantifica-
tion technology, an USCA can be designed
Fig. 7.5 Principle of SAE. Using high-amplitude
sound pressure, the shell of the microcapsules
breaks and the air bubble disintegrates rapidly.
The sudden disappearance of the microbubble is
misinterpreted by the ultrasound machine as a
quick movement and color-coded in the image.
This pseudo-Doppler effect is known as SAE.
Additionally, microcapsules coated with gold na-
noparticles (to give increased contrast for identifi-
cation in the electron micrograph) have been in-
vestigated in rat liver with electron microscopy be-
fore and after application of ultrasound (yellow ar-
rows). Only fragments of shells can be identified
in tissue post-ultrasound.
and optimized appropriately (i.e., with re-
gard to gas filling, size, shell thickness,
elasticity and pressure stability) utilizing
the flexible two-step process, which is
perfectly suitable for SPAQ.
Now the question is: How can we make
the bubbles smart in a way that they can
find targets inside a living organism?.
The answer can be found again in the
principles of colloid and interface science.
It is well known that a surface modifica-
tion of polymeric nanoparticles with biolo-
gically important antigens leads to parti-
cles which are useful for testing the pres-
ence of compliment antibodies in human
blood. After an ex vivo mixing of blood to-
gether with the antigennanoparticle con-
jugates, large, visible aggregates are
formed due to the interparticle connection
via relevant antibodies.
Different coupling strategies have been
established to manufacture these antigen
nanoparticle conjugates for the commonly
called aggregation assays [25]. One of the
most important is the [1-ethyl-3-(3-di-
methylaminopropyl) carbodiimide] (EDC)-
activated reaction which forms an amide
bond between a primary amine of the bio-
logically active substance and a carboxy
group on the surface of particles. Thus, all
we need to do is to introduce carboxy
groups on the surface of the polymer-stabi-
lized gas-filled microbubbles.
In the case of polybutylcyanoacrylate mi-
crobubbles a little caustic soda solution is
needed to perform a hydrolysis of the ester
groups just on the particle surface and the
free carboxy groups can be used for the
coupling of targeting ligands [15, 26]. To
be more flexible in animal experiments we
have preferred the coupling of streptavidin
to the microcapsules. Biotinylation of anti-
bodies is well established and the affinity
of biotin to streptavidin is outstanding. To
Fig. 7.6 SPAQ to enable the quantification of SAE
signals of static USCAs with high accuracy. The
principle of SPAQ is to create consecutive ultra-
sound images with a defined overlap. Thereby
only microcapsules lying out of the overlapping
region are subjects of SAE. The degree of the
overlap determines the thickness of the SAE imag-
ing layer. As a consequence of the higher spatial
resolution of this method, single gas-filled micro-
capsules can be discriminated and therefore quan-
tified even in high concentrations.
prepare the final formulation for in vitro
and in vivo tests, incubation of the biotiny-
lated antibodies together with streptavidin-
coated microbubbles for about 10 min at
room temperature is sufficient to produce
the ready-to-use formulation.
Streptavidin and antibody loading can
be quantified by a FACS assay. It could be
shown that the streptavidin loading per
microcapsule can be adjusted from several
hundreds to about 1 million.
The first in vitro investigations using a
flow-chamber assay proved the fact that
target-specific microbubbles tag on the tar-
get even at high flow rates [2729] and an
antibody loading of not more than 10
5
per
bubble is absolutely sufficient. Alternative
coupling strategies to streptavidinbiotin
have also been tested successfully [26, 30].
Different groups have shown that micro-
bubbles can be coated with target-specific
antibodies and could demonstrate the effi-
cacy in vivo [3136] with respect to a bina-
ry yes or no result. However, quantitative
information about the amount of arrested
USCA has only just recently been ob-
tained.
Based on the microcapsules and cou-
pling strategy described above, the in vivo
feasibility was first demonstrated in mice
and dogs. Using an i.v. administered L-se-
lectin ligand-specific USCA, active lymph
node targeting could be performed in both
species [37]. Exemplarily, impressive re-
sults based on this investigation are pre-
sented in Fig. 7.7.
Furthermore, it could be shown that the
induction of the microbubble-based SAE
effect in color Doppler and the rapid blood
clearance of nontargeted USCAs by cells
of the reticuloendothelial system within
3060 min, depending on the dose used,
yields a sensitive detection method and a
strong signal-to-noise ratio of targeted mi-
crocapsules. Moreover, it provides the op-
portunity of repeated investigations within
one session using USCAs targeted to the
same or other endothelial cell receptors.
This was the first time that active USCA
targeting has been demonstrated in lymph
nodes under normal conditions after i.v.
administration.
L-selectin ligand-specific USCA could be
a candidate for an indirect method of lym-
phography for the safe and less-invasive
ultrasonic identification of lymph nodes,
e.g., when performing a biopsy. Lymph
node-targeted microbubbles can be de-
tected easily with any ultrasound device
that has color Doppler capabilities.
After targeting healthy lymph nodes
without further quantification of the re-
sults, the SPAQ technique has been tested
successfully in an animal model of inflam-
mation [3840]. By conjugating microbub-
bles with antibodies to intercellular adhe-
sion molecule (ICAM)-1 and vascular cell
adhesion molecule (VCAM)-1, the authors
were able to depict and quantify ICAM-1
and VCAM-1 in rat autoimmune encepha-
lomyelitis (EAE), an established inflamma-
tory disease model of human multiple
sclerosis (MS) [41, 42]. Additionally, after
treatment with methylprednisolone, the
measured number of targeted anti-ICAM-1
and VCAM-1-microcapules (in spinal cord
and brain) was significantly less (P<0.01)
compared to untreated animals. This re-
sult is extremely important, because this
breakthrough highlights for the first time
that a target-specific USCA can be suitable
to act as a reporter of efficacy with re-
spect to a therapeutic effect.
The workflow of a routine SPAQ proce-
dure and in vivo images (through the skull
in living animals) regarding ICAM-1 target-
ing together with the quantitative result are
presented in Figs. 7.8 and 7.9. Additionally,
to validate the results regarding monitoring
of therapy with the SPAQ technique, a first
quantitative ex vivo/in vivo correlation has
been performed [40, 43].
In order to round off the picture, the
sounds of antibodies were recently re-
corded in several tumor models. For exam-
ple, the expression of extradomain B (ED-
B)-fibronectin (FN) in tumors could be
measured in appropriate animal models.
ED-B-FN is an angiogenesis-specific tar-
get exclusively found in the area of newly
formed blood vessels [4447]. A target-spe-
cific cyanine dye with single-chain antibod-
ies directed against ED-B-FN was reported
by Neri et al. [48] (see also Part V, Chapter
6). In addition, near IR fluorescent FN
analogs have recently been successfully ap-
Fig. 7.7 Two-dimensional color Doppler ultra-
sound images show a dog lymph node treated
with administered L-selectin ligand-specific USCA.
(A) Four selected images from the first scan. (B)
Image from the second scan of the same lymph
node, which was performed immediately after the
first scan. The images in (A) show typical micro-
bubble-based SAE signals in the paracortex,
whereas no signals are visible in the second scan
(B), which proves the destruction of microcap-
sules during the first scan. (C) The SAE signal
distribution in (A) corresponds immunohisto-
chemically to the high endothelial venule location.
(Reproduced with permission. 2004 Radiology
Society of North America, Inc.)
plied for optical imaging of tumor angio-
genesis in mice [49, 50]. A new affinity-
matured recombinant antibody single-
chain fragment called AP39 has been used
(see also Part V, Chapter 2).
The same antibody fragment with prov-
en specificity in vivo based on optical
imaging was used to manufacture AP39
microbubble conjugates. In a validated rat
MTLN3 tumor model, a dose of just
510
8
microbubbles kg
1
body weight of
AP39microbubble conjugates has been
administered. (Equal to about 810
16
mol
kg
1
! Compared to about 510
8
mol kg
1
antibodydye conjugates in the case of op-
tical imaging or approximately the same
10
8
mol kg
1
dose for radiolabeled com-
pounds.) Applying the SPAQ technique
just 15 min post-injection [compared to an
imaging window of 4 (earliest point) to
24 h (maximum contrast-to-background ra-
tio) in case of optical imaging [50]] the
amount of 36461 target-specific microbub-
bles could be quantified within a tumor of
15 mm size. The control experiment under
equal conditions, but using an isotype con-
trol antibody, leads to just 798 statistically
arrested bubbles in the target tumor.
To give a final impression of the molec-
ular imaging capabilities of the SPAQ
technology, a typical result of the tumor-
imaging study described above is given in
Fig. 7.10.
Considering the results on targeting
healthy, inflammatory and carcinogenic tis-
sue, it is quite realistic to state that, depend-
ing on the antibody linked to the surface of
the gas-filled microcapsules, the technique
Fig. 7.8 The workflow of a routine SPAQ proce-
dure. (Above left) A rat is placed in the SPAQ de-
vice. Four basic steps are illustrated. (1) Two-di-
mensional scans of the brain through the skull.
(2) Generating a 3-D dataset. (3) Definition of a
region of interest. (4) Data analysis. Subsequent
investigation of the same region of interest identi-
fies the amount of circulating USCAs (blood
pool), which can be subtracted from the amount
of microcapsules arrested in the target.
can be used to quantify the expression of
any accessible antigen expressed on the lu-
minal surface of endothelial cells. Thus,
the combination of target-specific USCAs
with SPAQ is therefore a promising tool
for the noninvasive and dynamic assess-
ment of disease-related molecules.
In our view, the technique will certainly
be first established as a molecular imag-
ing-based drug discovery and development
tool in preclinical practice. However, long-
term possible clinical applications can be
foreseen in safe and less-invasive ultra-
sound identification of certain tissue (e.g.,
lymph nodes), inflammatory regions of
MS or the sensitive detection of small tu-
mors. Also cardiovascular indications are
conceivable due to the more vascular,
blood-pool character of the USCAs. Finally,
the great advantages of dynamic and quan-
titative assessment should lead to the use
of target-specific USCAs as a disease-stag-
ing tool and as reporters of efficacy for
evidence-based patient selection to reach
Fig. 7.9 Quantitative results of SPAQ investiga-
tions in accordance to the workflow described in
Fig. 7.8. Bars show mean values and standard de-
viations (n=4) of SAE signals expressed as acous-
tic counts of ICAM-1 targeted ultrasound contrast
agents at the blood brain barrier. The values are
registered from the entire brain of healthy rats
and rats with adoptive transfer experimental auto-
immune encephalomyelitis (EAE-rats), respec-
tively. For further experimental details please refer
to Reinhardt et al [40]. Based on different investi-
gations five important theses could be verified:
(1) proof of principle regarding molecular imag-
ing, depiction and reproducible quantification of
target-specific microcapsules; (2) proof of specific-
ity by blocking the target with high doses
(200 lg) of anti-ICAM-1 antibodies prior to the ad-
ministration of anti-ICAM-1 conjugated USCA; (3)
proof of sensitivity since glycocorticosteroids
like methylprednisolone are known to cause a
down regulation of adhesion molecules, the treat-
ment of EAE-rats with Urbason (50 mg/kg b.w.,
daily injection for 3 days) results in a significant
decrease in acoustic counts; (4) target-specific
USCA as reporters of efficacy both experiments,
target blockade and cortison treatment, demon-
strate the potential of targeted microbubbles to-
gether with SPAQ to monitor and quantify a ther-
apeutic effect; (5) ex vivo/in vivo correlation all
validation experiments were performed ex vivo on
isolated tissue (in a water tank) (graph left), while
additional experiments were carried out on living
anesthetized animals through the intact skull. The
comparison (graph right) shows a first quantita-
tive ex vivo/in vivo correlation as depicted based
on the untreated experimental groups. Further-
more, the ICAM blockade experiment yields the a
similar number of acoustic counts as the control
group of healthy rats investigated with ICAM-1 tar-
geted USCA as well.
the visionary 100% responder rate a
challenging, but worthwhile long-term
aim.
7.6
Drug Delivery: The Magic Bullet
A perfect drug delivery system would be
noninvasive, detectable from outside the
body and should be able to target a dis-
eased region (see also Part VI, Chapter 1).
Furthermore, the release should be trig-
gered and quantified by external means.
At the beginning of the last century Paul
Ehrlich the founder of chemotherapy
introduced the idea of the magic bullet
(Zauberkugel) [51] for medical applica-
tions. His vision of a receptor-specific
drug, which acts in a highly specific man-
ner for a disease without any side-effects,
can be seen as the start of the search for
advanced drug delivery systems.
For example, nanosized polymeric drug
carriers for targeting and controlled release
have been extensively studied in the past.
In this case, a nanoparticle or capsule acts
like a container for a pharmacologically ac-
tive agent. Passive and active targeting can
be attained by carefully choosing the size
and surface modifications of the carrier
(see also Part VI, Chapter 8). Convention-
ally, drug release can be controlled via de-
sorption of surface-bound drugs, diffusion
through the particle matrix or the capsule
wall, or matrix erosion. In addition,
smart release can be achieved by using
smart polymers (pH or temperature sensi-
tive) or, more interestingly, by applying an
external stress to the drug carrier. As men-
tioned above, if the drug carrier is appro-
priately designed, release can be induced
by diagnostic ultrasound.
The first patents on the combined diag-
nostic and therapeutic use of USCAs were
Fig. 7.10 In-vivo measurement of ED-B-FN in
MTLN3 tumor of rats with AP39microcapsule
conjugates using SPAQ. (Above) Two-dimensional
color Doppler image 15 min after administering
unspecific USCAs. Only four SAE signals are visi-
ble in the target region. (Below) The same model,
but 15 min after injection of target-specific micro-
capsules. Significantly more SAE (182 signals at a
SPAQ resolution of 50 lm) are detectable. The in-
vestigation of the total tumor, which is about 15
mm in diameter, leads in this special case to an
amount of 36 461 arrested microparticles. As-
sumed the tumor is spherical (and thus occupies
a volume of about 1.8 mL), 20 633 contrast conju-
gates mL
1
tissue have been quantified. This
shows a practical example of the outstanding sen-
sitivity of the method.
filed by Stein et al. [52] in 1990 and Weit-
schies et al. [53] in 1994, and early descrip-
tions of an ultrasound contrast-based
micromachine were published by Ishi-
hara [54] in 1996. In the same year, Unger
[55] presented a similar concept. Two years
later Frinking et al. [56] published initial
in vitro results on the forced release of an
encapsulated compound from a micro-
spherical carrier by high-intensity ultra-
sound. However, the applied spray-drying
process showed several drawbacks (i.e.,
just 5% w/w loading in the shell, no accu-
rate control of acoustical properties of the
resulting microspheres, 20% immediate
release of surface-bound drug, etc.), but it
was the first published proof-of-principle
in vitro.
Independently, at the same time we
have developed a novel encapsulation pro-
cedure [15]. The starting material is the
well-defined microsphere dispersion pro-
duced via the two-step process. These mi-
crocapsules are mixed with the drug sub-
stance of interest and heated above a cer-
tain temperature the critical flooding
temperature (CFT) where the drug solu-
tion migrates through the now leaky shell
into the capsules. (Several experiments
have shown that the CFT is determined by
the nature of polymer and its molecular
weight.) Subsequent cooling stops the
flooding process immediately and a care-
fully performed final freeze-drying proce-
dure leads to solidification of drug inside
the microcapsule.
Optionally, depending on the solubility
of the drug in water, the filling CFT pro-
cess can be repeated several times and it
has been shown that the drug loading can
be performed above 100% (w/w), which
means that the shell thickness is more
than double sized. (Of course, the filling
influences the acoustical properties, but
this can be considered in the planning of
the two-step process to form the starting
microcapsules with weaker shells).
Formulations which are appropriate for
in vivo applications have been successfully
tested in vitro with respect to drug release
induced by diagnostic ultrasound. A typi-
cal release profile is shown in Fig. 7.11.
Unfortunately, the CFT process is lim-
ited to drugs with molecular weights smal-
ler than 80 kDa. Alternatively, encapsula-
tion of material with higher molecular
weight, e.g., reporter plasmids (pDNA), in
gas-filled microspheres can be performed
by spray-drying or double-emulsion proce-
dures [57]. Due to the energy impact in
the case of double-emulsion procedures,
for example, it is common to protect and
stabilize the pDNA with polycations such
as polylysine or polyethyleneimine. Fig.
7.12 schematically summarizes some of
the encapsulation procedures.
Another strategy for drug delivery with
gas-filled polymer-stabilized microbubbles
is based on a surface binding of drugs,
especially drugs with a high molecular
weight. In addition to covalent and adsorp-
tive coupling, linking can be easily per-
formed by using electrostatic interactions.
A further look into the literature on col-
loid and interface science demonstrates
that ionic interactions of oppositely
charged polyions lead to extremely stable
complexes [58]. This is also true for the in-
teractions of charged surfaces and poly-
ions. Considering that genetic material
(naked DNA and plasmids, and also the
surface of several types of virus) is in cer-
tain cases negatively charged, a complex
can be formed by mixing with a positively
charged polyion (e.g., polylysine). In addi-
tion to the protection and stabilization ef-
fect, such ionic complexes have also
shown a better transfection rate than
naked DNA (e.g., for synthetic nonviral
vector systems) [59] (see also Part VI, Chap-
Fig. 7.11 Typical drug release profile of gas- and
drug-filled microcapsule formulations. Percent re-
lease of an encapsulated model drug in physiolog-
ical buffer media as a function of time. Storage
under stirring for about 100 min at 378C shows
that just a baseline concentration of free drug is
measurable. By applying diagnostic ultrasound to
a certain volume of the dissolution media, a
strong release can be induced. Similar results can
be obtained by increasing the temperature above
the CFT of the capsules.
Fig. 7.12 Overview of some technical approaches
for the preparation of drug-loaded USCAs. Ap-
proaches I and II describe encapsulation tech-
niques based on solvent evaporation of a water/
oil/water emulsion and the CFT procedure, respec-
tively. Additionally, surface coupling strategies are
depicted in Approach III.
ters 6 and 7). Furthermore, complexes with
a surplus of cationic polyelectrolytes lead
to overall positively charged conjugates
which can be attached again via electro-
static interactions to the surface of car-
boxy-modified gas-filled microcapsules [15,
26]. In an adaptation to the concept of
gene guns based on gold nanoparticles,
which are coated with DNA and can be in-
jected using air pressure through the skin
into surface cells [60], these pDNA-com-
plex-coated microbubble conjugates may
be called noninvasive micro-gene guns.
7.7
Ultrasound, Microbubbles and Gene Delivery:
Noninvasive Micro-Gene Guns
In gene therapy, efficient and target-site-
specific in vivo gene delivery is a major
challenge. Free DNA without any delivery
system has been found to be highly sus-
ceptible to nuclease degradation [61] and is
rapidly cleared from the plasma when in-
jected i.v. [62]. Although viral-based deliv-
ery systems such as adenoviruses [63] and
retroviruses [64] efficiently introduce
genes, they suffer from drawbacks in
terms immunogenicity and toxicity (see
also Part I, Chapter 7 and Part II, Chapter
7). The immunogenicity of viral vectors
limits repeated use of the delivery system.
Therefore, nonviral gene delivery systems
such as cationic lipids, liposomes and
polymeric microspheres [59, 6567] have
been increasingly investigated as alterna-
tives to viral vectors due to their potential
advantages, such as ease of preparation
and scale-up, as well as their relative safety
and lack of immunogenicity (see also Part
VI, Chapters 6, 7, and 8). Their disadvan-
tages may be the lower transfection effi-
ciency and the transient nature of transfec-
tion compared to that of viral vectors. Cur-
rently, one of the most important obstacles
to viral and nonviral gene delivery systems
is the lack of organ and cell specificity (see
also Part VI, Chapter 5).
In order to overcome the drawbacks of
viral vector systems and previously investi-
gated nonviral systems, researchers fo-
cused their interest on microbubbles at
the end of the last century. Unger et al. de-
veloped an USCA for gene delivery, where
the gene material is entrapped in the cen-
ter of the microbubbles [68]. These micro-
bubbles are stabilized by a so-called soft
shell consisting of a lipid bilayer. The first
in vivo studies have shown that ultrasound
treatment of the heart of rabbits led to
gene expression after i.v. administration of
encapsulated pDNA containing a marker
gene [69]. Shohet et al. [70] demonstrated
that albumin-stabilized microbubbles
coated with adenovirus direct transgene ex-
pression to the myocardium of rats after
their treatment with diagnostic ultrasound.
Additionally, diagnostic ultrasound can
even promote and enhance released pDNA
uptake by surrounding cells [71, 72] (e.g.,
Sonoporation
TM
does not even use micro-
bubbles at all).
Overall, the major advantages seen in
using USCAs as a gene delivery system
are:
(1) Protection of the gene material in the
blood stream against nuclease degra-
dation.
(2) Site-specific and spatiotemporally con-
trolled delivery of gene material.
(3) Monitoring of the release procedure
(SAE).
(4) Use of the ultrasound-mediated sono-
poration effect.
Most of the published reports on ultra-
sound-facilitated gene delivery experiments
were focused on finding an effective meth-
od for the genetic treatment of ischemic dis-
eases, mainly by intramuscular injection of
DNA and/or microbubbles [7378]. A re-
view on the use of USCAs for gene delivery
in cardiovascular medicine has been pub-
lished recently by Bekeredjian et al. [79].
Only a very few researchers are investigat-
ing this approach for tumor gene therapy
using marker genes that do not exert a ther-
apeutic effect such as b-galactosidase and
luciferase, mainly by using intratumoral
substance administration [80, 81].
More recently, pDNA gas-filled micro-
capsule constructs have been investigated
in combination with diagnostic ultrasound
for ultrasound-mediated tumor gene deliv-
ery. Investigations have been performed in
different rodent tumor models [82].
The authors used the model plasmid
pUT651 containing the Escherichia coli
LacZ gene for b-galactosidase to demon-
strate the feasibility in CC531 liver tumors
of rats. The pUT651-containing USCAs
were administered in a first preliminary
experiment intra-arterially and in a second
experiment i.v. with simultaneous sonica-
tion of the tumors, both in a small num-
ber of animals.
Furthermore, the potential medical im-
pact of this delivery system has been
tested in Capan-1 (human pancreas adeno-
carcinoma)-bearing nude mice, using the
plasmid pRC/CMV-p16 (containing the tu-
mor suppressor gene p16). The tumor
suppressor gene p16, which plays an im-
portant role in anoikis, is deleted in Ca-
pan-1 cells. The outcome of this study is
summarized based on immunohistochem-
ical evaluation and tumor volume dou-
bling time in Figs. 7.13 and 7.14, respec-
tively.
As a result, a clear expression of the
pDNA was found in tumors of rats treated
with a combination of pUT651containing
USCA and ultrasound, while relevant con-
trols showed a significantly lower expres-
sion of the marker gene. Additionally, the
therapeutic effect of p16 was measured as
an increase of the tumor volume doubling
time. The controlled ultrasound-triggered
release of the expression vector for the tu-
mor suppressor gene p16 from USCAs
leads to a strongly significant (P <0.002)
tumor growth inhibition in comparison
with controls. This is the first time that a
reliable tumor treatment effect was shown
after i.v. injection of p16-containing USCA,
and site- and time-controlled DNA release
using a common diagnostic ultrasound de-
vice. In conclusion, it could be demon-
strated that the combination of gene-mate-
rial-carrying USCA and ultrasound pro-
vides a novel and effective approach for
nonviral gene therapy-based cancer treat-
ment.
A further improvement in this advanced
drug delivery approach can be expected by
the ability to target microbubbles via tar-
geting moieties to different tissues or re-
gions of disease (see also Part V, Chapters
2 and 6). This would allow site-specific de-
livery and release of therapeutic agents di-
rectly on the surface of the relevant cells.
While a bubble increases and subsequently
implodes within a very short time, the im-
plosion results in so-called liquid jet
streams at up to 400 km h
1
. The jets hit
against the surface of the cell and make
its membrane reversible leaky for a very
short period of time. Utilizing these jet
streams, the drug can be shot like a bullet
into the cell, which again underlines the
gene gun character of the microcapsules.
Furthermore, magic bullet-like proper-
ties can be derived from several attempts
at manipulation. Dayton et al. [83] have
shown that primary and secondary ra-
diation force can be used to manipulate
flowing contrast agents by ultrasound. Pri-
mary radiation forces may be used to di-
rect bubbles to a specific region of the ves-
7.7 Ultrasound, Microbubbles and Gene Delivery: Noninvasive Micro-Gene Guns 1319
sel wall where the flow velocity is lower.
Bubbles can even be stopped in a flow by
primary radiation forces at certain acoustic
pressures and high pulse repetition fre-
quencies [84]. Additionally, secondary ra-
diation force may be used to form aggre-
gates of bubbles in the location to be
treated. Finally, a SAE pulse induces the
release of encapsulated and/or adsorbed
drugs. In all the cases described, the
amount of released drug can be monitored
with SPAQ due to the fact that the concen-
tration of the released drug is proportional
to the encapsulation efficiency and num-
ber of ruptured microcapsules.
7.8
Summary: Ultrasound Theranostics Building
a Bridge between Therapy and Diagnosis
Imagine you are a 30-year-old female. You
take a blood test and receive a DVD that
contains your genetic make-up. Using a
Fig. 7.13 b-Galactosidase staining in CC531 liver
tumors to evaluate the gene transfection effi-
ciency. (AC) Beta-galactosidase activity after in-
tra-arterial administration and (D) after i.v. sub-
stance administration. As indicated by the arrows,
poor activity was seen with naked pUT651 and no
ultrasound treatment (A), moderate activity was
seen with naked pUT651 and ultrasound treat-
ment (B), and strong activity was seen with en-
capsulated pUT651 and ultrasound treatment (C).
(D) Strong activity of b-galactosidase 3 days after
i.v. administration of encapsulated pUT651 and ul-
trasound treatment. (Reproduced with permission.
2005 Radiology Society of North America, Inc.)
web service, you find that you are predis-
posed to an aggressive form of breast can-
cer. Your primary care physician refers you
for an ultrasound exam that uses a new
angiogenesis target-specific contrast agent.
It shows signs that new microvasculature
is being created in the left breast, charac-
teristic for early tumor growth.
A follow-up exam using a mixture of
new ultrasound contrast agents, which are
suitable for multiplexing the evidence of
several targets at the same time, confirms
based on a quantitative finding that a cou-
ple of specific proteins are overexpressed,
fueling aggressive tumor growth. Thus,
the malignancy is proven by a quantitative
evaluation on a concert of tumor-specific
biomarkers.
A therapeutic drug encapsulated in gas-
filled microcapsules targeted to one of the
most overexpressed vascularly accessible
protein is administered and released di-
rectly at the tumor, triggered from outside
of the body via ultrasound. The amount of
released drug can be monitored due to the
visibility of each single disrupting micro-
capsule and the knowledge of its drug
loading. After a sufficient concentration of
therapeutic has reached the target, the
treatment will be stopped.
Finally, a follow-up ultrasound exam
with target-specific USCAs confirms that
the overexpression is under control and
the vasculature creation has returned to
normal. [Scenario adapted from GE-
Healthcare [85] 2005.]
Yes, this sounds like science fiction, but
the use of SPECT or PET/CT (and some
early magnetic resonance imaging applica-
tions) (see also Part V, Chapters 4 and 5)
7.8 Summary: Ultrasound Theranostics Building a Bridge between Therapy and Diagnosis 1321
Fig. 7.14 Gene therapy effects confirmed by the
tumor volume doubling times. Evaluation intervals
of the tumor treatment study over three intervals.
Tumor volume doubling times of each group are
presented for each evaluation interval. T=treat-
ment group (encapsulated pDNAp16 with ultra-
sound treatment), C1=control 1 (encapsulated
pDNAp16 without ultrasound treatment),
C2=control 2 (encapsulated pDNAneo with ultra-
sound treatment), C3=control 3 (0.9% NaCl with
ultrasound treatment). (Reproduced with permis-
sion. 2005 Radiology Society of North America,
Inc.)
in combination with biochip-related diag-
nostic tools (see also Part I, Chapter 3 and
Part V, Chapter 8) and highly specific anti-
body-based radiodiagnostics and radiother-
apeutics (see also Part V, Chapters 2 and
6) makes this Find, Fight and Follow!
scenario close to clinical practice. Today,
interdisciplinary teams are professionally
developing tomorrows technology plat-
forms for ex vivo diagnostic, molecular
imaging and molecular medicine, and that
is the key exploiting the synergies be-
tween different disciplines of life sciences!
Although the largest number of target-
ing agents is available in the field of radio-
pharmaceuticals (and more targeted agents
are in the approval process), the main bot-
tleneck in an integrative theranostics
approach for clinical application is the lack
of available and approved targeting con-
trast agents and advanced drug delivery
systems.
The role of microbubbles in the Find,
Fight and Follow! triad is not near a clini-
cal routine as yet. However, preclinical re-
sults reported so far are more than pro-
mising.
State-of-the-art ultrasound is a high-reso-
lution imaging modality. Ultrasound is a
safe and well-established tool to image mor-
phology, and provides information about
diseases at the anatomical level. In combi-
nation with contrast agents and the SEA-
based sensitive particle acoustic quantifica-
tion technique, findings can be confirmed
based on actual figures. The concentration
of microbubble contrast agent in the region
of interest can be quantified with outstand-
ing sensitivity (down to single contrast
agent signals), and give maximum preci-
sion in an automatic scan process that
works independently of the examiner and
is highly reproducible. Recent progress in
the manufacture of contrast agents enables
us to design gas-filled microcapsules with
characteristic acoustic properties and a cer-
tain critical fragility, and enables the cou-
pling of targeting moieties. Target-specific
USCAs have proven their suitability for mo-
lecular imaging in many indications, and
can selectively track down molecules and
cell structures to be able to establish proof
of disease at a very early stage. Furthermore,
target molecules and therapeutic effects can
be visualized and quantified to make deci-
sions on highly individualized treatment
(see also Part I, Chapter 2). Considering
the strong sensitivity of ultrasound to small
variations of the structure of USCAs, even
multiplexing of microcapsules due to cer-
tain resonance behavior or critical fragility
is possible.
Additionally, the USCA can act as a con-
tainer to deliver drugs like small mole-
cules and biopharmaceuticals, and even
gene delivery is possible. Finally, con-
trolled release regarding site, time and
amount of drug can be realized and visual-
ized.
We think that the described technique
will be first established as a molecular
imaging-based drug discovery and develop-
ment tool in preclinical practice. However,
possible clinical applications of the tech-
netium of ultrasound can be foreseen in
the safe and less-invasive ultrasound iden-
tification of certain tissue (e.g., lymph
nodes), inflammatory regions or the sensi-
tive detection of small lesions. Also, as
stated before, cardiovascular indications
are conceivable due to more vascular,
blood-pool character of the USCAs.
The described advanced drug delivery
qualifications are being investigated in pre-
clinical settings, but considering that poly-
meric nano- and microparticle drug deliv-
ery systems are currently under investiga-
tion in several clinical trials, the step to ul-
trasound contrast-based drug delivery is
just a small one.
The cover of Modern Biopharmaceuticals
shows a sketch of a nanorobot working
within the blood stream. The artist in-
spired by Asimov possibly thought about
the diagnostic and therapeutic capabilities
of Ehrlichs vision of the magic bullet.
We believe that the ultrasound theranos-
tics concept is currently one of the most
promising candidates to realize Paul Ehr-
lichs vision in the new millennium.
References
1 Adapted from: Moby Dick by Herman Mel-
ville; Ch. 134. Penguin Modern Classics, Lon-
don, originally 1851.
2 Kaminski MS, Zelenetz AD, Press OW, et al. J
Clin Oncol 19, 39183928, 2001.
3 Wiseman GA, White CA, Stabin M, et al. Eur
J Nucl Med 27, 766777, 2000.
4 Vesselle H, Vesselle H, Grierson J, et al. Clin.
Cancer Res 8, 33153323, 2002.
5 Drevs J, Muller-Driver R, Wittig C, et al. Can-
cer Res. 62, 40154022, 2002.
6 Wells PNT. Biomedical Ultrasonics. Academic
Press, London, 1977.
7 Tachibana K, Tachibana S. Echocardiography
18, 323328, 2001.
8 Nelson TR, Pretorius DH. Ultrasound Med Biol
24, 12431270, 1998.
9 Ophir J, Parker KJ. Ultrasound Med Biol 15,
319333, 1989.
10 Lord Rayleigh (Strutt JW). Philos Mag 34, 94
98, 1917.
11 de Jong N, Hoff L, Skotland T, Bom N, Ultra-
sonics 30, 95103, 1992.
12 de Jong N, Hoff L, Ultrasonics 31, 175181,
1993.
13 Hoff L. Ultrasonics 34, 591593, 1996.
14 Frinkink PJA, de Jong N. Ultrasound Med Biol
24, 523533, 1998.
15 Briel A. J Pharm Pharmacol 56 (Suppl), 288,
2004.
16 Budde U, Briel A, Rling G, et al. Patent
DE19925311.0, 1999 (and US 6652782 B1).
17 Reinhardt M, Fritzsch T, Heldmann D, Siegert
J, Patent WO 93/25241, 1993.
18 Uhlendorf V, Hoffmann C. Proc IEEE Ultraso-
nics Symp 3, Cannes, France, 15591562,
1994.
19 Uhlendorf V, Scholle FD, Reinhardt M. Ultra-
sonics 38, 8186, 2000.
20 Tiemann K, Pohl C, Schlosser T, et al. Ultra-
sound Med Biol 26, 11611167, 2000.
21 Blomley M, Albrecht T, Cosgrove D, et al.
Radiology 210, 409416, 1999.
22 Hauff P, Fritzsch T, Reinhardt M, et al. Invest
Radiol 32, 9499, 1997.
23 Reinhardt M, Hauff P, Briel A, Schirner M.
Proc 3th Annu Meet Soc Molecular Imaging, St.
Louis, MO, p. 148, 2004.
24 Reinhardt M, Hauff P, Briel A, et al. Invest
Radiol 40, 27, 2005.
25 Hermanson GT. Bioconjugates. Academic
Press, New York, 1996.
26 Briel A, Rling G, Hauff P, et al. Patent: DE
10013850.0, 2000 (and WO 01/68150).
27 Joseph S, Olbrich C, Kirsch J, Hasbach M, Briel
A, Schirner M, Pharm Res 21, 920926, 2004.
28 Takalkara AM, Klibanova AL, Rychaka JJ,
Lindner JR, Leya K, J. Controlled Release 96,
473482, 2004.
29 Joseph S, Schlehnsog M, Groenewold T,
Quandt M, Olbrich C, Briel A, Schirner M,
Biosens Bioelectron 20, 18291835, 2005.
30 Olbrich C, Mossmayer D, Willuda J, Schmidt
W, Schirner M. Presented at Controlled Re-
lease Society, Miami, FL, 2005.
31 Lanza GM, Wallace KD, Scott MJ, et al. Circu-
lation 94, 33343340, 1996.
32 Klibanov AL, Hughes MS, Marsh JN, et al.
Acta Radiol. 412 (Suppl), 113120, 1997.
33 Unger E, Metzger P, Krupinski E, et al. Invest
Radiol 35, 8689, 2000.
34 Lindner JR, Song J, Christiansen J, et al. Cir-
culation 104, 21072112, 2001.
35 Leong-Poi H, Christiansen J, Klibanov AL, et
al. Circulation 107, 455460, 2003.
36 Ellegala DB, Leong-Poi H, Carpenter JE, et al.
Circulation 108, 336341, 2003.
37 Hauff P, Reinhardt M, Briel A, et al. Radiology
231, 667673, 2004.
38 Murer M, Linker R, Hauff P, Reinhardt M,
Briel A, et al. J Neurol 249 (Suppl 1), I/57,
2002.
39 Hauff P, Reinhardt M, Murer M, Linker R,
Briel A, Schirner M. Proc 3rd Annu Meet Soc
for Molecular Imaging, St. Louis, MO, p. 150,
2004.
40 Reinhardt M, Hauff P, Briel A, Linker RA,
Gold R, Rieckmann P, Becker G, Toyka K,
Murer M, Schirner M. Neuroimage, in press
2005.
References 1323
41 Gold R, Hartung H-P, Toyka KV. Mol Med To-
day 6, 8891, 2000.
42 Archelos JJ, Hartung H-P. Mol Med Today 3,
310321, 1997.
43 Hauff P, Reinhardt M, Murer M, Linker R,
Briel A, Toyka KV, Schirner M. Presented at
6th Int Symp on Ultrasound Contrast Imag-
ing, Tokyo, 2004.
44 Birchler M, Neri G, Tarli L, Halin C, Viti F,
Neri D, J Immunol Methods 231, 239248,
1999.
45 Tarli L, Balza E, Viti F, et al. Blood 94, 192
198, 1999.
46 Viti F, Tarli L, Giovannoni L, Zardi L, Neri D.
Cancer Res 59, 347352, 1999.
47 Santimaria M, Moscatelli G, Viale GL, et al.
Clin Cancer Res 9, 571579, 2003.
48 Neri D, Carnemolla B, Nissim A, et al. Nat
Biotechnol 15, 12711275, 1997.
49 Lich K, J Pharm Pharmacol 56 (Suppl), 272,
2004.
50 Licha K, Perlitz C, Hauff P, et al. Proc Soc Mol
Imaging 264, 2004.
51 Dieterle W (Director). Dr. Ehrlichs Magic Bul-
let [A movie about Paul Ehrlichs live. ]. War-
ner Bros, Los Angeles, CA, 1940.
52 Stein M, Heldmann D, Fritzsch T, Siegert J,
Rling G. Patent US 6264959 B1 based on
application 07/536373, 1990.
53 Weitschies W, Heldmann D, Hauff P, Fritzsch
T, Stahl H. Patent US 6068857 based on appli-
cation 08/605174, 1994.
54 Ishihara K. Proc 2nd Int Micromach Symp, 69
75, 1996.
55 Unger E, et al. Patent US 5542935, 1996.
56 Frinking PJA, Bouakez A, de Jong N, ten Cate
FJ, Keating S. Ultrasonic 96, 709712, 1998.
57 Seemann S, Hauff P, Schultze-Mosgau M,
Lehmann C, Reszka R. Pharm Res 19, 250
257, 2002.
58 Antonietti M, Conrad J, Thnemann AF, Mac-
romolecules 27, 6007, 1994.
59 Hart SL. Exp Opin Ther Patents 10, 199208,
2000.
60 Williams RS, Johnston SA, Riedy M, DeVit
MJ, McElligott SG, Sanford JC. Proc Natl Acad
Sci USA 88, 27262730, 1991.
61 Kawabata K, Takkura Y, Hashida M, Pharm
Res 12, 825830, 1995.
62 Yoshida M, Mahato RI, Kawabata K, Takakura
Y, Hashida M. Pharm Res 13, 599603, 1996.
63 Matsuse T, Teramoto S. Curr Ther Res Clin
Exp 61, 422434, 2000.
64 Kurian KM, Watson CJ, Wyllie AH. J Clin
Pathol, Mol Pathol 53, 173176, 2000.
65 Delfine P, Guillaume C, Floch V, et al. J
Pharm Sci 89, 629638, 2000.
66 Simoes S, Slepuskin V, Pires P, Gaspar R,
de Lima MCP, Duzgunes N. Biochim Biophys
Acta Biomembr 1463, 459469, 2000.
67 Jones DH, Corris S, McDonald S, Clegg JCS,
Farrar GA. Vaccine 15, 814817, 1997.
68 Unger EC, Hersh E, Vannan M, McCreery T.
Echocardiography 18, 355361, 2001.
69 Unger EC, McCreery TP, Schweitzer R. Pre-
sented at Macromolecular Drug Delivery Conf,
Breckenridge, CO, p. 20 (abstr), 1999.
70 Shohet RV, Chen S, Zhou Y-T, et al. Circula-
tion 101, 25542556, 2000.
71 Anwer K, Kao G, Proctor B, et al. Gene Ther 7,
18331839, 2000.
72 Bao S, Thrall BD, Miller DL. Ultrasound Med
Biol 23, 953959, 1997.
73 Taniyama Y, Tachibana K, Hiraoka K, et al.
Gene Ther 9, 372380, 2002.
74 Schratzberger P, Krainin JG, Schratzberger G,
et al. Mol Ther 6, 576583, 2002.
75 Li T, Tachibana K, Kuroki M, Kuroki M. Radi-
ology 229, 423428, 2003.
76 Lu QL, Liang HD, Partridge T, Blomley MJ.
Gene Ther 10, 396405, 2003.
77 Yamashita Y, Shimada M, Tachibana K, et al.
Hum Gene Ther 13, 20792084, 2002.
78 Bekeredjian R, Chen S, Frenkel PA, Grayburn
PA, Shohet RV. Circulation 108, 10221026,
2003.
79 Bekeredjian R, Grayburn PA, Shohet RV. J
Am Coll Cardiol 45, 329335, 2005.
80 Miller DL, Bao S, Gies RA, Thrall BD. Ultra-
sound Med Biol 25, 14251430, 1999.
81 Manome Y, Nakamura M, Ohno T, Furuhata
H. Hum Gene Ther 11, 15211528, 2000.
82 Hauff P, Seemann S, Reszka R, et al. Radiol-
ogy, in press 2005.
83 Dayton PA, Morgan KA, Klibanov AL, Bran-
denburger GH, Nightingale K, Ferrara KW.
IEEE Trans Ultrason Ferr Freq Con 44, 1264
1277, 1997.
84 Dayton PA, Klibanov AL, Brandenburger GH,
Ferrara KW. Ultras Med Biol 25, 11951201,
1999.
85 Scenario adapted from GE-Healthcare (2005).
http://egems.gehealthcare.com/geCommu-
nity/nmpet/interest_groups/molecular_imag-
ing/education/intro.jsp
Abstract
The earliest possible diagnosis of a disease
is imperative for an efficient therapy. This
simple fact is true for virtually any type of
disease, and is best documented for tumor-
ous conditions. At present, the single pro-
tein markers used exhibit high false-nega-
tive and/or false-positive rates, and the hope
for more reliable DNA- and RNA-based
screening tools has not been fulfilled. As a
consequence, research activities have
turned back to proteins as the real key
players in pathological processes. Over the
past few years, comparative protein profil-
ing by surface-enhanced laser desorption/
ionization time-of-flight mass spectrometry
(SELDI TOF-MS) has become acknowl-
edged as a promising means of detecting
specific and predictive protein patterns that
reflect certain stages of cancer, neurological
disorders, and infectious diseases. The SEL-
DI-based ProteinChip
System fulfils the

needs of the new proteomic era by enabling
comparative protein profiling by Expression
Difference Mapping analysis of several hun-
dreds of samples per day on a single, tech-
nology platform, using software support
for the construction of multi-marker predic-
tive models. The Interaction Discovery Map-
ping
TM
platform is introduced as the next
methodical step for antibody-based assays
and investigations into protein binding
partners of possible importance in the diag-
nosis of conditions and the development of
biopharmaceuticals.
Abbreviations
AFP alpha-fetoprotein
EAM energy-absorbing mole-
cules
HCG human chorionic gonado-
tropin
HPV human papillomavirus
HR-PAC host response protein
amplification cascade
1325
ISBN: 3-527-31184-X
8
Development of Multi-marker-based Diagnostic Assays
with the ProteinChip
System
Andreas Wiesner
Getting Insight Sense the Urgency for Early Diagnostics
ICAT isotope-coded affinity tag
ID identification
ITIH-IV inter alpha trypsin inhibi-
tor IV
LDH lactate dehydrogenase
NSE neuron-specific enolase
PAP prostatic acid phospha-
tase
PCA principal components
analysis
POMC pro-opiomelanocortin
PSA prostate-specific antigen
ROC receiver operating charac-
teristics
RT-PCR reverse transcription-poly-
merase chain reaction
SELDI TOF-MS surface-enhanced laser
desorption/ionization
time-of-flight mass
spectrometry
SEND surface-enhanced neat de-
sorption
TOF-MS time-of-flight mass spec-
trometry
VEGF vascular endothelial
growth factor
8.1
The Urgency of Earlier Diagnosis
The earlier a disease is diagnosed, the more
efficient a therapy can be. This is true for vir-
tually any type of disease, including infec-
tious conditions [1], Alzheimers disease as
well as other dementias [2, 3], and is best
documented for tumorous diseases [4],
where the current compilation of interna-
tionally acknowledged markers in human
blood [5] contains not more than a handful
of proteins: Prostate-specific antigen (PSA),
prostatic acid phosphatase (PAP), CA 125,
carcinoembryonic antigen (CEA), alpha-feto-
protein (AFP), human chorionic gonadotro-
pin (HCG), CA 19-9, CA 15-3, CA 27-29, lac-
tate dehydrogenase (LDH), and neuron-spe-
cific enolase (NSE). Most of these markers
were first described long ago PSA in
1971, CA 125 in 1981, CEA in 1955, AFP
in 1956, HCG in 1927, and CA 15-3 in
1984 [6] and are mainly used for prognostic
purposes. Applications in tumor diagnosis
are very limited, especially for the early can-
cer stages [4], where the success rates are dis-
appointingly low [7]. The survival rates of tu-
mor patients diagnosed late in disease pro-
gression and suffering from regionally or
distantly spread tumors have shown little
improvement over the past 30 years. Despite
huge research efforts, only a few targeted
therapeutics, such as Herceptin
(trastuzu-
mab), which is directed against the HER-2
receptor of breast cancer cells [8, 9] (see Part
I, Chapter 5) are in use and in general, non-
specific cytotoxic agents are administered
with limited success (see Part V, Chapter
6). It is clear that only an earlier diagnosis
could improve the situation, and this will
be certainly easier to achieve than the devel-
opment of better targeted methods for the
treatment of advanced-stage cancers.
The early detection of cervical cancer is an
encouraging example for the beneficial re-
sult of early diagnosis by less invasive
screening methods [4]. A cytological assay
to detect the very early stages of cervical can-
cer was developed in the mid-twentieth cen-
tury, and countries that included the pap
smear assay in their national screening
programs were able drastically to reduce
the incidence of this disease. Later, when
human papillomaviruses (HPVs) were rec-
ognized as causing the cancer by persistent
infection, augmentation of the pap smear
with molecular tests for HPV detection re-
sulted in a powerful tool for early diagnosis.
Many more such positive examples are ur-
gently needed to save more lives, and to
earn back the patients trust in the benefits
of diagnostics and healthcare [10].
System 1326
8.2
Proteins are Best Choice Again
During the 1980s, the disappointments of
single protein markers forced the greater
part of the scientific community to turn to
the field of genomics, with the aim being
to develop more efficient diagnostic tools
[11]. In most cases, reverse transcription-
polymerase chain reaction (RT-PCR) was
used to detect tumor-specific DNA altera-
tions, cell-free and mitochondrial tumor
DNA, tumor-associated viral DNA, and tu-
mor RNA. Due to the high but not always
error-free amplification of minute
amounts of DNA or RNA, as well as the dif-
ficulty of predicting the stability of RNA,
false-positive results were rather common.
Furthermore, data reproducibility was diffi-
cult to achieve, and results from different
laboratories were seldom comparable. In
addition, the development of microsatellite
analysis of tumor DNA could not be suc-
cessfully established in the diagnosis of
bladder, lung or colorectal cancers. The de-
tection of tumor-associated DNA hyper-
methylation is another methodical PCR
variation which has shown promising re-
sults, but has not been established in the
clinical area, mainly because the levels of
false-positive and/or false-negative results
were too high [12]. Furthermore, interpreta-
tion of the data is often difficult. For exam-
ple, DNA is released by not only malignant
but also benign cells; hypermethylation is
not only tumor-specific but is also age-re-
lated; consequently, all of the reported sen-
sitivity and/or specificity values require ma-
jor improvements to be made in order for
the test to be useful [1317]. Taken together,
improved nucleic acid-based methods will
certainly play some role in the future of tu-
mor diagnostics, but they are far from ful-
filling the promise shown at the beginning
of the genomic era.
Without doubt, one of the main reasons
for this lack of success by genomic meth-
ods is the fact that, on the DNA and RNA
level, we are too far away from the actual
physiologically active gene products, the
proteins. The statement . . . DNA makes
RNA makes protein. . . is simply wrong
and misleading as it implies a carbon-
copy-like process for the information path-
way from DNA to messenger RNA and
the final protein product. The statement
even entitled a slide in the 1993 Nobel lec-
ture of Phillip A. Sharp, who received the
prize for the detection of split genes and
RNA splicing. He described the process as
similar to . . . the work that a film editor
performs: the unedited film is scrutinized,
the superfluous parts are cut out and the
remaining ones are joined to form the
completed film. In such a way the fibro-
nectin gene can code for up to 20 different
fibronectin protein variants with different
locations and functions [18].
However, even when the final sequence
and quantity of mRNA is known, it is often
not possible to predict the protein composi-
tion because of a poor correlation between
mRNA and protein expression levels [19
21]. Furthermore, the exact knowledge of
the finally processed m-RNA sequence will
not help to predict the protein sequence.
Many proteins are additionally spliced into
smaller pieces before becoming biologically
active [22]. For example, this phenomenon
is well known for pro-opiomelanocortin
(POMC)-derived peptides in the hypothala-
mus [23], where a 32 kDa precursor protein
can be spliced into several different hor-
mones, each with a different range of activ-
ity. Thus, the primary sequence of the pro-
tein does not depend on the DNA sequence
or even the m-RNA sequence alone. Vir-
tually any protein can be subject to a num-
ber of post-translational modifications such
as phosphorylation, acylation, glycosylation
and/or other types of covalent alterations.
Many of these modifications are reversible,
and determine the localization, structure
and binding behavior of the protein, in turn
resulting in other sophisticated regulatory
activities of possible pathological impor-
tance [24].
Consequently, DNA tells what could
happen, and RNA what might happen
but only proteins can tell what actually
happens. Thus, real progress in the devel-
opment of new diagnostic tools and thera-
peutic strategies will depend mainly on
the new insights coming from investiga-
tions into proteins, their local concentra-
tion, structure, function, and interaction
[25, 26].
8.3
Current Tools
for Protein Biomarker Detection
On the basis of mRNA processing and
post-translational modifications, the esti-
mated number of different human pro-
teins is about 500000 more than 15
times higher than the estimated number
of coding genes in humans. Although only
a fraction of these proteins are present at
any given time in any particular location,
it will be a special challenge to catalogue
their structures and functions, to describe
their actual concentrations differing by or-
ders of magnitude, and to detect their
crosstalk-activities in complexes [2729].
Identifying, cataloging and functionally de-
scribing all human proteins will be infi-
nitely more challenging than it was to se-
quence the human genome [30].
On the other hand, realizing that the
overall protein pattern of tissues or body
fluids of a given individual is very similar
to that of others from the same species
(and is even more similar when indivi-
duals are of the same gender, age, and
physiological state), the situation appears
less complicated from the diagnostic point
of view [31]. With regard to biomarker dis-
covery, it is more promising to focus on
the differences between individual protein
patterns than to try to achieve a complete
protein mapping of all the components.
Learning more about which proteins are
unique to a certain physiological state will
allow new diagnostic assays to be devel-
oped, and to target the search for new bio-
pharmaceuticals [32].
In analytical terms, ease of use and reli-
able instrumentation with quantitative cap-
ability and high sample throughput, com-
bined with appropriate software tools for
handling the data output, are needed. Only
then can comparative profiling studies
lead to the discovery and validation of new
biomarkers. The system of two dimen-
sional polyacrylamide gel electrophoresis
(2D-PAGE) where proteins are separated
first by their isoelectric point and subse-
quently by their molecular weight indeed
contributed greatly to our understanding
of the wide variety of proteins in a given
sample. However, proteins in the peptide
range as well as those of high hydrophobi-
city or of extreme isoelectric points are dif-
ficult to separate, are thus typically ne-
glected, and result in a loss of potentially
interesting proteins. Furthermore, the 2D-
PAGE approach is time-consuming, diffi-
cult to standardize, and is in no way suited
to the rapid screening of higher sample
numbers, not even in the hands of experts
in this method [33, 34]. Similarly, column
chromatography in all its variations is
well suited for purification purposes, but
not for the efficient and reproducible com-
parative analysis of dozens or hundreds of
samples per day. More recently, an isotope-
coded affinity tag (ICAT) procedure was
developed to overcome the difficulties in
System 1328
protein quantitation [35]. Here, two protein
samples are differentially labeled on their
cysteine residues, whereby the labeling
molecules contain different isotopes. The
proteins are then cleaved proteolytically,
undergo HPLC-separation, and are finally
analyzed using mass spectrometry, when
the differently labeled peptides can be rec-
ognized by a specific shift in mass.
Although this is a quite sophisticated
approach, there are several practical draw-
backs, caused mainly by incomplete label-
ing or artifact generation. Therefore, tag-
less extraction-retentate chromatography
methods have been considered to be a
more promising alternative [36].
Protein micro-arrays equipped with spe-
cific capture molecules such as antibodies,
fragments thereof, peptides, or other bait
molecules have been developed in recent
years [37, 38]. Here, the challenges lay not
only in the generation of libraries of cap-
ture molecules that are able to mirror the
variety of proteins (and the slightly modi-
fied versions of each single type!) in a giv-
en sample, but also in the experimental
design where all problems known to be
associated with protein interaction experi-
ments and labeling procedures must be
taken into account. This task is much
more difficult to accomplish than it was
with DNA microarrays, and it will take
many more years before those types of mi-
cro-arrays are available as robust, afford-
able tools, and unpredicted screening will
be very difficult to achieve [39].
8.4
The ProteinChip
System at a Glance
The ProteinChip System was developed by
Ciphergen Biosystems, Inc. (www.cipher-
gen.com) to meet the demands of assay
8.4 The ProteinChip

Fig. 8.1 The different types of ProteinChip Arrays.
The spots on the chromatographic ProteinChip Ar-
rays possess hydrophobic, cationic, anionic, metal
ion affinity, or hydrophilic properties. These
chemical surfaces are best suited for protein ex-
pression profiling studies. Another series of Pro-
teinChip Arrays have pre-activated biological sur-
faces designed for coupling of biomolecules onto
the spots with applications in antibodyantigen
assays, receptorligand interaction studies and
DNA-protein binding experiments.
development in the post-genomic era.
Based on the patented SELDI process [40
42], it combines two formerly well-estab-
lished methods solid-phase chromatogra-
phy and time-of-flight mass spectrometry
(TOF-MS) into one easy-to-use system
for carrying out Expression Difference
Mapping applications on a single inte-
grated platform [4346].
The basic procedure for Expression Dif-
ference Mapping studies with the Pro-
teinChip System is quite simple. Virtually
any type of protein-containing solution can
be applied to the spots of the ProteinChip
Arrays. These spots present either chroma-
tographic surfaces with certain physico-
chemical characteristics (hydrophobic, cat-
ionic, anionic, metal ion-presenting or hy-
drophilic), or are pre-activated for the cou-
pling of capture molecules (protein, DNA,
or RNA) prior to sample loading (Fig. 8.1).
Typically, for Expression Difference Map-
ping experiments, chromatographic sur-
faces are used. The sample requirement is
low (110 lg total protein per spot), and
the sample volume can be freely chosen
from 0.5 lL up to around 300 lL. After a
short incubation period, unbound proteins
and any contaminants on the spot surface
are washed away. A solution of energy-ab-
sorbing molecules is applied to each of the
spots, and the ProteinChip Array is ready
for the analysis in the ProteinChip Reader,
a highly sensitive laser desorption/ioniza-
tion TOF-MS instrument. The results are
initially visualized in a graph, with the mass
System 1330
Fig. 8.2 Sample analysis in the ProteinChip Read-
er. After loading the ProteinChip Array into the
ProteinChip Reader, a laser beam is directed onto
the sample on the spot. Thereby protons are
transferred onto the peptides and proteins that
are subsequently accelerated by electromagnetic
fields through a flight tube. The time-of-flight is
inversely proportional to the molecular mass. This
results in a spectrum being generated where the
molecules in the sample are represented in a
graph with the mass to charge ratio of the native
compounds on the x-axis and the corresponding
signal intensity on the y-axis.
8.4 The ProteinChip

Fig. 8.3 The four steps of sample analysis. Despite covering a wide range of application areas,
the basic methodical procedure of sample analysis with the ProteinChip System is very similar
in all the different types of experiments.
to charge ratio of the sample components
plotted on the x-axis, and the corresponding
signal intensities plotted on the y-axis
(Fig. 8.2). ProteinChip Software enables
the user to control and influence the auto-
mated detection process, and also incorpo-
rates a wide range of univariate and multi-
variate software tools for the comparative
analysis of higher numbers of samples.
Thus, the entire Expression Difference
Mapping technique can be achieved in a
fast, four-step procedure (Fig. 8.3).
The ProteinChip System Series 4000 En-
terprise AutoBiomarker Edition, comprised
a ProteinChip Reader with autoloader ca-
pacity to automatically run simulta-
neously up to 168 bar-code-labeled Pro-
teinChip Arrays with, in total, 1344 sam-
ples, a customized Biomek
2000 liquid
handling robot and CiphergenExpress
TM
Data Manager Software for automated
sample tracking and advanced data analy-
sis together with the Biomarker Pat-
terns
TM
Software for the direct develop-
System 1332
Fig. 8.4 Hardware components of the ProteinChip
System. The ProteinChip Arrays (1) are arranged
in bioprocessors to match the 96-well microtiter
plate format (2). Samples can be applied using a
customized Biomek
2000 liquid-handling robot

(3). The ProteinChip System Series 4000 Enter-
prise Edition (4) automatically runs up to 168 bar-
code labeled ProteinChip Arrays with, in sum,
1344 samples at once.
ment of diagnostic assays on the same
platform (Figs. 8.4 and 8.5).
8.5
Distinctions of the SELDI Process
As described in Section 8.4, only those
proteins which interact actively with the
spot surfaces are analyzed in the Pro-
teinChip Reader, since all other compo-
nents are washed off in advance. Such
washing provides one of the most obvious
advantages of the SELDI process that
certain components (e.g., salts or deter-
gents) which commonly cause problems
with other analytical tools are removed.
Additionally, as each analysis is also
linked to an on-spot fractionation step, the
complexity of the samples is reduced, thus
providing much higher chances for the de-
tection of lower concentrated markers.
Normally, a number of different array
types are combined with washing steps of
varying stringencies to retain different pro-
tein subsets from the original samples.
Equivalent arraybuffer combinations of
the different sample groups can then be
analyzed comparatively to decipher the for-
merly unknown markers. With complex
samples such as serum, pre-fractionation
is preferable as it enhances the total num-
ber of signals resolved. Subsequent bio-
marker purification is easier to achieve be-
cause, in addition to the exact molecular
Fig. 8.5 Software-enabled assay development. The
CiphergenExpress
TM
Software with Data Manager
and Biomarker Analysis Modules is a relational
database system with a client-server architecture
designed for automated sample tracking and ad-
vanced data analysis. After identification and se-
lection of clusters for meaningful data reduction,
univariate and multivariate data analysis tools are
used for the detection of single and multiple
markers. ROC (Receiver Operating Characteris-
tics), AUC (Area Under the Curve), and PCA (Prin-
cipal Components Analysis) calculations are im-
plemented together with appropriate data visuali-
zation formats such as cluster diagrams, ROC
plots, heat maps and 3-D projection of PCA data.
The Biomarker Patterns
TM
Software is the key
component for the direct development of multi-
marker-based diagnostic assays.
weight, the basic physico-chemical charac-
teristics of the protein of interest can be
elucidated from its binding behavior under
certain washing conditions. Last but not
least active interaction of the analyte
with the spot surface ensures another im-
portant feature of the ProteinChip technol-
ogy, namely its ability to quantify individu-
al protein levels in a given sample; this is
a prerequisite for the successful use of any
technology in comparative profiling. As
the proteins are equally distributed on the
spot surfaces and the laser positioning is
conducted very exactly, the resulting signal
intensities correspond very well to the con-
centration of the proteins (Fig. 8.6).
8.6
The Pattern Track
TM
Process: From Biomarker
Discovery to Assay Development
Many recent publications have provided
proof of the suitability of the ProteinChip
System in the discovery and validation of
new biomarkers for a wide range of dis-
eases. This is true for ovarian [47], prosta-
tic [48], pancreatic [49] and head and neck
[50] cancer, as well as for a large number
of other tumorous diseases [7]. Similarly,
successful studies have been reported for
Alzheimers disease [51], viral [52], bacteri-
al [53], and parasitic [54] infections. The
same applies to investigations about possi-
ble new markers to monitor the effect of
drug treatment [5557], and also to predict
transplant rejections [58, 59]. The Pro-
teinChip System is currently used by more
System 1334
Fig. 8.6 Quantitation capability. The ProteinChip
System allows the relative and absolute quantita-
tion of proteins. In this example, interleukin-8 (IL-
8) was spiked into human serum and captured on
an antibody-coated ProteinChip Array. After analy-
sis, the resulting signal intensities correspond
very well to the concentration of the proteins.
than 1000 customers worldwide in phar-
maceutical and biotechnology laboratories,
as well as in academic and governmental
institutes. As diverse as the research areas
might be, one result is common to most
of them: Multimarker panels have been
discovered and were found to be superior
over single marker-based traditional assays.
For the elucidation of such patterns, a spe-
cialized tool the Biomarker Patterns Soft-
ware has been developed. This super-
vised learning program performs multi-
variate analyses and is able to classify the
processed data by the use of specially
adapted algorithms [60]. It creates tree-like
structured decision diagrams by splitting
the original data set (parent node) in two
sub-groups (child nodes) of highest possi-
ble purity. Each child node then becomes
a parent node at the time of creation, and
can be the origin of a new split. The pro-
gram calculates which signals (i.e., pro-
teins) are best suited to act as splitters,
whereby the splitting rules define what in-
tensity (i.e., protein abundance) a signal
must have to be sorted into one of the
groups (Fig. 8.7). After the initial learning
period where the tree building takes place,
the prediction success of the resulting
model can be tested with new sets of data.
The Biomarker Patterns Software is a true
prediction tool, enabling the user to con-
duct multivariate analysis on a profes-
sional level without the need to be a spe-
cialist in statistical evaluations. In contrast
to other classifiers, such as neural net-
works or nearest-neighbor calculations, the
models are easier to understand and each
sample can be tracked down the tree [61].
Even with all these developments in
technology, the importance of well-de-
signed marker discovery studies should
not be forgotten. For example, in the past
the success of tumor marker studies has
often been hampered by uncertainties in
tumor grading [62] and a lack in generally
accepted guidelines for reporting the re-
sults [63]. Furthermore, questions about
the numbers needed to screen [64] and
how to define surrogate end points in can-
cer research [65] are still under discussion.
In other words, there exists an urgent
need for generally accepted rules in mar-
ker discovery to assure the best possible
outcome from the corresponding projects.
A detailed review addressing this topic can
be found elsewhere [66].
In the Pattern Track Process, only mar-
ker proteins of proven importance will be-
come identified. In most cases, the protein
of interest needs to be purified or at least
enriched for subsequent identification ex-
periments [e.g., 6770]. Mini-spin col-
umns, microplate-formatted chromato-
graphic devices or traditional preparative
scale columns are suitable for separation.
One of the advantages of the SELDI pro-
cess is that the best-suited chromato-
graphic material and the correct elution
conditions can be predicted from the bind-
ing behavior of the protein on the Pro-
teinChip Arrays [7173]. Optimized separa-
tion materials have been developed by Bio-
Sepra, the Process Division of Ciphergen
Biosystems, Inc. For identification, the pu-
rified or enriched protein is cleaved proteo-
lytically, and the masses of the resulting
peptide fragments are determined using
the ProteinChip Reader. Afterwards,
searches in public databases are used for
the identification by peptide mass finger
printing. For unequivocal identifications
and partial sequencing of the protein, the
ProteinChip Arrays can be read in high-re-
solution tandem-MS instruments [74]. For
this purpose, an exchangeable UV laser de-
sorption ion source, the ProteinChip Inter-
face PCI 1000, is directly linked to a tan-
dem-MS system (Tandem MS QSTAR
;
from Sciex-MDS/Applied Biosystems).
TM
Process: From Biomarker Discovery to Assay Development 1335
System 1336
Fig. 8.7 The Biomarker Patterns
TM
Software. This
supervised learning program performs multivari-
ate analyses and creates tree-like structured deci-
sion diagrams by splitting the original data set
into sub-groups of highest possible purity. The
program calculates which signals (i.e., proteins)
are best suited to act as splitters, and the splitting
rules define what intensity (i.e., protein abun-
dance) a signal must have to be sorted into one
of the groups. The resulting model can be tested
with new data sets from blinded samples to verify
the reliability of the model under realistic condi-
tions. In the example given, four markers are part
of the model. According to the rules, the spec-
trum from the left-hand side is recognized as de-
rived from a tumor patient, whereas the one on
the right-hand side represents the normal state.
Recently, a new type of ProteinChip Ar-
ray was commercialized to further facili-
tate the analysis of peptide mass finger-
prints by Surface-Enhanced Neat Desorp-
tion (SEND) technology for protein identi-
fication (ID). In SEND ID, the energy-
absorbing molecules (EAM) are incorpo-
rated in the surface chemistry of the array.
Thus, separate EAM addition is not
needed and the chemical noise introduced
by the application of EAM is significantly
reduced, enabling low molecular-weight
species to be detected with greatly im-
proved signal-to-noise ratio, whilst also
minimizing the variability associated with
matrix addition. The spot surface contains
a hydrophobic interaction functional group
to enable the on-spot clean-up of samples.
Most other approaches to this application
require clean-up of the sample in a small
chromatography column in addition to
pre-mixing with matrix. By avoiding these
steps with SEND, the work flow is simpli-
fied and the losses of sticky peptides are
minimized. In addition, the suppression
of interfering matrix signal results in supe-
rior sensitivity and improved sequence cov-
erage.
The knowledge of the marker identities
will not only enable to transfer the assay
on antibody-coupled ProteinChip Arrays
[66] but also will support a better under-
standing of the underlying biology.
Furthermore, by coupling the marker pro-
tein and capturing physiological binding
partners, secondary markers might be
found. Those interaction experiments can
be performed either with pre-activated af-
finity beads and elution on chromato-
graphic arrays, or directly on pre-activated
ProteinChip Arrays. The Interaction Dis-
covery Mapping platform complements
the earlier established Expression Differ-
ence Mapping application. As with the en-
tire ProteinChip System, both methods
profit from the successful combination of
chromatographic principles with mass
spectrometric detection methods.
8.7
Protein Variants as Disease Markers
Many of the studies cited above revealed a
very interesting phenomenon namely,
that a number of the markers belong to
quite common proteins known to be of
higher abundance in biological fluids. At
first glance it seemed difficult to recognize
them as specific indicators for a certain
disease. However, from the exact molecu-
lar weights and identities delivered by the
SELDI process, it became clear that these
common proteins exhibit truncations or
other types of covalent modifications
which make them different from the re-
spective corresponding proteins in the
other patient groups. The ability to detect
and to quantify those protein variants is
another advantage of the SELDI-process,
and one which could not be achieved by
any traditional approach.
For example, such variants are reported
for serum amyloid A in nasopharyngeal
cancer [75] and renal cancer [76], as well
as the vascular endothelial growth factor
(VEGF) in bronchial cancer and for apoli-
poprotein A1, transthyretin and inter alpha
trypsin inhibitor IV (ITIH-IV) in ovarian
cancer [47]. In contrast, in other areas of
research (e.g., for Alzheimers disease), the
importance of the different truncated ver-
sions of the amyloid-beta peptides has al-
ready been recognized and used to moni-
tor the pathology of the disease [77]. The
diagnostic value of protein variants repre-
sents a new insight of promising perspec-
tives.
According to the traditional hypothesis,
marker proteins are released into the
bloodstream by tumor cells, and must
reach a certain concentration to become
detectable. It follows that the tumor must
have reached a reasonable size to produce
sufficient marker proteins above a certain
threshold. However, tumor cells release
not only passively floating proteins but
also active enzymes which modify and
cleave host proteins. Inflammation and
cancer progression are very closely related
[78], with cytokines and proteases acting in
concert [79, 80]. Proteases released by tu-
mors attack the surrounding host proteins
[81], and enzyme activity seems to be cor-
related to the aggressiveness of the lesions
[82]. Furthermore and this is of special
interest for diagnostic purposes the pro-
tease composition exhibits a very high
specificity when compared between differ-
ent types of tumor cells and normal cells
[83]. Thus, tumor-released enzymes could
act as generators of specifically modified
marker proteins. In this way, an amplifica-
tion process could be postulated with a
few enzymes that produce a large amount
of modified host proteins as putative bio-
markers, and these markers could signal
the presence of tumors at a very early
stage (Fig. 8.8). Ciphergen has applied for
patents on this concept, which it terms the
Host Response Protein Amplification Cas-
cade (HR-PAC), and on the use of the
approach in diagnostic testing.
8.8
Conclusion and Outlook
Over the past few years, the ProteinChip
System has proven to be a valuable tool
for the discovery and validation of newly
detected protein biomarker patterns. It is
used by more than 1000 customers world-
wide, including academic and governmen-
System 1338
Fig. 8.8 The Host Response Protein Amplification
Cascade (HR-PAC). According to this new theory,
disease-specific enzymes act as generators of spe-
cifically modified marker proteins which can serve
as valuable biomarkers by signaling a disease in a
very early stage. The covalently modified marker
proteins appear in biological fluids, become de-
tected by SELDI-based ProteinChip technology,
and can be used in multimarker assays to predict
clinical conditions.
tal institutes as well as pharmaceutical
companies. The SELDI process with its
unique combination of chromatographic
principles and mass spectrometric detec-
tion meets the challenges of the new
proteomic era in diagnostic assay develop-
ment. The Pattern Track process includes
all steps from biomarker discovery to assay
development on one single experimental
platform. All of the assays developed thus
far are still of research grade, and have
not been released for commercial use. Cur-
rently, multi-institutional studies are pro-
ceeding to evaluate the clinical robustness
of SELDI-based multi-marker detection
with the necessary scientific diligence be-
fore making it available for the public.
These studies will help in the establish-
ment of diagnostic assays, which in turn
will foster the development of modern bio-
pharmaceuticals. For up-to-date informa-
tion regarding these ongoing activities,
please refer to http://www.ciphergen.com.
References
1 Pfaller, M. A. Emerging Infectious Diseases 2001,
7(2), 312318.
2 Blennow, K., Hampel, H. The Lancet Neurology
2003, 2, 605613.
3 Blennow, K. J. Int. Med. 2004, 256, 224234.
4 Etzioni, R., Urban, N., Ramsey, S., McIntosh,
M., Schwartz, S., Reid, B., Radich, J., Ander-
son, G., Hartwell, L. Nat. Rev. Cancer 2003,
3(4), 243252.
5 National Cancer Institute Cancer facts Tumor
markers, http://cis.nci.nih.gov/fact/5_18.htm,
1998.
6 Bidart, J. M., Thuillier, F., Augereau, C., Cha-
las, J., Daver, A., Jacob, N., Labrousse, F., Voi-
tot, H. Clin. Chem. 1999, 45(10), 16951707.
7 Wiesner, A. Curr. Pharm. Biotechnol. 2004, 5,
4567.
8 Jarvinen, T. A., Liu, E. T. Breast Cancer Res.
Treat. 2003, 78(3), 299311.
9 Jarvinen, T. A., Liu, E. T. Comb. Chem. High
Throughput Screen 2003, 6(5), 455470.
10 Moynihan, R., Smith, R. Br Med J 2002,
324(7342), 859860.
11 Goessl, C. Curr. Med. Chem. 2003, 10(8), 691
706.
12 Simon, R. Br. J. Cancer 2003, 89, 15991604.
13 Johnson, P. J., Lo, Y.M. Clin. Chem. 2002,
48(8), 11861193.
14 Bernard, P. S., Wittwer, C. T. Clin. Chem. 2002,
48(8), 11781185.
15 Ransohoff, D. F. Science 2003, 299(5613),
16791680.
16 Laird, P. W. Nat. Rev. Cancer 2003, 3(4), 253
266.
17 Sidransky, D. Nat. Rev. Cancer 2002, 2(3),
210219.
18 Lim, L. P., Sharp, P. A. Mol. Cell. Biol. 1998,
18(7), 39003906.
19 Gygi, S. P., Rochon, Y., Franza, B. R., Aeber-
sold, R. Mol. Cell. Biol. 1999, 19(3), 1720
1730.
20 Chen, G., Gharib, T. G., Huang, C. C., Taylor,
J. M., Misek, D. E., Kardia, S. L., Giordano,
T. J., Iannettoni, M. D., Orringer, M. B., Ha-
nash, S. M., Beer, D. G. Mol. Cell. Proteomics
2002 1(4), 304313.
21 Anderson, L., Seilhamer, J. A. Electrophoresis
1997, 18(3-4), 533537.
22 Wallace, C. J. Protein Sci. 1993, 2(5), 697705.
23 Pritchard, L. E., Turnbull, A. V., White, A.
J. Endocrinol. 2002, 172(3), 411421.
24 Hanahan, D., Weinberg, R.A. Cell 2000,
100(1), 5770.
25 Pandey, A., Mann, M. Nature 2000, 405(6788),
837846.
26 Service, R. F. Science 2001, 294(5549), 2074
2077.
27 Gavin, A. C., Bosche, M., Krause, R., et al. Na-
ture 2002, 415(6868), 141147.
28 Kumar, A., Snyder, M. Nature 2002, 415(6868),
123124.
29 Gavin, A. C., Superti-Furga, G. Curr. Opin.
Chem. Biol. 2003, 7(1), 2127.
30 Tyers, M., Mann, M. Nature 2003, 422(6928),
193197.
31 Yarmush, M. L., Jayaraman, A. Annu. Rev.
Biomed. Eng. 2002, 4, 349373.
32 Srinivas, P. R., Srivastava, S., Hanash, S.,
Wright Jr., G. L. Clin. Chem. 2001, 47(10),
19011911.
33 Gygi, S. P., Corthals, G. L., Zhang, Y., Rochon,
Y., Aebersold, R. Proc. Natl. Acad. Sci. USA
2000, 97(17), 93909395.
34 Rabilloud, T. Proteomics 2002, 2(1), 310.
References 1339
35 Gygi, S. P., Rist, B., Gerber, S. A., Turecek, F.,
Gelb, M. H., Aebersold, R. Nat. Biotechnol.
1999, 17(10), 994999.
36 Weinberger, S. R., Viner, R. J., Ho, P. Electro-
phoresis 2002, 23(18), 31823192.
37 Weinberger, S. R., Morris, T. S., Pawlak, M.
Pharmacogenomics 2000, 1(4), 395416.
38 Jenkins, R. E., Pennington, S. R. Proteomics
2002, 1(1), 1329.
39 Kodadek, T. Trends Biochem. Sci. 2002, 27(6),
295300.
40 Weinberger, S. R., Davis, S., Makarov, A.,
Thompson, S., Purves, R., Whittal, R. M., in:
Encyclopedia of Analytical Chemistry (Meyers,
R. A., Ed.), John Wiley & Sons Ltd, Chichester,
2000, pp. 1191511984.
41 Weinberger, S. R., Dalmasso, E. A., Fung, E. T.
Curr. Opin. Chem. Biol. 2001, 6(1), 8691.
42 Tang, N., Tornatore, P., Weinberger, S. R. Mass
Spec. Rev. 2004, 23, 3444.
43 Yip, T. T., Lomas, L. Technol. Cancer Res. Treat.
2002, 1(4), 273280.
44 Issaq, H. J., Conrads, T. P., Prieto, D. A., Tiru-
malai, R., Veenstra T. D. Anal. Chem. 2003,
75(7), 148A155A.
45 Issaq, H. J., Veenstra, T. D., Conrads, T. P.,
Felschow, D. Biochem. Biophys. Res. Commun.
2002, 292(3), 587592.
46 Chapman, K. Biochem. Soc. Trans. 2002, 30(2),
8287.
47 Zhang, Z., Bast, R. C. Jr., Yu, Y., et al. Cancer
Res. 2004, 64(16), 58825890.
48 Adam, B. L., Qu, Y., Davis, J. W., et al. Cancer
Res. 2002, 62(13), 36093614.
49 Koopmann, J., Zhang, Z., White, N., et al.
J. Clin. Cancer Res. 2004, 10, 860868.
50 Wadsworth, J. T., Somers, K. D., Cazares, L. H.,
et al. Cancer Res. 2004, 10, 16251632.
51 Carrette, O., Demalte, I., Scherl, A., et al. Pro-
teomics 2003, 3, 14861494.
52 Poon, T. C. W., Chan, K. C. A., Ng, P.C., et al.
Clin. Chem. 2004, 50 (8), 14521455.
53 Gravett, M. G., Novy, M. J., Rosenfeld, R.G., et
al. JAMA 2004, 292, 462469.
54 Papadopoulos, M. C., Abel, P. M., Agranoff,
D., et al. The Lancet 2004, 363, 13581363.
55 Xiao, Z., Luke, B. T., Izmirlian, G., et al. Can-
cer Res. 2004, 64, 29042909.
56 Boot, R. G., Verhoek, M., de Frost, M., et al.
Blood 2004, 103, 3339.
57 Ginestse, C., Ho, L., Pompl, P., Bianchi, M.,
Pasinetti, G. M. (2003) Drugs 2003, 63
(Suppl.1), 2329.
58 Schaub, S., Rush, D., Wilkins, J., et al. J. Am.
Soc. Nephrol. 2004, 15, 219227.
59 Clarke, W., Silverman, B. C., Zhang, Z., Chan,
D. W., Klein, A. S., Molmenti, E. P. Ann. Surg.
2003, 237, 660665.
60 Fung, E. T., Enderwick, C. Biotechniques, Suppl.
Computational Proteomics 2002, 32, S34S41.
61 Adam, B. L., Qu, Y., Davis, J. W., et al. Cancer
Res. 2002, 62(13), 36093614.
62 Hayes, D. F. Recent Results Cancer Res. 1998,
152, 7185.
63 Riley, R. D., Abrams, K. R., Sutton, A. J., Lam-
bert, P. C., Jones, D. R., Heney, D., Burchill,
S.A. Br. J. Cancer 2003, 88(8), 11911198.
64 Rembold, C. M. Br. Med. J. 1998, 317(7154),
307312.
65 Schatzkin, A., Gail, M. Nat. Rev. Cancer 2002,
2(1), 1927.
66 Fung, E. Preclinica 2004, 2(4), 253258.
67 Zhang, L., Yu, W., He, T., et al. Science 2002,
298(5595), 9951000.
68 Uchida, T., Fukawa, A., Uchida, M., Fujita, K.,
Saito, K. J. Proteome Res. 2002, 1(6), 495499.
69 Diamond, D. L., Zhang, Y., Gaiger, A., Smith-
gall, M., Vedvick, T. S., Carter, D. (2003) J. Am.
Soc. Mass. Spectrom. 2003, 14(7), 760765.
70 Shiwa, M., Nishimura, Y., Wakatabe, R., Fuka-
wa, A., Arikuni, H., Ota, H., Kato, Y., Yamori,
T. Biochem. Biophys. Res. Commun. 2003,
309(1), 1825.
71 Santambien, P., Brenac, V., Schwartz, W.,
Boschetti, E., Spencer, J. (2002) Genet. Eng.
2002, 22, 13.
72 Weinberger, S. R., Boschetti, E., Santambien,
P., Brenac, V. J. Chromatogr. B Analyt. Technol.
Biomed. Life Sci. 2002, 782(1-2), 307316.
73 Shiloach, J., Santambien, P., Trinh, L., Schap-
man, A., Boschetti, E. J. Chromatogr. B Analyt.
Technol. Biomed. Life Sci. 2003, 790(1-2), 327
336.
74 Reid, G., Gan, B. S., She, Y. M., Ens, W., Wein-
berger, S., Howard, J.C. Appl. Environ. Micro-
biol. 2002, 68(2), 977980.
75 Cho, W. C. S., Yip, T. T. C., Yip, C., et al. Clin.
Cancer Res. 2004, 10, 4352.
76 Tolson, J., Bogumil, R., Brunst, E., et al. Lab.
Invest. 2004, 84, 845856.
77 Lewczuk, P., Esselmann, H., Groemer, T. W.,
et al. Biol. Psychiatry 2004, 55, 524530.
78 Coussens, L. M., Werb, Z. Nature 2002,
420(6917), 860867.
79 Van Damme, J., Struyf, S., Opdenakker, G.
Semin. Cancer Biol. 2004, 14(3), 201208.
System 1340
80 Fowlkes, J. L., Winkler, M. K. Cytokine Growth
Factor Rev. 2002, 13(3), 277287.
81 DeClerck, Y. A., Mercurio, A. M., Stack, M. S.,
et al. Am. J. Pathol. 2004, 164(4), 11311139.
82 Mahmood, U., Weissleder, R. Mol. Cancer
Ther. 2003, 2(5), 489496.
83 Nuttall, R. K., Pennington, C. J., Taplin, J.,
Wheal, A., Yong, V. W., Forsyth, P.A., Edwards,
D.R. Mol. Cancer Res. 2003, 1(5):333345.
References 1341
Abstract
Early detection of lung cancer by metabolic
profiling of human breath with ion mobili-
ty spectrometry (IMS) dream or reality?
Volatile metabolites occurring in exhaled
air are correlated directly to different kinds
of diseases. Some metabolites are biomar-
kers, i.e., acetone is related to diabetes, ni-
tric acid to asthma and ammonia to hepa-
titis, whereas other arise from bacteria. In
the present chapter, IMS coupled to a mul-
ti-capillary column as a pre-separation unit
is used to identify and quantify volatile
metabolites occurring in human breath
down to the nanogram and picogram per
liter range of analytes. The spectra ob-
tained from patients suffering from
chronic obstructive pulmonary disease and
pneumonia are discussed in detail.
Furthermore, IMS chromatograms of me-
tabolites of Serratia marcescens, Enterobacter
aerogenes and Escherichia coli are compared.
In addition, the effect of drug delivery on
a patient showing angina lateralis is pre-
sented as an example to show the potential
of the method developed in the field of de-
tection of pathways, effective dosage and
decisions of effective time intervals to deli-
ver biopharmaceuticals. The aim of the
studies is to introduce the investigation of
metabolites in human breath as a method
for the early recognition of selected dis-
eases and monitoring of the effectiveness
of drug delivery based on IMS data.
Abbreviations
COPD chronic obstructive pulmonary
disease
GC gas chromatography
IMS ion mobility spectrometry
MCC multi-capillary column
MS mass spectrometry
RIP reactant ion peak
VOC volatile organic compound
9.1
Introduction
The general aim of this chapter is to con-
tribute to the establishment of a fast and
low-cost device for human breath analysis
in addition to investigations of blood and
urine as a non-invasive standard method
in hospitals and point-of-care centers for
different medical applications. On the ba-
sis of miniaturized ion mobility spectro-
metry (IMS), the full procedure, including
sampling, pre-separation and identification
of metabolites in human exhaled air, will
be developed and implemented with re-
1343
9
Early Detection of Lung Cancer: Metabolic Profiling of Human Breath
with Ion Mobility Spectrometers
ISBN: 3-527-31184-X
spect to future application in hospitals. Me-
tabolic profiling of human breath of healthy
individuals and those suffering from differ-
ent diseases, in particular chronic obstruc-
tive pulmonary disease (COPD) and pneu-
monia, will be considered.
In the medical community it is well
known that humans exhale volatile meta-
bolites which potentially carry important
information about their health status.
Thus, successful and fast detection of po-
tential products of different metabolic pro-
cesses becomes attractive, especially if the
detection limits of the spectrometric meth-
ods used are low enough and the instru-
ments become available at moderate price
levels to be used as standard methods in
hospitals or point-of-care centers. The vi-
sion of the authors is to contribute to the
use of human breath as a carrier of infor-
mation of the health status of the body in
addition to human blood and urine.
Human exhaled breath contains numer-
ous volatile metabolites derived from both
endogenous metabolism and external ex-
posure to vapors/gases. Approximately 200
different analytes have been detected in
human breath; some are correlated to var-
ious common disorders like diabetes, heart
disease and lung cancer [114]. The com-
position of different constituents in re-
spired air is mostly representative for
blood-borne concentrations. Such analytes
were detected through gas exchange at the
bloodbreath interface in the lungs [15].
Thus, the presence and also the quantita-
tive variations of specific metabolites in ex-
haled air are directly linked to volatile or-
ganic compounds (VOCs) occurring in the
blood. However, as the blood contacts all
parts of the human body, including dis-
eased tissues or organs, the point of origin
is not fixed in all cases. Furthermore, me-
tabolites derived from local bacterial infec-
tions in the airway system can also be de-
tected directly in human breath. Currently,
a number of different techniques are used
in the field of breath analysis. The most
popular sampling method seems to be the
use of Teflon bags to collect human
breath. Also sorbent- or cryo-trapping of
exhaled air followed by desorption into an
analytical instrument is widely used. In
particular, mass spectrometric methods are
used for the identification of the metabo-
lites. Mostly, chromatographic pre-separa-
tion is also applied. Gas chromatography/
mass spectrometry (GC/MS) is used in the
majority of cases reported [16] (see also
Part V, Chapter 8). Overall, this is a rather
time-consuming process with numerous
different steps. All may lead to a signifi-
cant loss of analytes [2, 17] and, some-
times, analytes may adsorb on the surface
of the bag [18].
In the literature, the major VOCs found
in the breath of healthy persons are
isoprene (10600 ppb
v
), acetone (1
2000 ppb
v
), ethanol (101000 ppb
v
) and
methanol (150200 ppb
v
). All are products
of the standard metabolic processes in the
human body [7].
Considering the analytical methods
mentioned above, the high moisture con-
tent of breath samples is a major problem.
In addition, mass spectrometers, currently
undergoing a process of miniaturization,
have the status of laboratory instruments,
which are comparatively large and expen-
sive, and offer an analysis time of 12 h,
depending on the sample preparation
steps necessary. Therefore, there exists a
need for instruments applicable for direct,
on-line analysis of the breath from a pa-
tient and delivering results nearly just in
time. In particular, in the case where the
number of steps for sample handling
could be minimized, no additional labora-
tory steps become necessary and no addi-
tional carrier gases of high purity (like ni-
9 Early Detection of Lung Cancer: Metabolic Profiling of Human Breath 1344
trogen or helium as used in GC/MS) are
required, on-line methods for effective
breath analysis procedures become attrac-
tive for point-of-care applications. In such
cases, investigations on humans could be
handled in hospitals by standard personal.
Recently, IMS devices have become com-
paratively small and effective devices to de-
termine traces of quantities of VOCs down
to the low ppb
v
range, especially in air
[19]. The major advantages of IMS are that
no vacuum system is required for opera-
tion and ambient air can be used as a car-
rier gas. Worldwide, more than 70000
units are in service, especially to detect
chemical warfare agents, narcotics and ex-
plosives, and many different instruments
are available on the market [20, 21]. It was
shown that VOCs are detectable at the na-
nogram, and sometimes picogram, per li-
ter levels in air [2237].
9.2
Material and Methods: IMS
For the measurements described below, a
custom-designed IMS equipped with a
63
Ni b-ionization source was used [22, 25].
The operational principles and details of
IMS are summarized in [21, 28], and
therefore only a brief description will fol-
low. The term IMS refers to the method of
characterizing analytes in gases by their
gas-phase ion mobility. Normally, the drift
time of ion swarms formed using suitable
ionization sources and electrical shutters
is measured. Ion mobilities are character-
istic for analytes, and can provide a means
for detecting and identifying vapors. The
drift velocity is related to the electric field
strength by the mobility. Therefore, the
mobility is proportional to the inverse drift
time, which will be measured at a fixed
drift length.
IMS combines both high sensitivity and
relatively low technical expenditure with
high-speed data acquisition. The time to
acquire a single spectrum is in the range
of 10100 ms. Thus, IMS is suitable for
process control, but due to the occurrence
of ionmolecule reactions and relatively
poor resolution of the species formed, gen-
erally not for the identification of un-
known compounds. Compared with MS,
the mean free path of the ions is smaller
than the dimensions of the instrument.
Therefore, an ion swarm drifting under
such conditions experiences a separation
process that is based on different drift ve-
locities of ions with different masses or
geometrical structures. Collection of these
ions on a Faraday plate delivers a time-de-
pendent signal corresponding to the mo-
bility of the arriving ions. Such an ion mo-
bility spectrum contains information on
the nature of the different trace com-
pounds present in the sample gas.
The most important parts are the ioniza-
tion/reaction region and the drift region of
the instrument (see Fig. 9.1). The external
and homogenous electric field will be es-
tablished within the drift tube using sev-
eral drift rings for stabilization. The carrier
gas will take sample molecules within the
ionization region. The ionization of the
analytes occurs by chemical ionization on
collisions of the analyte with ionized car-
rier gas molecules by application of radio-
active ionization sources. A so-called drift
gas will flow from the Faraday plate to-
wards the ionization region. Normally, if
the shutter is closed, no ions can reach
the drift region. The drift gas will protect
the drift region and no neutral analyte
molecules should enter the drift region. If
the shutter is held closed, all analyte mole-
cules, neutrals and ions will be washed
out of the gas outlet. During the shutter
opening time, some ions will enter the
drift region. During several collisions with
the surrounding gas molecules, a steady
drift velocity will be reached. If no chemi-
cal reactions occur, in the ideal case total
separation will be reached at the Faraday
plate. The time-dependent voltage or cur-
rent during a time interval measured from
the half of the shutter opening pulse is
called the ion mobility spectrum. Using
air as carrier gas, the carrier gas molecules
are normally ionized by the b-particles di-
rectly. In the present case, positive ions
are under consideration. These primary
positive ions of the carrier gas (called reac-
tion or reactant ions) will undergo differ-
ent chemical reactions with the analyte
ions to form so-called product ions by pro-
ton transfer, nucleophilic attachment, hy-
dride abstraction, etc. Charge transfer and
proton abstraction could also occur.
All of the parts of the IMS which are in
contact with the analytes are formed from
inert materials. The relevant parameters of
the IMS are summarized in Tab. 9.1.
Fig. 9.1 Working principle of IMS.
Table 9.1 Main parameters of
63
Ni-IMS.
Ionization source
63
Ni b-radiation source, 510 MBq
Length of the drift region 12 cm
Electrical field strength 326 Vcm
1
Drift voltage 4 kV
Shutter opening time 300 ls
Drift and sample gas synthetic air
Drift gas flow 100 mL min
1
Sample gas flow 150 mL min
1
(optimized for breath analysis)
Temperature 258C (ambient)
Pressure 101 kPa (ambient)
To realize an effective pre-separation of
the rather complex mixtures occurring in
exhaled air, a 17-cm long polar multi-capil-
lary column (MCC; OV-5; Sibertech, Novo-
sibirsk, Russia) made by combining ap-
proximately 1000 capillaries with an inner
diameter of 40 lm and a film thickness of
0.2 lm was coupled to the
63
Ni-IMS [21,
22, 25].
The total column diameter of 3 mm al-
lows operation with a carrier gas flow up
to 150 mL min
1
, which is the optimum
flow rate for IMS. In addition, the effective
separation of humidity is one major advan-
tage of the MCC used [38].
The heating of the column is indispens-
able for the reproducibility of the chroma-
tographic results. To achieve comparable
retention times, the MCC was held at
308C during breath analysis. To realize iso-
thermal separation, a simple heating con-
struction is needed, which means a con-
cise size decline of the instrument.
In the sampling process, a subject blows
through a mouthpiece coupled with a
brass adapter designed at ISAS to a Teflon
tube (1/4 in.; Bohlender, Lauda, Germany),
which is connected to a 10-mL stainless
steel sample loop of an electric six-port
valve (Nalco; Macherey-Nagel, Dren, Ger-
many). By switching the six-port valve,
breath is transported by the carrier gas
from the sample loop into the MCC. The
separated substances can be directly ana-
lyzed by IMS. Therefore, the results can
be obtained within 10 min depending on
the separation time of the compounds.
This construction enables direct and rapid
sampling at a known breath volume. A
schematic drawing showing sampling and
detection using MCC-
63
Ni-IMS is shown
in Fig. 9.2.
9.3
Results and Discussion
Two typical spectra of acetone in air at
concentrations of 100 and 500 ng L
1
are
shown in Fig. 9.3. The two peaks observed
arise from air [reactant ion peak (RIP)] at
17.69 ms and the analyte acetone at
19.59 ms. The ionization of acetone is real-
ized by charge transfer from the air to the
Fig. 9.2 Diagram of the breath sampling unit and adaptation to MCC-
63
Ni-IMS.
acetone molecule. Thus, it becomes comes
clear that with higher concentration of the
analyte in air, the RIP will decrease and
the acetone-related peak will increase. In
general, the peak position is relevant for
the identification procedure and the peak
area is related to the concentration of the
analyte.
On the other hand, it is shown in
Fig. 9.3 that much higher concentrations
than 500 ng L
1
of the analyte acetone in
air will not be ionized effectively. In such
a case not enough further reactant ions of
the analyte, which are needed for effective
ionization of the molecule, are available.
Here, it should be noted that IMS is suit-
able for trace gas analysis.
The combination of a chromatographic
column with the IMS enables multidimen-
sional data analysis, and allows peak iden-
tification by the use of chromatographic
data (retention times) and also by use of
the specific ion mobility data (arrival time
at the Faraday plate) of the ions formed
from the analytes. Therefore, retention
times of the compounds separated by the
MCC and drift time values of the analytes
are shown in so-called IMS chromato-
grams. The MCC will reduce the number
of interactions in the ionization region of
the IMS as the formation of clusters and
avoids concurrent charge discharge trans-
fer by time-delayed entrance of the ana-
lytes. Thus, the effective ionization of
nearly all analytes in a given sample is or-
ganized by selection of proper GC and ef-
fective temperature programming of the
chromatographic column or keeping the
temperature constant, as in the present
case.
Fig. 9.4 shows the results of investiga-
tions of exhaled air of a healthy human
containing ammonia, acetone and ethanol
peaks, and showing also the RIP as major
signals. Smaller portions will not be con-
sidered here. In addition, the mobilities of
the ions calculated from the drift time of
the ion peak K
0
values (for details, see
Fig. 9.3 Ion mobility spectra of acetone in air (100 and 500 ng L
1
).
[21]) are included in the inlet, showing a
single spectrum at a single, fixed retention
time.
It is also clearly visible in Fig. 9.4 that
the humidity, which can reach really high
values, particularly when considering ex-
haled air, is separated effectively by using
the MCC. During the first 20 s or so high
values of humidity enter the ionization re-
gion of the IMS. In general, the MCC will
reduce the influence of humidity as ac-
quired. On the other hand, the main inter-
est will focus on an interval for higher
drift times of about 17 ms at any retention
time, as shown in the following section.
9.4
Clinical Study
In cooperation with the Lung Hospital in
Hemer, on-site measurements were carried
out with the exhaled air of 40 subjects, in-
cluding 22 patients suffering from differ-
ent pulmonary lung infections. Breath
samples of 18 different healthy persons
were also analyzed on-line using
MCC-
63
Ni-IMS. The full analysis was al-
ways performed in the same room, where
room air was determined before each of
the breath measurements. To reduce the
risk of cross-contamination from other
sources, subjects had not drunk, eaten or
smoked for at least 2 h before the breath
measurements.
An IMS chromatogram from the ex-
haled air of a patient suffering bacterial
lung infection is shown in Fig. 9.5. The
room air values are subtracted in all cases,
as discussed above. The colors refer to dif-
ferent peak height levels. The peaks identi-
fied are labeled with the mobility values
K
0
calculated from the drift time of the
ions of the analytes, and also taking into
account the values of the electric field
strength in the drift tube, the pressure
and the temperature conditions during the
experiment (for further details, see [21]).
Fig. 9.4 IMS chromatogram of human exhaled air inlet: single spectrum
at a fixed retention time of 3 s showing signals of the major constitutions
ammonia, ethanol and acetone.
As mentioned above, the formation of
so-called reactant (arising from the carrier
gas air) and product ions (formed by the
analytes) in IMS can be altered by mois-
ture. Therefore, separation of water mole-
cules is of great importance when working
with breath samples. Figure 9.5 shows that
the high breath moisture content (mobility
values K
01
=2.11 cm
2
Vs
1
) could be sepa-
rated effectively using the MCC. The sig-
nals correlated to moisture are always lo-
calized at the beginning of the IMS chro-
matogram. Thus, moisture has no disturb-
ing effect for peaks showing longer
retention times. In the topographic plot of
the exhaled air of the healthy person, acet-
one was identified based on its reduced
mobility value (K
04
=1.78 cm
2
Vs
1
). Acet-
one has a similar retention time as water,
but possesses a stronger proton affinity.
This is enough for acetone to be ionized
via ionmolecule reactions.
Comparing the two IMS chromatograms
of Figs. 9.4 and 9.5, the differences are
conspicuous. In Fig. 9.5, the two-dimen-
sional plot exhibits additional peaks. Two
major peaks occur at the retention times
of 28 s (K
02
=1.95 cm
2
Vs
1
) and 36 s
(K
03
=1.77 cm
2
Vs
1
). Both are probably de-
gradation products of antibiotics or other
drugs, because they were also found in the
breath of several persons who had taken
similar antibiotics. The peak at the posi-
tion of 50 s and K
05
=1.42 cm
2
Vs
1
was
also analyzed from the exhaled air of other
patients suffering from pneumonia infec-
tion. Two other larger analytes with lower
mobility values (K
06
=1.39 cm
2
Vs
1
and
K
07
=1.28 cm
2
Vs
1
) and longer retention
times (t
ret6
=244 s and t
ret7
=330 s) were of-
ten detected in patients with bacterial in-
fection and airway inflammation.
The exhaled air of a patient with COPD
is shown in Fig. 9.6. Two major peaks oc-
cur. The peak at lower retention times (P-
Co01-01) is identically to peaks found in
Staphylococcus aureus cell cultures (B-Sa-
01). The second peak at higher retention
and drift times, indicating a higher-molec-
ular-weight substance, is not correlated
with S. aureus. No other bacteria were
identified in additional microbiological in-
vestigations. However, it becomes obvious
that metabolites of bacteria occurring in
the lung should be taken into considera-
tion. Therefore, investigations of emissions
of bacteria were carried out in addition. As
a reference, Fig. 9.7 shows the IMS chro-
Fig. 9.5 IMS chromatogram of
the breath of a patient suffering
from a lung infection (K
0
refers
to the mobility values calculated
for each major peak).
matogram of volatile emissions obtained
from S. aureus. A single major peak (B-Sa-
01) with an ion mobility of 1.86 cm
2
Vs
1
and a retention time of 33 s (B-Sa-01) is
clearly visible.
To demonstrate the rather complex situa-
tion between different diseases and bacteria
involved, Fig. 9.8 shows the IMS chromato-
gram of a COPD patient with different infec-
tions. There are five, clearly separated, major
peaks. One among them, called P-Co02-04,
is identical to the peak B-Ec-05 arising from
Escherichia coli (see Fig. 9.9). With regard to
E. coli, five major peaks could also be recog-
nized. In addition to the rather small vola-
tiles (B-Ec-01, B-Ec-02 and B-Ec-03), two
peaks [K
0
=1.33 cm
2
Vs
1
(B-Ec-04) and
K
0
=128 cm
2
Vs
1
(B-Ec-05)] occur at larger
retention times (t
retB-Ec04
=129 s and t
retB-
Ec05
=327 s). As mentioned, peaks B-Ec-05
and P-Co02-04 are identical if the mobility
and not the drift time scale is considered, be-
Fig. 9.6 IMS chromatogram
(right) of a patient suffering
from COPD and two typical
single spectra (below).
cause of small changes in experimental con-
ditions affecting the mobility values.
To demonstrate the analytical potential
of IMS considering the direct detection of
the effectiveness of drug delivery, Fig. 9.10
shows two IMS chromatograms of a pa-
tient suffering on angina lateralis. The
main peak at K
0
=1.51 cm
2
Vs
1
and reten-
tion time t
ret
=48 s is considered as being
related to angina lateralis. On the other
hand, the peak was also identified as re-
lated to Enterobacter aerogenes (B-Eb04).
Note, the three additional peaks are cor-
related with E. coli (B-Ec-03, B-Ec-04 and
B-EC-05). The first IMS chromatogram
shows the situation before drug delivery
(Amoxibeta T 1000 containing amoxicillin
trihydrate). The other one was obtained
after 72 h of continuous amoxicillin trihy-
drate delivery once per day. It becomes ob-
vious that the major peak related to angina
lateralis and E. aerogenes decreases. In ad-
dition, in the special case of the antibiotic
delivered to the patient, E. coli peaks are
also reduced. Two peaks at higher reten-
tion times disappeared completely. Only
one peak remains after 3 days of applica-
tion of amoxicillin. The related single
Fig. 9.7 IMS chromatogram (left)
of S. aureus and the correspond-
ing single spectrum (below).
spectra for the main peak related to E.
aerogenes are shown in Fig. 9.11 together
in a single plot. The effect of the amoxilin
causes a peak height reduction at drift
times of 22.37 ms and retention time of
46 s during the first 3 days of application
of amoxicillin, as summarized in Fig. 9.12.
As expected, the signal intensity reduction
follows an exponential decrease. Thus, an
effective way of quickly determining the
effectiveness of drug delivery and/or of
actual use of drugs within IMS measure-
ments should be considered in more detail
in the future for different cases and specif-
ic metabolites in exhaled air.
Fig. 9.8 IMS chromatogram (above) and related single spectra (below) of a patient
suffering infections and COPD.
To verify the characteristic peaks of vola-
tile metabolites detected, further measure-
ments will be carried out including a larg-
er number of patients and healthy per-
sons. Data will be processed using statisti-
cal methods to clarify the assignment
between the diseases and the characteristic
pattern of the IMS topographic plots. In
the near future, the results should be con-
firmed by comparative analysis using mass
spectroscopy to identify the nature of the
metabolites detected by IMS.
9.5 Conclusions
By coupling IMS and a MCC for pre-sepa-
ration, investigations were carried out to
directly detect volatile metabolites in hu-
man exhaled air. The total analysis was re-
Fig. 9.9 IMS chromatogram of E. coli (above) and related single spectra (below).
alized within time intervals of less than
500 s. The system constructed shows suffi-
cient sensitivity to detect the metabolites
directly by showing characteristic IMS
chromatograms (peak height diagrams).
The exhaled air is introduced directly into
the MCC using a sample loop. Results of
investigations on patients suffering from
pneumonia show different metabolites in
comparison to healthy persons.
Further investigations using MS meth-
ods should allow the identification of the
metabolites found by IMS. The possibili-
ties for on-site and short-time analysis
using air as a carrier gas and working at
ambient pressure are the most important
benefits of the technique developed.
Using MCC-IMS directly in the diagnos-
tic clinic allows results to be obtained
within minutes, delivers additional infor-
Fig. 9.10 IMS chromatograms of the exhaled breath of a patient with
angina lateralis signals of E. coli and E. aerogenes were identified
(surrounded in white and red, respectively) before and 3 days after the
start of amoxicillin therapy.
mation for the therapeutic strategy and fa-
cilitates the building of databases for sev-
eral diseases. The general aim of the stud-
ies could introduce the investigation of vol-
atile metabolites in human breath as a
marker for the early recognition of se-
lected diseases, e.g. lung cancer, on the ba-
sis of IMS data. Furthermore, especially
when considering biopharmaceuticals, me-
tabolic profiling using different kinds of
IMS opens the possibility to also directly
investigate metabolic processes (including
single reactions and interactions of bio-
pharmaceuticals at the cell level up to the
entire human body), using human breath
as the carrier of information.
Fig. 9.11 Effect on the antibiotic on the peak height in an ion
mobility spectrum.
Fig. 9.12 Decrease of the signal (peak area) of the specific peak
considered in the ion mobility spectrum of a patient suffering angina
lateralis during drug delivery (amoxicillin) for 3 days.
Acknowledgments
Financial support of the Bundesminister-
ium fr Bildung und Forschung and the
Ministerium fr Wissenschaft und For-
schung des Landes Nordrhein-Westfalen,
and the funding of the common project by
G. A. S. Gesellschaft fr Analytische Senso-
rensysteme mbH is gratefully acknowl-
edged. The possibility for the analysis in
the Lung Hospital in Hemer and the help
of the assistants and doctors should be
mentioned with thanks. The authors wish
to express their special thanks to Mrs.
Oberdrifter for realizing laboratory sup-
port, Dr. Litterst and Dr. Westhoff for help-
ful discussions and organizing the clinical
measurements (all at the Lung Hospital
Hemer), and Mrs. Gssgen and Mrs. Sei-
fert (both at ISAS, Dortmund) for support
in the operation of the ion mobility spec-
trometer and preparation of the data han-
dling.
References
1 Phillips, M. Method for the collection and as-
say of volatile organic compounds in breath.
Anal. Biochem. 247, 272278, 1997.
2 Phillips, M., Gleeson, K., Hughes, J. M.,
Greenberg, J., Cataneo, R. N., Baker, L.,
McVay, W. P. Volatile organic compounds in
breath as markers of lung cancer: a cross-sec-
tional study. Lancet 353, 19301933, 1999.
3 Phillips, M., Greenberg, J. Ion trap detection
of volatile organic compounds in alveolar
breath. Clin. Chem. 38, 6065, 1992.
4 Rooth, G., Lund, M. D., stenson, S. Acetone
in alveolar air, and the control of diabetes.
Lancet 11021104, 1966.
5 Benolt, M. F., Davidson, W. R., Lovett, A. M.,
Nacson, S., Ngo, A. Breath analysis by atmo-
spheric pressure ionization mass spectrome-
try. Anal. Chem. 55, 805807, 1983.
6 Cheng, W.-H., Lee, W.-J. Technology develop-
ment in breath microanalysis for clinical diag-
nosis. J. Lab. Clin. Med. 133, 218228, 1999.
7 Fenske, J. D., Paulson, S. E. Human breath
emissions of VOCs. J. Air Waste Manag. Ass.
49, 594598, 1999.
8 Grote, C., Pawliszyn, J. Solid-phase microex-
traction for the analysis of human breath.
Anal. Chem. 69, 587596, 1997.
9 Hansel, A., Jordan, A., Holzinger, R. Proton
transfer reaction mass spectrometry: on-line
trace gas analysis at the ppb level. Int. J. Mass
Spectrom. Ion Proc. 149, 609619, 1995.
10 Jansson, B. O., Larsson, B. T. Analysis of or-
ganic compounds in human breath by gas
chromatography mass spectrometry. J. Lab.
Clin. Med. 74, 961966, 1969.
11 Lindinger, W., Hansel, A., Jordan, A. On-line
monitoring of volatile organic compounds at
pptv levels by means of proton-transfer-reac-
tion mass spectrometry (PTR-MS) medical ap-
plications, food control and environmental re-
search. Int. J. Mass Spectrom. Ion Proc. 173,
191241, 1998.
12 Manolis, A. The diagnostic potential of breath
analysis. Clin. Chem. 29, 515, 1983.
13 Ruzsanyi, V., Sielemann, S., Baumbach, J. I.
Determination of VOCs in human breath
using IMS. Int. J. Ion Mobility Spectrom. 5,
4548, 2002.
14 Stein, V. B., Narang, R. S., Wilson, L., Aldous,
K. M. A simple, reliable method for the deter-
mination of chlorinated volatile organics in
human breath and air using glass sampling
tubes. J. Anal. Toxicol. 20, 145150, 1996.
15 Pleil, J. D., Lindstrom, A. B. Exhaled human
breath measurement method for assessing ex-
posure to halogenated volatile organic com-
pounds. Clin. Chem. 43, 723730, 1997.
16 Gordon, S. M., Szidon, J. P., Krotoszynski,
B. K., Gibbons, R. D., ONeill, H. J. Volatile or-
ganic compounds in exhaled air from patients
with lung cancer. Clin. Chem. 31, 12781282,
1985.
17 Pleil, J. D., Lindstrom, A. B. Collection of a
single alveolar exhaled breath for volatile or-
ganic compounds analysis. Am. J. Ind. Med.
28, 109121, 1995.
18 Wilson, H. R. Breath analysis: physiological
basis and sampling techniques. Scand. J. Work
Environ. Health 12, 174192, 1986.
19 Li, F., Xie, Z., Schmidt, H., Sielemann, S.,
Baumbach, J. I. Ion mobility spectrometer for
online monitoring of trace compounds. Spec-
trochim. Acta B 57, 15631574, 2002.
References 1357
20 Spangler, G. E., Carrico, J. P., Campbell, D. N.
Recent advances in ion mobility spectrometry
for explosives vapor detection. J. Test. Eval. 13,
234240, 1985.
21 Baumbach, J. I., Eiceman, G. A. Ion mobility
spectrometry: arriving on site and moving be-
yond a low profile. Appl. Spectrosc. 53, 338
354, 1999.
22 Sielemann, S. Detektion flchtiger orga-
nischer Verbindungen mittels Ionenmobili-
ttsspektrometrie und deren Kopplung mit
Multi-Kapillar-Gas-Chromatographie. Disserta-
tion, University of Dortmund, 1999.
23 Sielemann, S., Baumbach, J. I., Pilzecker, P.,
Walendzik, G. Detection of trans-1,2-dichlor-
oethene, trichloroethene and tetrachloroethene
using multi-capillary columns coupled to ion
mobility spectrometers with UV-ionization
sources. Int. J. Ion Mobility Spectrom. 2, 1521,
1999.
24 Sielemann, S., Xie, Z., Schmidt, H., Baum-
bach, J. I. Determination of MTBE next to
benzene, toluene and xylene within 90 s using
GC/IMS with multi-capillary column. Int. J.
Ion Mobility Spectrom. 4, 6973, 2001.
25 Xie, Z., Ruzsanyi, V., Sielemann, S., Schmidt,
H., Baumbach, J. I. Determination of pentane,
isoprene and acetone using HSCC-UV-IMS.
Int. J. Ion Mobility Spectrom. 4, 8891, 2001.
26 Xie, Z., Sielemann, S., Schmidt, H., Li, F.,
Baumbach, J. I. Determination of acetone, 2-
butanone, diethyl ketone and BTX using
HSCC-UV-IMS. Anal. Bioanal. Chem. 372,
606610, 2002.
27 Baumbach, J. I., Sielemann, S., Xie, Z.,
Schmidt, H. Detection of the gasoline compo-
nents methyl tert-butyl ether, benzene, to-
luene, and m-xylene using ion mobility spec-
trometers with a radioactive and UV ioniza-
tion source. Anal. Chem. 75, 14831490, 2003.
28 Eiceman, G. A., Karpas, Z. Ion Mobility Spec-
trometry. CRC Press, Boca Raton, FL, 1994.
29 Eiceman, G. A., Leasure, C. S., Vandiver, V. J.
Negative ion mobility spectrometry for se-
lected inorganic pollutant gases and gas mix-
tures in air. Anal. Chem. 58, 7680, 1986.
30 Eiceman, G. A., Nazarov, E., Miller, R. A. A mi-
cro-machined ion mobility spectrometermass
spectrometer. Int. J. Ion Mobility Spectrom. 3,
1527, 2000.
31 Eiceman, G. A., Nazarov, E., Miller, R. A., Kry-
lov, E. V., Zapata, A. M. Micro-machined planar
asymmetric ion mobility spectrometer as a
gas chromatographic detector. Analyst 127,
466471, 2002.
32 Eiceman, G. A., Shoff, D. B., Harden, C. S.,
Snyder, A. P., Martinez, P. M., Fleischer, M. E.,
Watkins, M. L. Ion mobility spectrometry of
halothane, enflurane, and isoflurane anes-
thetics in air and respired gases. Anal. Chem.
61, 10931099, 1989.
33 Hill, C. A., Thomas, C. L. P. A pulsed corona
discharge switchable high resolution ion mo-
bility spectrometermass spectrometer. Analyst
128, 5560, 2003.
34 Hill, H. H., Eatherton, R. L. Ion mobility spec-
trometry after chromatography accomplish-
ments, goals, challenges. J. Res. Natl Bureau
Stand. 93, 425426, 1988.
35 Hill, H. H., Jr., Siems, W. F., St. Louis, R. H.,
McMinn, D. G. Ion mobility spectrometry.
Anal. Chem. 62, 1201A1209A, 1990.
36 St. Louis, R. H., Siems, W. F., Hill, H. H., Jr.
Detection limits of an ion mobility detector
after capillary gas chromatography. J. Microcol-
umn Sep. 2, 138145, 1990.
37 St. Louis, R. H., Hill, H. H., Jr. Ion mobility
spectrometry in analytical chemistry. Crit. Rev.
Anal. Chem. 21, 321355, 1990.
38 Vautz, W., Sielemann, S., Baumbach, J. I. De-
termination of terpenes in humid ambient air
using ultraviolet ion mobility spectrometry.
Anal. Chim. Acta 513, 393399, 2004.
ISBN: 3-527-31184-X
Part VI
Advanced Application Routes for Biopharmaceuticals
Abstract
The administration of biopharmaceuticals
to humans is most challenging, since a
variety of different conditions in the body
may interfere with the delivery of biophar-
maceuticals to their target, and may have
a negative impact on drug stability. Drug
delivery systems will serve as a measure to
overcome these issues, and will enhance
the applicability of biopharmaceuticals to
safe and efficacious therapy. Moreover, the
less frequent application of drug delivery
systems compared to conventional dosage
forms is likely to increase patient compli-
ance. In addition to well-known biologicals
such as insulin, antibodies, and luteinizing
hormone-releasing hormone (LHRH) ago-
nists, an increasing number of new biolog-
ical entities (NBEs) will soon enter the de-
velopment phase, and will very likely have
to be applied in drug delivery systems.
The principles known from drug delivery
systems for small molecules will be trans-
ferred to the development of drug delivery
systems with new specific features for
these NBEs. These systems will aim at the
protection of diverse biopharmaceuticals
(e.g., peptides, proteins and enzymes, as
well as DNA and RNA) against chemical
and enzymatic degradation. The applica-
tion of delivery systems is also intended to
alter the pharmcokinetics of administered
drugs by targeting to specific tissues, and
to realize extended circulation times for in-
creased biological half-lives. In addition, a
molecule-mediated uptake of the biophar-
maceuticals into specific cells can be
achieved. To accomplish this aim, princi-
ples such as polymer-based delivery sys-
tems, drugpolymer conjugates and a
broad variety of lipid-based vesicles are
employed in the form of dispersed and
matrix systems. The choice of delivery sys-
tem will be made according to the proper-
ties of the biopharmaceutical and its mode
of action, as well as the intended route of
administration. As of today, delivery sys-
tems for inhalational, nasal, oral, and
transdermal systemic administration of
biopharmaceuticals have been described,
in addition to the most common route of
1361
1
Advanced Drug Delivery Systems for Biopharmaceuticals
ISBN: 3-527-31184-X
Getting Inside Quest for the Best and How to Improve Delivery
parenteral administration of solutions and
implants. More recently, totally new ideas
for the delivery of NBEs, such as feedback-
controlled pumps and ceramic devices,
have emerged. Looking ahead, the main is-
sues will be the successful introduction of
new drug delivery-based medications into
clinics, associated with the preparation for
large-scale production under GMP condi-
tions. The development of new biocompa-
tible materials will be a crucial factor,
since the increasing number of NBEs with
their manifold properties will require a
broad portfolio of formulation possibilities
to meet the specific demands with regard
to route of administration and mode of ac-
tion of certain biopharmaceuticals.
Abbreviations
5-ASA 5-amino salicylic acid
ANAs antinuclear autoantibodies
DDS drug delivery systems
factor
GRAS generally regarded as safe
HGR human growth hormone
HPMA N-(-2-hydroxypropyl)methacryl-
amide
IBD inflammatory bowel disease
LHRH luteinizing hormone-releasing
hormone
LMM low molecular mass
MLV multilamellae vesicles
NBEs new biological entities
NCEs new chemical entities
PEG poly(ethylene glycol)
PLG poly(d,l-lactide-co-glycolide)
PLGA poly(d,l-lactide-co-glycolide)
poly(PGA) poly (l-glutamic acid)q
poly-SMA poly(styrene-co-maleic acid
anhydride)
PTH parathyroid hormone
RES reticuloendothelial system
SAIB sucrose acetate isobutyrate
SCID severe combined immuno-
deficiency
S-CT salmon calcitonin
SLN solid lipid nanoparticles
SMANCS styrene maleic anhydride
SUV small unilamellar liposomes
TPP alkaline tripolyphosphate
TTS transdermal therapeutic system
1.1
Introduction
In recent years, the successful administra-
tion of biomolecules to patients has proved
to be one of the most challenging areas of
the development of biopharmaceuticals,
despite our knowledge of the functions of
different absorption barriers in the body.
This situation is reflected by the compara-
tively small number of biopharmaceuticals
available commercially, and which are
mainly administered parenterally, by sub-
cutaneous or intravenous injection. Cur-
rently, depot systems to deliver peptide
hormone analogs and peptide or antibody
solutions and vaccines are the only mar-
keted formulations of this type.
The increasing number of new biologi-
cal entities (NBEs) which are currently en-
tering the development phase will raise
the need for advanced drug delivery sys-
tems that allow biopharmaceuticals to real-
ize their whole potential in the treatment
of diseases. Drug delivery systems, rang-
ing from modified release tablets to sus-
tained release implants and transdermal
patches, are widely used for small mole-
cules. These systems may be a starting
point for advanced drug delivery systems
which are modified with regard to the
somewhat multi-layered requirements of
modern biopharmaceuticals. This chapter
provides an overview of the diverse drug
delivery strategies that have been pursued
in order to overcome or circumvent ab-
sorption barriers, and how such delivery
systems will affect the pharmacokinetics,
stability, degradation, efficacy, and effi-
ciency of the respective incorporated bio-
pharmaceuticals. The sites of administra-
tion of the respective carrier systems are il-
lustrated schematically in Fig. 1.1.
1.2
Challenges for the Administration
Due to the similarity between biopharma-
ceuticals and natural substances both in
the human body and the biological envi-
ronment, many mechanisms and pro-
cesses exist which protect against the ef-
fects of xenobiotics. Epithelial and mucosal
cells, both in the body cavities and the gas-
trointestinal tract serve as efficient barriers
against the absorption of many proteins
and peptides, as well as virus particles and
cells. Thus, major efforts must be underta-
ken to provide significant bioavailability,
though this is the realm of formulation.
After successful absorption, a biological
entity must reach its specific target in or-
der to be efficacious.
The fate of (biopharmaceutical) drugs in
the body is illustrated schematically in
Fig. 1.2. An understanding of absorption
and elimination of a drug in an individual
patient, when related to the pharmacologi-
cal effects of a given quantity of that drug,
makes it possible to determine a dose that
will be clinically effective, but with mini-
mal toxicity.
Although such relationships are well de-
fined for many classical drugs, this
approach has not been utilized to any
great extent for new biological entities. In
this chapter, pharmacokinetic considera-
tions will be discussed with particular re-
gard to degradation, half-life, and target-
specific distribution.
1.2.1
Enzymatic Degradation
Degradation seems to be one of the main
obstacles in achieving good bioavailability
for biopharmaceuticals. Proteases are pre-
sumed to be the major metabolic pathway
for peptide and protein drugs [1], as they
are ubiquitous in the body, on both the lu-
menal and systemic sides of the mucosa
[2]. Several approaches have been used to
overcome this degradation, including the
co-administration of protease inhibitors,
pro-drugs, conjugates [3], and drug deliv-
ery systems.
Yamamoto et al. [4] showed that 0.01%
aprotinin (a serine protease inhibitor) re-
duced the metabolism of insulin and
proinsulin in homogenates of albino rabbit
buccal mucosa, which otherwise would
have occurred at 7080% within 2.5 hours.
Moreover, Lehr et al. suggest that polycar-
bophil, a bioadhesive polymer, may protect
some peptides from proteolysis, though
the mechanism of this is unknown [5].
Others [6] have developed a series of pro-
drugs for peptides, with the aim of over-
coming the metabolic barrier imposed by
different peptidases. Stable prodrugs
proved to be N-hydroxymethylated deriva-
tives of the assessed dipeptides Gly-L-Leu
and Gly-L-Ala [6].
1.2.2
Half-life of Biopharmaceuticals
In pharmacokinetic terms, the half-life is
an expression of the relationship between
F
i
g
.
1
.
1
A
d
m
i
n
i
s
t
r
a
t
i
o
n
s
i
t
e
s
f
o
r
a
d
v
a
n
c
e
d
d
r
u
g
d
e
l
i
v
e
r
y
s
y
s
t
e
m
s
f
o
r
b
i
o
p
h
a
r
m
a
c
e
u
t
i
c
a
l
s
,
a
n
d
t
h
e
i
r
l
o
c
a
t
i
o
n
i
n
t
h
e
s
y
s
t
e
m
i
c
c
i
r
c
u
l
a
t
i
o
n
.
the volume of distribution and clearance.
Since the organs of elimination can only
clear drug from the blood or plasma that
is in direct contact with that organ, the
time-course of a drug in the body will de-
pend upon both the volume of distribution
and clearance.
The half-life is a useful kinetic parame-
ter for therapeutic molecules, as it defines
the dosing interval at which drugs should
be administered. Half-life also describes
the time required to reach steady state, or
to decay from steady-state conditions after
a change in the administration regimen.
However, as an indicator of the processes
involved in drug elimination or distribu-
tion the half-life has only limited value. It
must be mentioned that, for many pro-
teins and peptides, the glomerular filtra-
tion rate is high, and has more impact on
the elimination of a biopharmaceutical
than does the hepatic first-pass effect [7].
One shortcoming of peptides is their
low metabolic stability. Endogenous pep-
tides are degraded by a variety of pro-
teases, and often have short in vivo half-
lives. In addition, the size and structure of
a peptide may have a significant impact
on its stability. Small linear peptides with
no structural modifications are rapidly de-
graded, and thus may have extremely short
half-lives (*25 min). Larger peptides/pro-
teins particularly those containing disul-
fide linkages are often more stable, but
the half-lives may still be less than 12 h
[8]. Although some pharmaceutical compa-
nies report candidates with dramatically im-
proved metabolic stability, for many pep-
tides either more frequent dosing and/or
the use of larger doses may be required to
obtain therapeutic peptide concentrations
in vivo. Thus, for biopharmaceuticals not
only degradation in the liver but also degra-
dation by various proteases determine the
in vivo half-life. The most common means
of altering the half-life is to use a drug deliv-
ery system, for example allowing extended
circulation times in the blood or sustained
release of the drug, as well as protecting
against degrading enzymes.
Fig. 1.2 Scheme of the pharmacokinetics of (biopharmaceutical) drugs.
1.2.3
Target-specific Distribution
The delivery of biopharmaceuticals to the
target site is often reported as the key to
efficient therapy [9, 10]. Molecule delivery
to targeted cells and specific compart-
ments within cells, as well as targeting to
specific organs, is of utmost importance
(see Part VI, Chapters 6 and 8). The deliv-
ery of particular biomolecules to a certain
target in vivo may usually only be achieved
by means of delivery systems such as lipo-
somes and colloidal particles of a certain
size [11, 12], or in conjunction with a
homing device such as an antibody or a
specific ligand [13] (see Part V, Chapter 6
and Part VI, Chapter 7). If the target-spe-
cific distribution of biological entities is
not accomplished, this has strong implica-
tions on the dose to be administered. For
example, there is an increased likelihood
of enzymatic degradation or a larger vol-
ume of distribution; hence, overall drug
intake during therapy can be increased tre-
mendously.
1.3
Drug Delivery Strategies
In order to overcome these problems, sev-
eral strategies have been developed, and
these will be outlined in the following sec-
tions.
As mentioned above, proteins are both
physically and chemically labile, and in
most cases an intravenous infusion would
be the only means of providing acceptable
blood levels over a desired time period.
Since this approach would be unacceptable
for most patients, two major strategies have
been devised to overcome theses challenges:
conjugation with other macromolecules;
and
encapsulation in various types of liquid
and solid microreservoir delivery sys-
tems.
1.3.1
Conjugates
A major challenge for the use of small
proteins (molecular weight <40 kDa) is
their short half-life due to a rapid renal
elimination. Linking proteins with water-
soluble polymers leads to water-soluble
conjugates with higher molecular weights,
and these remain for much longer periods
in the body. For example, in rabbits the
half-life of a dextran and soybean trypsin
conjugate with a molecular weight of
127 kDa is about 10-fold longer compared
to the plain protein of 27 kDa [14]. Addi-
tionally, receptor-mediated uptake into the
cells of the reticuloendothelial system
(RES) is prevented, and this also leads to a
longer half-life [15, 16]. Moreover, there
are additional benefits of peptide and pro-
teinpolymer conjugations, such as in-
creased resistance against enzymatic de-
gradation and reduced immunogenicity
[17]. In some cases, selective accumulation
in the target area can be achieved, and this
is of major interest in cancer therapy. For
example, compared to normal tissue, tu-
mor vessels can take up conjugates with a
molecular weight >40 kDa. In addition,
the lymphatic system in these areas is un-
able to provide a full drainage function,
and this may lead to an enhanced perme-
ability and retention effect (EPR) of high-
molecular weight compounds. Small mole-
cules are not accumulated as they are able
to diffuse back into the blood circulation
[18].
1.3.1.1 Conjugates with
Poly(ethylene glycol) (PEG)
The first steps to use protein conjugates
were taken during the late 1960s. Whilst
at the time proteins derived via recombi-
nant DNA were not available, the initial
experiences were made with insulin and
asparaginase, and the risk of immuno-
genic reaction was high. The research
teams of the day first suggested that com-
bining proteins with hydrophilic polymers
might reduce the imunogenicity of pro-
teins. PEG and poly(propylene glycol) were
used as infusion ad juvants to prevent the
occurrence of lipid embolism in major ves-
sel surgery. Davis and his colleagues first
used PEGylation to show that a decreased
immunogenicity in animals could be
achieved, in addition to an extended circu-
latory life for the drug [19] (see Part VI,
Chapter 2). Details on the technology of
PEGylation were recently described [16,
20]. PEG is a highly water-soluble and flex-
ible polymer, with the molecules providing
a very narrow polydispersity (MW/
MN*1.01). Moreover, the short PEGs
(<1 kDa) are non-toxic and generally re-
garded as safe for a variety of pharmaceu-
tical products (i.e., those approved by the
regulatory authorities). The first-generation
conjugates consisted mainly of protein
plus one linear monomethoxy PEG. Subse-
quently, a variety of conjugation chemis-
tries was used, starting with amine conju-
gation. Indeed, the second generation of
PEGylated proteins used a much wider
variety of PEG chemistry for cysteine mod-
ification, oxidized carbohydrates, or N-ter-
minus and reversible PEGylation, as well
as heterobifunctional PEG chemistry.
In addition, branched PEG molecules
with a range of molecular weights (5 to
40 kDa) were included such that it was
possible to create a tailor-made system ac-
cording to the physico-chemical properties
of the protein [21]. However, since biologi-
cal properties such as cytotoxicity, hemato-
toxicity, carcinogenicity, teratogenicity,
complement activation, and immunogeni-
city are dependent upon molecular weight,
they may change when the respective con-
jugates are prepared [22]. Despite these
shortcomings, a range of conjugates has
been approved for the treatment of severe
diseases during the past 14 years (Table
1.1).
1.3.1.2 Conjugates with
N-(-2-hydroxypropyl)methacrylamide
(HPMA) Co-polymer
The first HPMA co-polymer conjugates
which are currently undergoing Phase I
and II clinical trials represent the careful
development of a purely rational drug
polymer conjugate design. As the HPMA
co-polymer had not been used previously
as a pharmaceutical agent, it was neces-
sary to conduct the entire process of pre-
clinical toxicology studies. For a conjugate
with doxorubicin, all data were collected to
show the safety of this novel polymeric
compound; this GLP study demonstrated
the details of the process [31]. Other conju-
gates with platinate [32], paclitaxel [33] or
camptothecin [34] are also currently under
investigation. As seen previously with
PEG, the variety of chemical possibilities
offers numerous design opportunities. Es-
ters and amides are used primarily to link
the HPMA co-polymer to the protein,
though malonate has also been described
as a suitable linker [32].
1.3.1.3 Conjugates with Poly(styrene-co-
maleic acid anhydride) [(poly(SMA)]
Another possible means of increasing the
half-life of a drug is by conjugation to
polymers that are able to attach to plasma
components, even when conjugated with
the drug. Although the molecular weight
of the polymer might be less than the re-
nal threshold, the resultant complex circu-
lates for as long as the attached natural
plasma components such as lipoproteins
or serum albumin. For example, SMA
with a molecular weight of only 1.5 kDa
leads to an increased plasma circulation of
anti-cancer drugs, due to the conjugate
being able to bind with plasma albumin.
These conjugates are also protected against
enzymatic degradation, and reduce the im-
munogenicity of the conjugated proteins
[35, 36]. At present, clinical trials for sev-
eral such compounds are ongoing.
1.3.1.4 Conjugates with poly
(L-glutamic acid) [poly(PGA)]
With poly(PGA), a biodegradable polymer
backbone is linked to drugs, used primar-
ily as chemotherapy. Conjugates with pa-
clitaxel (Cell Therapeutics Inc./Chugai
Pharmaceutical Co. Ltd.) showed good re-
sults in pre-clinical studies when adminis-
tered either alone or combined with radia-
tion [37]. A number of clinical studies are
also currently under way to determine effi-
cacy against various types of solid tumor
[3841].
1.3.1.5 Other Polymer Conjugates
Today, new polymers are continually being
sought as a means of introducing im-
proved peptide carriers to the clinics (see
Table 1.1 Polymerprotein conjugates available commercially, or in clinical development
Compound Name Status
(year to market)
Indication Reference
PEG-adenosine
deaminase
Adagen 1990 SCID syndrome 23
SMANCS Zinostatin,
Stimalmer
1993 (Japan) Hepatocellular carcinoma 24
PEG-L-asparaginase Oncaspar 1994 Acute lymphoblastic
leukemia
25, 26
PEG-a-interferon
2b
PEGINTRON
TM
2000 Hepatitis C 27
PEG-a-interferon
2b
PEGINTRON
TM
Various
clinical trials
Cancer, multiple sclerosis,
HIV/AIDS
27
PEG-a-interferon 2a, PEGASYS 2002 Hepatitis C 28
PEG-HGR Pegvisomant 2002
(approved EU)
Acromegaly 29
PEG-G-CSF PEGfilgrastim,
Neulasta
TM
2002 Prevention of neutropenia
associated with cancer
chemotherapy
30
EU, European Union; G-CSF, granulocyte colony-stimulating
factor; HGR, human growth hormone; HIV, human immuno-
deficiency virus; PEG, polyethylene glycol; SCID, severe com-
bined immunodeficiency; SMANCS, styrene maleic anhydride;
TNF, tumor necrosis factor (modified from Ref. [22]).
Part VI, Chapter 3). The quest for conju-
gates which circulate in the bloodstream
for longer periods of time, and accumulate
in the target tissue but which can also be
eliminated when necessary, has not yet
been fulfilled. Selective polymer degrada-
tion caused by enzymes or by a change in
pH have been investigated with the aim of
optimizing the half-life of conjugates [42,
43]. In this respect, cancer research is a
leading therapeutic area, as the relation-
ship of risk to benefit must be assessed in
a different manner than for other indica-
tions. Hence, it is in this area that new
compounds are most likely to be evalu-
ated.
1.3.2
Drug Delivery Systems (DDS)
Conjugates represent a very elegant and
promising means of protecting bio-com-
pounds. However, they create new chemi-
cal entities (NCEs) which have to undergo
the full regulatory process to reach the
clinic or the market, respectively. Many ga-
lenic approaches have been set up to solve
the problems of high degradation and rap-
id elimination by providing a type of envel-
ope to the protein or peptide. Some of
these have led to injectable depot formula-
tions, but for many DDS the intention is
to apply the drug via other routes than pa-
renteral administration, so that treatment
is more convenient. For less severe dis-
eases, this approach might be necessary to
improve the patients compliance with
therapy. Most of the DDS discussed in the
following section are not dedicated solely
to biopharmaceuticals, but have been
proved to be useful for a range of sensi-
tive compounds.
1.3.2.1 Implants
Although implants normally require mi-
cro-surgery, some can be applied using
large needles (16 gauge). Most implants
cannot be adapted to special needs, and so
a weight-based dosing (i.e., a defined
quantity of drug per kg body weight) is
not possible [46]. Recently developed meth-
ods involve titanium osmotic pump sys-
tems, which are implanted subcutaneously.
These pumps have been developed to deli-
ver leuprolide, a peptide to suppress tes-
tosterone in prostate cancer patients [44],
although they may also be useful for pro-
tein delivery in other indications [45].
Other implanted pump systems have been
designed which can be refilled by inserting
a needle into the pump reservoir. These
systems may be suitable for weight-based
dosing, but they require invasive surgical
procedures for their implantation and re-
moval.
1.3.2.2 Microreservoir Delivery Systems
for Injection
The advantage of a biodegradable delivery
system is a single procedure (no removal)
and suitability for injection with a conven-
tional needle and syringe [46]. Clearly, in-
jection site reactions to the matrices re-
quire extensive examination to ensure that
the matrix material is compatible with the
injection site and the protein [47]. It is es-
sential that pilot toxicology studies of the
biodegradable matrix be performed before
the development of matrices for protein
delivery.
1.3.2.3 Liposomes
Vesicles are spherical constructions of one
to several lamellae consisting of lipid bilayer
membranes separating an aqueous inner
core from the aqueous continuous phase.
The molecules consist of a hydrophilic head
and two hydrophobic tails that arrange
themselves spontaneously into vesicles.
The hydrophilic regions are directed to-
wards the aqueous phases, whereas the hy-
drophobic regions build the hydrophobic
core of the membrane. The structure of a li-
posome is shown schematically in Fig. 1.3.
Liposomes are artificial vesicles consist-
ing of phospholipid, cholesterol, and glyco-
lipids. Small unilamellae liposomes (SUV)
are smaller than 50 nm in size, and the
more lamellae included in the liposome
membrane, the larger the size of the lipo-
some. Multilamellae vesicles (MLV) can at-
tain a diameter from 100 nm to 1000 nm;
their structure was first demonstrated by
Bangham in 1965 [48].
Liposomes can provide a type of envel-
ope to protect solutions of biologicals
against chemical degradation. The many
application opportunities of liposomes are
based on their ability to associate with
other molecules. Hydrophilic molecules
can be incorporated, amphiphilic and ionic
molecules can be adsorbed at the surface,
and lipophilic compounds can be dissolved
within the membrane. The preparation of
liposomes can be performed using a vari-
ety of methods. Depending on the proper-
ties of the drug, high-pressure homogeni-
zation, extrusion or emulsion methods
with organic dissolution components can
be used. The choice of membrane lipids,
the number of lamellae, and the size of
the liposomes determine the loading ca-
pacity for drug molecules as well as the re-
lease pattern.
The encapsulation of drug molecules
also leads to delivery systems, especially in
the case of hydrophilic compounds where
a permeation barrier with depot effect is
provided. The permeation velocity is con-
trolled by the properties of the mem-
branes, as well as by the lipophilicity and
size of the incorporated drug. Even large
molecules are released slowly in the body,
but unfortunately this also occurs under
storage conditions. As liposomes do not
have a solid surface, an equilibrium is
built up between incorporated or adhered
drug and free drug molecules, and this
can lead to burst effects when liposome
dispersions are diluted.
Liposomes appear to be suitable delivery
systems due to their non-toxic and/or nat-
ural components, though their instability
against the acid pH in the stomach and
gastrointestinal enzymes limits their use
to parenteral and topical applications [49].
Amphotericin B, daunorubicin, and doxor-
ubicin were commercially available liposo-
mal parenteral drugs during the early
1990s [50], while liposomal preparations of
heparin and econazole were registered for
dermal application. Research investiga-
tions on purified and defined lipids offered
a basis for the development of the first li-
posomal vaccine formulation with inacti-
vated hepatitis A virions by the Swiss
authorities, and this was registered in
1994. Newer approaches have also used
liposomes as vectors for gene delivery (see
Part VI, Chapter 7).
Fig. 1.3 Schematic view (section) of a liposome
from phospholipids, which form a bilayer envelop-
ing a aqueous inner space. In this case, a small
unilamellar vesicle (SUV) is shown.
Various strategies have been developed
to create site-specific liposomes. Bioactive
linkers such as glycopeptides [51] and anti-
bodies [52, 53] have been used to create
site-specific systems. Here again, cancer
therapy has taken the leading role in de-
veloping new systems, and specific anti-
bodies for most known tumors are now
available. For example, antibodies against
colorectal cancer, ovarian cancer and pros-
tate cancer which recognize the specific
antigens of these tumors can be linked
with liposomes filled with anticancer
drugs. Some antibodies are or will be
available at a relatively low cost because
they can be produced by recombinant en-
gineering [54]. In anticancer research,
monoclonal antinuclear autoantibodies
(ANAs) with nucleosome-restricted speci-
ficity have shown special promise [55].
ANAs are specific against a variety of tu-
mors rather than a single type of tumor,
as are most anticancer antibodies. One
subclass of ANAs is non-pathogenic, and
some of these can differentiate between
cancer cells and healthy cells [56]. Further
examination of the ANAs might well lead
to new, specific delivery systems [17].
It has also been shown that simple car-
bohydrates can be bound to liposomes by
hydrophobic interactions. Unfortunately,
these bonds are not sufficiently stable, and
this may lead to desorption on dilution
and mechanical agitation, or destabiliza-
tion of the liposomal bilayer after coagula-
tion or peptization of the polysaccharides
[57]. However, in general polysaccharide
coatings stabilize the vesicular constructs
physico-chemically in bioenvironments
and make them site-specific. Consequently,
they can be used as anchor molecules to
target liposomes to specific organs and
cells [58]. The most promising results were
obtained by coating the liposomal surface
with natural or hydrophobized polysacchar-
ides (e.g., dextran, amylopectin, mannan)
or their palmitoyl or cholesteroyl deriva-
tives by covalent anchoring [59, 60].
Although liposomes are comprised of rec-
ognized non-toxic components, they cannot
provide ideal storage or release parameters
due to drug equilibrium inside and outside
the vesicles. Solid microparticles and nano-
particles have been investigated as a means
of overcoming this problem, and offer the
possibility of building physically and chemi-
cally stable molecules.
1.3.2.4 Polymeric Microspheres
and Nanospheres
Biodegradable hydrophobic microspheres
Several approaches have been attempted to
encapsulate peptides and proteins in solid
polymer capsules or matrices, and many
examples of in vitro data have been pub-
lished [61]. Among the biodegradable poly-
mers, poly(d,l-lactide-co-glycolide (PLGA)
may be the best known attempt to over-
come physical and chemical instability of
bio-compounds. This is achieved by build-
ing up a matrix that is degraded slowly in
the body and provides slow and controlled
release systems [6265]. Meanwhile, a vari-
ety of polymers is available which can be
used as biodegradable matrices for protein
delivery, including polyanhydrides, poly-
ester derivatives [6669], collagen [70], and
hyaluronic acid derivatives [71] (see Part
VI, Chapter 6). In general, there are two
types of microsphere:
The compound is either dissolved or dis-
persed within the material that builds a
matrix for the release (microparticle).
There may be a drug-loaded core sur-
rounded by a polymeric shell (microcap-
sule).
Possible distributions of the drug in micro-
spheres are shown schematically in Fig. 1.4.
Microspheres can be made by dispersing
or dissolving the drug in the polymer solu-
tion, and adding a second liquid that is
not a non-solvent for the polymer but pro-
vides a limited solubility for the polymer
solvent. There are two common methods
of producing microparticles: the solvent
deposition method; and the solvent eva-
poration method.
With the solvent deposition method, the
drug and matrix polymer/macromolecule
are dissolved in one organic solvent phase
that is miscible with water. This solution
is dispersed in an aqueous phase, and the
organic solvent diffuses into the continu-
ous water phase; this leads to the precipi-
tation of polymer and drug (deposition).
The deposition process involves desolva-
tion of the polymers, co-acervation, and fi-
nally precipitation to ultra-fine, solid drug-
loaded polymer particles [72].
In particle production by solvent eva-
poration, the drug and polymers are dis-
solved (solvation) in one organic solvent
phase that is not miscible with water. The
drug-loaded polymer solution is dispersed
in a water phase, yielding an oil in water
(O/W) emulsion. The organic solvent parti-
tions into the aqueous phase from which
it is removed by evaporation [73]. Usually,
double-emulsion techniques are used to
prepare the microcapsules. Water/oil/water
(W/O/W) or oil/oil/water (O/O/W) systems
are prepared by dissolving or dispersing
the compound in an aqueous or lipid me-
dium. The polymer solution is added, and
a first emulsion step results in a system
with drug-containing droplets dispersed in
the polymer solution. For the second
emulsion step, an aqueous phase is added.
The system is stirred, which leads to a
double emulsion (Fig. 1.5). The polymer
solvent mixes with the outer aqueous
phase, the saturation solubility of the poly-
mer is exceeded, and the polymer partici-
pates as a shell around the drug-loaded
core [74]. Electron micrographs of two dif-
ferent microcapsule formulations are
shown in Fig. 1.6; these consist of a biode-
gradable polymer, a biocompatible surfac-
tant, and the model drug.
All aqueous microsphere dispersions
must be dried after the capsules are com-
pletely hardened. Freeze-drying is a com-
mon method of gaining stable powders
that can be stored and applied. It is impor-
tant to minimize the amount of resting
water in the formulation, as not only the
drug can be hydrolyzed but also the poly-
mer. Polyesters and polyanhydrides both
degrade by hydrolysis catalyzed by water;
consequently, the release mechanism
usually depends on drug diffusion as well
as on polymer degradation.
Although these systems protect biologi-
cal drugs against enzymes and degradation
due to environmental conditions, other pa-
rameters must be carefully evaluated when
preparing microspheres. The elaboration
of these systems with hydrophobic poly-
Fig. 1.4 Drug distribution in microparticles. Left: dissolved in polymer matrix.
Center: dispersed in polymer matrix. Right: as a solid core in a polymer shell.
mers requires the use of aggressive condi-
tions such as organic solvents, emulsion
methods (e.g., sonication) or heat-produc-
ing stirring systems (e.g., Ultra Turrax)
that compromise the stability of the encap-
sulated molecule. Furthermore, the pH in-
side the capsules may be intolerably low
due to starting degradation of the polymer
(e.g., PLGA). Additional excipients must
be used to protect the biological com-
pound against its protective shell. For ex-
ample, alkaline buffer salts are added to
keep the pH around the drug within a
physiological range.
Solid lipid nanoparticles (SLN) SLN com-
bine the advantages of traditional particu-
late drug carriers, but simultaneously
avoid some of their major disadvantages.
Like polymeric particles, they provide a
modification of the release profile due to
the solid state of the particle matrix and
chemical stabilization that is, the protec-
tion of drugs incorporated into the solid
Fig. 1.5 Preparation of microcapsules with a double
emulsion technique. A primary water-in-oil (W/O)
emulsion (step 1) serves as the oil phase for
secondary emulsification step with a water phase
containing a hydrophilic surfactant (step 2). The
result is a multiple water-in-oil-in-water (W/O/W)
emulsion which leads to drug-loaded microparti-
cles after drying.
Fig. 1.6 SEM micrograph of PLG-microcapsules containing human
serum albumin (left) and cytochrome C (right) as model drug.
matrix against chemical degradation (e.g.,
hydrolysis). Similar to emulsions and lipo-
somes, SLN possess a low toxicity of the
excipients. As physiological lipids and exci-
pients of generally regarded as safe
(GRAS) status can be used, SLN show a
good tolerability [75]. Additionally, the pos-
sibility of large, industrial-scale production
by high-pressure homogenization is a ma-
jor advantage over most polymeric micro-
particles. SLN are produced by dispersing
a melted lipid or a melted mixture of lip-
ids in an aqueous solution. After the prep-
aration of a pre-emulsion (lipid droplet in
the range of micrometer size), a second
emulsification step is applied in a high-
pressure homogenizer. The resulting
emulsion contains nano-sized lipid drop-
lets in the aqueous phase, and is cooled
under gentle stirring to avoid aggregation.
The resulting nanoparticles can be freeze-
dried and stored either under ambient or
freezing conditions, depending on the in-
corporated drug [76]. Although this tech-
nique seems more suited to very hydro-
phobic compounds, proteins can also be
incorporated and delivered without any
damage to the drug [77].
Polymeric solutions building in situ implants
A major challenge is the limited dose con-
centration of microspheres due to drug-
loading constraints. Dispersions are un-
likely to contain more than 15% (m/m)
particles that usually provide a maximum
drug load of less than 20%. Considering
an injection volume of 12 mL, not more
than 60 mg drug substance can be applied.
This is not a concern for low drug dose in-
dications such as growth factors [64, 65,
78] or vaccines [79], but it may be an ob-
stacle for the application of other thera-
peutic peptides.
Less limited in dose, and less complex
with regard to the manufacturing process,
are the recent advances in the develop-
ment of injectable gel systems. These
might allow for a greater dose concentra-
tion. One factor for enhanced compliance
is the fairly low viscosity of these systems,
which allows application via a smaller di-
ameter needle (2527 G versus 2123 G
for microspheres). In these systems,
poly(d,l-lactide-co-glycolide) (PLG) or su-
crose acetate isobutyrate (SAIB) dissolved
in hydrophobic solvents [8082] are used
as the drug carrier. The solvents remain
with the gel after injection, and slowly dis-
solve into the tissues. Hydrophilic solvents
are also used that rapidly leave the system,
yielding a solid (PLG) or semi-solid (SAIB)
implant [80, 83, 84]. Other aqueous sys-
tems use co-polymers of lactide/glycolide
and PEG, and form a viscous gel (hydro-
gel) at physiological temperatures [66]. Pro-
tein release from these hydrogels depends
on diffusion through the gel, so the effect
of polymer degradation is less important
than in other systems. One issue with in-
jectable gel systems is the impact of injec-
tion volume and the choice of administra-
tion site on the morphology of the depot
and the effect on the drug-release profile.
Hydrophilic microspheres and nanospheres
In contrast to the hydrophobic polymers
discussed above, hydrophilic polymers of-
fer a less challenging environment for pro-
teins and peptides. Alginate and gelatin
have been used as capsule materials for
various compounds, but chitosan has
achieved particular interest within the past
few years. This deacetylated form of the
polysaccharide chitin, which can be ob-
tained from crustacean shells, offers sev-
eral physico-chemical advantages. Chitosan
is positively charged and can therefore re-
act with negatively charged macromole-
cules. Its interactive forces, as well as its
solgel transition stages, make chitosan an
interesting polymer for the delivery of
challenging labile macromolecular com-
pounds [8587]. Although chitosan has
been known since the 19th century, it was
found to be suitable for drug delivery only
in 1984 [88]. Again, cancer research had a
pioneering role as 5-fluoruracil was the
first incorporated drug. Several crosslink-
ing agents were used since the first-used
glutaraldehyde had a negative influence on
cell viability and reduced the principally
low toxicity of chitosan. Today, the gelati-
nation of chitosan on contact with polyan-
hydrides is used for gelatinations. For the
formation of nanoparticles, an alkaline tri-
polyphosphate (TPP) phase is mixed with
an acidic chitosan phase. Nanoparticles
form immediately upon mixing through
inter- and intramolecular linkages develop-
ing between the phosphate groups of TPP
and the amino groups of chitosan. Parti-
cles can be made from pure chitosan, but
modifications can be made by involving
other hydrophilic polymers or macromole-
cules such as polyethylene oxide or poly-
propylene oxide [89]. Thereby, the delivery
properties of the particles are changed as
well as their susceptibility to interact with
biological surfaces. The coating of chitosan
particles with PEG led to a less positive
surface charge of the particles, and im-
proved their biocompatibility [90]. Due to
the mild conditions required for their for-
mation, these chitosan nanoparticles
seemed to be very promising systems for
the association and delivery of sensitive
macromolecules. Proteins such as bovine
serum albumin (BSA), tetanus toxoid,
diphtheria toxoid [89] and the peptide in-
sulin [91] have been associated to these na-
noparticles.
The ability of chitosan nanoparticles to
encapsulate these molecules was seen to
be very high. To obtain the desired final
protein loading, it was sufficient to adjust
the amount of protein added to the formu-
lation. Protein loading reached values as
high as 50% (50 mg protein per 100 mg
nanoparticles), which is a much higher
loading capacity than in any other nano-
particulate protein carrier [85]. The use of
chitosan nanoparticles as a permeation en-
hancer and for nasal delivery are discussed
below.
1.3.3
Non-invasive Administration
Whilst the delivery systems described so
far have all been applied parentally, pa-
tients would appreciate more convenient
formulations that could be taken perorally,
or that at least do not require a painful in-
jection.
Several approaches were made to admin-
ister peptides and proteins by another
route, but they all require a certain ability
of the formulation to overcome the natural
barriers of the body that is, the mucosal
membranes (oral, buccal, rectal, vaginal,
nasal or inhalational application) or the
skin (transdermal systems) (see Part VI,
Chapter 5). Very often, some chemical
help is needed to establish suitable condi-
tions for non-invasively administered for-
mulations.
1.3.3.1 Penetration Enhancers
Permeation enhancers are auxiliary agents
that make membranes less impermeable
for drug molecules. Several working mech-
anisms are used:
Mucolytic compounds may reduce the
thickness of the mucus layers of muco-
sal membranes. N-Acetylcysteine, a well-
known mucolyticum against strong
cough, fluidizes the mucus by cleaving
the disulfide bonds that connect the mu-
cus glycoproteins [92]. However, the use
of sulfhydryl compounds must be care-
fully assessed with the drug molecule as
the disulfide bonds of the peptide will
also be cut. Thus, there is clearly no
problem when there are no disulfide
bonds within the molecule, but when
the conformation inhibits access of the
disulfide bonds to the sulfhydryl com-
pound, the protein may also remain
stable. Mucolytic enzymes such as tryp-
sin, bromelain or papain can be used
the same way as sulfhydryl compounds.
These provide an identical mode of ac-
tion, and therefore have the same re-
strictions [93].
Low molecular mass (LMM) permeation
enhancers include various classes of
substances. All molecules that interact
with the membrane lipids or proteins
and lead to membrane perturbation, in
principle, be applied and may be helpful
to guide drugs through the cells of the
mucosa. However, the paracellular route
seems more attractive because the drug
molecules cannot be degraded within
the cells. The tight junctions that cover
about 0.2% of the intestinal mucosa can
be opened by several small molecules
(Table 1.2). However, although only a
small area of the mucosa is harmed re-
versibly by these penetration enhancers,
intestinal damage or toxic side effects
caused by systemic circulation of the en-
hancers may occur [94].
The third class of permeation enhancers
are the polymeric enhancers. In compar-
ison to the LMM, these provide addi-
tional advantages as they can often also
be used as a particle-forming carrier. Ad-
ditional mucoadhesive properties keep
the delivery systems in the area of drug
adsorption [108]. The best known exam-
ple for cationic polymers is chitosan.
The enhancing effect for poorly absorb-
able drugs has been shown in cell mod-
els, as well as in rats. It is assumed that
Table 1.2 Use of LMM absorption enhancers on different tissues
[95].
Absorption enhancer Peptide drug used Site of
administration
Reference
Polyoxyethylene-24-cholesterol ether Octreotide Oral 96
Sodium caprate Insulin Oral 97
Sodium taurodihydrofusidate Insulin Vaginal 98
Phosphato-dihydrofusidate Leucine enkephalin Vaginal 99
Sodium glycodeoxycholate Buserelin Buccal 100
Sodium salicylate Insulin Rectal 101
Na
2
EDTA Insulin Rectal 102
n-Lauryl-b-maltopyranoside Insulin Colon 103
Zot Insulin Ileum 104
Phospholipids Desmopressin Caco-2 105
Taurodeoxycholate Insulin Nasal 106
Sodium tauro-24,25-dihydrofusidate Calcitonin Nasal 107
Polyoxyethylene-20-stearyl ether Gonadorelin Ocular 101
the interaction between the positively
charged amino groups of chitosan and
the negatively charged sialic groups of
membrane-bound glucoproteins open
the tight junctions [109]. In order to op-
timize the molecule, derivatives were in-
troduced. For example, N-trimethyl-chit-
osan can also be used for intrajejunal
administration as it is soluble at pH val-
ues above pH 1.5, in contrast to chitosan
[110].
Anionic polymeric permeation enhancers
have a different effect. They do not inter-
act directly with the membrane, but bind
and deplete Ca
2+
ions from the extracellu-
lar cell medium, which leads to an open-
ing of the tight junctions [111]. The prop-
erties of anionic polymers have also been
improved. By immobilization of free sulf-
hydryl groups onto various polymers, their
permeation-enhancing effect on hydrophi-
lic compounds is strongly increased [112].
In addition, thiolated polymers (thiomers)
show improved mucoadhesive properties
which allows them to concentrate in the
area of drug absorption [113].
1.3.3.2 Prodrugs and Analogs
Alteration of the physico-chemical proper-
ties of the molecules seems to be a labor-
ious, but promising, approach [114]. Com-
parable to the various kinds of conjugates
discussed above, they require the synthesis
of a new chemical entity (NCE), whilst
changes in chemical stability, solubility, li-
pophilicity configuration and enzyme liabi-
lity can also be achieved that facilitate
non-invasive administration [115].
1.3.3.3 Nasal Administration
The nasal cavity is generally a good ab-
sorption site for lipophilic drugs. Pharma-
cokinetic profiles often identical to those
obtained after an intravenous injection
have been observed, and bioavailabilities
can approach 100%. However, due to lim-
ited membrane permeability, the nasal ab-
sorption of polar drugs and especially
high molecular-weight polar drugs such as
peptides and proteins is limited. Polar
drugs with molecular weights <1000 Da
will generally pass the membrane using
the paracellular route through the tight
junctions between the cells since, although
tight junctions are dynamic structures and
can open and close to a certain degree
when needed, the mean size of the chan-
nels is less than 10 and the transport of
larger molecules is much more limited
[116]. Larger peptides and proteins can
pass the nasal membrane using an endo-
cytotic transport process, but the amounts
have been shown to be significantly lower
[117]. Additionally, the high mucociliary
clearance rapidly removes all substances
and particles that are not mucoadhesive,
with their half-life of clearance typically
about 1520 min [118]. Another barrier
that cannot be neglected is enzymatic ac-
tivity within the nasal cavity. Exopeptidases
can cleave peptides at their N and C termi-
ni (mono- and diaminopeptidases), and
endopeptidases attack internal peptide
bonds (e.g., serine, cysteine) [119]. How-
ever, in the past few years several formula-
tions have reached the market (Table 1.3).
Meanwhile, new formulations are under
development aiming for a better exploita-
tion of this route of administration as it of-
fers the advantage of a rapid onset of ac-
tion. Novel targets such as reaching the
brain or the cerebrospinal fluid are also
under discussion [120]. One promising
way to administer peptides and proteins
via the nose would be delivery systems
that protect the drugs from the enzymes,
show mucoadhesive properties, and open
the tight junctions. This is all fulfilled by
chitosan solutions or micro-/nanoparticles
that are currently under investigation for
this route of application. In general, it has
been shown that chitosan powder formula-
tions whether in the form of micro-
spheres or powders can in many in-
stances provide a better absorption-pro-
moting effect than chitosan solutions
[120]. These delivery systems seem to be
very effective in delivering small molecules
as well as peptides. Nasal bioavailabilities
of around 20% or more have been ob-
tained in clinical trials for peptides such
as leuprolide (1300 Da), salmon calcitonin
(S-CT, 3500 Da) and parathyroid hormone
(PTH, 4000 Da) across the nasal mem-
brane. For the luteinizing hormone-releas-
ing hormone (LHRH) analogue, goserelin,
it was shown in a sheep model that a chit-
osan microsphere formulation provided
bioavailabilities in the order of 40% rela-
tive to an intravenous injection [121]. For
some drugs, such as PTH, a chitosan pow-
der formulation is also the most appropri-
ate formulation due to stability problems
encountered with PTH solution formula-
tions.
1.3.3.4 Pulmonary Administration
Among the non-invasive routes, pulmo-
nary delivery shows the greatest promise
because of the higher protein bioavailabili-
ty compared to transdermal or oral deliv-
ery. Depending upon the protein, bioavail-
ability for this route of application has
been reported to range from *10% to
50% in animals [122]. The lung absorption
of proteins may be limited by proteases
that reduce the overall bioavailability [123].
Studies of pulmonary insulin delivery sug-
gest that the overall bioavailability in hu-
mans is in the order of 1015% (2030%
relative to subcutaneous injection), which
is much lower than the published bioavail-
ability in animals (57%). This shows that
animal experiments using this route of de-
livery are clearly not predictive of the hu-
man experience. A major concern for the
delivery of proteins to the lungs is clinical
toxicology in this tissue. Proteins such as
growth factors and cytokines have a local
effect on the tissues which might lead to
severe side effects. Another concern is that
of dose level. For example, monoclonal an-
tibodies may pose less of a safety risk, but
doses for systemic administration are typi-
cally in the region of 2 mg protein per kg
body weight, making pulmonary delivery
not feasible due to the mass requirements
of more than 100 mg per dose. Also,
weight-based dosing is difficult to achieve
with pulmonary delivery because normally
a single-dose configuration is used. This
means that patients would need to use dif-
ferent dosage sizes and multiple adminis-
trations in order to achieve the desired
dose. However, novel devices in combina-
tion with particles that are designed to
reach the deep lung are under develop-
ment. The developed dispersed spheres in-
clude a liquid drug carrier (droplets) [124]
as well as solid and porous microparticles
[122, 125]. The droplets and the solid parti-
Table 1.3 Nasal delivery of peptides and proteins
Nasal formulation with Company
Salmon calcitonin Novartis
Desmopressin Ferring
Buserelin Aventis
Nafarelin Searle
PTH In clinical trials
Leuprolide In clinical trials
Insulin In clinical trials
Interferon In clinical trials
PTH, parathyroid hormone.
cles each have a diameter of about 1
3 lm, which is the typical diameter of par-
ticles that can reach the deep lung. The
porous particles are about 10 lm in size
but, due to their lower density, they pro-
vide the same aerodynamic properties as
the small particles. Particle formulations
with insulin, interferons, and other pro-
teins are currently in clinical trials with
these systems (see Part VI, Chapter 4).
Similar to nasal delivery, rapid protein
absorption and a fast onset of peak serum
levels are also observed after pulmonary
delivery. This type of exposure pattern is
preferred for some indications, but could
also cause unwanted side effects. If a con-
stant serum level is desired, then twice
daily dosing may be needed to achieve the
required serum exposure profile. The de-
velopment of slower-absorbing dosage
forms requires the use of slower-dissolving
particles or controlled release formula-
tions. By using insoluble insulin in large
porous particles, a longer duration of insu-
lin exposure (12 h) is observed in rats. In-
deed, it is suggested that this method may
yield a long-acting insulin preparation
[125]. Alternatively, controlled release sys-
tems such as biodegradable nanospheres
or microspheres can provide a longer expo-
sure time and lower peak serum levels
[126, 127]. However, with regard to the
current limitations of rapid absorption and
bioavailability, the pulmonary delivery of
peptides and proteins may be limited to
drugs that require low doses and do not
show side effects as a result of high peak
serum levels or local tissue reactions.
1.3.3.5 Buccal Administration
The oral mucosa can be divided into a
non-keratinized area consisting of the floor
of the mouth (sublingual), the buccal mu-
cosa (cheeks), and a keratinized area com-
prising the gum (gingiva), the palatal mu-
cosa, and the inner side of the lips. The
mucous membranes have a total area of
100 cm
2
, and show differences in struc-
ture, thickness and blood flow depending
on their location within the oral cavity.
The buccal mucosa consists principally of
two components: the epithelium, and the
underlying connective tissue. The interface
between these two layers is formed by the
basal complex. The rapid turnover of the
epithelial cells is an important feature of
the oral cavity as it makes drug absorption
more difficult due to continually changing
permeability characteristics [128].
Among the mucosal routes, peptide
transport through the buccal mucosa was
found to be much less sensitive to degra-
dative enzymes than nasal, vaginal, and
rectal administration [129, 130]. The buccal
mucosa seems to be deficient in proteases
such as pepsin, trypsin and chymotrypsin;
these are present in gastric and intestinal
secretions and are known to contribute to
peptide hydrolysis. However, the enzy-
matic barrier includes exo- and endopepti-
dases. Aminopeptidases (exopeptidases)
present in the buccal mucosa and carboxy-
peptidase were shown to be primarily re-
sponsible for protein degradation [129,
131]. Some of the strategies to overcome
the buccal membrane have already been
discussed, including the addition of per-
meation enhancers and the composition of
drug conjugates. However, site-specific de-
livery systems have also been developed.
Three types of non-attached drug delivery
systems can be defined: fast-dissolving ta-
blets; chewing-gum; and microporous hol-
low fibers. These all release the drug in
the mouth, but the lack of intimate contact
with the mucosa might not be favorable
for peptide absorption due to possible en-
zymatic degradation in saliva. Although
some interesting results have been re-
ported, non-attached buccal mucosal drug
delivery has many drawbacks due to the
local physiological environment for ex-
ample, the presence of saliva and the in-
take of foods or liquids. Bioadhesive deliv-
ery systems are immobilized and therefore
provide an intimate contact between the
drug and the buccal mucosa. A high drug
concentration is maintained at the adsorp-
tive surface during the application time,
and enzymatic degradation of the drug is
low due to limited contact with the saliva.
This concept has been applied to different
systems including sustained release ta-
blets, semi-solid dosage forms, films, and
patches [132]. A variety of polymers has
been described as being suitable for buccal
patches and hydrogel-based delivery sys-
tems. Polymers derived from natural
sources such as agarose, gelatin, and the
cellulose derivatives hyaluronic acid and
chitosan have been used [132]. Synthetic
mucoadhesive biocompatible polymers
such as polyacrylates and co-polymers of
polyacrylic acid were reported to be feasi-
ble [133]. Furthermore, a stabilizing pro-
teolysis-inhibitory effect was observed with
some polymers, such as polycarbophil
[134] and various Carbopols [135].
1.3.3.6 Oral Application
Clearly, the most accepted and perhaps
most convenient application form is the
peroral tablet or capsule. The environment
of the gastrointestinal tract is quite chal-
lenging for proteins and peptides, as they
must be protected against enzymes in the
stomach as well as in the small intestine.
Moreover, when colon delivery is desired,
the microflora of the large intestine must
also be taken into account. Furthermore,
the changing pH along the transition
route and the epithelial tight junctions in
the gut limit the degree of freedom of the
manifold formulation principles (Table
1.4).
Surface active agents with bioadhesive
microspheres or nanospheres have been
discussed in the previous section. As these
systems are suited to the application on
mucosal membranes, they are also used in
peroral formulations. Furthermore, the for-
mulation of proteins together with so-
called carriers has been developed during
the past few years. These carriers range
from N-acylated a-amino acids [136] to aro-
matic amides and sulfamides [137]; the
mechanism of the latter system remains
the subject of investigation. Carriers do
Table 1.4 Oral formulation principles
Variation of absorption mechanism Enzyme inhibition
Carriers
Enhancers
Prodrugs
Targeting transporters
Mucoadhesion
Formulation principles Microparticles
Nanoparticles
Emulsions
Liposomes
Variation of adsorption site Buccal delivery
Colon delivery
Delivery to the small intestine
not permeabilize the epithelium of the
small intestine, as do surfactant permeation
enhancers, neither do they form a regular
encapsulation shell, as in microsphere or
nanosphere systems. It is assumed that car-
rier molecules in high concentrations in-
duce a conformational change from a
folded, native structure to a partially un-
folded or molten globule state in significant
populations of protein molecules. It is pos-
sible that this conformation is in some
way responsible for the improved protein
absorption following oral administration
[137]. However, bioavailability was found
to be considerably higher (up to 100-fold
for insulin) in animals with carriers com-
pared to bioavailability without carriers,
though the total uptake was still less than
0.1% for the tested proteins. This leads to
feasibility problems, as for example the dai-
ly dose of recombinant human growth hor-
mone (rhGH) for an average pediatric pa-
tient might reach 1.5 g [46], which is unfa-
vorable both with regard to cost and patient
compliance. Nevertheless, carrier-mediated
delivery has shown promise for the admin-
istration of heparin [138].
1.3.3.7 Colon Delivery
Peptides and proteins are also interesting
potential candidates for colon-specific drug
delivery. Several strategies have been devel-
oped to target the release of drugs to the
colon. The formation of prodrugs has al-
ready been discussed; for example, sulfasa-
lazine, olsalazine and ipsalazide are pro-
drugs of 5-amino salicylic acid (5-ASA)
that have been developed as colon-specific
delivery systems to treat inflammatory
bowel disease (IBD) [139]. The diverse mi-
croflora and various enzymes present in
the colon have recently been exploited to
release drugs in the colon. Although the
large intestine is a potential site for the ab-
sorption of drugs, there are several diffi-
culties in delivering drugs effectively to
the colon. Changing pH conditions and
the long, patient-dependent transition time
during the passage of drug formulations
from mouth to colon lead to many techni-
cal difficulties. Thus, safe and predictable
drug delivery to the colon remains a chal-
lenge. However, new hope for the effective
targeting of drugs to the colon has been
provided recent developments in pharma-
ceutical technology, including the coating
of drugs with bacteria-degradable, pH-sen-
sitive polymers, or embedding drugs in
bacteria-degradable matrices. The use of
enteric coating is the most common meth-
od. The change in pH that varies continu-
ously down the gastrointestinal tract is
used as a trigger to release a drug in the
colon. The most commonly used pH-de-
pendent coating polymers are methacrylic
acid co-polymers, known as Eudragits R
(Rhm Pharmaceuticals, Darmstadt, Ger-
many), and especially Eudragit RL and Eu-
dragit RS. These polymers contain ioniz-
able carboxyl groups, which means that
the coating remains intact in the acidic en-
vironment of the stomach. However, as
the pH increases in the small intestine the
coating begins to dissolve in the alkaline
medium, causing the drug to be released
[140]. Polysaccharide and azopolymer coat-
ings have been used as carriers for colon-
specific targeting, because they are stable
in the stomach and small intestine, but
are degraded by the colonic bacteria [141].
Polysaccharidases are bacterial enzymes
which are available in sufficient quantity
to be used in colon targeting of drugs. Fol-
lowing this approach, various polysacchar-
ides such as amylose guar gum, dextran,
pectin, inulin, and chitosan have been in-
vestigated for colon-specific drug release.
[141]. Of course, bioadhesion has also been
proposed as a means of improving the per-
formance and extending the mean resi-
dence time of colonic drug delivery sys-
tems [142, 143].
It seems necessary to combine several of
these different approaches within one for-
mulation to utilize the accumulation of di-
verse enhancing effects to reach acceptable
bioavailability via the oral route (see Table
1.4), as has been achieved with the bioad-
hesive, particle-forming polymer chitosan
that is able to open tight junctions.
Today, the oral delivery of protein and
peptides has graduated from the proof-of-
concept stage to the product-development
stage. Until now, over 100000 human do-
sages with macromolecules have been de-
veloped, and this represents the range of
potential candidates for oral delivery. The
future of products developed for the oral
administration of macromolecules will de-
pend on the major multi-national pharma-
ceutical companies, which have the re-
sources for production, development, and
marketing. Clearly, when such companies
investigate such technology, the proof-of-
concept can quickly be turned into a com-
mercial success. Thus, the first oral prod-
ucts to reach the market-place might in-
clude oral versions of heparin, insulin, cal-
citonin, and PTH [144]. Although the time
frame for a commercial launch of oral
macromolecules can be estimated to be at
least 35 years, the probable continued
clinical success with these initial com-
pounds will doubtless be the signal to be-
gin the development of many other impor-
tant therapeutic products. Nevertheless, it
is likely that effective oral formulations for
peptides and proteins will remain highly
compound-specific [145].
1.3.3.8 Transdermal Systems
The skin represents the ultimate barrier
for proteins and peptides. Indeed, one of
the main functions of the skin is to pre-
vent access to the body of environmental
substances that might be potentially toxic
or infective. Since the days of Cleopatra
taking a bath in asss milk, man has at-
tempted to deliver drugs via the skin.
However, it was not until 1980 that the
first true transdermal therapeutic system
(TTS) entered the market as a patch re-
leasing scopolamine against nausea [146].
Currently, several TTS with small mole-
cules are available commercially, but mo-
lecular weight of the drug molecules is re-
stricted to masses below 100 Da [147].
Nevertheless, novel techniques were devel-
oped which attempted to overcome the
major hurdle of the outermost skin layer,
the stratum corneum. Various drugs may
be applied by means of transfersomes,
needle-free injections and methods using
physical resources such as iontophoresis
[148] (electroporation) and ultrasound [149]
(sonoporation).
Transfersomes are special ultra-adaptable
(mixed) lipid aggregates that provide a
high deformability. This flexibility in shape
is achieved through a positive feedback
mechanism that allows the transfersome
to adjust to the ambient conditions [150].
When a transfersome is exposed to vari-
able local stress, that enforces aggregate
elongation, the individual aggregate com-
ponents begin to separate into different re-
gions of the membrane. The more soluble
ingredient(s) accumulate in the areas of
greatest local deformation and push the
less-soluble ingredient(s) into the least de-
formed aggregate areas. It is therefore
favorable to use mixed lipid vesicles, in-
cluding one or more poorly soluble ingre-
dients and at least one more-soluble am-
phipathic ingredient (which can form
much smaller spheroid self-aggregates). In
a non-deformed quasi-spherical aggregate,
all components are distributed at random,
whereas in a deformed aggregate the
most-soluble molecules have a higher af-
finity to accumulate in the most strongly
curved parts of the vesicles.
By reconfiguring their shape, transfer-
somes can travel through the more hydro-
philic parts of the very hydrophobic stra-
tum corneum [151]. Drugs of various hy-
drophilicities and molecular weight have
already been incorporated into these ultra-
deformable carriers, and have proved able
to cross the skin. For example, glycody-
namic measures in patients with type 1
diabetes mellitus showed good results in
glucose reduction in the blood after epicu-
taneous administration of transfersomes
containing insulin [152].
Other techniques of transporting large
molecules across the skin involve micro-le-
sions of the stratum corneum. Electropora-
tion and sonoporation generate transcellu-
lar or intercellular hydrophilic pathways
through the stratum corneum, which re-
main open for a day or longer depending
on the method used and the width of a
channel [153]. With transdermal iontophor-
esis, charged solutes are transported by
electric repulsion from an electrode. Addi-
tionally, the flow of electric current may
enhance permeability of the stratum cor-
neum, and electro-osmosis may influence
the transport of non-ionic compounds and
larger polar peptides. The erythema asso-
ciated with iontophoresis usually resolves
within 24 hours of treatment [154]. Besides
minor uncomfortable side effects such as
involuntary muscle contractions at the in-
stant of the electric pulse, severe reactions
such as osteomyelitis, dysphagia, pharyn-
gocutaneous fistulas and intense general
wound breakdown have been described
[155]. Iontophoresis seems to be useful for
enhancing the penetration of peptides with
a molecular weight up to 7 kDa, but it has
been reported to have no effect on larger
molecules [156]. To date, sonophoresis has
been investigated for 50 years, and
although encouraging results for the deliv-
ery of insulin have been reported, further
research is necessary to answer all ques-
tions with regard to safety issues [157].
Although efforts at developing miniature
and relatively inexpensive power sources
will be important for patient use and con-
venience, the enhancing effect of low-fre-
quency ultrasound on protein permeability
suggests a possible future application of
this approach for needleless, painless in-
jections of proteins.
Painless injections can also be realized
by air pressure-driven advices, and so-
called pen systems are already available
for insulin. Although these still involve the
use of needles, the pressure-driven applica-
tion is favored by patients as a painless al-
ternative [158].
Furthermore, microneedle systems have
recently been the subject of investigation.
Needles with a diameter of between 2 and
10 lm can be used to overcome the stra-
tum corneum, leaving only very small
holes that are comparable to those caused
by the techniques mentioned above. The
needles can be solid and sharp (solid nee-
dle), hollow and sharp (hollow needle), or
hollow and blunt (hollow tubes). They
have a length of about 50100 lm, which
is not long enough to reach the pain adap-
tors in the deeper skin. When assembled
in an 33 mm array of about 400 needles,
the total diameter of the microneedles is
similar to a 30-G needle, and this provides
a suitable application tool. Initial trials
with the application of insulin have shown
promising results [159]. Examples of mi-
croneedles are shown in Fig. 1.7 (from
Refs. [160, 161]).
1.4
Outlook
The ideal delivery system for all routes of
administration should release its contents
only at a desired region of absorption,
whereas the drug delivery system attaches
by specific interaction with surface deter-
minants specific for that region. The deliv-
ery system should travel independently of
the transitory constraints that are typical
for the route of administration. As yet,
such a delivery system is unavailable, but
it would benefit peptide and protein drugs
as well as other poorly absorbed drugs
[162]. The use of novel injection devices in
the parenteral administration of proteins
has today become more common for daily
protein administration (e.g., for insulin
and rhGH). During the next few years,
this will continue to be the quickest and
least expensive new delivery system for
commercialization. Depot delivery systems
already provide opportunities to improve
patient compliance through fewer injec-
tions, and these systems might also in-
crease the viability of local delivery for
drugs which are not well tolerated systemi-
cally, or are degraded too rapidly to provide
therapeutic drug levels after systemic ad-
ministration.
Among the non-invasive routes, pulmo-
nary delivery has shown the greatest pro-
mise because of its greater protein bio-
availability compared to transdermal or
oral delivery. Phase III clinical trials of
pulmonary insulin delivery are ongoing,
and the results will undoubtedly demon-
strate the capability of that method. The
development of improved non-invasive
routes for the delivery of biomolecules
must continue to allow the biopharmaceu-
ticals to contribute to a significant extent
in the treatment of diseases where, at
present, no sufficient medication is avail-
able. Moreover, although the main princi-
ples of the different routes of administra-
tion are well known, major development
efforts must be made to provide the re-
quired delivery systems for these promis-
ing new biopharmaceutical drugs.
Finally, the advanced drug delivery sys-
tems to be used for the new biological en-
tities will have to make the transition from
research to market with the associated is-
sues, including product stability, large-
scale production of both drugs and formu-
lations, and also the regulatory require-
Fig. 1.7 Left: Microscopic view of an array of microneedles
shown next to the tip of a typical hypodermic needle [160]
(Copyright (2003) National Academy of Sciences, U.S.A.).
Middle and right: SEM micrographs of Hypodermic micro-
needle design (reprinted from [161], Copyright (2003), with
permission from Elsevier).
ments for these mainly macromolecular
and thus inhomogeneous drugs and de-
livery systems [163]. Only when these
problems have been overcome the full po-
tential of biopharmaceutical therapeutics
can be realized.
References
1 Gloff CA, Benet LZ: Pharmacokinetics and
protein therapeutics, Adv Drug Del Rev 1990,
4, 359386.
2 Yamahara H, Lee VHL: Drug metabolism in
the oral cavity, Adv Drug Del Rev 1993, 12,
2540.
3 Caliceti P, Veronese FM: Pharmacokinetic and
biodistribution properties of poly(ethylene gly-
col)protein conjugates, Adv Drug Del Rev
2003, 55, 12611277.
4 Yamamoto A, Hayakawa E, Lee VHL: Insulin
and proinsulin in mucosal homogenates of
the albino rabbit: implications in peptide de-
livery from normal routes, Life Sci 1990, 47,
24652474.
5 Lehr CM: (1991) Bioadhesive Drug Delivery
Systems for Oral Application. Ph.D. Thesis,
Leiden University, Netherlands.
6 Bundgaard H, Moss J: Prodrug of peptides. 6.
Bioreversible derivatives of thyrotropin-releas-
ing hormone (TRH) with increased lipophili-
city and resistance to cleavage by TRH-specific
serum enzymes, Pharm Res 1990, 7, 885892.
7 Bocci V: Catabolism of therapeutic proteins
and peptides with implications for drug deliv-
ery, Adv Drug Deliv Rev 1990, 4, 149169.
8 Latham PW: Therapeutic peptides revisited,
Nat Biotechnol 1999, 17, 755757.
9 Opalinska JB, Gewirtz AM: Nucleic-acid thera-
peutics: Basic principles and recent applica-
tions, Nat Rev Drug Disc 2002, 1, 503514.
10 Gewirtz AM, Stein CA, Glazer PM: Facilitat-
ing oligonucleotide delivery: helping antisense
deliver on its promise. Proc Natl Acad Sci
USA 1996, 93, 31613163.
11 Roerdink FH, Regts J, Van Leeuwen B,
Scherphof GL: Intrahepatic uptake and pro-
cessing of intravenously injected small unila-
mellar vesicles in rats. Biochim Biophys Acta
1984, 770, 195202.
12 Turner NC, Wright NA: The response to in-
jury. In: McGee JOD, Isaacson PG, Wright
NA (Eds.), Oxford Textbook of Pathology.
1992, pp. 351390, Oxford University Press,
Oxford, UK.
13 Moghimi SM, Hunter AC, Murray JC: Long-
Circulating and Target-Specific Nanoparticles:
Theory to Practice, Pharmacol Rev 2001, 53,
283318.
14 Takakura Y, et al.: Control of pharmaceutical
properties of soybean trypsin inhibitor by con-
jugation with dextran. II: Biopharmaceutical
and pharmacological properties, J Pharm Sci
1989, 78, 219222.
15 Veronese FM, Harris JM: Introduction and
overview of peptide and protein pegylation,
Adv Drug Deliv Rev 2002, 54, 453456.
16 Harris JM, Chess RB: Effect of pegylation on
pharmaceuticals, Nat Rev Drug Disc 2003, 2,
214221.
17 Torchilin VP, et al: Peptide and protein drug
delivery to and into tumors: challenges and
solutions, DDT 2003, 8, 259266.
18 Maeda H, et al.: Mechanism of tumor-targeted
delivery of macromolecular drugs, including
the EPR effect in solid tumor and clinical
overview of the prototype polymeric drug
SMANCS, J Control Release 2001, 74, 4761.
19 Davis FF: The origin of pegnology, Adv Drug
Del Rev 2002, 54, 457458.
20 Veronese FM, Harris JM (Eds.): Special issue:
Peptide and protein PEGylation, Adv Drug
Del Syst 2002, 54, 453609.
21 Roberts MJ, Bentley MD, Harris JM: Chemis-
try for peptide and protein PEGylation, Adv
Drug Deliv Rev 2002, 54, 459476.
22 Duncan R: Pharmacokinetic and biodistribu-
tion properties of poly(ethylene glycol)protein
conjugates, Nat Rev Drug Disc 2003, 2, 347360.
23 Levy Y, et al.: Adenosine deaminase deficiency
with late onset or recurrent infections:
response to treatment with polyethylene glycol
modified adenosine deaminase, J Pediatr
1988, 113, 312317.
24 Maeda H, Konno T. In: Maeda H, Edo K, Ishi-
da N (Eds.), Neocarzinostatin: The Past, Pres-
ent, and Future of an Anticancer Drug.
pp. 227267, Springer, Berlin, 1997.
25 Graham ML: PEGASPARAGINASE: a review
of clinical studies, Adv Drug Deliv Rev 2003,
55, 12931302.
26 Muss HB, Spell N, et al.: A phase II trial of
PEG-L-asparaginase in the treatment of non-
Hodgkins lymphoma, Invest New Drugs
1990, 8, 125130.
References 1385
27 Wang Y-S, et al.: Structural and biological
characterisation of pegylated recombinant in-
terferon a-2b and its therapeutic implications,
Adv Drug Deliv Rev 54, 2002, 547570.
28 Reddy KR, Modi MW, Pedder S: Use of pegin-
terferon, a-2a (40KD) (Pegasys
) for the treat-

ment of hepatitis C, Adv Drug Deliv Rev
2002,54, 571586.
29 Mukherjee A, et al.: How to optimise pegviso-
mant treatment of acromegaly safely?, Endo-
crine Abstracts 2002, 4, OC8.
30 Kinstler O, et al.: Mono-N-terminal poly(ethy-
lene glycol)-protein conjugates, Adv Drug De-
liv Rev 2002, 54, 477485.
31 Duncan R, Coatsworth JK, Burtles S: Preclini-
cal toxicology of a novel polymeric antitumour
agent: HPMA copolymerdoxorubicin (PK1),
Hum Exp Toxicol 1998, 17, 93104.
32 Gianasi E, et al.: HPMA copolymers platinates
containing dicarboxylato ligands. Preparation,
characterisation and in vitro and in vivo evalu-
ation, J Drug Targeting 2002, 10, 549556.
33 Meerum Terwogt JM, et al.: Phase I clinical
and pharmacokinetic study of PNU166945,
a novel water soluble polymer-conjugated pro-
drug of paclitaxel, Anticancer Drugs 2001, 12,
315323.
34 Schoemaker NE, et al.: A phase I and pharma-
cokinetic study of MAG-CPT, a water soluble
polymer conjugate of camptothecin, Br J Can-
cer 2002, 87, 608614.
35 Mu Y, et al.: Bioconjugation of laminin pep-
tide YIGSR with poly(styrene co-maleic acid)
increases its antimetastatic effect on lung me-
tastasis of B16-BL6 melanoma cells, Biochem
Biophys Res Commun 1999, 255, 7579.
36 Maeda H: SMANCS and polymer-conjugated
macromolecular drugs: advantages in cancer
chemotherapy, Adv Drug Deliv Rev 2001, 46,
169185.
37 Li C, Yu DF, Newman RA, Cabral F, Stephens
LC, Hunter N, Milas L, Wallace S: Complete
regression of well-established tumors using a
novel water-soluble poly(L-glutamic acid)-pacli-
taxel conjugate, Cancer Res 1998, 58, 2404
2409.
38 Kudelka AP, Verschraegen CF, Loyer E, Wal-
lace S, Gershenson DM, Hon J, Ho L, Gar-
zone PD, Warner M, Bolton MG, Kavanagh
JJ: Preliminary report of a phase I study of es-
calating dose PG-paclitaxel (CT-2103) and
fixed dose cisplatin in patients with solid tu-
mors, Proc Am Soc Clin Oncol 2002, 21, 2 Pt
2, Abs 2146.
39 Garzone PD, Mitchell P, Bolton MG, Nemu-
naitis J: Preliminary report of a phase I study
of combination chemotherapy using CT-2103
and carboplatin in patients with solid tumors,
Proc Am Soc Clin Oncol 2002, 21, 2 Pt 2, Abs
2161.
40 Sabbatini P, Brown J, Aghajanian C, Hensley
ML, Pezzulli S, Flaherty C, Lovegren M, Funt
S, Warner M, Mitchell P, Bolton MG, et al: A
phase I/II study of PG-paclitaxel (CT-2103) in
patients (pts) with recurrent ovarian, fallopian
tube, or peritoneal cancer, Proc Am Soc Clin
Oncol 2002, 21, 1 Pt 2, Abs 871.
41 Harper H, Marsland T, Mitchell P, Infortuno
J, Bolton MG: Phase II study of first line che-
motherapy using CT-2103 in patients with
non-small cell lung cancer who are >70 years
of age or who have PS=2, Proc Am Soc Clin
Oncol 2002, 21, 2 Pt 2, Abs 2685.
42 Hreczuk-Hirst D, Chicco D, German L, Dun-
can R: Dextrins as potential carriers for drug
targeting: Tailored rates of dextrin degradation
by introduction of pendant groups, Int J
Pharm 2001, 230, 5766.
43 Tomlison R, Heller J, Duncan R, Brocchini S:
Pendent chain functionalised polyacetals that
display pH-dependent degradation: A platform
for the development of novel polymer thera-
peutics, Macromolecules 2002, 35, 473480.
44 Fowler JE, Flanagan M, Gleason DM, Klim-
berg IW, Gottesman JE, Sharifi R: Evaluation
of an implant that delivers leuprolide for one
year for the palliative treatment of prostate
cancer, Urology 2000, 55, 639642.
45 Stevenson CL, Tan MM: Solution stability of
salmon calcitonin at high concentration for
delivery in an implantable system, J Peptide
Res 2000, 55, 129139.
46 Cleland JL, Daugherty A, Mrsny R: Emerging
protein delivery methods, Curr Opin Biotech-
nol 2001, 12, 212219.
47 Anderson JM, Langone JJ: Issues and perspec-
tives on the biocompatibility and immunotoxi-
city evaluation of implanted controlled release
systems. J Control Release 1999, 57, 107113.
48 Bangham A, et al.: Diffusion of univalent ions
across lamellae of swollen phospholipids, J
Mol Biol 1965, 13, 238252.
49 Schubert R. In: Mller R, Hildebrand G (Eds),
Liposomes in drug formulations. Pharmazeu-
tische Technologie Moderne Arzneiformen,
pp. 219237. WV GmbH, Stuttgart, 1998.
50 Ghyczy M. In: Mller R, Hildebrand G (Eds.),
Drugs with phosphatidylcholine and lipo-
somes: Development, assessment, perspec-
tives, Pharmazeutische Technologie Moderne
Arzneiformen, pp. 207218. WV GmbH,
Stuttgart, 1998.
51 Yamaouchi, et al.: Effects of sialic acid deriva-
tive on long circulation time and tumor con-
centration of liposomes, Int J Pharm 1995,
113, 141148.
52 Wright S, et al.: Antibody-directed liposomes
as drug-delivery vehicles, Adv Drug Deliv Rev
1989, 3, 343389.
53 Maruyama K, et al.: Characterization of in vivo
immunoliposomes targeting to pulmonary en-
dothelium, J Pharm Sci 1990, 74, 978984.
54 Jurcic JG, et al.: Monoclonal antibody therapy
of cancer. Cancer Chemother Biol Response
Modif 1997, 17, 195216.
55 Iakoubov, LZ, Torchilin VP: A novel class of
antitumor antibodies: nucleosome-restricted
antinuclear autoantibodies (ANA) from
healthy aged nonautoimmune mice, Oncol
Res 1997, 9, 439446.
56 Iakoubov L, et al.: Anti-nuclear autoantibodies
of the aged reactive against the surface of tu-
mor but not normal cells, Immunol Lett 1995,
47, 147149.
57 Sunamoto J, et al.: Protein anchored and poly-
saccharide anchored liposomes as drug car-
riers, CRC Crit Rev Ther Drug Carrier Syst
1986, 2, 117136.
58 Vias S: Potential of polysaccharide anchored
liposomes in drug delivery, targeting and im-
munization, J Pharm Pharmacol Sci 2001, 4,
138158.
59 Sunamoto J, et al.: Improved drug delivery to
target specific organs using liposomes an-
chored with polysaccharides. In: Chiellini E,
Giusti P (Eds.), Polymers in Medicine,
pp. 157168. Plenum Press, NY, 1984.
60 Akiyoshi K, et al.: Cell specificity of polysac-
charide derivatives on liposomal surface,
Chem Lett 1990, 473, 19901992.
61 Cleland JL, Langer R: Formulation and deliv-
ery of proteins and peptides: design and devel-
opment strategies. In: Cleland JL, Langer R
(Eds.), Protein Formulations and Drug Deliv-
ery. Washington, DC: ACS Symposium Series,
pp. 121. American Chemical Society, 1994.
62 Bartus RT, Tracy MA, Emerich DF, Zale SE:
Sustained delivery of proteins for novel thera-
peutic agents, Science 1998, 281, 11611162.
63 Langer R: Biomaterials in drug delivery and
tissue engineering: one laboratorys experi-
ence, Accounts Chem Res 2000, 33, 94101.
64 Benoit JP, Faisant N, Venier-Julienne MC,
Menei P: Development of microspheres for
neurological disorders: from basics to clinical
applications, J Control Release 2000, 65, 285
296.
65 Duenas ET, Park A, Daugherty A, Kahn J, Ko-
walski J, Cuthbertson A, Cleland JL: Develop-
ment of poly-(D,L-lactide-co-glycolide) micro-
sphere formulations containing recombinant
human vascular endothelial growth factor to
promote local angiogenesis, J Control Release
2001, 69, 1324.
66 Jeong B, Choi YK, Bae YH, Zentner G, Kim
SW: New biodegradable polymers for inject-
able drug delivery systems, J Control Release
1999, 62, 109114.
67 Jung T, Breitenbach A, Kissel T: Sulfobuty-
lated poly(vinyl alcohol)- graft-poly(lactide-co-
glycolide)s facilitate the preparation of small
negatively charged biodegradable nano-
spheres, J Control Release 2000, 67, 157169.
68 Breitenbach A, Li YX, Kissel T: Branched bio-
degradable polyesters for parenteral drug de-
livery systems, J Control Release 2000, 64,
167178.
69 Heller J, Barr J, Ng SY, Shen HR, Schwach-
Abdellaoui K, Einmahl S, Rothen-Weinhold A,
Gurny R: Poly(ortho esters) their develop-
ment and some recent applications, Eur J
Pharm Biopharm 2000, 50, 121128.
70 Maeda M, Tani S, Sano A, Fujioka K: Micro-
structure and release characteristics of the
minipellet, a collagen-based drug delivery sys-
tem for controlled release of protein drugs,
J Control Release 1999, 62, 313324.
71 Moriyama K, Ooya T, Yui N: Hyaluronic acid
grafted with poly(ethylene glycol) as a novel
peptide formulation, J Control Release 1999,
59, 7786.
72 Fessi H, Puisieux F, Devissaguet, JP, Am-
mouri, N, Benita S: Nanocapsule formation
by interfacial polymer deposition following
solvent displacement, Int J Pharm 1989, 55,
R1.
73 Tice TR, Gilley RM: Preparation of injectable
controlled-release microcapsules by a solvent-
References 1387
evaporation process. J Control Release 1985,
2, 343357.
74 Blanco D, Alonso MJ: Development and char-
acterization of protein-loaded poly(lactide-co-
glycolide) nanospheres, Eur J Pharm Bio-
pharm 1997, 43, 287294.
75 Mller RH: Solid Lipid nanoparticles. In: Ml-
ler RH, Hildebrand GE (Eds.), Pharmazeu-
tische Technologie: Moderne Arzneiformen,
Chapter 21, WV GmbH Stuttgart, 1997.
76 Mller RH, Lucks JS: Arzneistofftrger aus
festen Lipidteilchen (Feste Lipidnanosphren
(SLN)), European Patent EP 0 605 497 B1,
1996.
77 Almeida AJ, Runge S, Mller RH: Peptide-
loaded solid lipid nanoparticles (SLN): influ-
ence of production parameters, Int J Pharm
1997, 149, 255265.
78 Laham RJ, Sellke FW, Edelman ER, Pearlman
JD, Ware JA, Brown DL, Gold JP, Simons M:
Local perivascular delivery of basic fibroblast
growth factor in patients undergoing coronary
bypass surgery: results of a phase I random-
ized, double-blind, Circulation 1999, 100,
18651871.
79 Johansen P, Men Y, Merkle HP, Gander B:
Revisiting PLA/PLGA microspheres: an analy-
sis of their potential in parenteral vaccination.
Eur J Pharm Biopharm 2000, 50, 129146.
80 Okumu FW, Daugherty A, Sullivan SA, Tipton
AJ, Cleland JL: Evaluation of SABER as a local
delivery system for rhVEGF formulation de-
sign and in vitro assessment, Proc Int Symp
Control Release Bioact Mater 2000, 27, 8034
8035.
81 Brodbeck KJ, Pushpala S, McHugh AJ: Sus-
tained release of human growth hormone
from PLGA solution depots, Pharm Res 1999,
16, 18251829.
82 Jain RA, Rhodes CT, Railkar AM, Malick AW,
Shah NH: Controlled release of drugs from
injectable in situ formed biodegradable PLGA
microspheres: effect of various formulation
variables, Eur J Pharm Biopharm 2000, 50,
257262.
83 Royals MA, Fujita SM, Yewey GL, Rodriguez
J, Schultheiss PC, Dunn RL: Biocompatibility
of a biodegradable in situ forming implant
system in rhesus monkeys, J Biomed Mater
Res 1999, 45, 231239.
84 Ravivarapu HB, Moyer KL, Dunn RL: Sus-
tained suppression of pituitary-gonadal axis
with an injectable, in situ forming implant of
leuprolide acetate, J Pharm Sci 2000, 89, 732
741.
85 Janes KA, Calvo P, Alonso MJ: Polysaccharide
colloidal particles as delivery systems for mac-
romolecules, Adv Drug Del Rev 2001, 8397.
86 Dodane V, Vilivalam D: Pharmaceutical appli-
cations of chitosan, Pharm Sci Technol Today
1998, 1, 246253.
87 Paul W, Sharma C: Chitosan, a drug carrier
for the 21st century, STP Pharma Sci 2000,
10, 522.
88 Ohya Y, Shiratani M, Kobayashi H, Ouchi T:
Release behavior of 5-fluorouracil from chito-
san-gel nanospheres immobilizing 5-fluorour-
acil coated with polysaccharides and their cell
specific cytotoxicity, Pure Appl Chem 1994,
A31, 629642.
89 Calvo P, Remunan-Lopez C, Vila-Jato JL, Alon-
so MJ: Novel hydrophilic chitosan-polyethy-
lene oxide nanoparticles as protein carriers, J
Appl Polym Sci 1997, 63, 125132.
90 Calvo P, Bougaba A, Appel M, Fattal E, Alonso
MJ, Couvreur P: Oligonucleotide-chitosan
nanoparticles as new gene therapy vector,
Proc World Meet APGI/APV 2 1998, 1111
1112.
91 Fernandez-Urrusuno R, Calvo P, Remunan-
Lopez C, Vila-Jato JL, Alonso MJ: Enhance-
ment of nasal absorption of insulin using
chitosan nanoparticles, Pharm Res 1999, 16,
15761581.
92 Bernkop-Schnrch A, Clausen AE: Biomem-
brane permeability of peptides: Strategies to
improve their mucosal uptake, Mini Rev Med
Chem 2002, 2, 295305.
93 Bernkop-Schnrch A, Giovanelli R, Valenta C:
Peroral administration of enzymes: Strategies
to improve the galenic of dosage forms for
trypsin and bromelain, Dev Ind Pharm 2000,
26, 115121.
94 Aungst B: Intestinal permeation enhancers,
J Pharm Sci 2000, 89, 429442.
95 Bernkop-Schnrch A, Clausen AE, Guggi D:
The use of auxiliary agents to improve the
mucosal uptake of peptides, Med Chem Rev
online 2004, 1, 110.
96 Drewe J, Fricker G, Voderscher J, Beglinger C:
Enteral absorption of octreotide: absorption
enhancement by polyoxyethylene-24-cholester-
ol ether, Br J Pharmacol 1993, 108, 298303.
97 Morishita M, Morishita I, Takayama K, Yoshi-
haru M, Nagai T: Site-dependent effect of
aprotinin, sodium caprate, Na
2
EDTA and so-
dium glycolate on intestinal absorption of
insulin, Biol Pharm Bull 1993, 16, 6872.
98 Richardson JL, Illum L, Thomas NW, Arti-
cles R: Vaginal absorption of insulin in the
rat: effect of penetration enhancers on insu-
lin uptake and mucosal histology, Pharm
Res 1992, 9, 8783883.
99 Sayani AP, Chun IK, Chien YW: Transmuco-
sal delivery of leucine enkephalin: stabiliza-
tion in rabbit enzyme extracts and enhance-
ment of permeation through mucosae,
J Pharm Sci 1993, 82, 11791185.
100 Hoogstraate AJ, Coos Verhoef J, Pijpers A,
van Leengoed LA, Verheijden JH, Junginger
HE, Bodde HE: In vivo buccal delivery of
peptide drug buserelin with glycodesoxycho-
late as an absorption enhancer in pigs,
Pharm Res 1996, 13, 12331237.
101 Aungst BJ, Rogers NJ: Site dependence of
absorption-promoting actions of laureth-p,
Na+salicylate, Na
2
EDTA, and aprotinin on
rectal, nasal, and buccal delivery of insulin,
Pharm Res 1988, 5, 305308.
102 Uchiyama T, Sugiyama T, Quan Y-S, Kotani A,
Okada N, Fujita T, Muranishi S, Yamamoto A:
Enhanced permeability of insulin across the
rat intestinal membrane of various absorption
enhancers: their intestinal mucosa toxicity
and absorption enhancing mechanism of n-
lauryl-beta-D-maltopyranoside, J Pharm
Pharmacol 1999, 51, 12411250.
103 Fasano A, Uzzau S: Modulation of intestinal
tight junctions by Zonula occludens toxin per-
mits enteral administration of insulin and
other macromolecules in an animal model,
J Clin Invest 1997, 99, 11581164.
104 Hovgaard L, Brondsted H, Nielsen HM:
Drug delivery studies in Caco-2 monolayers.
II. Absorption enhancer effects of lysopho-
sphatidylcholines, Int J Pharm 1995, 114,
141149.
105 Gordon GG, Moses AC, Silver RD, Flier JS,
Carey MC: Nasal absorption of insulin: En-
hancement by hydrophobic bile salts, Proc
Natl Acad Sci USA 1987, 82, 74197423.
106 Lee WA, Ennis RD, Longenecker JP, Bengts-
son P: The bioavailability of intranasal calci-
tonin in healthy volunteers with and without
a permeation enhancer, Pharm Res 1994,
11, 747750.
107 Chiou GC, Shen ZF, Li BH: Effects of per-
meation enhancers BL-9 and Brij-78 on ab-
sorption of four peptide eyedrops in rabbits,
Zhongguo Yao Li Xue Bao 1992, 13, 201.
108 Lueen, HL, Rentel, CO, Kotze AF, Lehr
CM, deBoer AG, Verhoef JC, Junginger HE:
Mucoadhesive polymers in peroral peptide
drug delivery. IV. Polycarbophil and chitosan
are potent enhancers of peptide transport
across intestinal mucosae in vitro, J Control
Release 1997, 45, 15.
109 Artursson P, Lindmark T, Davis SS, Illum L:
Effect of chitosan on the permeability of
monolayers of intestinal epithelial cells
(Caco-2), Pharm Res 1994, 11, 13581361.
110 Thanou M, Verhoef JC, Marbach P, Jungin-
ger HE: Intestinal absorption of octreotide:
N-trimethyl chitosan chloride (TMC) amelio-
rates the permeability and absorption prop-
erties of the somatostatin analogue in vitro
and in vivo, J Pharm Sci 2000, 89, 951.
111 Clausen AE, Bernkop-Schnrch, A: Thio-
lated carboxymethylcellulose: in vitro evalua-
tion of its permeation enhancing effect on
peptide drugs, Eur J Pharm Biopharm 2001,
51, 2532.
112 Bernkop-Schnrch A, Clausen AE, Hnatys-
zyn M: Thiolated polymers: synthesis and in
vitro evaluation of polymercysteamine con-
jugates, Int J Pharm 2001, 226, 185194.
113 Bernkop-Schnrch A, Schwarz V: Polymers
with thiol groups: A new generation of mu-
coadhesive polymers? Pharm Res 1999, 16,
876881.
114 Woodley JF: Peptidase enzymes of GIT; bar-
riers to peptide delivery, but potential for
controlled release, Crit Rev Ther Carrier Syst
1994, 11, 14961500.
115 Samanen J, Wilson G, Smith PL, Lee CP,
Bondinell W, Ku T, Rhodes G, Nichols A:
Chemical approaches to improve the oral
bioavailability of peptidergic molecules,
J Pharm Pharmacol 1996, 48, 119135.
116 McMartin C, Hutchinson LEF, Hyde R, Pe-
ters GE: Analysis of structural requirements
for the absorption of drugs and macromole-
cules from the nasal cavity, J Pharm Sci
1987, 76, 535540.
117 Inagaki M, Sakakura Y, Itoh J, Ukai K,
Miyosh Y: Macromolecular permeability of
the tight junction of the human nasal muco-
sa, Rhinology 1985, 23, 213221.
118 Soane RJ, Frier M, Perkins AC, Jones NS,
Davis SS, Illum L: Evaluation of the clear-
References 1389
ance characteristics of bioadhesive systems
in humans, Int J Pharm 1999, 178, 5565.
119 Lee VHL: Enzymatic barriers to peptide and
protein absorption, CRC Crit Rev Ther Drug
Carrier Syst 1988, 5, 6997.
120 Illum L: Nasal drug delivery possibilities,
problems and solutions, J Control Release
2003, 87, 187198.
121 Illum L, Watts P, Fisher AN, Jabbal-Gill I,
Davis SS: Novel chitosan based delivery sys-
tems for nasal administration of a LHRH-
analogue, STP Pharm 2000, 10, 8994.
122 Patton JS: Pulmonary delivery of drugs for
bone disorders, Adv Drug Deliv Rev 2000,
42, 239248.
123 Shen Z, Zhang Q, Wei S, Nagai T: Proteoly-
tic enzymes as a limitation for pulmonary
absorption of insulin: in vitro and in vivo in-
vestigations, Int J Pharm 1999, 192, 115
121.
124 Gonda I: The ascent of pulmonary drug de-
livery, J Pharm Sci 2000, 89, 940945.
125 Vanbever R, Ben-Jebria A, Mintzes JD, Lan-
ger R, Edwards DA: Sustained release of in-
sulin from insoluble inhaled particles, Drug
Dev Res 1999, 48, 178185.
126 Kawashima Y, Yamamoto H, Takeuch H,
Fujioka S, Hino T: Pulmonary delivery of in-
sulin with nebulized DL-lactide/glycolide
copolymer (PLGA) nanospheres to prolong
hypoglycemic effect, J Control Release 1999,
62, 279287.
127 Morimoto K, Katsumata H, Yabuta T, Iwana-
ga K, Kakemi M, Tabata Y, Ikada Y: Gelatin
microspheres as a pulmonary delivery sys-
tem: evaluation of salmon calcitonin absorp-
tion, J Pharm Pharmacol 2000, 52, 611617.
128 Veuillez F, Kalia YN, Jacques Y, Deshusses
J, Buri P: Factors and strategies for improv-
ing buccal absorption of peptides, Eur J
Pharm Biopharm 2001, 51, 93109.
129 Wearly LL: Recent progress in protein and
peptide delivery by noninvasive routes, Crit
Rev Ther Drug Carrier Syst 1991, 8, 331
394.
130 Lee VHL, Kashi SD, Patel RM, Hayakawa E,
Inagaki K: Mucosal peptide and protein de-
livery: proteolytic activities in mucosal homo-
genates, Proc Int Symp Control Release
Bioact Mater 1987, 14, 2324.
131 Stratford RE, Lee VHL: Aminopeptidase ac-
tivity in homogenates of various absorptive
mucosae in the albino rabbit: implications
in peptide delivery, Int J Pharm 1986, 30,
7382.
132 Jacques Y, Buri P: Les systmes thrapeu-
tiques bioadhsifs application buccale: as-
pects pharmacocintiques, pharmacologi-
ques et cliniques, chapter II. In: Muqueuse
Buccale et Systmes Therapeutiques Bioad-
hsifs, Medecine et Hygine, Genve, 1998,
4778.
133 Shojaei AH, Li X: Mechanisms of buccal
mucoadhesion of novel copolymers of acrylic
acid and polyethylene glycol monomethy-
lether monomethacrylate, J Control Release
1997, 47, 151161.
134 Li C, Bhatt PP, Johnston TP: Transmucosal
delivery of oxytocin to rabbits using a mu-
coadhesive buccal patch, Pharm Dev Technol
1997, 2, 265274.
135 Wang W: Oral protein drug delivery, J Drug
Target 1996, 4, 195232.
136 Leone-Bay A, Santiago N, Achan D, Chaudh-
ary K, DeMorin F, Falzarano L, et al.: N-acy-
lated -amino acids as novel oral delivery
agents for proteins, J Med Chem 1995, 38,
42634269.
137 Stoll BR, Leipold HR, Milstein S, Edwards
DA: A mechanistic analysis of carrier-
mediated oral delivery of protein therapeu-
tics, J Control Release 2000, 64 217228.
138 Leone-Bay A, Paton DR, Weidner JJ: The de-
velopment of delivery agents that facilitate
the oral absorption of macromolecular
drugs, Med Res Rev 2000, 20, 169186.
139 Reddy SM, Sinha VR, Reddy DS: Novel oral
colon-specific drug delivery systems for
pharmacotherapy of peptide and nonpeptide
drugs, Drugs Today 1999, 35, 537580.
140 Spitael J, Kinget R: Solubility and dissolu-
tion rate of enteric polymers. Acta Pharm
Technol 1979, 25, 163168.
141 Chourasia MK, Jain SK: Polysaccharides for
colon targeted drug delivery, Drug Deliv
2004, 11, 129148.
142 Chickering DE, Mathiowitz E: Bioadhesive
microspheres. I. A novel electrobalance-
based method to study adhesive interactions
between individual microspheres and intest-
inal mucosa, J Control Release 1995, 34,
251262.
143 Chickering DE, Jacob JS, Mathiowitz E:
Bioadhesive microspheres. 2. Characterisa-
tion and evaluation of bioadhesion involving
hard, bioerodible polymers and soft-tissue.
Reacting Polymers 1995, 25, 189206.
144 Goldberg M, Gomez-Orellana I: Challenges
for the oral delivery of macromolecules, Na-
ture Rev Drug Discov 2003, 2, 289295.
145 Lee HJ: Protein drug oral delivery: the recent
progress, Arch Pharm 2002, 25, 572584.
146 Hadgraft J, Guy RH: Transdermal drug de-
livery: developmental issues and research in-
itiatives. New York: Marcel Dekker, 1989.
147 Potts RO, Guy RH: Predicting skin perme-
ability, Pharm Res 1992, 9, 663669.
148 Green PG, Hinz RS, Kim A, et al: Ionto-
phoretic delivery of a series of tripeptides
across the skin in vitro, Pharm Res 1991; 8,
11211127.
149 Menon GK, Bommannan DB, Elias PM:
High-frequency sonophoresis: permeation
pathways and structural basis for enhanced
permeability. Skin Pharmacol 1994, 7, 130
139.
150 Cevc G, Schtzlein A, Blume G: Transder-
mal drug carriers: basic properties, optimiza-
tion and transfer efficiency in the case of
epicutaneously applied peptides. J Control
Release 1995; 36, 316.
151 Cevc G, Blume G: New, highly efficient for-
mulation of diclofenac for the topical, trans-
dermal administration in ultradeformable
drug carriers, Transfersomes, Biochim Bio-
phys Acta 2001, 1514, 191205.
152 Cevc G: Transdermal drug delivery of insu-
lin with ultradeformable carriers, Clin Phar-
macokinet 2003, 34, 461474.
153 Cevc G: Drug delivery across the skin, Exp
Opin Invest Drugs 1997, 6, 18871937.
154 Kanikkannan N: Iontophoresis-based trans-
dermal delivery systems, Biodrugs 2002, 16,
339347.
155 Gothelf A, Mir LM, Gehl J: Electroche-
motherapy: results of cancer treatment using
enhanced delivery of bleomycin by electro-
poration, Cancer Treatment Rev 2003, 29,
371387.
156 Turner NG, Ferry L, Price M, et al.: Ionto-
phoresis of poly-L-lysines: the role of molec-
ular weight? Pharm Res 1997; 14, 1322
1331.
157 Kost J: Ultrasound-assisted insulin delivery
and noninvasive glucose sensing, Diabetes
Tech Therapeutics 2002, 4, 489497.
158 Graff MR, McClunahan MA: Assessment by
patients with diabetes mellitus of two insu-
lin pen delivery systems versus a vial and
syringe, Clin Ther 1998, 20, 486496.
159 Martano W, Davis SP, Holiday NR, Wang J,
Gill HS, Prausnitz MR: Transdermal delivery
of insulin using microneedles in vivo.
Pharm Res 2004, 21, 947952.
160 McAllister DV, Wang PM, Davis SP, Park
JH, Canatella PJ, Allen MG, Prausnitz MR:
Microfabricated needles for transdermal de-
livery of macromolecules and nanoparticles:
fabrication methods and transport studies,
Proc Natl Acad Sci USA 2003, 100, 13755
13760.
161 Mukerjee EV, Collins D, Isseroff RR, Smith
RL: Microneedle array for transdermal bio-
logical fluid extraction and in situ analysis,
Sensors and Actuators A: Physical, 2004,
114(23), 267275.
162 Kompella UB, Lee VHL: Delivery systems
for penetration enhancement of peptide and
protein drugs: design considerations, Adv
Drug Deliv Rev 2001, 46, 211245.
163 Niu C-H, Chiu Y-Y: FDA Perspective on pep-
tide formulation and stability issues,
J Pharm Sci 1998, 87 (11), 13311334.
References 1391
Abstract
Formed through the fusion of Inhale Ther-
apeutic Systems (San Carlos, CA), Shear-
water Corporation (Huntsville, AL) and
Bradford Particle Design (Bradford, UK),
Nektar offers a suite of leading drug deliv-
ery technologies that encompasses mole-
cule engineering including advanced con-
jugation with poly(ethylene) glycol (PEG,
i.e., (PEGylation), particle engineering
comprising both pulmonary particle tech-
nology and supercritical fluid technology,
and advanced drug delivery solutions for
oral, injectable and pulmonary administra-
tion. PEGylation has come into its own as
a powerful approach for enhancing the
properties of biopharmaceuticals. There
are six marketed PEG protein products uti-
lizing this technology and many more cur-
rently in clinical trials. Benefits which can
be achieved through application of PEGyla-
tion include extended circulation lifetime,
improved biodistribution, decreased im-
munogenicity, increased solubility, de-
creased proteolytic degradation and greater
stability of the drug product on storage. In
this chapter, we will not cover details of
the entire field; instead, we will focus on
applications which have led to marketed
products. In doing so, we will contrast
early, first-generation approaches to PEGy-
lation with current, second-generation
technology, and discuss improvements in
properties of the products as well as clini-
cal benefits which result from application
of current reagents and methods. Finally,
we will discuss formulation properties of
PEG drug products compared to those of
the native biopharmaceutical products.
Abbreviations
PAGE polyacrylamide gel
electrophoresis
CV-N cyanovirin
dimPEG dimethoxyPEG
G-CSF granulocyte colony-
stimulating factor
HPLC high-performance liquid
chromatography
1393
ISBN: 3-527-31184-X
2
Poly(ethylene) Glycol Conjugates of Biopharmaceuticals
in Drug Delivery
Pathfinder New Ways for Peptides, Proteins and Co
IFN interferon
LHRH luteinizing hormone
releasing hormone
MALDI-TOF matrix-assisted laser
desorption/ionization
time of flight
mPEG methoxyPEG
PEG poly(ethylene) glycol
rh recombinant human
SDS sodium dodecylsulfate
2.1
Introduction
Conjugation of poly(ethylene) glycol (PEG)
to proteins has evolved from the pioneering
work of Davis [1] in the 1970s to its present
position as a valuable tool for enhancement
of the properties of protein biopharmaceuti-
cals. Numerous such conjugates are mak-
ing their way through clinical trials and
six are currently marketed. In addition, the
first PEGylation of an antisense oligonu-
cleotide is a PEGylated aptamer (pegaptanib
sodium; Macugen
TM
) which was recently
approved for the treatment of macular de-
generation. As is usual with emerging tech-
nologies, the materials and techniques in
use have not been static, and there have
been significant improvements in the qual-
ity of the polymer, reagents, coupling ap-
proaches and purification methods; these
factors will be discussed here. Advances in
biotechnology have led to an abundance of
proteins with potential for therapeutic ap-
plications. The therapeutic performance of
many of these proteins requires improve-
ment, with problems such as short circula-
tion lifetime, proteolytic degradation, low
solubility, immunogenicity and non-opti-
mal biodistribution presenting barriers to
development. Conjugation of proteins with
PEG is now a clinically proven technology
for enhancing the therapeutic performance
of biopharmaceuticals. As is appropriate
for this book, we will focus on the applica-
tion of PEGylation technology in the thera-
peutic protein field only and will not in-
clude the increasingly interesting small-
molecule-drug PEGylation area. We will de-
scribe the properties of PEG while noting
properties which are desirable for produc-
ing high-quality PEG reagents for protein
conjugation. We will further describe the
chemistry of PEG reagents and conjugation
of these reagents to proteins. Rather than
presenting an exhaustive review, we will
place emphasis on chemistries which have
resulted in clinically proven drug products.
Excellent detailed reviews on reagents and
protein PEGylation [24] have recently been
published. We will also discuss formulation
of PEG protein drug products a topic that
has not been previously reviewed.
2.2
The Polymer
PEG is a polymer comprised of ethylene-
oxy units and, in a common form, two ter-
minal hydroxyl groups. It is prepared by
reaction of ethylene oxide with a base such
as sodium hydroxide in an anionic ring-
opening polymerization reaction:
O
[
[
Base
n H
2
C
[
CH
2

HO(CH
2
CH
2
O)
n1
CH
2
CH
2
OH
While the above type of PEG is useful in
many applications, the most valuable PEG
as a raw material in PEG protein applica-
tions is methoxyPEG (mPEG), prepared by
reaction of methanol or methoxyethanol
with ethylene oxide under basic condi-
tions:
O
[
[
Base
CH
3
OCH
2
CH
2
OH+nH
2
C
[
CH
2
CH
3
O(CH
2
CH
2
0)
n+1
OH
Since it is impossible to stop the polymer-
ization at a particular value of n, the poly-
mer does not consist of a single strand of
a single molecular weight, but instead has
a distribution of molecular weights charac-
terized by its polydispersity, M
w
/M
n
, where
M
w
is the weight average molecular weight
and M
n
is the number average molecular
weight. mPEGs in the molecular weight
range of 540kDa are used in marketed
PEG proteins. Polydispersity ranges from
about 1.01 in mPEGs in the 5-kDa range
to about 1.1 for the higher-molecular-weight
PEGs. Narrow polydispersity is a desirable
feature of a PEG raw material and reproduc-
ibility in the average molecular weight is
also essential. A second important property
of mPEG as a protein conjugation raw ma-
terial is diol content. PEG diol is a bypro-
duct in the mPEG polymerization process
arising from the presence of water in the re-
action mixture as a competitor with methox-
yethanol in the ethylene oxide ring-opening
process. Some commercial mPEGs contain
diol in concentrations as high as 15%.
When diol-containing mPEG is converted
to a reactive form for protein conjugation,
the diol is also converted and since the
two forms cannot generally be separated
in a reagent-purifying process, both react
with the protein. Reaction of the protein
with the diol-derived reagent leads to an un-
desirable crosslinking reaction which both
lowers yield and complicates purification.
Very low diol PEG can be prepared by ex-
haustive methylation of benzyloxyPEG to
obtain a mixture of benzyloxymPEG and
dimPEG (from PEG diol) [5]. The benzyl
group is then removed by reduction to yield
a mixture of mPEG and dimPEG. When the
mPEG is converted to a reactive form for
protein PEGylation, the dimPEG remains
unchanged and is also unreactive toward
proteins. The inert dimPEG is then easily
removed in the PEG-protein purification
process.
PEG is a white, amorphous powder
melting in the 608C range. It has the
rather unusual property of being highly
water soluble, and soluble in certain or-
ganic solvents such as methylene chloride,
chloroform and dimethylsulfoxide, while
being insoluble in ethyl ether or hexane.
These interesting properties can be uti-
lized to an advantage in the preparation
and purification of PEG derivatives, partic-
ularly PEG reagents and PEG small-mole-
cule drugs. For example, one can conduct
a reaction in methylene chloride and puri-
fy the product by precipitation in ether.
Such a sequence results in small-molecule
reagents remaining in solution with PEG
product being insoluble and recoverable by
filtration [6].
In vivo, PEG is a benign material, and is
approved for oral, parenteral drugs, as well
as in cosmetics and foods [7]. In aqueous
media, it is hydrated and has two to three
water molecules associated with each ethy-
leneoxy unit [8]. The latter is thought to be
an important factor in its nonimmuno-
genic properties. The value of PEG in drug
delivery is primarily due to the fact that
when conjugated to another molecule such
as a drug, many of the properties of PEG
are transferred to the conjugate (see also
Part VI, Chapter 1).
In this chapter, the molecular weight of
PEG will be denoted in Daltons (Da) or
kilodaltons (kDa). For example, a PEG
with a molecular weight of 20000 daltons
will be referred to as PEG 20000 daltons
of PEG 20-kDa or just PEG 20000. Each
ethylene glycol molecule in PEG has a mo-
lecular weight of 44 daltons. This would
mean that PEG 20-kDA would have ap-
proximately 455 ethylene glycol units.
When two PEG chains each of 20-kDa are
linked together to make a branched re-
agent, the nomenclature is PEG2 40-kDa.
2.3
Safety and Disposition of PEG
The safety of PEGs has been studied for
over four decades. Numerous animal stud-
ies have been performed with PEGs of dif-
ferent molecular weights, when adminis-
tered through different routes, and when
attached to lipids, nanoparticles and mac-
romolecules. A review of the literature
suggests that PEGs are safe when admi-
nistered parenterally.
Yamaoka et al. investigated the in vivo
distribution of linear free
125
I-labeled
PEGs with molecular weights of 6, 20, 50
and 170 kDa after i.v. administration to
mice. A two-compartment model was used
to illustrate the pharmacokinetic distribu-
tion of PEG from blood to key organs such
as the heart, lung, liver, spleen, kidney,
gastrointestinal tract and thyroid gland [9].
The time course of a PEG substrate, in
vivo, follows a two-phase system. The first
phase is the rapid early distribution of
PEG into various organs (a phase) and the
second phase is the slow elimination
phase (b phase). The following equations
are used to calculate the relevant pharma-
cokinetic parameters applicable to this
model:
For i.v. delivery:
C
p(t)
= Ae
at
Be
bt
For s.c., i.p. and i.t. delivery:
C
p(t)
= Ae
at
Be
b
(A B)e
Kat
or
C
p(t)
= [D(a K
21
)[=[V
c
(a b)[e
at
[D(K
21
b)[=[V
c
(a b)[e
bt
C
(t)
is the concentration of PEG in blood
at time t; D is the dose; K
a
, K
12
, K
21
and
K
e
are first-order rate constants;
(a+b) =(K
12
+K
21
+K
e
); V
c
is the volume of
distribution of the central compartment;
and A and B are parameters of the equa-
tion that allow to determine the y-inter-
cepts of each phase.
Fig. 2.1 shows that as the molecular
weight of PEG increases, the renal clear-
ance decreases, the in vivo residence time
increases and the volume of distribution
remains relatively constant. Mass balance
calculations 4 h post-dose showed that the
polymer distributed to the liver (1.65
2.88%), kidney (0.290.68%) and gastroin-
testinal tract (2.8713.53%). The majority
of the polymer was located in the blood
and excrements. The cut-off molecular
weight of globular proteins for glomerular
filtration is about 60000 Da. However, in
the case of non-ionic, coiled polymers like
PEG, the cut-off could be as low as about
30000 Da. These PEGs of lower molecular
weight have a higher vascular permeability
and are excreted primarily in urine. In a
second study of Yamaoka et al. [10], the
distribution of 6- and 50-kDa PEGs was
measured following s.c., i.m. and i.p. in-
jections. An increase in molecular weight
of PEG resulted in a decrease in the rate
of absorption (K
a
), distribution (K
12
, K
21
)
and excretion (K
e
). However, the rate of
absorption of PEG following an i.p. injec-
tion remained constant at approximately
0.48 h
1
, suggesting that clearance from
the peritoneum cavity is relatively constant
irrespective of the size and type of poly-
mer. Studies with rat Kuppfer cells showed
that phagocytosis in the liver played a ma-
jor role in the disposition of high-molecu-
lar-weight PEGs. These pinocytosis experi-
ments showed that the PEG uptake in-
creased with increasing molecular weights
of above 50000 Da.
Acute and subchronic toxicology studies
of PEG have been performed in a number
of animal species. In one very early study,
rabbits (n=49/group) received 1g of PEG
300, 400, 1450, 3350 and 6000 i.v. for 6 days
a week and for 5 weeks [11]. One death was
observed in each of the high-dose groups.
In another study, a 10% PEG 4000 solution
in normal saline was injected to beagle dogs
178 times for 12 months at doses of 10, 30
and 90mgkg
1
. There were no adverse ob-
servations reported in terms of body weight,
hematology and gross examinations of key
organs [12]. The primary route of elimina-
tion was through the kidneys with some
amount detected in the feces. In a distribu-
tion study, poly(PEG 2000-[
14
C]lysine) was
prepared and injected i.v. and i.p. into
CD1 male mice. In vivo, the polymer retains
its backbone, and excretion was primarily
through the kidneys and biliary tract. There
were no signs of excessive accumulation in
the liver, spleen or kidney. When the nonra-
dioactive parent polymer was used in an
acute toxicology study, there were no signs
of toxicity at doses up to 10 gkg
1
, after clin-
ical observations and histopathology of tis-
sues were monitored [13]. In a small clinical
safety study, six men were i.v. injected with
PEG 6000. Sixty-three percent of the dose
was recovered in urine in the first 1 h and
96% in 12 h [14]. Additional studies, not
published, have shown that PEG and PEG
copolymers can be safely dosed i.v. and s.c.
at doses as high as 2 gkg
1
in rats and dogs.
Some minor skin lesions and irritations
were observed at the site of repeated injec-
tions.
2.4
PEG Reagents and Conjugation
In some applications such as coupling
PEG to a small molecule drug bearing a
carboxylic acid, mPEG-OH or HO-PEG-
OH may be used directly. For protein ap-
plications, however, mPEG-OH is not di-
rectly useful and the terminal OH must
first be converted to a functional group
Fig. 2.1 Pharmacokinetic parameters as a function of PEG molecular weights,
when administered i.v. to mice. Key: t
1/2b
, min (n); K
e
, min (
); V
c
, mL (`),
AUC, % dose h mL
1
(*).
which will react under mild conditions,
usually in aqueous media, with one or
more nucleophiles present on the protein.
The majority of these activated PEGs fall
into one of several classes: (1) acylating re-
agents, (2) alkylating reagents and (3)
thiol-reactive reagents.
2.4.1
Acylating Reagents
The most commonly used reagents of this
class are succinimidyl-activated PEG car-
boxylic acids or succinimidyl (or 1-benzo-
triazolyl) carbonate-activated mPEGs as
shown in Fig. 2.2.
Succinimidyl-activated PEG carboxylic
acids generally react with little or no selec-
tivity with amino groups such as lysines
and N-terminal amines on proteins to
form stable amide linkages [15].
In the case of mPEG N-succinimidyl succi-
nate, the reaction also leads to a stable
amide linkage to protein amino groups,
but the linker to PEG bears an ester group
which is hydrolyzed in vivo by esterases
[16].
Fig. 2.2 Common PEG reagents in use for protein PEGylation.
(mPEG SC) mPEG 1-benzotriazoyl carbonate (mPEG BTC)
O
O
|
mPEGCO + Protein-NH
2
O
O
|
mPEGCNH-Protein
This process results in a residual tag which
becomes a potential hapten attached to pro-
tein amines. This PEGylation technology
was used in the first two marketed PEG pro-
tein products, PEG asparaginase [17] (On-
caspar
, Enzon), for treatment of lympho-

blastic leukemia (see also Part II, Chapter
6), and PEG adenine deaminase [18] (Ada-
gen
, Enzon), for the treatment of severe

combined immunodeficiency disease, using
multiple attachment of 5-kDa mPEG N-suc-
cinimidyl succinate.
Both succinimidyl and 1-benzotriazolyl
carbonyl mPEG [19] also react with protein
amines resulting in stable carbamate or
urethane linkages.
Such PEGylation reactions are generally per-
formed at a pHabove 7, but when conducted
under acidic conditions, succinimidyl carbo-
nyl mPEGhas also led to significant PEGyla-
tion on the imidazole group of histidine [20]:
Such a derivative is labile in the presence
of water and leads to stability problems
which prevent solution formulation of the
drug product.
2.4.2
Alkylating Reagents
Protein amino groups are alkylated by PEG
aldehydes under reducing conditions. Since
a-amino groups on protein lysines are more
basic than are the a-amino groups, selective
PEGylation can be achieved on the protein
N-terminus by conducting the reaction at
a pH of about 5 [21].
H
[
mPEGCH
2
CH
2
C=O
mPEG-propionaldehyde
+Protein-NH
2
NaBH3CN
mPEGCH
2
CH
2
CH
2
NH-Protein
This technique has been used to selectively
PEGylate the N-terminus of granulocyte
colony-stimulating factor (G-CSF) (see also
Part VIII, Chapter 3) using 20-kDa mPEG-
propionaldehyde. mPEG-acetaldehyde has
also been used as an alkylating agent, but
it is less stable than is the mPEG-propio-
naldehyde derivative. One approach that
increases the utility of mPEG-acetaldehyde
involves the more stable PEG-acetaldehyde
diethyl acetal [22]. This derivative can be
hydrolyzed under acidic conditions to yield
a soluble aldehyde hydrate, which is then
reacted in situ with the protein under re-
ducing conditions.
O
CH
3
(OCH
2
CH
2
)
n
OCCH
2
CH
2
CO
O
| |
O O
+Protein-NH
2
ester hydrolyses
CH
3
(OCH
2
CH
2
)
n
OCCH
2
CH
2
CNH-Protein
| |
O O
ON
O
CH
3
(OCH
2
CH
2
)
n
OCO Protein-NH
2
|
O
O
CH
3
(OCH
2
CH
2
)
n
OCNH-Protein
|
O
ON
mPEGOC OCN
N
CH
2
Protein
O
|
mPEG-tresylates have been used in non-
selective alkylation of protein amino groups.
Ideally, this occurs as a simple nucleophilic
displacement reaction; however, at higher
pH, an eliminationaddition side-reaction
can occur [23]. Use of this reagent has not
yet led to a marketed protein product.
O
|
mPEGOSCH
2
CF
3
|
O
mPEG-tresylate
+Protein-NH
2
mPEG-NH-Protein
2.4.3
Thiol-reactive Reagents
Several PEG reagents are available for site-
specific PEGylation with protein cysteine
thiols. mPEG-maleimide reacts to form a
stable sulfide linkage via a 1,4-addition re-
action of the thiol to the a, b-unsaturated
site on the maleimide moiety [24].
Thiol groups on proteins can also be
linked to PEG via a disulfide bond. Either
mPEG-OPSS of mPEG-thiolsulfonate is ef-
fective for this purpose.
mPEGOPSS+ProteinSH
ProteinSSmPEG
O
|
mPEGSSCH
3
+ProteinSH
|
O
ProteinSSmPEG
2.5
Biopharmaceutical Conjugates
Peptide and protein therapeutics often
have short in vivo half-lives due to proteo-
lysis and renal clearance. Carbohydrate re-
ceptor clearance mechanisms and immune
system responses can also contribute to a
reduction in half-life. PEGylation of pro-
teins does the following:
1. Increases the apparent overall molecular
weight such that renal clearance of the
protein conjugate is reduced.
2. Protects the protein from degradation by
proteases and masks the carbohydrate
receptor clearance mechanisms.
O
O
mPEG-maleimide
mPEG-OPSS
O
|
mPEGSSCH
3
|
O
mPEG-methane thiolsulfonate
mPEGN
mPEGSS
N
O
O
O
mPEGN
+ProteinSH
mPEGN
S-Protein
O
3. Allows for increased blood circulation
and hence longer plasma half-lives and
bioavailability.
4. Shields the protein from the immune
system, thereby avoiding an antigen
antibody response.
5. Improves the solubility and frequently
the stability of the protein [24].
PEGylation results in the shielding of a
macromolecule from an enzyme active site
or a receptor-binding area. This results in
a loss of some bioactivity. The PEG mole-
cule is mobile and hence the active sites
are not totally shielded, but available for re-
ceptor binding. Site-specific engineering of
the protein sequence allows for selective PE-
Gylation while maintaining a high degree of
bioactivity. This approach is gaining interest
in biopharmaceutical drug discovery.
The first two PEGylated proteins to be
marketed in the early 1990s utilized first-
generation PEG conjugation technologies
and reagents [17, 18]. Even though they
served small niche markets, these products
opened the avenues for new-generation
PEG reagents, and also confirmed that
PEG was biologically compatible and safe
even after long-term use. Both of these
products contained several PEG molecules
each 5000 Da per protein molecule, a term
referred to as random PEGylation. The
first PEGylated protein in clinical trials
was PEG-adenosine deaminase (Adagen
)
for the treatment of severe combined im-
munodeficiency disease caused by a defi-
ciency of adenosine deaminase or bubble
boy disease. This compound has 1117
PEG molecules per molecule of protein
[25]. The second product is PEG-L-aspara-
ginase (Oncaspar
) used to treat acute

lymphoblastic leukemia. The native pro-
tein from Escherichia coli (135000Da) has
been used since the 1960s for treating leu-
kemia. However, many patients developed
hypersensitivity. Despite the multiple at-
tachment of PEG, the conjugated drug is
as effective as the native drug and is less
immunogenic when administered once
every 2 weeks [26].
2.5.1
Interferon (IFN)-a
The first PEGylated IFN drug approved for
the treatment of hepatitis C was IFN-a2b,
now marketed under the trade name PEG-
Intron
by Schering-Plough. Early work

with IFN-a2b was done using mPEG 12-
kDa SC [27]. The PEG was attached to the
protein at pH conditions of 5.410. When
the pH was acidic, the PEG predominantly
attached at the N-terminus. When the pH
was raised above 8, the PEGylation was
random and the e amino groups of lysine
were conjugated. Different positional iso-
mers were identified and tested in vitro
using a cytopathic assay. PEGylation oc-
curred on the histidine, lysine, tyrosine,
cysteine and serine residues. In all, 14
monoPEGylated positional isomers were
identified. All monoPEGylated conjugates
were less active than native IFN in the cy-
topathic assay, but the reduction in activity
of the isoform produced at pH 6.5 was sig-
nificantly less than the other monoPEGy-
lated isoforms [28]. The final active phar-
maceutical ingredient was about 95%
mono-conjugated, but contained a total of
14 isoforms. Half of the final mono-conju-
gate was PEGylated at His34, but con-
tained a mixture of isoforms as previously
described. IFN-a2b has a total of three his-
tidines (His7, His34 and His57). Carbox-
yalkylation at His7 and His57 was not de-
tected by the methods used in these stud-
ies, and this may be due to poor solvent
accessibility. Despite an overall reduction
in bioactivity to 28% when compared to
native IFN-a2b, the biotherapeutic value of
the PEGylated protein has been demon-
strated in the clinic with good efficacy and
no dose-limiting adverse effects. The phar-
macokinetics of the Intron A
versus
PEG-Intron
are summarized in Tab. 2.1.

The second PEGylated IFN drug to be ap-
proved was IFN-a2a, marketed under the
trade name Pegasys
by Roche Pharmaceu-
ticals. Initially this protein was conjugated
via a urea linkage with a mPEG 5-kDa re-
agent. An elegant cation-exchange high-per-
formance liquid chromatography (HPLC)
method was developed to separate the
monoPEGylated species into 11 isoforms
[29]. Peptide mapping was used to deter-
mine that all lysine sites on the protein
were conjugated, but not the N-terminus.
The mono-conjugates were tested in vitro,
but retained only 640% of the antiviral spe-
cific activity of native protein. The antiproli-
ferative activity of individual conjugates was
about 18.6 pM when compared to 1.7 pM
for the native. The potencies of the conju-
gated molecules were not significantly dif-
ferent and clinical development activities
were pursued. In 1994, Phase II clinical
trials with PEG 5-kDa IFN-a2a were halted
after once a week dosing failed to show suf-
ficient improvement in efficacy over unPE-
Gylated IFN given 3 times a week. Model-
ing of the pharmacodynamics coupled with
the clinical trial data suggested that this
first-generation molecule had to be admi-
nistered at least twice weekly in order to
have sufficient advantage over the nonPE-
Gylated counterpart [30]. The apparent mo-
lecular weight of the conjugate should be
about 50-kDa in order to avoid rapid renal
Table 2.1 Summary of clinical pharmacokinetic parameters of PEGylated biopharmaceuticals
Compound Dose Route n t
1/2
CL V
d
F
(%)
L-Asparaginase
[64]
1000 U m
2
i.v.
infusion
27 20 h 2196 mL m
2
day
1
2336 mL m
2
100
PEG 5-kDa
asparaginase
1000 U m
2
i.v.
infusion
27 357 h 128 mL m
2
day
1
2093 mL m
2
100
PEG 5-kDa adeno-
sine deaminase
15 U kg
1
i.m. 6 36 days
PEG 5-kDa
visomant
20 mg s.c. 28 to
36 mL h
1
7 L 57
Filgrastim 3.45
11.5 lg kg
1
s.c. 3.5 h 0.50.7 mL
min
1
kg
1
150 mL kg
1
PEG 20-kDa
filgrastim
s.c. 379 1580 h
IFN-a2a 36 MIU i.m. 3.78.5 h 2.143.62
mL min
1
kg
1
0.22
0.74 L kg
1
80
PEG2 40-kDa
IFN-a2a
50140 h 94 mL/h 812 L
IFN-a2b 5 MIU m
2
i.m.,
s.c.
12 23 h 154 mL h
1
kg
1
PEG
12-kDa IFN-a2b
1 lg kg
1
s.c. 2260 h 22 mL h
1
kg
1
0.99 L kg
1
PEG aptanib
sodium
0.253 mg intra-
vitreal
15 614 days 100
filtration. A series of conjugates were pre-
pared in order to study the effect of PEG ar-
chitecture on rat pharmacokinetics using
higher-molecular-weight, second-generation
reagents [31]. A single linear PEG 20-kDa
was as effective as the same PEG size in a
branched architecture. Both the diPEGy-
lated and mono-branched 40-kDa total con-
jugates showed some improvement over the
20-kDa conjugates. A single large PEG was
preferred when PEGylation occurs at multi-
ple sites [32]. A branched PEG had a smaller
volume of distribution, was more resistant
to proteolytic digestion, and had improved
pH and thermal stabilities relative to linear
PEG conjugates [31]. PEGylation of IFN-a2a
with a 40-kDa branched H-hydroxysuccini-
mide ester resulted in the formation of a
stable amide linkage to the protein. The pu-
rified mono-conjugate was comprised pri-
marily of four positional isomers, Lys31,
Lys121, Lys131 and Lys134. These four at-
tachment sites accounted for around 94%
of the conjugate. The remaining 6% occur
at Lys70 and Lys83. The N-terminal cysteine
which is disulfide bonded to Cys98 was not
PEGylated. The resultant conjugate was
much less heterogeneous than the earlier
5-kDa conjugate which had 11 isoforms.
The in vitro bioactivity of the 40-kDa conju-
gate was only 7% of native protein. How-
ever, the in vivo pharmacokinetics and effi-
cacy were superior to native control [33].
The half-life of the 40-kDa conjugate was
80 h compared to 38 h for native protein.
A summary of the pharmacokinetic data
can be found in Tab. 2.1.
Product and clinical comparisons have
been made between Pegasys and PEG-In-
tron when both drugs are given s.c. as
multiple-dose injections. The profile for
Pegasys at steady-state plateaus at around
1250 pgmL
1
over a 7-day period. On the
other hand, PEG-Intron exhibits a peak
and drop-off behavior that is less remark-
able when compared to native IFN. PEG-
Intron and Pegasys are both effective
drugs at appropriate doses which compen-
sate for the reduction in bioactivity due to
the PEG shielding. However, PEG Intron
and Intron A are biologically indistin-
guishable in their immunotherapeutic pro-
files [28]. Pegasys has a 10-fold reduction
in the incidence of antibodies seen clini-
cally, IFN-a=15% and Pegasys=1.5% [33].
Some of the differences between these two
PEGylated IFN compounds are summa-
rized in Tab. 2.2.
Table 2.2 Properties of Pegasys and PEG-Intron [25, 33, 65]
PEG-Intron Pegasys
C
max
after 1544 h C
max
after 7296 h
Therapeutic plasma concentration for 4872 h Therapeutic plasma concentration for 168 h
Mean elimination half-life around 40 h
(2260 h)
Renal clearance rate is 100-fold reduced when
compared to native
Around 5-fold greater mean half-life than
Intron A
Volume of distribution 614 L
Renal clearance rate is 7-fold reduced when
compared to native
Volume of distribution greater than 20 L
2.5.2
IFN-b
IFN-b is sold commercially under the trade
names Betaseron
(Berlex, Schering AG),

Rebif
(Serono) and Avonex
(Biogen)
[25] for the treatment of multiple sclerosis.
The latter is administered intramuscularly
once per week, while the former two are
administered 3 times per week s.c. Several
groups have explored PEGylation as an
approach to improving the pharmacoki-
netics profile of IFN-b.
Pepinsky and colleagues explored PEGy-
lation as a solution to the relatively short
half-life exhibited by native IFN-b [34].
Based on structural data [35], the N-termi-
nus was readily accessible to solvent.
PEGylation was accomplished by reductive
alkylation of the N-terminal amine at pH
6 with excess 20-kDa mPEG aldehyde.
Sodium cyanoborohydride was used to re-
duce the intermediate Schiff base forming
a stable secondary amine linkage, con-
firmed by peptide mapping of the purified
conjugate. Steric hindrance minimized
multiple PEGylation contamination. The
native protein (22.5-kDa) coupled with a
20-kDa PEG was of sufficient size to pro-
vide a 5-fold improvement in serum half-
life and a 10-fold reduction in clearance
without loss of bioactivity. The volume of
distribution was lower with the PEGylated
compound than for native protein, consis-
tent with the large conjugate being re-
stricted to the vascular compartment [36].
A single free thiol, Cys17 is present on
native IFN-b, making it a candidate for
thiol-specific PEGylation. Attempts to put
a single large PEG at the site led to low
yields and steric hindrance coupled with
the lower reactivity of larger-molecular-
weight PEGs prevented this strategy from
becoming a viable scalable process. To
overcome the combined issues of reactivity
and steric hindrance, a two-step approach
to the PEGylation was devised as illus-
trated in Fig. 2.3. A small reactive PEG
OPSS 2-kDa reagent specific for cysteine
was first reacted with the protein. This
provided a sterically unhindered site for at-
tachment of a large 40-kDa branched PEG.
Additional variation on this approach
using hetero-bifunctional reagents in addi-
tion to homo-bifunctional reagents has
been explained in patent filings [37]. This
process also yielded a PEG-conjugate that
retained full bioactivity and had an im-
proved pharmacokinetics profile relative to
the nonPEGylated counterpart.
2.5.3
Pegvisomant
Pegvisomant (Somavert
, Pfizer) [25] is a
recombinant analog of human growth hor-
mone that is first in the class of drugs
called growth hormone receptor antago-
nists. It specifically counteracts excess
growth hormone in the treatment of acro-
megaly. The protein component is 22-kDa.
Due to its relatively small size, the native
protein was cleared via the kidneys and/or
growth hormone receptor internalization
with a serum half-life of about 30min [38].
With an average of 46 PEG 5-kDa bound
to the protein, Pegvisomant still requires
daily injections and 97% of patients in a
12-month trial had normal levels of insu-
lin-like growth factor I, a clinical marker
for efficacy. Summary of the pharmacoki-
netic data can be found in Tab. 2.1.
2.5.4
G-CSF
Traditional random amine PEGylation is
typically performed at pH 79. As de-
scribed in the preceding examples, ran-
dom PEGylation yields multiple PEG-iso-
mers and multiple isoforms per isomer.
Kinstler et al. compared PEGylation strate-
gies for G-CSF [39]. Advantages in both
bioactivity and stability were observed in
preparations of N-terminally PEGylated G-
CSF compared to internal lysine PEGyla-
tion (see also Part VIII, Chapter 3). The
chemistry used to create the N-terminal
conjugate also affected final conjugate sta-
bility.
Random PEGylation was done at pH 8.0
using excess PEG 6-kDa succinimidyl carb-
oxymethyl. This resulted in a mixture of
mono-, di- and unreacted native recombi-
nant human (rh) G-CSF. Three forms of
monoPEGylated rhG-CSF were isolated by
ion-exchange chromatography and further
purified by size-exclusion chromatography.
Table 2.3 lists the observed sites of PEGyla-
tion with their relative yields. Peptide map-
ping identified the species as N-terminus,
Lys35 and Lys41 PEG-isoforms with rela-
tive yields of 3: 2: 1. The remaining ly-
sines, 17 and 24, did not appear to be ap-
Table 2.3 Site of G-CSF PEGylation and relative yields
Site of modification Relative PEGylation
yields
Relative bioactivity
(%)
Prolonged
in vivo activity
N-terminus 3 68 Yes
Lys35 2 56 Yes
Lys41 1 21 No
Fig. 2.3 PEGylation of IFN-b using a two-step approach to target a buried cysteine residue.
preciably PEGylated. PEGylated conjugates
were tested in an in vitro mitogenic assay.
The N-terminally modified peptide re-
tained 68% bioactivity compared to 50 and
21%, respectively, for Lys35 and Lys41. In
vivo testing in hamsters confirmed perfor-
mance of the N-terminal, and Lys35-conju-
gated isoforms showed some prolongation
in acceptable white blood cell counts after
a single s.c. dose (as measured by area un-
der the curves). However, the Lys41 conju-
gate was less active than the control.
PEG-aldehyde reagents can be used for
N-terminal and lysine side-chain PEGyla-
tion of proteins. Selectivity for PEGylation
at the N-terminus of the protein can be en-
hanced by performing the PEGylation at low
pH as explained earlier. This strategy was
used to exploit the differences in pK
a
values
between the a amino group (pK
a
7.8) com-
pared to the e amino group (pK
a
10.1) on
the side-chains of lysine residues of G-CSF
[40]. Using low pH 5 conditions and so-
dium cyanoborohydride as the reducing
agent, a mono PEG 6-kDa proprionalde-
hyde conjugate was obtained with a 71%
monoPEGylated conjugate yield, 28% mul-
tiPEGylated and below 1% native protein
after cation-exchange purification. Peptide
mapping confirmed the PEG was located
at the N-terminus of the mono-conjugated
species. The two strategies to produce N-
terminally modified G-CSF via the mPEG
6-kDa succinimidyl carboxymethyl and
mPEG 6-kDa propionaldehyde both yield-
ed similar products as verified by analyti-
cal analysis. However, stability studies at
458C indicated that the acylated mPEG-
rhG-CSF made via the succinimidyl car-
boxymethyl route degraded faster than the
mPEG-rhG-CSF made via the aldehyde re-
ductive amination route. The primary
mechanism of sample degradation was ag-
gregation. Over an 8-week accelerated de-
gradation study, the alkylated conjugate
(aldehyde route) was approximately 5-fold
less aggregated than the acylated conju-
gate. Acylation results in formation of a
stable amide bond, but the conjugate is
not charged. The experimentally deter-
mined pI values were consistent with the
predicted charge reduction in that the pI
dropped from 6.1 for the alkylated conju-
gate to pI 5.7 for the acylated conjugate.
Additional conjugates using higher-mo-
lecular-weight PEGs led to the 2002 launch
of Amgens marketed product Neulasta
TM
(pegfilgrastim), a N-terminally PEGylated
conjugate of recombinant methionyl G-
CSF (filgrastim, Neupogen
) [25] used for

treating granulocyte depletion during che-
motherapy (see also Part VIII, Chapter 3).
Neulasta
TM
has a protein component of
around 19-kDa and a single 20-kDa PEG
covalently bound to the N-terminal.
Whereas nonPEGylated Neupogen
(fil-
grastim) requires daily injections adjusted
to patient body weight, Neulasta
TM
is a
standard dose given once a chemotherapy
cycle. The improved pharmacokinetic pro-
file of Neulasta compared to Neupogen is
due primarily to a reduction in renal filtra-
tion (Tab. 2.1).
2.5.5
Cyanovirin (CV-N)
The examples presented earlier were at-
tempts to attach PEG to proteins either by
random approaches or by N-terminal and
lysine directed efforts. One can genetically
engineer a reduction in the number of
lysines in order to reduce the number of
PEG-isoforms. Thiol PEGylation of pro-
teins is a much more attractive option
since proteins often have one or very few
free sulfhydryl groups available.
CV-N is an 11-kDa protein originally iso-
lated from the cyanobacterium Nostoc ellip-
sosporum [41]. This anti-HIV agent specifi-
cally targets N-linked high-mannose oligo-
saccharides on the viral envelope of HIV-1
[42]. As such it falls into the category of
entry and fusion inhibitors. It is effective
against various strains of HIV-1, HIV-2,
FIV (feline), SIV (simian) and other glyco-
sylated envelope viruses at nanomolar con-
centrations. It has also shown activity
against Ebola virus [43], although at con-
centrations around 10-fold higher than for
HIV. Because of its bacterial origin, the
protein was expected to be immunogenic
and to have a relatively short half-life due
to its small size. The protein contains five
lysines in addition to the N-terminus
which could be PEGylated. Some work has
been conducted [44] showing that the
PEGylation of lysines resulted in loss of
activity. This was verified by a published
structure that showed that the lysines were
proximal to the two mannose-binding sites
[45, 46]. The absence of free cysteines in
the native protein allowed for a specific
mutation of a cysteine for glutamine 62.
This created a single site at which PEG 5-,
20- and 30-kDa maleimides could be at-
tached. All of the conjugates had in vitro
antiviral activity comparable to AZT. The
30-kDa PEG-CV-N conjugate was tested in
a preliminary toxicology study in mice and
was found to be less toxic than the native
CV-N. In addition, when the 20- and 30-
kDa conjugates were examined for immu-
nological response in mice, as shown in
Fig. 2.4, the results showed that PEGyla-
tion of the protein provided a substantial
reduction in antibody titers [47].
2.6
PEGylation of Peptides
PEGylation of peptides has also been ex-
plored in attempts to increase half-life and
solubility. Peptides are often subject to pro-
tease inactivation and, due to their small
size, are readily cleared by the kidneys. At-
tachment of a very large PEG to a small
peptide sufficient to protect against renal
filtration may completely mask the in vivo
activity. However, a pro-drug large PEG-
conjugate could slowly hydrolyze under
physiological conditions and provide an ef-
ficient way of making these peptides bio-
available.
Luteinizing hormone releasing hormone
(LHRH), a decapeptide secreted by the hy-
pothalamus, is capable of inducing the re-
lease of luteinizing hormone and follicle-
stimulating hormone. Both antagonists
and agonists of LHRH have been synthe-
sized for various uses in contraception and
Fig. 2.4 Immunologic response in mice
to dosing with mPEG-MAL conjugated
or unconjugated CV-N(Q62C).
treatment of hormone-dependent disor-
ders. The structure of antide, a potent
LHRH antagonist, is shown in Fig. 2.5
[48]. The peptide is poorly soluble in phys-
iological buffers, and has poor bioavailabil-
ity and irreproducible pharmacokinetics.
A novel strategy for the preparation of
PEG-LHRH analog conjugates having a
PEG moiety covalently bound exclusively
to the serine residue of a LHRH analog
was desired. However, in order to attach a
PEG only to the serine OH, the nitrogen
group at isopropyl lysine required protec-
tion. In addition, the serine OH group
was sterically hindered. To circumvent
these issues, the strategy diagrammed in
Fig. 2.5 was employed [49]. The N-isopro-
pyl-Lys8 residue was first protected and, to
circumvent steric hindrance of the serine
OH group, a glycine spacer was attached
via an ester. The glycine amine was next
PEGylated with a 20- or 40-kDa PEG. Gen-
tle deprotection of the tBOC followed with-
out disrupting the PEG-antide ester link-
age. The half-life for the cleavage of the
conjugate was 5.6 h in pH 7.2 buffer at
378C.
2.7
Formulations of PEGylated
Biopharmaceuticals
The design and development of biopharma-
ceutical dosage forms requires understand-
ing of the physical and chemical properties
of the drug substance (the PEGylated mole-
cule versus that of the non-modified mole-
cule), the effect of the body on the effective-
ness of the drug and its dosage form, and
the biological factors that affect the drug
and its availability at the site of action. PE-
Gylated biopharmaceuticals are delivered
primarily as injections since their high mo-
lecular weights exclude them from being
bioavailable as oral and dermal products.
Studies in rats and dogs have shown that
the oral bioavailability of PEGs is between
Fig. 2.5 Antide structure [48].
79 and 100% for oligomers up to 600 Da.
The number drops to less than 2% when
the molecular weight is above 1000 Da [50].
The key to preparing a stable formula-
tion is the completion of definitive prefor-
mulation studies. PEGylation affects the
surface properties of the protein and pep-
tides, and this, in turn, affects the binding
properties, bioactivity, the in vivo half-life
and the immunogenicity of the molecule.
When the glucose binding protein, conca-
navalin A was conjugated with up to five
mPEG 5-kDa nitrophenol carbonate units,
the glucose binding constant of the mole-
cule increased from 7338 to
2589333 M
1
[51]. However, when the
number of PEG chains per concanavalin
molecule was greater than 5, the binding
constant decreased to 173130 M
1
sug-
gesting a complete masking of the protein.
PEGylation also affects the formulation
properties of biopharmaceuticals. It has a
large exclusion volume in aqueous solu-
tions due to significant molecular mobility
and hydration (two to three water molecules
per ethylene oxide unit) [52]. This large ex-
clusion zone acts to reduce protein aggrega-
tion at interfaces, reduce interactions on
surfaces, and reduce the need for the use
of carrier proteins and/or protein stabili-
zers. The reduction of aggregation is aptly
demonstrated by the study of b-amyloids
[53]. Felix demonstrated that by conjugating
a PEG to the C-terminus of bAP(14)-NH
2
,
the time required to form amyloid fibers (an
aggregation function) was significantly in-
creased. Another example is a case study
with CWK
18
, a 20-amino-acid synthetic pep-
tide. PEG 5-kDa vinyl sulfone and PEG 5-
kDa orthopyridyl disulfide were reacted
with the free cysteine residue to create
PEG-peptide conjugates that were non-re-
ducible and reducible, respectively [54].
When bound to plasmid DNA, the resulting
condensates had mean diameters and f po-
tentials of 8090 nm and +10 mV, respec-
tively, when tested at concentrations be-
tween 0.05 and 2mgmL
1
. In comparison,
the native peptide has a diameter of 60nm
and a potential of +25mV. However, when
the concentration increased from 0.05 to
0.5 mgmL
1
, the particle size increased to
400nm with visible evidence of flocculates.
Another example of preformulation is
PEG-staphylokinase (SY161) (see also Part
II, Chapter 1) synthesized with a PEG 5-
kDa maleimide. The effect of buffer
strength (20100 mM), buffer type (phos-
phate, citrate, and carbonate), sodium
chloride concentration (62.5250 mM) and
pH (59) on the conformation and stability
of the protein was studied using a two-lev-
el full factorial design [55]. Circular dichro-
ism was used to evaluate the secondary
structure of the protein. Stability toward
unfolding was investigated using high sen-
sitivity differential scanning calorimetry.
DePEGylation, aggregation and protein
loss were evaluated using size-exclusion
HPLC with on-line light scattering. SY161
was found to have the highest conforma-
tional stability at pH 7.0 where the net sur-
face charge is minimal (pI 6.8). Phosphate
and citrate buffers were preferred to carbo-
nate buffers. At basic pH, the molecule
first dePEGylates and then aggregates,
leading to a loss of potency. Low pH and
high ionic strength did lead to some
change in ellipticity, but high pH resulted
in unfolding, dePEGylation and aggrega-
tion. The increase in total ionic strength
(from buffer and salt) resulted in an in-
crease in stability of protein due to a mech-
anism of preferential hydration. The pres-
ence of electrolytes such as sodium chloride
was also demonstrated in another example
of stability of a PEGylated protein. When
the unstable monomeric form of brain-de-
rived neurotrophic factor was N-terminal
PEGylated with a PEG 20-kDa reagent, the
rate of protein degradation was accelerated.
However, when 150 mM of sodium chloride
was incorporated into the formulation, im-
proved conformational and thermodynamic
stability was achieved [56]. Another prefor-
mulation observation was made in-house
with the PEG-biphalin [57]. The effect of
buffer strength (25100 mM), buffer type
(phosphate and acetate) and pH (59) on
the stability of the octapeptide was studied.
Fig. 2.6 shows that the PEG 2-kDa conju-
gate of biphalin was chemically stable be-
tween a pH of 5.8 and 7.0. Alkaline condi-
tions led to a loss of compound by dePEGy-
lation and peptide degradation.
PEGylation alters the conformational sta-
bility of proteins like hemoglobin and
brain-derived neurotrophic factor [58].
PEGylation of hemoglobin with a vinyl sul-
fone reagent reduced the loss of structure
induced by lyophilization, resulting in
phase separation among excipients. Hemo-
globin favors the dextran phase in PEG/
dextran partition experiments, with a parti-
tion coefficient of 0.3. After PEGylation,
the conjugate favors the PEG phase with a
coefficient of 3.1. Similar observations
were reported with bovine serum albumin,
granulocyte macrophage colony-stimulat-
ing factor and IgG [59].
PEG and its conjugates are not compati-
ble with other hydrophilic polymers such
as polyacrylic acid and polyvinyl alcohol.
The incompatibility is related to competi-
tive affinity for water molecules through
hydrogen bonding. Most formulations of
PEG-proteins are prepared as solutions or
lyophilized powders. The following deci-
sion tree (Scheme 1) illustrates the formu-
lation options that are available.
Table 2.4 lists the recipes of formulated
marketed biopharmaceuticals. When for-
mulated as lyophilized sterile powders, the
recipe calls for buffer salts (phosphates or
acetates), bulking and osmotic agents
(mannitol, sucrose or sorbitol), and stabili-
zers (glycine, human albumin, or polysor-
bates). On the other hand, sterile solutions
contain buffers (pH 4.07.3), osmotic ad-
justors (sodium chloride or sorbitol), and
stabilizers and preservatives (benzyl alco-
hol). IFN-a2b is present in Schering
Ploughs Intron A and PEG-Intron. The
former utilizes a protein stabilizer, human
albumin, while the PEG-Intron formula-
tion does not. The PEG-conjugate provides
a simpler formulation; it also reduces the
potential risk of blood-transmitted diseases
caused by infectious viruses and often liv-
ing pathogens that may be present in the
Fig. 2.6 Effect of pH on the aqueous
stability of PEG-biphalin.
plasma-derived excipients (see also Part I,
Chapter 6, Part II, Chapter 3 and Part VII,
Chapter 1). Elimination of human-derived
plasma products also removes a potential
regulatory delay by reducing overall risk
(see also Part VII, Chapter 4). Finally, re-
moving the human albumin from the for-
mulation reduces the complexity of the
analytical separation and characterization
of the finished product. The formulation
of the Roche first-generation native prod-
uct (Roferon
, Roche) and the second-gen-

eration PEGylated product (Pegasys
,
Roche) have almost identical formulations
as stable solutions. Only the counter ion
of the buffering system was changed. Sim-
ilar observations can be made for Amgens
filgrastim and pegfilgrastim formulations.
Pegvisomant (Somavert, Pfizer) is supplied
as a sterile, white lyophilized powder in-
tended for subcutaneous injection after re-
constitution with 1 mL of Sterile Water for
injection USP. This product is similar to
the PEG-L-asparaginase (Oncaspar) and
PEG-adenosine deaminase (Adagen).
These three randomly conjugated prod-
ucts have poor long-term stability as solu-
tions and the stability of each positional
isomer has not been demonstrated.
2.8
Analysis of PEG-conjugates
The analysis of PEG-conjugates poses a
number of analytical challenges. Typically
a PEGylated protein reaction yields a mix-
ture of varying numbers of PEGs bound to
the protein and attachment occurs at mul-
tiple sites. The simplest method for analy-
sis of a reaction mixture is frequently by
sodium dodecylsulfatepolyacrylamide gel
electrophoresis (SDS-PAGE). A qualitative
estimate of native protein compared to 1-
mers, 2-mers, etc., can be easily visualized
by staining of the gel. The large hydrody-
namic volume of the PEG will make the
PEG-protein conjugate appear larger than
it really is when protein molecular weight
standards are used. The polydispersity of
the PEG will make protein bands more
broad than protein only. Thus, even with a
gel scanner or densitometer, quantitation
of PEG isomers by SDS-PAGE is difficult.
What is possible, however, is a visualiza-
tion of the number of different PEG-iso-
mer species, since it is often easier to see
a laddering of the PEG-isomers on a gel
than it is to develop an accurate HPLC
separation method.
Small quantities of PEG-conjugate are
often purified by size exclusion chroma-
Scheme 1
T
a
b
l
e
2
.
4
C
o
m
p
o
s
i
t
i
o
n
o
f
P
E
G
y
l
a
t
e
d
b
i
o
p
h
a
r
m
a
c
e
u
t
i
c
a
l
s
p
r
o
d
u
c
t
f
o
r
m
u
l
a
t
i
o
n
s
[
2
5
]
A
c
t
i
v
e
I
F
N
-
a
2
b
I
F
N
-
a
2
a
A
d
e
n
o
s
i
n
e
L
-
A
s
p
a
r
a
-
P
e
g
v
i
s
o
-
P
e
g
a
p
t
a
-
F
i
l
g
r
a
s
t
i
m
(
r
h
-
G
C
S
F
)
d
e
a
m
i
n
a
s
e
g
i
n
a
s
e
m
a
n
t
n
i
b
I
n
t
r
o
n
A
P
E
G
-
I
n
t
r
o
n
R
o
f
e
r
o
n
P
E
G
a
s
y
s
A
d
a
g
e
n
O
n
c
a
s
p
a
r
S
o
m
a
v
e
r
t
M
a
c
u
g
e
n
N
e
u
p
o
g
e
n
N
e
u
l
a
s
t
a
M
o
l
e
c
u
l
a
r
w
e
i
g
h
t
o
f
a
c
t
i
v
e
(
D
a
)
1
9
2
7
1
1
9
2
7
1
1
9
0
0
0
1
9
0
0
0
1
3
6
0
0
0
2
1
9
9
8
1
0
1
3
5
1
8
8
0
0
1
8
8
0
0
C
o
n
c
e
n
t
r
a
t
i
o
n
o
f
a
c
t
i
v
e
3
,
5
,
1
0
M
I
U
d
o
s
e
1
7
4
2
2
2
l
g
v
i
a
l
1
0
.
0
1
1
0
.
0
3
3
m
g
0
.
1
8
2
5
0
I
U
7
5
0
I
U
1
0
,
1
5
,
2
0
U
v
i
a
l
1
0
.
3
m
g
d
o
s
e
1
0
.
3
0
.
4
8
m
g
v
i
a
l
1
6
m
g
M
o
l
e
c
u
l
a
r
w
e
i
g
h
t
o
f
P
E
G
(
k
D
a
)
1
2
4
0
5
5
5
4
0
2
0
H
u
m
a
n
a
l
b
u
m
i
n
(
m
g
)
1
.
0
G
l
y
c
i
n
e
(
m
g
)
2
0
1
.
3
6
S
o
d
i
u
m
p
h
o
s
p
h
a
t
e
,
d
i
b
a
s
i
c
,
a
n
h
y
d
r
o
u
s
(
m
g
)
2
.
3
1
.
1
1
5
.
5
8
5
.
5
8
1
.
0
4
0
.
0
5
8
S
o
d
i
u
m
p
h
o
s
p
h
a
t
e
,
m
o
n
o
-
b
a
s
i
c
(
m
g
)
0
.
5
5
1
.
1
1
1
.
2
0
1
.
2
0
0
.
3
6
0
.
0
6
9
S
o
d
i
u
m
c
h
l
o
r
i
d
e
(
m
g
)
7
.
2
1
8
.
0
8
.
5
0
8
.
5
0
0
.
8
A
m
m
o
n
i
u
m
a
c
e
t
a
t
e
(
m
g
)
0
.
7
7
S
o
d
i
u
m
a
c
e
t
a
t
e
,
t
r
i
h
y
d
r
a
t
e
(
m
g
)
2
.
6
2
A
c
e
t
a
t
e
(
m
g
)
0
.
5
9
0
.
9
4
0
.
3
5
S
o
d
i
u
m
(
m
g
)
0
.
0
3
5
0
.
0
5
6
0
.
0
2
A
c
e
t
i
c
a
c
i
d
0
.
0
5
S
u
c
r
o
s
e
(
m
g
)
5
9
.
2
M
a
n
n
i
t
o
l
(
m
g
)
3
6
S
o
r
b
i
t
o
l
5
0
.
0
8
0
.
0
3
0
.
0
P
o
l
y
s
o
r
b
a
t
e
2
0
(
m
g
)
0
.
0
2
P
o
l
y
s
o
r
b
a
t
e
8
0
(
m
g
)
0
.
0
7
4
0
.
0
5
0
.
2
0
.
0
4
0
.
0
6
4
B
e
n
z
y
l
a
l
c
o
h
o
l
(
m
g
)
1
0
.
0
1
0
.
0
S
t
e
r
i
l
e
w
a
t
e
r
f
o
r
i
n
j
e
c
t
i
o
n
0
.
5
m
L
1
m
L
1
m
L
1
m
L
9
0
l
L
1
1
.
6
m
L
0
.
6
m
L
D
e
s
c
r
i
p
t
i
o
n
s
t
e
r
i
l
e
c
l
e
a
r
s
o
l
u
t
i
o
n
s
t
e
r
i
l
e
p
o
w
d
e
r
s
t
e
r
i
l
e
,
c
o
l
o
r
-
l
e
s
s
t
o
l
i
g
h
t
y
e
l
l
o
w
s
o
l
u
t
i
o
n
s
t
e
r
i
l
e
,
c
o
l
o
r
-
l
e
s
s
t
o
l
i
g
h
t
y
e
l
l
o
w
s
o
l
u
-
t
i
o
n
(
p
H
6
.
0
)
s
t
e
r
i
l
e
,
c
o
l
o
r
l
e
s
s
s
o
l
u
t
i
o
n
(
p
H
7
.
3
)
s
t
e
r
i
l
e
,
c
o
l
o
r
-
l
e
s
s
s
o
l
u
t
i
o
n
(
p
H
7
.
3
)
s
t
e
r
i
l
e
p
o
w
d
e
r
s
t
e
r
i
l
e
,
c
o
l
o
r
-
l
e
s
s
s
o
l
u
t
i
o
n
(
p
H
6
7
)
S
t
e
r
i
l
e
c
l
e
a
r
s
o
l
u
t
i
o
n
(
p
H
4
.
0
)
S
t
e
r
i
l
e
c
l
e
a
r
s
o
l
u
t
i
o
n
(
p
H
4
.
0
)
tography. If the differences in molecular
weight are sufficient, it is possible to sepa-
rate the 1-mers from conjugates contain-
ing two or more PEGs bound per protein.
The apparent molecular weight of a PEG-
protein conjugate determined by this
method will give an overestimate of the
true molecular weight if protein standards
are used for calibration. If PEG polymers
alone are used as standards, the apparent
molecular weight of the conjugate will be
smaller than the true size. PEG-CV-N was
used as an example to demonstrate this
observation (Tab. 2.5).
An accurate determination of molecular
weight can be done by matrix-assisted
laser desorption/ionization time of flight
(MALDI-TOF) analysis. Fig. 2.7 demon-
strates the effect of polydispersity on the
MALDI-TOF analysis of mPEG 5-kDa
MAL CV-N. Each peak is separated by 44
mass units associated with the ethyleneoxy
units of the polymer. The polydispersity in-
creases with an increase in the size of the
PEG. The PEG-conjugate is also polydis-
perse and the resolution of the peaks de-
creases as the molecular weight of either
the PEG or the protein increases. The pro-
Table 2.5 Molecular weight analysis of PEG-CV-N conjugates using linear PEG standards (in Da)
Sample Theoretical Gas-phase
chromatography
MALDI-TOF
CV-N 11000
PEG 5-kDa CV-N 16000 7328 16452
PEG 20-kDa CV-N 31000 27655 32109
PEG 30-kDa CV-N 41000 38001 42787
Fig. 2.7 MALDI-TOF analysis of mPEG 5-kDa MAL-CV-N Reaction mixture.
tein conjugate signal decreases with in-
creasing molecular weight. MALDI-TOF
cannot be used for quantitation, but can
be used to accurately determine the true
molecular weight of not only purified con-
jugates, but of some species in a reaction
mixture. The MALDI-TOF data gives an
average molecular weight reflecting the
polydispersity of the original PEG. During
the course of purification, the absolute
molecular weight may change slightly, if a
fraction is collected that selectively re-
moves high- or low-molecular-weight spe-
cies in the tail of a preparative peak. Even
so, an accurate assessment of the number
of PEGs bound to a peptide or protein can
accurately be determined by this method.
No single analytical method will provide
all the qualitative and quantitative answers
during PEG-biopharmaceutical character-
ization. Size-exclusion/gel-filtration and
SDS-PAGE may separate 1-mer from 2-
mers and higher molecular weights. How-
ever, neither method usually clearly re-
solves positional isoforms. Ion-exchange or
reverse-phase chromatography methods
have been used to address this issue. Re-
verse phase was used to separate the 14
isoforms of IFN-a2b [28] and the 11 iso-
forms of IFN-a2a [29]. Typically, reverse
phase is used for separation of the pep-
tides generated during peptide mapping.
Some disadvantages of PEG conjugation
can be noted, such as the PEG masking of
the protein from desired charge interac-
tions on ion exchange or hydrophobic in-
teractions on reverse-phase resins. PEGyla-
tion of lysine residues reduces the overall
charge on the protein as observed in the
G-CSF example, but the observed effect in
chromatography is not always predictable.
PEGylation of a neutral cysteine residue as
in the case of CV-N itself does not affect
the calculated pI of the protein, but the
PEG may mask otherwise exposed charged
areas thus altering the behavior toward ion
exchange resins during purification. An
example of a reverse phase purification of
a CV-N conjugate is given in Fig. 2.8.
Historically, various colorimetric meth-
ods were used to determine the number of
modified lysine groups [60]. These meth-
Fig. 2.8 Reverse-phase purification of PEG 30-kDa MAL CV-N (Q62C).
ods require several milligrams of protein
and are subject to interference by free
PEG. Fluorimetric methods that only re-
quire nanogram quantities of protein have
also been developed for quantitation of ly-
sine residues [61]. Cysteine modification
has been performed with Ellmans reagent
[62]. These methods, however, do not
reveal relative yields of individual iso-
forms. The HPLC method used to deter-
mine the relative ratios of monoPEGylated
G-CSF conjugates (Tab. 2.3) is commonly
used in development prior to regulatory
filing [63].
2.9
Summary and Future Outlook
The discovery and development of new
proteins, peptides, antisense oligonucleo-
tides and antibodies for the treatment of
disease will increase in the next decade.
With these advances, drug delivery will al-
ways be a product development hurdle
that scientists must address and PEGyla-
tion will continue to be an attractive for-
mulation option for the foreseeable future.
As patents expire, competition in the sup-
ply of PEG reagents and in the develop-
ment of PEGylated biogenerics will also
grow. While PEGylation of proteins is be-
coming a mature area, we expect continu-
ing discoveries to increase the value of this
approach in drug delivery. New approaches
to site-specific PEGylation are expected
and new reversible PEGylation methods
are likely. Increased application of PEG in
targeted drug delivery will also improve ef-
ficacy and reduce side-effects of biophar-
maceuticals.
References
1 Abuchowski, A., Van Es, T., Palzuk, N. C., Da-
vis, F. F., Alteration of immunological proper-
ties of bovine serum albumin covalent attach-
ment of polyethylene glycol, J. Biol. Chem.
252, 35783581, 1977.
2 Morpurgo, M., Veronese, F. M., Conjugates
of peptides and protein polyethylene glycols,
Methods Mol Biol. 283, 4570, 2004.
3 Pasut, G., Guiotto, A., Veronese, F. M., Pro-
tein, peptide, and non-peptide drug PEGyla-
tion for therapeutic application, Expert Opin.
Ther. Patents 14(5), 136, 2004.
4 Shorr, R. G. L., Bentley, M., Zhao, S., Parker,
R., Whittle, B., Poly(ethylene glycol) conjuga-
tion of protein and small molecule drugs, in
Drug Delivery Systems in Cancer Therapy
(D. M. Brown, Ed.), Humana Press, Totowa,
NJ, 2003, pp. 137174.
5 Bentley, M. D., Harris, J. M., Kozlowski, A.,
Heterobifunctional derivatives of poly(ethylene
glycol) and methods for their preparation,
US patent 6,448,369, 2002.
6 Harris, J. M., Introduction to biotechnical and
biomedical applications of poly(ethylene gly-
col), in Poly(ethylene glycol) Chemistry, Biotech-
nical and Biomedical Applications (J. M. Harris,
Ed.), Plenum Press, New York, 1992, pp. 114.
7 Working, P., Newman, M. S., Johnson, J., Cor-
nacoff, J. B., Safety of Poly(ethylene glycol)
and poly(ethylene glycol) derivatives, in
Poly(ethylene glycol) Chemistry and Biological
Applications (J. M. Harris and S. Zalipsky,
Eds.), ACS Symposium Series 680, ACS,
Washington, DC, 1997, pp. 4557.
8 Antonsen, K. P., Hoffman, A. S., Water struc-
ture of PEG solutions by differential scanning
calorimetry measurements, in Poly(ethylene
glycol) Chemistry: Biotechnical and Biomedical
applications (J. M. Harris, Ed.), Plenum Press,
New York, 1992, pp. 1528.
9 Yamaoka, T., Tabata, Y., Ikada, Y., Distribution
and tissue uptake of poly(ethylene glycol) with
different molecular weights after intravenous
administration to mice, J. Pharmac. Sci. 83,
601606, 1994.
10 Yamaoka T., Tabata, Y., Ikada, Y., Fate of water
soluble polymers administered via different
routes, J. Pharmac. Sci. 84, 349354, 1995.
11 Smyth, H., Carpenter, C., Shaffer, C., The tox-
icity of high molecular weight polyethylene
glycols: chronic oral and parenteral adminis-
References 1415
tration, Am. Pharmac. Ass. Sci Ed. 36, 157
160, 1947.
12 Carpenter, C., Woodside, M., Kinkead, E., King,
J., Sullivan, L., Response of dogs to repeated
intravenous injection of polyethylene glycol
4000 with notes on excretion and sensitization,
Toxicol. Appl. Pharmacol. 18, 3540, 1971.
13 Nathan, A., Zalipsky, S., Ertel, S., Agathos, S.,
Yarmush, M., Kohn, J., Copolymers and poly-
ethylene glycol: a new family of functional
drug carriers, Bioconjugate Chem. 4, 5462,
1993.
14 Shaffer, C., Critchfield, F., Chemical and phys-
ical examination of some polyethylene glycols,
Am. Pharmac. Ass. Sci Ed. 39, 340344, 1950.
15 Harris, J. M., Kozlowski, A., Polyethylene gly-
col and related polymers monosubstituted
with propionic or butanoic acids and func-
tional derivatives thereof for biotechnical
applications, US patent 5672662, 1997.
16 Carter, M. C., Meyerhoff, M. E., Instability of
succinyl ester linkages in O2-monosuccinyl
cyclic AMP-protein conjugates at neutral pH,
17 Abuchowski, A., Kazo, G. M., Verhoest, Jr.,
C. R., Van Es, T., Kafkewitz, D., Nocci, M. L.,
Viau, A. T., Davis, F. F., Cancer therapy with
chemically modified enzymes. I. Antitumor
properties of polyethylene glycol-asparaginase
conjugates, Cancer Biochem. Biophys. 7, 175
186, 1984.
18 Herschfield, M. S., Buckley, R. H., Greenberg,
M. L., Melton, A. L., Schiff, R., Hatem, C.,
Kurzberg, J., Markert, M. L., Kobayashi, R. H.,
Kobayashi, A. L., Abuchowski, A., Treatment
of adenosine deaminase deficiency with poly-
ethylene glycol-modified adenosine deami-
nase, N. Engl. J. Med. 316, 589596, 1987.
19 Dolence, E. K., Hu, C., Tsang, R., Sanders, C.
G., Osaki, S., Electrophilic polyethylene oxides
for the modification of polysaccharides, poly-
peptides (proteins) and surfaces, US patent
5650234, 1997.
20 Wang, Y., Youngster, S., Bausch, J., Zhang, R.,
McNemar, C., Wyss, D. F., Identification of
the major positional isomer of pegylated inter-
feron alpha-2b, Biochemistry 39, 1063410640,
2000.
21 Kinstler, O. B., Brems, D. N., Lauren, S. L.,
Paige, A. G., Hamburger, J. B., Treuheit, M. J.,
Characterization and stability of N-terminally
pegylated rhG-CSF, Pharmac. Res. 13, 996
1002, 1995.
22 Bentley, M. D., Harris, J. M., Poly(ethylene)
glycol aldehyde hydrates and related polymers
and applications in modifying amines, US
patent 5990237, 1999.
23 Gais, H. J., Rupert, S., Modification and im-
mobilization of proteins with polyethylene gly-
col tresylates and polysaccharide tresylates:
evidence suggesting a revision of the coupling
mechanism and structure of the polymer
polymer linkage, Tetrahedron Lett. 36, 3837
3838, 1995.
24 Shen, X., N-maleimidyl polymer derivatives,
US patent 6602498, 2003.
25 Thw PDR
Electronic Library, 2004, pub-

lished by Thomson PDR, Montvale, NJ, USA.
26 Greenwald, R. B., Choe, Y. H., McGuire, J.,
Conover, C. D., Effective drug delivery by
PEGylated drug conjugates, Adv. Drug Deliv.
Rev. 55, 217250, 2003.
27 Wylie, D. C., Voloch, M., Lee, S., Liu, Y.-H.,
Cannon-Carlson, S., Cutler, C., Pramanik, R.,
Carboxylated histidine is a pH-dependent
product of PEGylation with SC-PEG, Pharmac.
Res. 18, 13541360, 2001.
28 Grace, M., Youngster, S., Gitlin, G., Sydor, W.,
Xie, L., Westreich, L., Jacobs, S., Brassard, D.,
Bausch, J., Bordens, R., Structural and biolog-
ic characterization of PEGylated recombinant
IFN-a2b, J. Interferon Cytokine Res. 21, 1103
1115, 2001.
29 Monkarsh, S. P., Ma, Y., Aglione, A., Bailon,
P., Ciolek, D., DeBarbieri, B., Graves, M. C.,
Hollfelder, K., Hanspeter, M., Palleroni, A.,
Porter, J. E., Russoman, E., Roy, S., Pan, Y.-C.
E., Positional isomers of monopegylated inter-
feron a-2a, isolation, characterization, and bio-
logical activity, Anal. Biochem. 247, 434440,
1992.
30 Nieforth, K. A., Nadeau, R., Paterl, I., Mould,
D., Use of an indirect pharmacodynamic stim-
ulation model of MX protein induction to
compare in vivo activity of interferon alfa-2a
and a polyethylene glycol-modified derivative
in healthy subjects, Clin. Pharmacol. Ther. 59,
636646, 1996.
31 Bailon, P., Palleroni, A., Schaffer, C. A.,
Spence, C. L., Fung, W.-J., Porter, J. E., Ehrlich,
G. K., Pan, W., Xu, Z.-X., Modi, M. W., Farid,
A., Berthold, W., Rational design of a potent,
long-lasting form of interferon: a 40kda
branched polyethylene glycol-conjugated inter-
feron a-2a for the treatment of hepatitis C,
Bioconjugate Chem. 12, 195202, 2001.
32 Bailon, P., Berthold, W., Polyethylene glycol-
conjugated pharmaceutical proteins, Pharmac.
Sci. Technol. Today 1, 352356, 1998.
33 Reddy, K. R., Modi, M. W., Pedder, S., Use of
peginterferon alfa-2a (40 KD) (Pegasys
) for
the treatment of hepatitis C, Adv. Drug Deliv.
Rev. 54, 571586, 2002.
34 Pepinsky, R. B., Lepage, D. J., Gill, A., Chakra-
borty, A., Vaidyanathan, S., Green, M., Baker,
D. P., Whalley, E., Hochman, P. S., Martin, P.,
Improved pharmacokinetic properties of a
polyethylene glycol-modified form of interfer-
on-b-1a with a preserved in vitro bioactivity, J.
Pharmac. Exp. Ther. 296, 10591066, 2001.
35 Karpusas, M., Nolte, M., Benton, C. B., Meier,
W., Lipscomb, W. N., Goelz, S., The crystal
structure of human interferon beta at 2.2-
resolution, Proc. Natl Acad. Sci. USA 94,
1181311818, 1997.
36 Pepinsky, B., Runkel, L., Brickelmaier, M.,
Whitty, A., Hochman, P., Newton, M. A., Poly-
mer conjugates of interferon beta-1a and uses,
US patent application 2003/0021765 A1, 2003.
37 Sawlivich, W., Polyol-IFN-beta conjugates,
Patent publication WO9955377, 2000.
38 Kopchick, J. J., Parkinson, C., Stevens, E. C.,
Trainer, P. J. Growth hormone receptor antago-
nists: discovery, development, and use in pa-
tients with acromegaly, Endocr. Rev. 23, 623
646, 2002.
39 Kinstler, O. B., Gabriel, N. E., Farrar, C. E., De-
Prince, R. B., N-terminally chemically modi-
fied protein compositions and methods, US
patent 5985265, 1999, and N-terminally che-
mically modified protein compositions and
methods, US patent 5824784, 1998.
40 Kinstler, O. B., Brems, D. N., Lauren, S. L.,
Paige, A. G., Hamburger, J. B., Treuheit, M. J.,
Characterization and stability of N-terminally
PEGylated rhG-CSF, Pharmac. Res. 13, 996
1002, 1996.
41 Boyd, M. R., Gustafson, K. R., McMahon, J. B.,
Shoemaker, R. H., OKeefe, B. R., Mori, T., Gu-
lakowski, R. J., Wu, L., Rivera, M. I., Lauren-
cot, C. M., Currens, M. J., Cardellina II, J. H.,
Buckheit, R. W., Nara, P. L., Pannell, L. K.,
Sowder II, R. C., Henderson, L. E., Discovery
of cyanovirin-N, a novel human immunodefi-
ciency virus-inactivating protein that binds
viral surface envelope glycoprotein gp120:
potential applications to microbicide develop-
ment, Antimicrob. Ag. Chemother. 41, 1521
1530, 1997.
42 Bolmstedt, A. J., OKeefe, B. R., Shenoy, S. R.,
McMahon, J. B., Boyd, M. R., Cyanovirin-N de-
fines a new class of antiviral agent targeting
N-linked, high-mannose glycans in an oligo-
saccharide specific manner, Mol. Pharmacol.
59, 949954, 2001.
43 Barrientos, L. G., OKeefe, B. R., Bray, M., San-
chez, A., Gronenborn, A. M., Boyd, M., Cyano-
virin-N binds to the viral surface glycoprotein,
GP1,2 and inhibits infectivity of Ebola virus,
Antiviral Res. 58, 4756, 2003.
44 Green, M. E., Roberts, M., Goodin, R., Boyd,
M. R., Mori, T., OKeefe, B. R., Site-specific
modification of a mutant cyanovirin-N with a
large poly(ethylene glycol) moiety yields a po-
tent anti-HIV agent with reduced toxicity in
mice, J. Controlled Rel. Soc. 403, 2003, poster
abstract.
45 Botos, I., OKeefe, B. R., Shenoy, S. R., Cartner,
L. K., Ratner, D. M., Seeberger, P. H., Boyd,
M. R., Wlodawer, A., Structures of the com-
plexes of a potent anti-HIV protein cyanovirin-
N and high mannose oligosaccharides, J. Biol.
Chem. 277, 3433634342, 2002.
46 Bewley, C. A., Gustafson, K. R., Boyd, M. R.,
Covell, D. G., Bax, A., Clore, M., Gronenborn,
A. M., Nat. Struct. Biol. 5, 571578, 1998.
47 Green, M. E., Roberts, M., Goodin, R., Zhang,
Y., Novak, M., Bossard, M. J., Case study for
site-specific peg-protein modification: peg-cya-
novirin-N, a potential anti-HIV agent, AAPS,
poster # W5048, 2003.
48 Folkers, K., Bowers, C. Y., LJungquist, A.,
Tang, P-F.L., Louisa, Kobota, M., Feng, D.-M.,
Effective antagonists of the luteinizing hor-
mone releasing hormone which release negli-
gible histamine, EP patent 377665, 1995.
49 El-Tayar, N., Zhao, X., Bentley, M. D., PEG-
LHRH analog conjugates, US patent
6433135B1, 2002.
50 He, H., Murby, S., Warhurst, G., Gifford, L.,
Walker, D., Ayrton, J., Eastmond, R., Rowland,
M., Species differences in size discrimination
in the paracellular pathway reflected by oral
bioavailability of poly(ethylene glycol) and D-
peptides, J. Pharmac. Sci. 87, 626633, 1998.
51 Kim, J., Park, K., Glucose binding property of
pegylated concanavalin A, Pharmac. Res. 18,
794799, 2001.
52 Roberts, M. J., Bentley, M. D., Harris, J. M.,
Chemistry for peptide and protein PEGylation,
Adv. Drug Deliv. Rev. 54, 459476, 2002.
References 1417
53 Felix, M., Site-specific poly(ethylene glycol)
PEGylation of peptide, in Poly(ethylene glycol)
Chemistry and Biological Applications (J. M. Har-
ris and S. Zalipsky, Eds.), ACS Symposium
Series 680, ACS, Washington, DC, 1997,
chap. 16, pp. 218238.
54 Kwok, K., McKenzie, D. L., Evers, D. L., Rice,
K. G., Formulation of highly soluble poly(ethy-
lene glycol)-peptide DNA condensates, J. Phar-
mac. Sci. 87, 9961003, 1999.
55 Bedu-Addo, F., Moreadith, R., Advant, S.,
Preformulation development of recombinant
pegylated staphylokinase SY 161 using statisti-
cal design, AAPS Pharm., Sci. 4, article 19,
2002, www.aaps.org..
56 Callahan, W., Narhi, L., Kosky, A., Treuheit,
M., Sodium chloride enhances the storage
and conformational stability of BDNF and
PEG-BDNF, Pharmac. Res. 18, 261266, 2001.
57 Huber, J. D., Campos, C. R., Egleton, R. D.,
Witt, K., Guo, L., Roberts, M. J., Bentley,
M. D., Davis, T. P., Conjugation of low molecu-
lar weight poly(ethylene glycol) to biphalin en-
hances antinociceptive profile, J. Pharmac. Sci.
92, 13771385, 2003.
58 Heller, M., Carpenter, J., Randolph, T., Confor-
mational stability of lyophilized PEGylated
proteins in a phase-separating system, J. Phar-
mac. Sci. 88, 5864, 1999.
59 Delgado, C., Malmsten, M., Van Alstine, J.,
Analytical partitioning of poly(ethylene glycol)
modified proteins, J. Chromatogr. 692, 263
272, 1997.
60 Habeeb, A. F. S. A., Determination of free ami-
no groups in proteins by trinitrobenzenesulfo-
nic acid, Anal. Biochem. 14, 328336, 1966.
61 Karr, L. J., Donnelly, D. M., Kozlowski, A., Har-
ris, J. M., Use of poly(ethylene glycol)-modified
antibody in cell extraction, Methods Enzymol.
228, 377390, 1994.
62 Riddles, P. W., Blakely, R. L., Zerner, B., Ell-
mans reagent: 5,5'-dithiobis(2-nitrobenzoic
acid) a reexamination, Anal Biochem. 94, 75
81, 1979.
63 Kinstler, O. B., Gabriel, N. E., Farrar, C. E., De-
Prince, R. B., N-terminally chemically modi-
fied protein compositions and methods, US
patents 5985265, 1999 and 5824784, 1998.
64 Ho, D. H., Brown, N. S., Yen, A., Holmes, R.,
Keating, M., Abuchowski, A., Newman, R.,
Krakoff, I., Clinical Pharmacology of Polyethy-
lene glycol-L-asparaginase, Drug Met. Dispos.
14, 349352, 1986.
65 Wang, Y.-S., Youngster, S., Grace, M., Bausch,
J., Bordens, R., Wyss, D. F., Structural and
biological characterization of pegylated recom-
binant interferon alpha-2b and its therapeutic
implications, Adv. Drug Deliv. Rev. 54, 547
570, 2002.
Abstract
Infectious diseases such as hepatitis C or tu-
berculosis are widespread among the hu-
man population. Despite many efforts, no
efficient treatment is available to date,
which makes the need to develop new vac-
cines obvious. The use of specific protein
or peptide subunits of the pathogens for
vaccination contributes to the design of ef-
fective and safe vaccines, and lowers the
costs of production. However, peptides
alone are in general not very immunogenic,
and require adjuvants to induce an ade-
quate immune response. We have devel-
oped two novel adjuvants IC30 and IC31
that strongly enhance the immune re-
sponse; both of these are based on the cat-
ionic drug delivery transport system. IC30,
a cationic poly-amino acid poly-L-arginine
was identified as an adjuvant that trans-
ferred in a highly efficient manner peptides
to antigen-presenting cells (APCs) in an in-
vestigation to use tumor antigens, as thera-
peutic vaccines in mice. This enhanced up-
take of peptides led subsequently to a
strongly improved peptide-specific T cell re-
sponse and a reduction in tumor growth.
The length of the poly-L-arginine molecule
and the negative charge of the peptides in-
fluence the uptake of the peptides. A thera-
peutic vaccine against hepatitis C has been
developed subsequently using IC30 formu-
lated with five synthetic peptides. Results
from clinical trials, both in healthy volun-
teers and chronically infected patients, will
be discussed. The success with IC30 has
prompted the search for further adjuvants
with even better characteristics. Cationic
antimicrobial peptides (CAMPs) are used
by the immune system as a defense mecha-
nism against infections by microbes.
Hence, although CAMPs have been used
as antibiotic therapeutics, it was not known
that they could function as adjuvants. The
adjuvant effect was first shown for an artifi-
cial CAMP, KLKL
5
KLK, when co-injected
with ovalbumin. Furthermore, it was shown
that, like IC30, KLKL
5
KLK enhances the as-
sociation of antigen to APCs and induces
the formation of an antigen depot at the site
of injection. The adjuvant properties of
KLKL
5
KLK could be further enhanced by
combination with ODN1a, a novel immu-
nostimulatory deoxynucleotide containing
repeats of deoxy-inosine/deoxy-cytosine.
This novel adjuvant, IC31, has the unique
capacity to stimulate T and/or B lympho-
cytes in vivo. In this chapter we will discuss
results of further pre-clinical models where
IC30 and IC31 have been tested in existing
and novel vaccines, together with details of
recent experiments that enlighten their
mechanisms of actions.
1419
3
Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems
ISBN: 3-527-31184-X
Abbreviations
AIDS aquired immunodeficiency syn-
drom
APC antigen-presenting cell
BCG bacillus Calmette-Gurin
vaccine
CAMP cationic antimicrobial peptide
EBV Epstein-Barr virus
FACS fluorescence-activated cell
sorting
GCP good clinical practice
GLP good laboratory practice
GM-CSF granulocyte-macrophage colony-
stimulating factor
HCV hepatitis C virus
HER2 human epidermal growth factor
receptor2
HSV herpes simplex virus
IFA incomplete Freunds adjuvant
IFN interferon
IL interleukin
MHC major histocompatibility com-
plex
ODN oligodeoxynucleotide
PAMP pathogen-associated molecular
patterns
TLR toll-like receptor
TNF-a tumor necrosis factor-alpha
TRP-1 tyrosinase-related protein-1
3.1
Vaccines and their Importance in the Fight
against Human Diseases
Edward Jenner first showed more than
200 years ago that infection with cowpox
virus could protect against human small-
pox. Amazingly, at that time nothing was
known about the pathogens causing dis-
eases, and it was only later when Robert
Koch discovered that infections were
caused by micro-organisms. These and
other discoveries enabled the production
of vaccines, and some diseases (e.g., small-
pox) have subsequently been eradicated.
Other infections such as diphtheria or po-
lio, although not yet eradicated, have be-
come very rare in the Western world, be-
cause of early childhood vaccination pro-
grams. Despite the tremendous success of
vaccination programs, many infectious dis-
eases such as hepatitis C or tuberculosis
are still widespread among the human
population. It is estimated that 170 million
people worldwide are infected with hepati-
tis C, and that one-third of the worlds
population is carrying pathogens causing
tuberculosis. Despite many efforts, no effi-
cient treatment of the above-mentioned
diseases is available to date, and the need
to develop new vaccines is clear.
As vaccines have dramatically reduced the
burden imposed by infectious disease on
the population in the developed world,
safety has become more important than
ever. This is explained by the fear that the
risk of experiencing adverse effects by vacci-
nation seems higher than the risk of con-
tracting the disease, since many diseases
have vanished due to vaccination. There-
fore, the safety of vaccines must be moni-
tored very carefully and at the highest stan-
dard, because most vaccines are given to
healthy individuals to prevent a disease.
Whilst adverse reactions caused by vaccina-
tion are not tolerated when healthy adults
and especially healthy infants are vacci-
nated, more tolerance is usually shown by
regulatory authorities and also by the public
when a vaccine or other medication is given
to sick individuals for therapeutic use. Ad-
verse events following vaccination must be
monitored in large clinical trials before a
vaccine can be used for routine vaccination.
Independent of the vaccine composition,
the ultimate goal of any vaccine must be
the induction of a protective specific im-
mune response. Two fundamentally differ-
ent types of adaptive immune response
can be initiated by vaccination:
The cellular response results in the genera-
tion of, for example, cytotoxic T cells aid-
ing the fight against intracellular patho-
gens such as viruses and certain bacteria.
The humoral response results in activa-
tion of B cells and subsequently in the
production of specific antibodies, which
bind to extracellular pathogens and in-
duce their destruction.
The induction of a specific and long-last-
ing immune response depends critically
on the composition of a vaccine: the anti-
gen (whole pathogen, subunits of patho-
gen) and (for most of the time) also an ad-
juvant that helps to increase the immuno-
genicity of the antigen.
Traditional vaccines contain whole or-
ganisms, composed of either live, inacti-
vated and/or attenuated pathogens. The
smallpox vaccine is an example of a vac-
cine containing a live related virus (bovine)
that is less virulent in humans than the
smallpox virus, but still similar enough to
prevent viral smallpox disease. In 1952, Jo-
nas Salk developed the first vaccine con-
taining an inactivated pathogen the polio
vaccine that made use of formalin-killed
polioviruses. To the present day, viruses
are often propagated in cell cultures (such
as human diploid fibroblast cell lines or
Vero cells) and are subsequently formalin-
inactivated. Another example of using an
inactivated microbe is the hepatitis A vac-
cine. Although killed microbes can be
used for vaccination, not all killed patho-
gen-based vaccines are capable of inducing
a strong immune response. The third
form of traditional vaccines is to use live-
attenuated pathogens; this vaccine type is
mainly used for viral vaccines.
During the past few decades, advances
in molecular biology and a better under-
standing of immunology have enabled the
development of subunit vaccines. One ex-
ample is that of an acellular Bacillus per-
tussis vaccine, which contains a detoxified
version of the pertussis toxin in combina-
tion with one or more B. pertussis antigens
such as the filamentous hemagglutinin,
pertactin or a fimbrial protein. The devel-
opment of this subunit vaccine was also
stimulated by local as well as systemic ad-
verse reactions caused by the vaccine con-
sisting of inactivated B. pertussis, which
has been used for many years. For other
bacterial pathogens for example, Strepto-
coccus pneumoniae, which causes invasive
bacterial infection mainly in children aged
less than 2 years and adults older than 65
years [1] polysaccharide vaccines contain-
ing purified capsular polysaccharide anti-
gens of multiple serotypes were developed
for vaccination. However, the existence of
more than 90 serotypes of Strep. pneumo-
niae prevents the achievement of protec-
tion against all clinical isolates by this
approach.
As polysaccharide vaccines were shown
to elicit only limited protection in infants
and young children, conjugated polysac-
charide vaccines were developed to provide
a more potent and sustained immune re-
sponse [2]. Many of these traditional vac-
cines target childhood diseases, and are
used in combinations for pediatric applica-
tions in order to reduce the number of in-
jections during the first years of life. Cur-
rently, vaccine combinations can prevent
against between three to six different dis-
eases, such as the measles-mumps-rubella
(MMR) vaccine or the diphtheria-tetanus-
pertussis combination, which may be ad-
ministered together with Haemophilus in-
fluenzae, hepatitis B, or poliovirus vac-
cines.
More recently, a recombinant non-infec-
tious subunit viral vaccine derived from
the hepatitis B surface antigen (Recombi-
vax HB) has been developed. The antigen
is produced in fermentation cultures of
Saccharomyces cerevisiae, and is therefore
free from human blood products. Hepati-
tis B is an inflammation of the liver
caused by the hepatitis B virus, and can be
very serious or even fatal. The virus is
usually spread by contact with infected
blood, though an infection can be pre-
vented by vaccination. The vaccine is
highly immunogenic, well-tolerated, and
possesses an excellent protective efficacy
that leads to immunity for 10 years.
To date, more than 20 infectious diseases
can be prevented by vaccination, and novel
vaccines are being developed every year.
However, despite the successful use of tradi-
tional vaccines for many diseases, many ill-
nesses are very difficult to prevent or can
not be prevented by vaccination at present.
Reasons for this may be the inability to
grow pathogens in vitro (e.g., Treponema pal-
lidum), the existence of numerous serotypes
(e.g., Strep. pneumoniae), or antigenic varia-
tion of a particular infectious agent (e.g.,
Neisseria sp.). Furthermore, whole inacti-
vated pathogens may cause severe adverse
effects (e.g., Streptococcus pyogenes) and live
attenuated pathogens may induce disease
upon vaccination, all of which makes tradi-
tional vaccine development against some
diseases very difficult, or impossible. Mi-
cro-organisms have in addition developed
sophisticated immune evasion mechanisms
[e.g., herpes simplex virus (HSV), human
immunodeficiency virus (HIV), Mycobacter-
ium tuberculosis, Plasmodium falciparum, Ep-
stein-Barr virus (EBV), and hepatitis C virus
(HCV)], which necessitates novel strategies
for immune therapy.
The lack of prophylactic vaccines for
life-threatening diseases such as AIDS and
cancer has given attention also to the de-
velopment of therapeutic vaccines. The
function of a therapeutic vaccine is not to
prevent a disease, but to reduce the symp-
toms caused by a disease or, in an ideal
case, to eliminate a pathogen and thereby
the disease altogether.
Targets for therapeutic vaccination are
chronic, severe infections where vaccina-
tion may lower the burden imposed on pa-
tients or eliminate the infectious agent.
One of many examples is HSV, which
causes life-long, recurrent chronic infec-
tions. Various vaccination approaches have
been evaluated to treat HSV infections,
and vaccines based on HSV-2 envelope gly-
coproteins could elicit cellular immunity
and have reached advanced-phase clinical
trials [3]. Therapeutic vaccines for the
treatment of HIV patients are also in the
progress, and some vaccines were shown
in Phase II trials to be capable of slowing
the progression of HIV infection and to
boost the immune response against the
AIDS virus (Immune Response Corpora-
tion, Carlsbad). Yet, despite the tremen-
dous efforts of many research groups, no
efficient therapeutic vaccination is cur-
rently available for the treatment of infec-
tions such as hepatitis C and AIDS.
Immunotherapy for the treatment of
cancer has received renewed interest dur-
ing the past decade, mainly because differ-
ent cancer treatments such as radiotherapy
or chemotherapy have the drawback of
killing all growing cells, regardless of
whether they are cancer cells or normal
cells. Therefore, several different
approaches are presently being undertaken
to identify novel vaccines containing tu-
mor antigens, which might enable more
specific treatments of cancer patients (see
also Part V, Chapter 6). However, the effec-
tiveness of these novel therapeutic vac-
cines and their safety has yet to be exam-
ined. One example of such a novel
approach to treat prostate cancer patients
is a vaccine containing the complex carbo-
hydrate molecule globo H, an antigen
present on prostate cancer cells, conju-
gated to keyhole limpet hemocyanin admi-
nistered with the adjuvant QS-21. The vac-
cine was examined over 26 weeks, and has
proven to be safe and capable of inducing
specific high-titer IgM antibodies against
globo H in prostate cancer patients with a
broad range of stages, and awaits further
validation in Phase II and III trials [4]. An-
other promising tumor antigen is HER2/
neu, which was found to be overexpressed
in breast cancer and other carcinoma cells,
and different HER2/neu peptides have
subsequently been included in cancer vac-
cines (see Part I, Chapter 5). The efficacy
of some of these peptides has been as-
sessed in patients with breast and ovarian
cancer. Although T cell responses against
the peptides were detected in immunized
patients, no clinical responses have yet been
described [5]. Many more approaches are
currently being examined for their safety
and efficacy; however, their potency for the
treatment of cancer has yet to be shown.
Besides the attention that has been giv-
en to the development of therapeutic vac-
cines in the fight against cancer, therapeu-
tic vaccines are also in development
against the use of addictive drugs such as
nicotine or for the treatment of autoim-
mune diseases [6].
3.2
Adjuvants: An Overview
3.2.1
The History of Adjuvants
Spurred on by shortcomings of traditional
vaccines, much attention has been given
to the development of alternative vaccines
using approaches based on recombinant
proteins, peptides, or DNA. The selection
of recombinant proteins or synthetic pep-
tides of the pathogen for vaccination can
not only focus and augment the effective
immune response, but may also greatly
improve the safety of a vaccine, especially
when adverse reactions have been ob-
served after vaccination with the complete
pathogen.
However, proteins and in particular
peptides are often not very immuno-
genic by themselves, and require the
help of adjuvants to induce an adequate
immune response. Approximately 80 years
ago, Ramon showed that the co-injection
of compounds as diverse as agar, tapioca,
starch, oil, and saponin with tetanus and
diphtheria toxin increased the antitoxin re-
sponse [7]. An adjuvant was originally de-
fined as a substance that increases the im-
mune response when co-injected with an
antigen. However, Ramon showed that this
effect can even be caused by the co-injec-
tion of breadcrumbs with diphtheria or te-
tanus toxin, which made the requirement
for a more stringent definition obvious. To
date, the most important requirements of
an adjuvant are safety and potency. Thus,
the adjuvant should not cause any severe
side effects, and it should consistently in-
crease the immune response towards co-
administered antigens, ultimately leading
to fewer and/or lower doses of the antigen.
In addition, the adjuvant should be a
stable component and it should not be im-
munogenic by itself. In traditional vac-
cines very often components of organisms
acted as adjuvants, and/or aluminum salts
were used. With the need to develop novel
subunit vaccines, it quickly emerged that
novel and more effective adjuvants were
also needed.
3.2.2
Modes of Action of Adjuvants
To date, highly heterogeneous substances
with various (natural) functions and/or
chemical structures have been identified
as adjuvants, which therefore show a
broad spectrum of different capabilities.
However, it is commonly accepted that ad-
juvants induce or enhance antigen-specific
immune responses via at least one of the
following steps:
Facilitated delivery of antigen to the sec-
ondary lymphoid organs for a sufficient
period of time (e.g., via depot formation
at injection site and/or enhanced uptake
of antigens by APCs).
Activation of APCs (e.g., via toll-like re-
ceptor, TLR) for the induction of co-
stimulatory molecules, which is a pre-re-
quisite for the induction of potent T cell
responses [8].
Immunity to different infections re-
quires a distinct immune reaction against
the pathogens invading the organism.
Therefore, the adjuvants and antigens that
comprise a vaccine must be selected for
the onset of the required specific immuno-
logical pathway. Vaccination leads to an
adaptive immune response, and most vac-
cines stimulate preferentially either a cel-
lular response or a humoral response. In
general, the first contact of the immune
system with pathogens is mediated by
cells of the innate immune system, where-
by dendritic cells play the most important
role as APCs. They are present in most tis-
sues and, upon stimulation, move rapidly
towards the infection or injection site in
order to capture viruses, bacteria, or vac-
cine antigens. In order to induce an adap-
tive immune response, antigens must be
efficiently translocated to lymphoid organs
such as the spleen and the lymph nodes;
only there can the stimulation of nave T
cells and B cells take place. After uptake
of antigens, the dendritic cells migrate to
the lymphoid organs, become changed in
their activation status, present the antigen
to nave T cells and B cells, and induce
their proliferation and differentiation into
effector cells. Depending on several factors
(e.g., the type of antigen, dose of antigen,
type of APC, cytokines produced by APCs),
it is reported that preferentially either a
type 1 or a type 2 adaptive immune re-
sponse is induced. Whilst a type 1 re-
sponse is predominantly required for the
generation of cytotoxic T cells, the optimal
activation of the innate immune system
as well as the production of certain im-
munglobulin (Ig) subclasses a type 2 re-
sponse is predominantly required for the
induction of humoral responses with re-
spective Ig subclasses. It is generally ac-
cepted that type 1 immune responses are
required for the defense of intracellular
pathogens as well as tumors, while type 2
immune responses are required for the de-
fense of extracellular pathogens. Cellular
type 1 responses are characterized by the
production of cytokines such as interleu-
kin (IL)-12 and interferon-c (IFN-c), and
cellular type 2 responses by cytokines IL-4
and IL-5. In mice, the IgG subclasses
IgG1 and IgG2 are representatives for type
2 and type 1 humoral immune responses,
respectively.
Adjuvants such as QS21 or CpG-ODN
have been described which induce a type 1
cellular immune response, whereas in
contrast, aluminum hydroxide was shown
to induce a type 2 immune response.
Thus, it is important to choose the correct
adjuvant(s) that is required to initiate a hu-
moral and/or cellular as well as type 1
and/or type 2 immune response by vacci-
nation. For an efficient stimulation of an
immune response capable of eliminating a
current infection or providing a long-term
response to prevent infections on recur-
rent encounters, it may be advantageous
to combine more than one adjuvant with
different capabilities, together with the
protective antigen(s).
3.2.3
Adjuvants Commonly Used for Human
Vaccination
Aluminum salts have been used for sev-
eral decades in vaccines which prevent
early childhood diseases, and have been
proven safe in this respect. Vaccines were
prepared using in situ precipitation in the
presence of the antigen, or by adsorbing
the antigen onto an aluminum gel [7]. The
different aluminum compounds were
thought to induce the formation of a depot
at the injection site from where the anti-
gen is slowly released, although recent re-
ports indicated that aluminum salts induce
only limited depot formation [9]. Although
aluminum compounds are highly effective
in a primary immunization, the immune
response is not higher in the secondary
immunization when compared to soluble
vaccines. Furthermore, despite their good
safety profile, aluminum salts were shown
to exert some adverse reactions such as
contact hypersensitivity, inflammation,
subcutaneous nodule formation, or an in-
creased level of IgE antibodies, which was
implied with allergic reactions. A further
drawback of aluminum salts is that they
are ineffective in combination with some
antigens [7]. The biggest limitation for the
use of aluminum salts for general vaccina-
tion strategies is their inability to induce a
cytotoxic T cell response, as they mainly
induce a type 2-like immune response
[10]. For production purposes, it should
also be emphasized that aluminum-adju-
vanted vaccines can be neither frozen nor
lyophilized, because these procedures
cause the adjuvant to aggregate and pre-
cipitate [7].
MF59 is a water-in oil-emulsion com-
prising squalene, Tween 80, and Span85.
It has been reported to stimulate humoral
and cellular immune responses in combi-
nation with several subunit antigens. It
has also been proven safe, and does not
cause any major adverse reactions. In clin-
ical trials it has already been used with in-
fluenza, HSV or HIV antigens, and has
been marketed as part of an enhanced in-
fluenza vaccine for the elderly. MF59 does
not form a depot at the injection site, but
targets macrophages and dendritic cells at
the site of injection and in lymph nodes.
It has also been shown that the level of cy-
tokine (IL-2, IL-4, IL-5, IL-6 and IFN-a)
production is increased [11].
Virosomes are tiny vesicles containing
viral hemagglutinin in addition to mem-
brane-derived phospholipids, and this en-
ables the virosomes to enter their target
cells via an endolysosomal pathway. After
endocytosis, the viral hemagglutinin medi-
ates membrane fusion with the endosome,
and specific antigens are released into the
cytoplasm of the cell. Depending on the
location of the antigen, whether on the
surface or encapsulated within the vesicle,
virosomes can stimulate humoral and/or
cellular immune responses. Once they
have delivered the antigens, the virosomes
are completely degraded within the cells.
During the 1990s, a virosomal hepatitis
A vaccine and in 2001, an influenza vac-
cine were developed by Berna Biotech by
applying the virosome technology.
To date, aluminum salts, MF59, and
virosomes are the only adjuvants which
have been used in licensed products,
though many more adjuvants (e.g., poly-L-
arginine, CpG-ODN) have been tested in
pre-clinical experiments and/or in clinical
trials and hold promise for the develop-
ment of novel vaccines. However, for a
number of novel adjuvants safety issues
have precluded their use in vaccine formu-
lations for humans.
3.3
Cationic Peptides as Novel Vaccine
Adjuvants
Peptide antigens are, in most cases, poorly
immunogenic in their own right, and
must be transported and exposed to im-
mune cells with high efficiency to induce
a specific and effective activation of the
immune system. Therefore, much effort
has recently been undertaken to develop
new and efficient delivery systems for pep-
tide antigens, and this has led to the devel-
opment of poly-cationic peptides as novel
vaccine adjuvants.
Early experiments showed that a trans-
ferrin-polycation complex transported bac-
terial DNA into cells [12]. Ions are taken
up by cells as an iron-transferrin complex
by receptor-mediated endocytosis. Prota-
mine or poly-lysine ligated by disulfide
bonds to transferring and mixed with a lu-
ciferase-encoding plasmid may bind the
DNA because of the cationic properties of
the complex [12]. Subsequently, avian ery-
throblasts and human K-562 cells were in-
cubated with the transferrin-polycation
peptide-DNA complex, and the complexes
were recognized and transported into the
cells by receptor-mediated endocytosis and
taken up into endosome-like intracellular
vesicles [12]. Treatment with chloroquine
(an agent that affects the endosomal pH)
enhanced the uptake considerably. In con-
trast to other transfection methods, the
transfection of cells with transferrin-
mediated endocytosis did not cause signifi-
cant cell death, because of the physiologi-
cal nature of the method and transfected
cells even differentiated normally [12].
Poly-cationic compounds have also been
used previously to transport proteins such
as heparin, albumin, or horseradish per-
oxidase into cells [13, 14].
3.3.1
Poly-L-Arginine and its Mode of Action as
an Adjuvant
The observations of an enhanced transport
of DNA or proteins into cells by poly-cat-
ionic amino acids were used in an attempt
to identify novel adjuvants capable of
transporting peptide antigens into cells.
An effective display of bacteria-, virus-,
or tumor-derived peptide antigens by APCs
to T lymphocytes will enhance the im-
mune response against infections or tu-
mors. Hence, the capability of translocat-
ing peptide antigens into APCs in the
presence of poly-cationic peptides was
studied systematically [15, 16]. This novel
method of transporting antigens into cells
was termed transloading. In these trans-
loading experiments, fluorescence-labeled
peptide antigens plus poly-cationic pep-
tides, as poly-L-arginine or poly-L-lysine,
were incubated with bone marrow-derived
APCs in vitro. The intracellular increase of
fluorescence was measured in the pres-
ence of both poly-cationic peptides, with
poly-L-arginine being more efficient. The
peptide delivery depended upon the degree
of polymerization of the poly-cationic pep-
tide, with a minimum chain length of 15
amino acids [15].
Further studies using either peptides (Fig.
3.1) or whole proteins (Fig. 3.2) showed that,
in the presence of poly-L-arginine the en-
hanced uptake of antigens by MHC class
II-positive APCs takes also place in vivo
(compare Fig. 3.1A versus Fig. 3.1B, where
peptide was injected alone). It could be also
shown that such antigen-charged APCs mi-
grate to draining lymph nodes (Fig. 3.2),
where activation of nave Tcells takes place.
It is largely unknown how poly-cationic
peptides promote the uptake of other com-
ponents into cells, but it has been proposed
that they permeabilize the cell membrane.
However, an adjuvant using such a mecha-
nism to deliver DNA or proteins into cells
might not be very useful, since the cells
can leak cellular components. Thus, the
mechanism enabling poly-L-arginine to
translocate DNA or proteins into cells was
examined and the release of cellular lactate
dehydrogenase (LDH) was measured fol-
lowing treatment with poly-L-arginine. No
release of LDH was observed, indicating
that poly-L-arginine might be taken up by
endocytosis, with the transloading effi-
ciency being greatly reduced for poly-L-argi-
nine at low temperatures, again confirming
Fig. 3.1 Poly-L-arginine enhances the uptake of
antigens by MHC class II-positive antigen-present-
ing cells in vivo. A peptide derived from listerioly-
sin from Listeria monocytogenes was labeled with
SFX (fluorescein, green fluorescent dye) and in-
jected in combination with poly-L-arginine (A) or
alone (B) subcutaneously into the flank of mice.
At 7 days after injection, cryosections of the injec-
tion sites were analyzed by confocal microscopy.
To define antigen-presenting cells, the cells were
counterstained with anti-MHC class II-Texas Red
(red fluorescent dye).
Fig. 3.2 Antigen-presenting cells loaded with anti-
gen in the presence of poly-L-arginine migrate to
draining lymphoid organs. Green-fluorescent pro-
tein was injected with poly-L-arginine three times
subcutaneously. At 5 days after the last injection,
cytospins of draining lymph node cells were ana-
lyzed by confocal microscopy. To define antigen-
presenting cells, the cells were counterstained
with anti-MHC class II-Texas Red (red fluorescent
dye).
the possibility that it is internalized by en-
docytosis [15].
It was discovered recently that the nucle-
ar transcription activator protein (Tat), en-
coded by HIV type 1, is a naturally occur-
ring macromolecule that enters cells. The
determination of structural requirements
revealed that the deletion of one arginine
residue from either the amino terminus or
the carboxyl terminus resulted in a signifi-
cant (80%) loss in transport capabilities.
Thus, the arginine content is primarily re-
sponsible for cellular uptake, and further-
more the presence of at least six arginine
residues is an important factor in this re-
spect. Significantly, conjugates containing
seven to nine arginine residues exhibited
better uptake than the natural Tat. Several
proteins attached covalently to HIV-1 Tat
have been delivered into cells, although
the detailed mechanism of cellular uptake
remains unknown [17, 18]. Recent co-local-
ization studies have shown that a nona-ar-
ginine (R
9
) is internalized by endocytosis
rather than by crossing the plasma mem-
brane, and that the delivery of molecules
into live mammalian cells involves binding
to cell surface heparan sulfate, since R
9
was incapable of entering living cells defi-
cient in heparan sulfate [19].
In addition to the enhanced uptake of
antigens by APCs, poly-L-arginine also ex-
erts its adjuvant effects via the formation
of a depot at the injection site (Fig. 3.3).
This effect, which is based on ionic inter-
action of the vaccine components, pro-
longs the availability of antigen in the
body. The consequences are most probably
Fig. 3.3 Poly-L-arginine induces the formation of a
depot at the injection site. Melanoma-derived pep-
tide TRP-2
181-188
labeled with SFX (fluorescein)
was injected subcutaneously into the flank of
mice. At 4 days after injection, pictures were taken
from the inner side of the skin at the injection
site, using a digital camera.
constant uptake by APCs, thus leading to
sustained priming of specific T cells and,
in turn, prolonged immune responses.
3.3.2
Poly-L-Arginine as Type I-inducing Adjuvant
for Peptide-based Vaccines
On the basis of these potent adjuvant prop-
erties, poly-L-arginine was developed as an
adjuvant for peptide-based vaccines. In a
number of pre-clinical settings, poly-L-argi-
nine was analyzed for its potency to induce
specific Tcell responses against peptides de-
rived from bacteria, viruses, or tumors.
As one model system in pre-clinical
studies, the efficiency of poly-L-arginine
for the treatment of melanomas (M3 mod-
el) and mastocytomas (P815 mastocytoma)
was examined. As no tumor peptide anti-
gens were known at that time for the M3
melanoma, four putative peptide antigens
containing H-2K
d
binding sites of the
known tumor antigens tyrosinase and of
tyrosinase-related protein-1 (TRP-1) were
selected by computer analysis and com-
bined for co-injection with poly-L-arginine
into mice [20]. For transfection of a P815
mastocytoma, a peptide (SYFPEITHI) de-
rived from the Janus kinase JAK1 was
used, because it was shown to be pre-
sented by 5% of all MHC class molecules
in P815 cancer cells.
It is generally believed that APCs play a
major role in initiating the cascade leading
to activation of tumor-specific T cells. As
APCs are present in high frequencies in
the skin (Langerhans cells), the subcuta-
neous route is highly suited to the efficient
targeting of APCs. Thus, mice were in-
jected with the peptide vaccine three times
subcutaneously and challenged one week
after the last vaccination by the adminis-
tration of tumor cells [21]. Tumor cells sta-
bly transfected with a GM-CSF plasmid
served as standard for vaccine efficiency,
enabling the comparison of a cytokine-se-
creting cellular vaccine with peptide vacci-
nation. Immunization with poly-L-arginine
and the JAK1-derived peptide provided
protection similar to the cellular granulo-
cyte-macrophage colony-stimulating factor
(GM-CSF) vaccine, whereas protection
against M3 melanoma was lower, which
may have been due to the arbitrary selec-
tion of the peptides. The results implied
that peptide vaccines can induce an im-
mune response comparable to that of a cy-
tokine-secreting cellular vaccine, if the
peptides are applied in adequate quanti-
ties. However, in contrast to cellular vac-
cines the peptide vaccines are inexpensive
to produce and chemically well defined.
With these novel vaccination regimens
at hand, it must be evaluated whether can-
cer can efficiently be cured by a combined
application of chemotherapy, radiation
therapy, cellular vaccination and/or pep-
tide- or protein-based vaccination. The ap-
plication of immunotherapy of cancer may
be envisaged for cancer patients who have
undergone a chemo- or radiation therapy
first to reduce the size of the tumor(s) to a
minimum before vaccination. The contin-
ued identification of tumor-associated or
tumor-specific peptides and proteins will
enable peptide or protein vaccines to be
applied to cancer patients. Those patients
who suffer from cancers for which no tu-
mor-derived antigens have been identified
could, alternatively, be treated with cellular
vaccines. The therapeutic vaccination
approach with either vaccine type could
hopefully prevent the recurrence of meta-
stases and possibly help patients in the
cure of this devastating disease.
As mentioned above, for many infec-
tious diseases no vaccine or vaccines with
low efficacy exists at present. An example
of this is tuberculosis, a condition which
is estimated to develop in 8 million people
each year, whilst in 2 million people it
proves to be fatal [22]. During the past few
decades, most people infected with tuber-
culosis have lived in developing countries,
and many have been simultaneously in-
fected with HIV, while others were sub-
stance abusers. However, more recently tu-
berculosis has been seen to be spreading
in the Western world a situation which
has led to renewed efforts for its control,
though as yet no efficient vaccination is
available. Currently, although the bacillus
Calmette-Gurin (BCG) vaccine is used for
vaccination in Third World countries, clini-
cal trials have shown an efficacy of only
about 50%, which makes the need for nov-
el tuberculosis vaccines clear [23]. Poly-L-
arginine has been successfully used in
combination with tumor-specific antigens
as a vaccine in mice, and was also able to
protect animals from tumor growth. Thus,
an investigation was made as to whether
poly-L-arginine has immunogenic capaci-
ties when mixed with peptide antigens de-
rived from M. tuberculosis. Consequent
studies in mice revealed that the co-injec-
tion of both compounds resulted in an in-
duction of peptide-specific, IFN-c-produc-
ing T cells (unpublished results) (Fig. 3.4).
Poly-L-arginine showed comparable im-
munostimulatory effects when used in
combination with peptide antigens derived
from different bacteria or viruses. Thus,
these results supported not only further
pre-clinical but also clinical studies to-
wards the development of novel antigenic
peptide-based vaccines using poly-L-argi-
nine as adjuvant. Such vaccines will offer
the benefits of stability, low costs, and rap-
id and easy preparation.
3.3.3
Clinical Experiences with Poly-L-arginine
The first vaccine where poly-L-arginine has
been applied in humans is a fully syn-
thetic therapeutic hepatitis C virus (HCV)
vaccine. This vaccine was named IC41,
and consists of a mixture of synthetic pep-
Fig. 3.4 Poly-L-arginine strongly induces peptide-
specific type 1 responses. A peptide mixture
(p180; p184; p185; p186; p187), where all pep-
tides were derived from M. tuberculosis, was in-
jected three times subcutaneously either in com-
bination with poly-L-arginine or IFA. At 7 days
after the third injection, the number of IFN-c-pro-
ducing cells was determined by ELISpot assay. For
restimulation, a peptide derived from Plasmodium
falciparum (p1774) was used.
tides representing conserved T cell epitopes
of HCVplus poly-L-arginine as a synthetic T
cell adjuvant. The aim of this therapeutic
approach is to restore a so-called type I T
cell response against HCV in chronically in-
fected patients. This response is typically
seen in about 15% of infected persons
who do not proceed to chronicity but can
clear HCV during the acute phase of infec-
tion. Since the pre-clinical experience with
poly-L-arginine described earlier demon-
strated its ability to induce type I immune
responses in animal models, this represents
a promising T cell adjuvant for peptide vac-
cines to treat HCV.
Various doses of the above-mentioned
vaccine have been tested in several clinical
trials comprising more than 200 subjects:
in an initial Phase I study, safety and pre-
liminary immunogenicity data of several
doses were obtained. Results from that
trial prompted the initiation of a dose-opti-
mization study comprising 128 healthy
volunteers in 10 different dose groups.
The study was a randomized, single-blind,
parallel-group, controlled study conducted
to assess dose optimization and safety of
the HCV peptide vaccine, IC41, in healthy
subjects and was conducted in one center
in Austria. In total, 128 subjects were ran-
domly assigned to receive one of seven dif-
ferent doses and ratios of HCV peptide
vaccine with poly-L-arginine, HCV peptide
vaccine alone, poly-L-arginine alone, or sa-
line solution. All subjects received four ad-
ministered vaccinations at monthly inter-
vals, with immunogenicity being assessed
at each of these time points and at 3
months after the last vaccination.
The T cell stimulatory efficacy of poly-L-
arginine was tested in a Phase II clinical
trial in chronic HCV patients who had not
responded to, or had relapsed from, stan-
dard interferon/ribavirin therapy. This in-
vestigation was a randomized, double-
blind study of HCV peptide vaccine, IC41,
and was conducted in 11 centers in Ger-
many, Austria, and Poland. Sixty patients
were assigned at random to receive either
HCV peptide vaccine with poly-L-arginine,
HCV peptide vaccine alone, or poly-L-argi-
nine alone. All patients received six vacci-
nations at monthly intervals, with immu-
nogenicity assessed at each time point and
at 3 and 6 months after the last vaccina-
tion.
As a first important result, these trials
confirmed the excellent safety profile of
completely synthetic peptides in general,
and poly-L-arginine in particular. Further-
more, several important lessons regarding
the activation of human T cells were
learned: in both studies, T cell responses
were assessed using [
3
H]-thymidine prolif-
eration and IFN-c ELIspot assays, and flu-
orescence-activated cell sorting (FACS).
These assays, which have been standard-
ized and validated at Intercell AGs Clini-
cal Immunology Laboratory, enable reliable
measurements of epitope-specific T cell re-
sponses induced by vaccination. All assays
were performed in compliance with Good
Laboratory Practice (GLP)/Good Clinical
Practice (GCP) requirements. Standardiza-
tion of the blood cell isolation procedure
at the different investigational sites led to
a high rate of evaluable assays. However,
due to the lack of inter-laboratory standard-
ization of T cell assays, comparison of the
results of this study with published data
from similar trials is difficult. Cryopre-
served blood cells were used, which may
have resulted in a possible underestima-
tion of T cell responses compared with as-
says that utilize fresh blood.
Healthy volunteer vaccine responder
rates in the peptide control group (66.7%)
were comparable to those in the verum
groups, but were lower than the maximum
responder rates obtained, confirming that
optimal induction of peptide-specific T
cells requires co-injection of peptide and
poly-L-arginine. In addition to the vaccine
responder rates, comparable T cell prolif-
eration responder rates were observed in
the verum groups and the peptide control
group. ELIspot responder rates, however,
were greatest in the verum groups. This
finding implies that co-administration of
the T cell adjuvant, poly-L-arginine, is not
necessarily required for a proliferative T
cell response in healthy subjects, but is re-
quired for induction of functional IFN-c-
secreting T cells, which is a key aspect in
most infectious disease or cancer indica-
tions.
In the Phase II study population of
chronic HCV patients, a slightly different
picture was obtained: in general, CD4
+
and CD8
+
T cell responses to IC41 pep-
tides were more frequent and more vigor-
ous in peripheral blood samples from
those patients who were immunized with
peptide and poly-L-arginine together, than
in samples from those patients immu-
nized with peptide or poly-L-arginine
alone, thereby confirming the requirement
of poly-L-arginine as a T cell adjuvant. Vac-
cine responder rates were approximately 2-
to 3-fold higher in the verum groups than
in the control groups (Fig. 3.5). T cell pro-
liferation responders were more numerous
in the verum groups (3060%) than in the
control groups (017%).
Most importantly however, IFN-c ELI-
spot responders were observed exclusively
in the verum groups (Fig. 3.6). These re-
sults demonstrated for the first time that
poly-L-arginine is able to induce type I re-
sponses, even in the setting of chronic
HCV infection in patients who could not
be cured by the IFN/ribavirin standard
therapy. Significant IFN-c ELIspot re-
sponses were detected against both HLA-
class II (recognized by helper T cells) and
HLA-class I (recognized by cytotoxic T
cells) peptides. Analysis of the data up to
and including those obtained at the second
immunological check (Visit 11) performed
6 months after the last immunization re-
vealed no important differences in immu-
nological responder rates compared to the
data obtained at Visits 1 to 10, indicating
sustainability of the IC41-induced immune
response. The study also disclosed that T
cell immunity against the virus can be
raised to a level not too different from that
induced in healthy vaccines. Thus, immu-
Fig. 3.5 Vaccine responder rates in chronically in-
fected hepatitis C virus (HCV) patients. Peptide/
poly-L-arginine (9/12; 75%) versus control groups
(3/12; 25% in both).
Fig. 3.6 Type 1 responder rates in chronically in-
fected hepatitis C virus (HCV) patients. Type 1
(IFN-c ELIspot) responders were identified in the
peptide/poly-L-arginine group (5/12; 42%), but
not in the peptide-only or poly-L-arginine-only
groups (0/12 in both).
nosuppression may not be as prevalent as
anticipated in patients. Nonetheless, it re-
mains to be elucidated how such T cell re-
sponses can be optimally applied to reduce
disease progression or to ameliorate symp-
toms, and eventually to clear the infection.
Taking these results together, poly-L-argi-
nine represents one of the first synthetic
T cell adjuvants, which has consistently
from in vitro experiments up to incurable
chronically infected patients been able to
induce and augment the desired type of
immune response. Its ease of manufac-
ture, excellent safety profile and its efficacy
even in difficult settings such as chronic
HCV infection make it promising new
tool in the continuing battle against infec-
tious diseases and cancer.
3.4
Cationic Antimicrobial Peptides (CAMP)
as Novel Adjuvants
Higher vertebrates have developed an in-
nate or natural immune response, as well
as an adaptive response, to fight off micro-
bial invasions. The innate immune re-
sponse is triggered immediately after mi-
crobes attack the organism. The immune
system targets common structures con-
served in many micro-organisms to fend
them off, and antimicrobial peptides
which are produced in large quantities at
sites of infection and inflammation are
ancient weapons in these defense mecha-
nisms [24].
The use of antimicrobial peptides as a
response to microbial invasion is common
to all animals and plants, and it has there-
fore been suggested these peptides play a
fundamental role in the evolution of multi-
cellular organisms. Antimicrobial peptides
are used in the defense against a wide
range of microbes including bacteria, fun-
gi, viruses, and protozoa, and to date more
than 500 different antimicrobial peptides
have been discovered and are registered in
the antimicrobial database of the Univer-
sity of Nebraska Medical Center. Indeed,
the sequence diversity is so large that the
same peptide sequence is rarely discovered
from different organisms, as the exposure
to different microbes is unique to each an-
imal or plant, depending on the environ-
ment in which they live [24]. However,
although antimicrobial peptides are very
diverse, different peptides clearly form
molecules with common clusters of hydro-
phobic and cationic amino acids. These
peptides are summarized in the group of
cationic antimicrobial peptides (CAMPs).
In mammals, four sub-groups have been
described: a-defensins; b-defensins; h-de-
fensins; and cathelicidins. CAMPs are se-
creted into internal body fluids or stored
in cytoplasmic granules of professional
phagocytes. Cathelicidins are stored for ex-
ample as inactive fusion proteins in the
granules of granulocytes and are activated
by enzymatic cleavage [25]. Defensins are
stored in the cytoplasmic granules of neu-
trophils, macrophages or intestinal Paneth
cells [26]. Some peptides, for example the
human b-defensins-1, are constitutively
synthesized, whereas other peptides such
as the human b-defensins-2 are synthe-
sized only upon induction.
Antimicrobial peptides target the mem-
branes of pathogens, which are composed
of a bilayer containing lipids with nega-
tively charged phospholipids. However, at
present it is not fully understood how anti-
microbial peptides are able to kill patho-
genic microbes, and consequently a variety
of different mechanisms has been sug-
gested. It has been proposed that antimi-
crobial peptides are able to change the
structure of the cell membrane by displac-
ing lipids, and sometimes even enter the
target cell [24]. It has also been suggested
that the antimicrobial peptides facilitate
the integration of a hydrolase into the cell
wall, leading to its degradation. It is
further speculated that antimicrobial pep-
tides kill bacteria by depolarization of the
bacterial membrane [24]. Other antimicro-
bial peptides, such as nisin, which is pro-
duced by lactococci, interacts with the
membrane-bound peptidoglycan precur-
sors lipid I and lipid II, with this interac-
tion presumably playing a role in a pore
formation process [27].
The development of bacterial resistance
against conventional antibiotics is very
common and often also very fast although,
surprisingly, resistance against antimicro-
bial peptides has rarely been described.
Most animals and plants attack pathogens
with several different antimicrobial pep-
tides, but as these peptides do not contain
conserved sequence motifs their destruc-
tion by proteases can be difficult. Some
antimicrobial peptides have been used pre-
viously for human therapy, mainly by topi-
cal administration, but many of these
must used at such high concentrations
that they have not passed safety regula-
tions. Despite these difficulties, several
antibacterial peptides are currently under-
going clinical trials, and some have been
licensed for therapeutic use in humans.
The skin, together with the gastrointesti-
nal and respiratory tracts, are the areas
which are constantly exposed to microbes
and, consequently, they are the main loca-
tions of synthesis of antimicrobial pep-
tides. Besides their direct antimicrobial ac-
tivity, antimicrobial peptides released from
circulating cells or induced in the epithelia
can also promote a response of the adap-
tive immune system. Different possible
modes of action have been suggested for
a- and b-defensins. It was suggested that
antimicrobial peptides enhance the recruit-
ment of immature dendritic cells and of
effector T cells to the site of infection. In
addition, defensins have been implied as
facilitating the uptake of antigens by im-
mature dendritic cells via complex forma-
tion and internalization through a recep-
tor. It has also been suggested that defen-
sins enhance the maturation of immature
dendritic cells either directly, or indirectly
by the induction of tumor necrosis factor
(TNF) or IL-1 [26]. However, an analysis of
the various functions of defensins or other
antimicrobial peptides is difficult, because
different antimicrobial peptides and also
chemokines have overlapping functions;
hence, the effects of antimicrobial peptides
on the adaptive immune system remain to
be investigated.
3.4.1
KLKL
5
KLK: An Artificial CAMP
Small antimicrobial peptides, which con-
tain approximately 3540 amino acids are
still too large to be used as synthetically
produced therapeutics. Thus, smaller pep-
tides that are effective in fending off bacte-
ria were sought, whereupon the peptide
RSLCLLHCRLK-NH2, corresponding to
amino acid position 717 of the antimicro-
bial peptide sapecin B from Sacrophaga
peregrine, was shown to possess significant
antimicrobial properties. After modifica-
tion of the original peptide, two un-
decapeptides RLKLLLLLRLK-NH2 and
KLKLLLLLKLK-NH2 (also referred to as
KLKL
5
KLK) were obtained, with each
having strong antimicrobial properties.
Whereas the natural peptide sapecin B is
toxic only towards Gram-positive bacteria,
the synthetic peptide proved to be effective
in defense against both Gram-positive and
Gram-negative bacteria, indicating that
modifications of a natural sequence can
improve efficiency and/or substrate speci-
ficity [28]. Both synthetic peptides contain
a basic region at both termini, with a cen-
tral hydrophobic region of five leucine res-
idues, which were shown to be essential
for the function of the peptide. Further-
more, both synthetic peptides are effective
as antimicrobial treatment against S. aure-
us, E. coli, and Candida albicans, whereby
the peptide KLKL
5
KLK possesses the larg-
er activity.
Soon after the initial discovery of the
antimicrobial activity of these peptides, it
was shown that KLKL
5
KLK interacts with
phospholipids vesicles resembling the
composition of the E. coli membranes and
S. aureus, and that the peptides have no
significant hemolytic activities to bovine
erythrocytes, indicating that their primary
target is the bacterial membrane [28]. It is
thought that interaction of the peptides
with phospholipids in the bacterial mem-
brane is essential for subsequent disrup-
tion of the electrochemical membrane po-
tential. This results in the loss of the abil-
ity of ATP synthesis and proline uptake,
which is postulated as being the cause of
death of the bacteria [29]. It has further
been suggested that the peptides form
multiple layers around the bacterial mem-
brane, and that the ionic interaction of
KLKL
5
KLK and the bacterial membrane is
a prerequisite for this layer formation [29].
In addition, the internal leucines of the
peptide might be necessary in order to
form a channel through the outer and in-
ner membrane, thereby causing diffusion
of low molecular-weight substances [29].
By using KLKL
5
KLK and its derivatives,
it was shown for the first time that an arti-
ficial peptide could prevent infections with
S. aureus in mice [30]. In addition to their
direct antibacterial effects, it is thought
that the chemotherapeutical activity of
KLKL
5
KLK and its derivatives is due to the
activation of neutrophils. Experiments
showed that the activation of neutrophils
by KLKL
5
KLK can be inhibited by pertus-
sis toxin, implying that it is mediated
through a G-protein-coupled receptor [31].
Subsequently, small quantities of calreticu-
lin, a 60-kDa protein that binds to
KLKL
5
KLK, were shown to be present on
the surface of plasma membranes of neu-
trophils [31]. Although calreticulin was
known to be a Ca
2+
binding molecular
chaperone in the membrane of the endo-
plasmic reticulum required for MHC class
I antigen processing, nothing was known
of its function on the cell surface. How-
ever, since antibodies directed to the N- or
C-terminus of calreticulin inhibit the acti-
vation of neutrophils by KLKL
5
KLK, an
important role is implicated for calreticu-
lin in this process.
3.4.2
KLKL
5
KLK as Type 2-inducing Adjuvant
Although natural CAMPs are utilized by
the human immune system in the battle
against the invasion of many micro-organ-
isms, they have also been shown to pos-
sess properties suitable for adjuvant action.
The discovery of natural CAMPs has sub-
sequently led to the development of artifi-
cial antimicrobial peptides to treat infec-
tions. Recently, it was shown that the arti-
ficial CAMP KLKL
5
KLK has also the po-
tency to induce adaptive immune
responses against co-injected antigens [32].
Mice were vaccinated with the protein
ovalbumin as a model antigen in combina-
tion with KLKL
5
KLK, and the immune re-
sponse was examined after repeated im-
munizations. The humoral response was
analyzed in sera of vaccinated mice by de-
termining ovalbumin-specific antibody pro-
duction (total IgG and subtypes IgG1 and
IgG2). A strong induction of total IgG was
observed after two immunizations, and
significant antibody titers were still detect-
able at 34 months after the last injection.
The determination of IgG-subtypes re-
vealed that in the presence of KLKL
5
KLK,
only ovalbumin-specific IgG1 antibodies
are induced, thus indicating the induction
of a type 2 response (Fig. 3.7). A compari-
son with the adjuvant alum showed that
very similar levels of total IgG and IgG1
are induced by KLKL
5
KLK. Antigen-specif-
ic production of the type 2 cytokine IL-5
further supports the notion that
KLKL
5
KLK might constitute a potent adju-
vant inducing type 2 cellular and humoral
immune responses (Fig. 3.8).
This property of KLKL
5
KLK encouraged
further investigations to determine its
mode of action. As mentioned earlier, the
Fig. 3.7 KLKL
5
KLK induces antigen-specific humoral type 2
responses. The protein ovalbumin (OVA) was injected twice (on
days 0 and 28) alone, or in combination with KLKL
5
KLK or alum.
Serum samples were taken from mice at day 26 (prime) or day
54 (boost) and analyzed for IgG1 by ELISA.
Fig. 3.8 KLKL
5
KLK induces antigen-specific cellular type 2
responses. The protein ovalbumin (OVA) was injected twice (on
days 0 and 28) alone or in combination with KLKL
5
KLK or alum.
Splenocytes of vaccinated mice were restimulated ex vivo with
OVA to determine the specific production of IL-5 (type 2
cytokine).
uptake of proteins or peptides by APCs
and their presentation is commonly
thought to be an essential step for stimula-
tion of the adaptive immune response. In
initial experiments, the artificial CAMP,
KLKL
5
KLK, was tested for its capability to
enhance the association of ovalbumin to a
monocyte-macrophage cell line in vitro
[32]. It was found that KLKL
5
KLK promotes
association of the APCs with ovalbumin,
and that this enhancement was dependent
on the concentration of KLKL
5
KLK.
The ability of KLKL
5
KLK to maintain
the antigen at the injection site was also
tested using color-labeled compounds [32].
Mice were vaccinated subcutaneously, and
the distribution of labeled ovalbumin in-
jected either alone or together with alum
or KLKL
5
KLK was compared. Injection of
ovalbumin alone did not result in any de-
pot formation, and the labeled compound
was barely detectable at 3 hours after injec-
tion. However, the co-injection of ovalbu-
min and alum resulted in a prolonged de-
pot formation which was still visible at 30
days after injection. The combination of
ovalbumin and KLKL
5
KLK induced also a
strong depot formation at the injection site
for 5 days, but the staining was reduced so
that only traces were visible at 30 days
after injection. The use of labeled antigen
and adjuvant allows the analysis of their
distribution throughout the animal after
vaccination as a matter of time. After in-
jection of labeled ovalbumin together with
labeled KLKL
5
KLK, neither could be de-
tected in secondary lymphoid organs such
as the spleen or lymph nodes after 1, 5, or
14 days. KLKL
5
KLK was detected in the
kidneys shortly after injection (24 h), but
the level had fallen after only 2 weeks.
These results indicate that KLKL
5
KLK ex-
erts its function mainly at the injection
site, and that it is subsequently discarded
directly from the depot to the kidneys [32].
Thus, although KLK and poly-L-arginine
induce different types of adaptive immune
response, their mode of action as adju-
vants (enhanced uptake of antigens, depot
formation at injection site) appear similar,
indicating a common mechanism of cat-
ionic peptide delivery systems. However,
more detailed analyses are needed to con-
firm this hypothesis.
3.5
Cationic Peptide Delivery Systems
in Combination with Other Adjuvants
Based on the promising data obtained
using cationic peptides as adjuvant in the
sense of antigen delivery systems in the
context of vaccines, the question arose as
to whether these systems could also be
used in combination with other adjuvants.
As in earlier experiments cationic peptides
were used to transport DNA molecules
into cells, it was clear that poly-L-arginine
should be tested in combination with oli-
godeoxynucleotides containing CpG-motifs
(CpG-ODN), which were described to be
immunostimulatory substances on their
own, and to analyze the immunostimula-
tory effect of the combined adjuvants.
3.5.1
Poly-L-arginine in Combination
with CpG-ODN
Immune cells have the ability to recognize
conserved pathogen-associated molecular
patterns (PAMP) such as motif CpG, lipo-
polysaccharide, lipoproteins, double-strand-
ed RNA, or flagellin. In contrast to mam-
malian DNA, bacterial DNA contains fre-
quent motifs of unmethylated CpG dinu-
cleotides. The CpG DNA motifs are
capable of activating the innate immune
system, the first step in the defense
against micro-organism invasion of the
body. The motifs induce the activation of
macrophages, dendritic cells or neutro-
phils, which in turn promotes phagocyto-
sis and subsequent elimination of the
pathogen. Recent data suggest that the re-
cognition of PAMPs depends on the acti-
vation of TLRs [33]. To date, 10 TLRs have
been identified, and each receptor is cap-
able of recognizing distinct PAMPs, specif-
ic for certain micro-organisms, indicating
that the immune system recognizes and
identifies an infection via these TLRs. The
receptor molecules possess a cytoplasmic
TIR (Toll/IL-1R) homology domain which
associates with an adaptor molecule called
MyD88 [33, 34]. MyD88 is capable of stim-
ulating further steps in the signaling cas-
cade, which eventually leads to the activa-
tion of JNK and NF-jB. The MyD88-de-
pendent pathway is common to all TLRs
except TLR3 and TLR4, which both signal
through a MyD88-independent pathway
[33]. CpG DNA was shown to bind to
TLR9, and it has been reported that
MyD88 knockout mice do not respond to
CpG DNA [34]. The innate immune re-
sponse is often initiated by phagocytosis of
pathogens by macrophages. Intact bacteria
can be recognized by macrophage recep-
tors, which initiate a response localized at
the plasma membrane. Whereas the acti-
vation of immune cells can be mediated
by contact with the cell wall constituents
of intact micro-organisms, bacterial DNA
cannot be detected by the immune system
until it is liberated from the pathogen
cells. The TLR9 signaling pathway is ini-
tiated in endosomes when bacteria have
been processed, and therefore early endo-
cytosis the formation and maturation of
endosomes containing CpG DNA and
TLR9 receptor is a prerequisite for the
onset of the immune response triggered
by bacterial DNA [34, 35].
Short DNA stretches containing un-
methylated CpG motifs (CpG-ODN) like
bacterial DNA are known to be potent in-
ducers of type 1-like immune responses,
as indicated by the predominant produc-
tion of IL-12 and IFN-c, but also IL-6 and
TNF-a. Although CpG-ODN has been de-
scribed as a very powerful inducer of the
immune response, side effects caused by
the strong induction of systemic IL-6 and
TNF-a have been reported in rodents [36,
37].
CpG-ODN and poly-L-arginine are both
described as inducing type 1 immune re-
sponses, and were suggested to be power-
ful adjuvants, though such a combination
has not been tested until recently. Because
CpG-ODN and poly-L-arginine have oppo-
site charges, it is easy to imagine that they
interact through their electrostatic attrac-
tion. The electrostatic interaction of CpG-
ODN and poly-L-arginine was confirmed
by agarose gel electrophoresis, which
showed that they form stable complexes.
In addition, a vaccine mixture of ovalbu-
min (OVA)-derived peptide together with
poly-L-arginine and CpG-ODN or a combi-
nation of peptide with either poly-L-argi-
nine or CpG-ODN was compared for their
ability to induce a peptide-specific T cell
response. When the numbers of peptide-
specific IFN-c-producing cells were com-
pared at 4 days after injection in mice, the
results indicated that poly-L-arginine and
CpG-ODN, when combined, induced
strongly enhanced peptide-specific im-
mune responses compared to peptide ap-
plication with either of the immunomodu-
lators alone (Fig. 3.9).
The potency of the poly-L-arginine/CpG-
ODN combination was confirmed by the
induction of a strong immune response
against different peptides derived from
mouse tyrosinase-related protein-2, a
mouse mastocytoma P815 peptide, and a
bacterial peptide derived from listeriolysin
of L. monocytogenes after co-injection into
mice. In addition, it was shown that very
small amounts of poly-L-arginine and
CpG-ODN (up to 100-fold lower than the
individual components) were sufficient to
induce a strong immune response.
Furthermore, the response was seen to be
remarkably prolonged by the combined ad-
ministration of poly-L-arginine and CpG-
ODN, as high numbers of antigen-specific
T cells were observed for at least 372 days
after single injection of the vaccine.
CpG-ODN, despite being a powerful in-
ducer of a type 1 response, was also
shown to elicit some potentially harmful
adverse effects in rodents by inducing
high levels of systemic IL-6 and TNF-a. In-
deed, in the sera of vaccinated mice, CpG-
ODN induced high levels of both cyto-
kines, but no such induction was caused
by a poly-L-arginine/CpG-ODN combina-
tion. Thus, the potentially harmful sys-
temic release of pro-inflammatory cyto-
kines induced upon injection of CpG-ODN
in mice can be prevented by co-administra-
tion of poly-L-arginine. In contrast, the de-
pot effect at the injection site seen upon
application of poly-L-arginine was effi-
ciently prolonged (for at least 92 days)
when poly-L-arginine and CpG-ODN were
co-injected. The slow release of antigens
from the injection site and their uptake by
APCs leads ultimately to the priming of T
cells, and hence to a long-lasting immune
response. Such depot formation could also
explain the inhibition of the potentially
harmful release of TNF-a and IL-6; more-
over, it might also be the reason for the ef-
fective immune response induced by small
quantities of poly-L-arginine/CpG-ODN
when injected into mice.
Taking all of these results together, the
use of poly-L-arginine as a cationic peptide
delivery system, when combined with a
second adjuvant such as CpG-ODN, may
represent an improved vaccine strategy for
humans in order to induce antigen-specific
type 1 immune responses.
Fig. 3.9 Poly-L-arginine/CpG-ODN-based peptide
vaccines induce strong peptide-specific immune
responses. Mice were injected with the ovalbumin
(OVA)-derived peptide OVA
257-264
alone, or in
combination with poly-L-arginine, CpG-ODN,
or poly-L-arginine/CpG-ODN. At 4 days after injec-
tion, draining lymph node cells were analyzed for
peptide-specific IFN-c production. The melanoma-
derived peptide TRP-2
181-188
was used as a control
for restimulation.
3.6
The Development of IC31 and Future
Prospects
Based on the concept of using a cationic
peptide delivery system in combination
with an immunostimulatory synthetic
DNA sequence for improved vaccination
strategies, we are currently developing a
novel adjuvant, called IC31. IC31 repre-
sents the combination of the artificial
CAMP KLKL
5
KLK, which is described as
inducing type 2 responses, and a novel im-
munostimulatory oligodeoxynucleotide,
called ODN1a and as derivative of poly
I : C based on repeats of desoxy-inosine/
desoxy-cytosine [38]. Poly I : C, a double-
stranded RNA molecule has been shown
to induce preferentially a type 1 immune
response by stimulating macrophages to
produce IL-12 and IFN-a; in addition, it in-
duces the maturation of in vitro-cultured
dendritic cells [39, 40]. Poly I : C has also
been shown to have protective effects in a
number of animals against several differ-
ent DNA and RNA viruses such as HSV,
rabies, and encephalomyocarditis virus,
and has also been used in several clinical
trials as an immunomodulator, which
caused no or little toxicity [41, 42]. How-
ever, since poly I : C is relatively unstable
and its length is very difficult to standard-
ize, we developed a stable and non-toxic
ODN of defined length containing deoxy-
inosine/deoxy-cytosine repeats (ODN1a).
We assumed that IC31 that is, the com-
bination of KLKL
5
KLK and ODN1a would
result in an effective adjuvant, and we are
currently investigating the potency of IC31
on the induction of adaptive immune re-
sponses. Initial results indicate that pep-
tide-specific type 1 cellular immune re-
sponses and protein-specific mixed type 1/
type 2 cellular and humoral immune re-
sponses are induced. This implies that
IC31 might be used successfully in the bat-
tle against extracellular pathogens, as well
as against intracellular pathogens or tumors
a property that is not fulfilled by other ad-
juvants such as alum or MF59.
3.7
Conclusions
Traditional vaccines have been used suc-
cessfully to reduce drastically disease and
mortality worldwide, or to eradicate an in-
fectious agent altogether. While the first
vaccines to be developed consisted of
whole microbes, there is a clear trend to-
wards the development of more defined
vaccines consisting of distinct antigens de-
rived from the individual pathogens. This
has been caused mainly by the limitations
and adverse effects of traditional vaccines
when applied to the prevention or treat-
ment of distinct infectious pathogens, and
became possible only as a result of ad-
vances made in the development of novel
biotechnological approaches and an under-
standing of the virulence mechanisms of
the relevant pathogenic organisms. Mod-
ern, defined vaccines contain specific anti-
gens of a pathogen such as recombinant
proteins or short peptides, which have
been shown to induce protective immune
responses. Although short antigenic pep-
tides have been used with success for im-
munization eliciting protective immune re-
sponses, it has emerged that peptide anti-
gens are by themselves not very immuno-
genic and require the help of adjuvants.
There are, at present, very few adjuvants li-
censed for human vaccination, and these
are still limited in their applicability, as
they preferentially support the induction
of either a type 1 or a type 2 immune re-
sponse. In addition, their safety profile can
be further improved, and the advantages
of combined adjuvant properties (e.g., de-
pot formation, enhanced uptake of anti-
gens by APCs, etc.) are not supported by
all of them.
Thus, we developed novel adjuvants
based on compounds of the cationic pep-
tide delivery system. Initially, we examined
the properties of poly-L-arginine as well as
KLKL
5
KLK, and determined that both cat-
ionic peptides can efficiently charge APCs
with antigen, thereby inducing a specific
and strong type 1 or type 2 immune re-
sponse, respectively, upon co-injection with
specific antigens. Moreover, poly-L-argi-
nine and KLKL
5
KLK were shown to
strongly induce the formation of a depot
at the injection site, which prolongs the
availability of antigens in the body, thereby
leading to improved immune responses.
All of the pre-clinical data obtained lent
support to the use of poly-L-arginine and
KLKL
5
KLK as promising adjuvants in hu-
man vaccines.
Poly-L-arginine has already been evalu-
ated in clinical trials, with results showing
it to be safe for human vaccination, as no
adverse effects have been observed. These
clinical trials have also provided evidence
that a vaccine consisting of poly-L-arginine
and HCV-specific peptides has the poten-
tial of being used as a therapeutic vaccine,
as chronic HCV patients were vaccinated.
Therefore, poly-L-arginine represents a
promising adjuvant which not only sup-
ports the induction of immune responses
for therapeutic vaccines, thereby reducing
the burden imposed by or cure a disease,
but also serving as a prophylactic vaccine
to prevent the development of disease fol-
lowing infection by a pathogen.
Based on our success in developing cat-
ionic peptides as adjuvants, we have
shown that their combination with immu-
nostimulatory DNA molecules leads to
powerful novel adjuvants that fulfill all re-
quirements for the induction of improved
immune responses. IC31 a combination
of the antimicrobial peptide KLKL
5
KLK
and an oligodeoxynucleotide containing
desoxy-inosine/desoxy-cytosine awaits ex-
amination in clinical trials, with the expec-
tation that it will induce stronger and
more effective immune responses against
a variety of pathogens. Clearly, IC31 holds
the promise that existing vaccines may be
improved by their combination with this
novel adjuvant, or that diseases for which
the development of vaccines has until now
been unsuccessful, might be prevented or
treated.
References
1 Dennehy, P. H. (2001) Active Immunization in
the United States: Developments over the past
decade. Clin. Microbiol. Rev. 14, 872908.
2 Rubin L. G. (2000) Pneumococcal vaccine. Pe-
diatr. Clin. North Am. 47, 269285.
3 Koelle, D. M. and Lawrence, C. (2003) Recent
Progress in Herpes Simplex Virus Immuno-
biology and Vaccine Research. Clin. Microbiol.
Rev. 16, 96113.
4 Slovin, S. F., Ragupathi, G., Adluri, S., Un-
gers, G., Terry, K., Kim, K., Spassova, M.,
Bornmann, W. G., Fazzari, M., Dantis, L., Ol-
kiewicz, K., Lloyd, K. O., Livingston, P. O., Da-
nishefsky, S. J. and Scher, I. H. (1999) Carbo-
hydrate vaccines in cancer: Immunogenicity
of a fully synthetic globo H hexasaccharide
conjugate in man. Proc. Natl. Acad. Sci. USA
96, 57105715.
5 Correa, I. and Plunkett, T. (2001) Update on
HER-2 as a target for cancer therapy: HER2/
neu peptides as tumour vaccines for T cell rec-
ognition. Breast Cancer Res. 3, 399403.
6 Malin, D., Alvarado, C. L., Woodhouse, K. S.,
Karp, H., Urdiales, E., Lay, D., Appleby, P.,
Moon, W. D., Ennifar, S., Basham, L. and Fat-
tom, A. (2002) Passive Immunization against
nicotine attenuates nicotine discrimination.
Life Sci. 70, 27932798.
7 Gupta, R. K. and Siber, G. R (1995) Adjuvants
for human vaccines current status, problems
and future prospects. Vaccine 13, 12631276.
References 1441
8 Schijns, V. E. J. C. (2003) Mechanisms of vac-
cine adjuvant activity: initiation and regulation
of immune response by vaccine adjuvants.
Vaccine 21, 829831.
9 Flarend, R. E., Hem, S. L, White, J. L., Elmore,
D., Suckow, M. A., Rudy, A. C. and Dandashli,
E. A. (1997) In vivo absorption of aluminium-
containing vaccine adjuvants using
26
Al. Vac-
cine 15, 13141318.
10 Zauner, W., Lingnau, K., Mattner, F. von Ga-
bain, A. and Buschle, M. (2001) Defined syn-
thetic vaccines. Biol. Chem. 382, 581595.
11 Valensi, J. P., Carlson, J. R. and Van Nest,
G. A. (1994) Systemic cytokine profiles in
BALB/c mice immunized with trivalent influ-
enza vaccine containing MF59 oil emulsion
and other advanced adjuvants. J. Immunol.
153, 40294039.
12 Wagner, E., Zenke, M., Cotton, M., Beug, H.
and Birnstiel, M. L. (1990) Transferrin-poly-
cation conjugates as carriers for DNA uptake
into cells. Proc. Natl. Acad. Sci. USA 94,
32623267.
13 Shen, W. C. and Ryser, H. J. (1978) Conjuga-
tion of poly-L-lysine to albumin and horserad-
ish peroxidase: a novel method of enhancing
the cellular uptake of proteins. Proc. Natl.
Acad. Sci. USA 75, 18721876.
14 Shen, W. C. and Ryser H. J. (1981) Poly(L-ly-
sine) has different membrane transport and
drug-carrier properties when complexed with
heparin. Proc. Natl. Acad. Sci. USA 78, 7589
7593.
15 Buschle, M., Schmidt , W., Zauner, W.,
Mechtler, K., Trska, B., Kirlappos, H. and
Birnstiel, M. L. (1997) Transloading of tumor
antigen-derived peptides into antigen-present-
ing cells. Proc. Natl. Acad. Sci. USA 94, 3256
3261.
16 Mattner, F., Fleitmann, J.-K., Lingnau, K.,
Schmidt, W., Egyed, A., Fritz, J., Zauner, W.,
Wittmann, B., Gorny, I., Berger, M., Kirlap-
pos, H., Otava, A., Birnstiel, M. L. and
Buschle, M. (2002) Vaccination with poly-L-ar-
ginine as immunostimulant for peptide vac-
cines: induction of potent and long-lasting T
cell responses against cancer antigens. Cancer
Res. 62, 14771480.
17 Mitchell, D. J., Kim, D. T., Steinman, L., Fath-
man, C. G. and Rothbard, J. B. (2000) Polyargi-
nine enters cells more efficiently than other
polycationic homopolymers. J. Peptide Res.
56, 318325.
18 Wender, P. A., Mitchell, D. J., Pattabiraman,
K., Pelkey, E. T., Steinman, L. and Rothbard,
J. B. (2000) The design, synthesis, and evalua-
tion of molecules that enable or enhance cel-
lular uptake: Peptoid molecular transporters.
19 Fuchs, S. M. and Raines, R. T. (2004) Pathway
for polyarginine entry into mammalian cells.
20 Buschle, M., Schmidt, W., Berger, M., Schaff-
ner, G., Kurzbauer, R., Killisch, I., Tiedemann,
J.-K., Trska, B., Kirlappos, H., Mechtler, K.,
Schilcher, F., Gabler, C. and M. L. Birnstiel.
(1998) Chemically defined, cell-free cancer
vaccines: use of tumor antigen-derived pep-
tides or polyepitope proteins for vaccination.
Gene Ther. Mol. Biol. 1, 309322.
21 Schmidt, W., Schweighofer, T., Herbst, E.,
Maas, G., Berger, M., Schilcher, F., Schaffner,
G. and Birnstiel, M. L. (1995) Cancer vaccines:
The interleukin 2 dosage effect. Proc. Natl.
Acad. Sci. USA 92, 47114714.
22 Dye, C., Scheele, S., Dolin, P., Pathania, V.,
and Raviglione, M. C. (1999) Consensus state-
ment. Global burden of tuberculosis: esti-
mated incidence, prevalence, and mortality by
country. WHO Global Surveillance and Moni-
toring Project. JAMA 282, 677686.
23 Colditz, G. A., Brewer, T. F. , Berkey, C. S., Wil-
son, M. E., Burdick, E., Fineberg, H. V. and
Mosteller, F. (1994) Efficacy of BCG vaccine in
the prevention of tuberculosis. Meta-analysis
of the published literature. JAMA 271, 698
702.
24 Zasloff, M. (2002) Antimicrobial peptides of
multicellular organisms. Nature 1, 389395.
25 Van tHof, W., Veerman, E. C. I., Helmerhorst,
E. J. and Amerongen, A. V. N. (2001) Antimi-
crobial peptides: properties and applicability.
Biol. Chem. 382, 597619.
26 Yang, D., Biragyn, A., Kwak, L. W. and Oppen-
heim, J. J. (2002) Mammalian defensins in im-
munity: more than just microbicidal. Trends
Immunol. 23, 291296.
27 Brotz, H., Josten, M., Wiedemann, I., Schnei-
der, U., Gotz, F., Bierbaum, G. and Sahl,
H.G. (1998) Role of lipid-bound peptidoglycan
precursors in the formation of pores by nisin,
epidermin and other lantibiotics. Mol. Micro-
biol. 30, 317327.
28 Alvarez-Bravo, J., Kurata, S. and Natori, S.
(1994) Novel synthetic antimicrobial peptides
effective against methicillin-resistant Staphylo-
coccus aureus. Biochem. J. 302, 535538.
29 Alvarez-Bravo, J., Kurata, S. and Natori, S.
(1995) Mode of action of an antibacterial pep-
tide, KLKLLLLLKLK-NH2. Biochem. J. 117,
13121316.
30 Nakajima, Y., Alvarez-Bravo, J., Cho, J.-H.,
Homma, K.-I., Kanegasaki, S. and Natori, S.
(1997) Chemotherapeutic activity of synthetic
antimicrobial peptides: correlation between
chemotherapeutic activity and neutrophil-acti-
vating activity. FEBS Lett. 415, 6466.
31 Cho, J.-H., Homma, K.-I., Kanegasaki, S. and
Natori, S. (1999) Activation of human neutro-
phils by a synthetic anti-microbial peptide,
KLKLLLLLKLK-Nh2, via cell surface calreticu-
lin. Eur. J. Chem. 266, 878885.
32 Fritz, J. H., Brunner, S., Birnstiel, M. L.,
Buschle, M., Gabain, A., Mattner, F. and Zau-
ner, W. (2004) The antimicrobial peptide
KLKLLLLLKLK induces adaptive immunity to
co-injected antigens. Vaccine 22, 32743284.
33 Takeda, K. and Akira, S. (2004) TLR signaling
pathways. Semin. Immunol. 16, 39.
34 Takeshita, F., Gursel, I., Ishii, K. J., Suzuki, K.,
Gursel, M. and Klinman, D. M. (2004) Signal
transduction pathways mediated by the inter-
action of CpG DNA with Toll-like receptor 9.
Semin. Immunol. 16, 1722.
35 Ahmad-Nejad, P., Hcker, H., Rutz, M.,
Bauer, S., Vabulas, R.M. and Wagner, H.
(2002) Bacterial CpG-DNA and lipopolysac-
charides activate Toll-like receptors at distinct
cellular compartments. Eur. J. Immunol. 32,
19581968.
36 Sparwasser, T., Miethke, T., Lipford, G.,
Borschert, K., Hacker, H., Heeg, K., Wagner,
H. (1997) Bacterial DNA causes septic shock.
Nature 386, 336337.
37 Sparwasser, T., Miethke, T., Lipford, G., Erd-
mann, A., Hacker, H., Heeg, K., Wagner, H.
(1997) Macrophages sense pathogens via DNA
motifs: induction of tumor necrosis factor-al-
pha-mediated shock. Eur. J. Immunol. 27,
16711679.
38 Schellack, C., Prinz, K., Egyed, A., Fritz, J.,
Wittmann, B., Ginzler, M., Swatosch, G., Zau-
ner, W., Kast, C., Akira, S., Gabain, A. V.,
Buschle, M., Lingnau, K. (submitted) IC31, a
novel adjuvant signaling via TLR9, induces
potent cellular and humoral immune re-
sponses.
39 Manetti, R., Annunziato, F., Tomasevic, L.,
Gianno, V., Parronchi, P., Romagnani, S. and
Maggi, E. (1995) Polyinosinic acid: polycy-
tidylic acid promotes T helper type 1-specific
immune responses by stimulating macro-
phage production of interferon-alpha and in-
terleukin-12. Eur. J. Immunol. 25, 26562660.
40 Cella, M., Salio, M., Sakakibara, Y., Langen,
H., Julkunen, I. and Lanzavecchia, A. (1999)
Maturation, activation, and protection of den-
dritic cells induced by double-stranded RNA.
J. Exp. Med. 189, 821829.
41 Simmler, M. C., Bruley-Rosset, M., Belpomme,
D. and Schwarzenberg, L. (1977) Clinical trial
of poly I-poly C as an immunity adjuvant and
an immunorestoration agent. Eur. J. Cancer
13, 463467.
42 Guggenheim, M. A. and Baron, S. (1977) Clin-
ical studies of an interferon inducer, polyribo-
inosinic-polyribocytidylic acid [poly (I)-poly
(C)], in children. J. Infect. Dis. 136, 5058.
References 1443
Abstract
The current advocacy of intensive insulin
therapy regimens involving multiple daily
subcutaneous injections places a heavy bur-
den of compliance on patients, and has
prompted interest in developing alternative,
less invasive routes of delivery. Various ef-
forts have been made to develop alternative
methods for administering insulin, among
which is the RapidMist
TM
Diabetes Man-
agement System. This is based on a proprie-
tary formulation technology that allows a
liquid, aerosolized pharmaceutical formula-
tion to be delivered accurately into the
mouth of the patient as a spray. This intro-
duces a high-velocity, fine-particle aerosol
(Oralin
TM
) into the patients mouth, thus in-
ducing a markedly increased deposition of
the preparation over the mucosal mem-
brane, the deposition being much larger
than occurs with conventional technology.
This rapidly moving, fine particle aerosol
is able to traverse the thin membrane so
that the insulin molecules are absorbed rap-
idly into the bloodstream (aided by absorp-
tion enhancers) and reach the peripheral
circulation within 10 min of application.
Studies conducted in patients with type 1
and type 2 diabetes showed clearly that
Oralin
TM
has a more rapid absorption and
metabolic control than subcutaneously in-
jected insulin. This novel, pain-free, oral in-
sulin formulation has notable attributes of
rapid absorption, a simple (user-friendly)
administration technique, precise dose con-
trol (comparable to injection within one
unit), and bolus delivery of drug. A simpli-
fied approach to pain-free prandial insulin
delivery offered by this technique will sig-
nificantly reduce the incidence of key com-
plications by allowing increased patient
compliance with consistent drug adminis-
tration to regulate patients blood glucose
levels. This chapter describes the recent re-
sults of clinical studies (in type 1 and type 2
diabetes patients) by comparing the efficacy
of Oralin
TM
with subcutaneously injected
insulin.
Abbreviations
FPG fasting plasma glucose
GADA GAD antibody
HbA
1c
glycated hemoglobin
ICA islet cell antibody
MDI metered dose inhaler
OHAs oral hypoglycemic agents
Oralin
TM
oral spray insulin
PPGI post-prandial glucose increment
RIA radioimmunoassay
UKPDS United Kingdom Prospective
Diabetes Study
1445
4
The Evolving Role of Oralin
TM
(Oral Spray Insulin) in the Treatment
of Diabetes using a Novel RapidMist
TM
Diabetes Management System
Pankaj Modi
ISBN: 3-527-31184-X
4.1
Introduction
Worldwide, the prevalence of diabetes is
increasing, with the US Centers for Dis-
ease Control and Prevention referring to
the condition as . . . the epidemic of our
times. According to the World Health Or-
ganization, the present number of dia-
betics worldwide is 135 million, and this is
expected to increase to 300 million by
2025.
The Diabetes Control and Complications
Trial (DCCT) and United Kingdom Pro-
spective Diabetes Study (UKPDS) demon-
strated incontrovertibly in individuals with
type 1 and type 2 diabetes that improve-
ment in glycemic control decreases the
risk of long-term microvascular complica-
tions and greatly improve diabetics quality
of life. It was also demonstrated that in-
tensive glycemic control reduced the over-
all risk of diabetic eye disease, kidney
damage, stroke, and mortality [14]. The
knowledge that intensive therapy in this
population is both safe and efficacious in
reducing the incidence of key complica-
tions is critically important to the manage-
ment of diabetes. However, the best way to
achieve tight glycemic control is not clear.
Although diet and exercise can improve
glycemic control early in the course of the
disease, oral medications often become the
mainstay of type 2 diabetes treatment [5
8]. In type 2 diabetes, metabolic control de-
teriorates in most patients when the dura-
tion of diabetes increases; such deteriora-
tion is most likely explained by a decrease
in insulin secretion. Thus, significant
number of patients with type 2 diabetes
cannot achieve tight glycemic control with
oral agents, and so need to be treated with
insulin either as a single agent or added
to an oral regimen [912]. The addition of
insulin therapy has been shown to result
in a significant decrease in fasting plasma
glucose (FPG) and HbA1c values, as well
as reduced insulin requirements [1219].
It is a well-known fact that many subjects
dislike needles, and often refuse to accept
injection therapy, thereby affecting their
compliance with insulin therapy. Patients
treated with two or more oral hypoglyce-
mic agents are subjected to an additive
risk of adverse events, and dose adjust-
ments may become complex when multi-
ple drugs are used. Taken together, the
above-mentioned facts provide a strong sci-
entific basis for incorporating the use of
insulin, as a routine practice, in the thera-
py of type 2 diabetes mellitus. On the
other hand, daily practice observed the re-
luctance of a vast majority of patients
affected with this disease to incorporate
frequent daily injections, and this has
prompted the search for non-invasive
methods of administration of insulin.
4.2
Rationale for Oralin
TM
Development
4.2.1
Oralin
TM
Delivery
TM
System
The search for an oral form of insulin has
been under way since Banting and Bests
original discovery of insulin. Oral insulin
would not only free diabetic patients from
some of the daily painful and inconvenient
injections, but would also provide a more
physiological route of administration.
The oral mucosa provides a near-ideal,
non-invasive portal of entry into the sys-
temic circulation, on the basis of four
main reasons. First, the oral cavity is rela-
tively permeable. Second, the oral mucosa
has a very rich blood supply, with many
superficial blood vessels, and this makes it
TM
(Oral Spray Insulin) in the Treatment of Diabetes 1446
a great access point to systemic circulation.
Third, the oral mucosa is a very robust
area, which shows short recovery times
after stress or damage. Fourth, the oral
mucosa offers an attractive surface area for
drug delivery. The low permeability, rich
blood supply, suitable and attractive sur-
face area and robustness of the oral cavity
provide for a very attractive route of ad-
ministration for systemic drug delivery.
When all of these excellent drug delivery
features are combined with the preference
of both patients and physicians for the oral
route as a delivery method, the buccal cavi-
ty becomes the ideal route for the adminis-
tration of insulin.
4.2.2
Development of the RapidMist
TM
Diabetes
Management System
A variety of efforts have been made to de-
velop such alternative methods for insulin
administration, among which is included
the RapidMist
TM
Diabetes Management
System (Fig. 4.1). The Advanced Rapid-
Mist
TM
System is defined as having a criti-
cal series of attributes: fast access to the
circulatory system; precise dosing control;
simple, self-administration procedure; and
bolus delivery of drug.
This system is based on a proprietary
formulation technology, which allows a liq-
uid pharmaceutical formulation to be de-
livered accurately into the mouth of the
patient via an aerosolized spray. This sys-
tem introduces a high-velocity, fine-particle
aerosol into the patients mouth, therefore
inducing a markedly increased deposition
of the preparation over the regional muco-
sa a deposition that is much larger than
that observed with conventional technol-
ogy. It is a well-known fact that the thin
oral membranes contain many superficial
blood vessels, and guards the ample sur-
face area in direct contact with the circula-
tion. Thus, a fast-moving, fine-particle
aerosol is able to traverse this thin mem-
brane (the droplets impact at speeds of
3545 m s
1
). When the insulin molecules
have penetrated these superficial thin
layers, they are rapidly absorbed into the
bloodstream (aided by absorption enhanc-
ers), and appear in the peripheral circula-
tion within 10 minutes of application. The
proposed mechanism of absorption of Ora-
lin
TM
is shown in Fig. 4.2.
4.3
The Benefits of Oralin
TM
Oralin
TM
provides a range of benefits for
the patient:
Needle-free and pain-free therapy: inten-
sive diabetes therapy requires at least
three to four injections per day. Oralin
TM
avoids the use of needles.
Rapid absorption: Oralin
TM
is absorbed
into the bloodstream faster than injected
insulin.
TM
1447
Fig. 4.1 The Oralin RapidMist
TM
delivery device.
Stability: Oralin
TM
is stable at room tem-
perature, and does not require refrigera-
tion.
Higher compliance: needle-free, pain-
free insulin therapy should increase pa-
tient compliance.
Quality of life: the small size of the de-
vice means that it is convenient to carry
anywhere, and to use comfortably in
public. The reduction in dosing time be-
fore a meal offers a more flexible life-
style. The improved compliance im-
proves the condition, which in turn
leads to a better quality of life.
This chapter focuses on the recent suc-
cesses achieved during development of the
Oralin RapidMist
TM
Diabetes Management
System in various clinical trials, and its fu-
ture potential as meal insulin (replace-
ment of subcutaneous injections) in pa-
tients with type 1 and type 2 diabetes.
4.4
The Preparation
and Pharmaceutical Properties of Oralin
TM
Oralin
TM
is prepared by dissolving the reg-
ular-acting human insulin crystals in water
at neutral pH, between 7.37.6. To this so-
lution are added the other ingredients,
which include glycerin, phenols, and stabi-
lizers to improve stability and facilitate
room-temperature storage, and absorption
enhancers to aid absorption through the
oral mucosa. The solution is mixed thor-
oughly and pH re-adjusted if necessary.
All components of the formulation are
FDA-approved chemicals for human con-
sumption and pharmaceutical use. The re-
sultant solution is then placed in an ano-
dized canister fitted with a proprietary me-
tered dose valve and charged with the non-
CFC propellant HFA-134a, using specially
designed aerosol equipment. The end
product is an aerosolized aqueous insulin
solution which is delivered via the Rapid-
Mist
TM
[modified metered dose inhaler
(MDI)] device.
Oralin
TM
is a tasteless, colorless, liquid
aerosol mist that does not cause any irrita-
tion, burning or discomfort in the mouth
TM
Fig. 4.2 Oralin
TM
proposed mechanism of absorption.
after repeated administrations. The formu-
lation is rapidly absorbed in the mouth
within 510 minutes of application, and its
onset of action is faster than that of subcu-
taneously injected insulin (e.g., Humu-
lin
). The insulin dose per puff (spray)

can be controlled either by adjusting the
insulin concentration in the formulation
or by the metered dose valve chamber.
4.4.1
Uniformity of Dose Delivery through
the RapidMist
TM
Device
The treatment of diabetes requires precise
dosing of insulin in order to avoid large
fluctuations in glucose levels after meals
and throughout the day. The RapidMist
TM
device is capable of delivering the precise
dose (comparable to injection within one
unit) required by diabetic patients. The
uniformity of dose delivery of the Rapid-
Mist
TM
device was determined by monitor-
ing the insulin content of individual puffs,
using a standard HPLC assay. The solution
from the device was sprayed into a fixed
volume of analysis buffer solution at a low
pH, the insulin was allowed to dissolve
with mixing and stirring of the flask con-
tents, and a fixed volume (50 lL) of solu-
tion was injected onto the HPLC column.
The vial was weighed before and after
TM
1449
Fig. 4.3 Uniformity of dose delivery through the RapidMist
TM
device.
each puff to ensure that puffs were deliv-
ered consistently. Analysis of each puff
(dose) showed quantitatively that the de-
vice is capable of delivering the required
dose, in a very precise manner, throughout
the life of the vial (Fig. 4.3).
4.4.2
Evidence of Buccal Absorption:
Tc99 Radio-label In-vivo Human Study
In order to establish, beyond reasonable
doubt, that quantitative deposition and ab-
sorption of Oralin
TM
occurs via the mouth
mucosa and not in the lungs, a gamma-
scintigraphy study was conducted by CRO,
Pharmaceutical Profile Ltd, using a radio-
labeled (Tc99) Oralin
TM
formulation. The
study was open-label, non-randomized in
format, and comprised seven healthy hu-
man volunteers. The Oralin
TM
device con-
sisted of a modified MDI device, modified
to deliver the radiolabeled insulin into the
patients mouth. The Tc99 Oralin
TM
for-
mulation was charged into the Rapid-
Mist
TM
device using proprietary equip-
ment. A simple procedure was followed to
administer the Tc99-labeled Oralin
TM
dose:
subjects were asked to position the device
in the mouth and to spray the Oralin
TM
formulation by depressing the device once.
Subjects were asked not to exhale or
breathe for 5 s in order to keep the dose in
the mouth, without expelling the mist
from the mouth during exhalation. The
procedure was repeated for the next dose.
At 5 min after the spray, subjects were
photographed with the gamma camera to
quantify distribution of the formulation in
the mouth, oropharynx, esophagus, stom-
ach, and lungs. As expected, no lung de-
position was observed; rather, most deposi-
tion occurred in the mouth, oropharynx,
and gastrointestinal areas (Fig. 4.4).
4.4.3
Formulation Safety: Toxicology Studies
The aim of this study was to examine and
report changes in the oral cavity epithelial
cell cytology (histopathological changes)
during 24-month chronic administration
of Oralin
TM
spray formulation, compared
to a placebo formulation. The secondary
objective was to examine hematological
changes and any hepatotoxicity.
This was a double-blind, parallel group
study conducted in 40 healthy beagle dogs
of mixed gender, age range 10 months to
2 years, and body weight 2025 kg. The
dogs were acclimatized for 2 weeks, and
fed a normal diet, three to four times daily.
Before randomization to the study, the dogs
were examined for their general well-being,
including a buccal mucosa examination, by
conducting a mouth biopsy of four different
sites (e.g., cheeks, upper and lower mouth
sections around the tongue, and under the
TM
Fig. 4.4 Gamma scintigraphy radiolabeled Tc99
study of Oralin
TM
deposition. Note deposition in
the mouth and gastrointestinal tract, but not in
the lungs.
tongue and the pallet). The electrocardio-
gam, vital signs, and a complete biochem-
ical profile were performed (including se-
rum electrolytes, liver function, hematology,
hemoglobin, blood counts and the kidney
function along with urinalysis).
The Oralin
TM
formulation was adminis-
tered three times daily before feeding in
40 dogs, and compared to a placebo for-
mulation spray (saline only) over a 24-
month period. In order to avoid low blood
sugar levels, the dogs were allowed to eat
10 minutes after receiving the Oralin
TM
spray or placebo.
The dogs were examined every 2 weeks
for their general well-being, and every 2
months for cytologic (histopathologic, i.e.,
epithelial scrapings) examination of the
oral cavity (buccal mucosa). A complete
blood count and biochemical profile (se-
rum electrolytes, EKG, arterial blood pres-
sure, urinalysis, liver and kidney func-
tions) were repeated every 6 months. After
24 months, the dogs remained in good
health, all having readily accepted daily
Oralin
TM
administration (three puffs, giv-
en three times daily) without evidence of
aversion or adverse effects. There were no
noticeable changes in their body weights
there was neither significant weight loss
nor gain and there were no changes in
biochemical profiles or liver and kidney
function.
Visual inspection of the oral cavities
(mouth) of the dogs did not reveal any evi-
dence of lesions, redness or changes in
the mucosal linings attributable to Ora-
lin
TM
use. Epithelial scrapings from each
site, for evaluation of the oral cavity, did
not reveal any evidence supportive of toxic
injury that is, there was no evidence of
epithelial structural changes or changes in
cell structure or mucosal linings.
Thus, Oralin
TM
oral spray formulation,
in comparison with placebo, was found to
be safe for chronic oral administration
over an extended time period.
4.4.4
Dose-ranging Euglycemic Clamp Study
in Subjects with Type 1 Diabetes
The primary aim of this study was to eval-
uate the dose-ranging effects, and pharma-
cokinetic and pharmacodynamic properties
of the Oralin
TM
formulation.
This was a single-center, randomized,
four-way, open-label, crossover comparison
of three different doses of Oralin
TM
on
three study days (five, ten and 20 puffs),
and on one occasion subcutaneous injection
of 0.1 unit kg
1
in 11 male or female sub-
jects with type 1 diabetes. The inclusion cri-
teria included: a FBG of 72180 mg dL
1
during the screening process, an absence
of other clinical anomalies except from
those derived from their metabolic condi-
tion that were relatively minor; a physical
examination without reasonably major clin-
ical abnormalities; a normal EKG; HbA1c
11% (normal range 4.56.2%); and body
mass index (BMI) <28 kg m
2
.
All subjects were assessed over four test
periods, 3 to 14 days apart, to determine
dose-ranging effects and pharmacody-
namics by euglycemic clamp, and insulin
pharmacokinetics over 6 hours. Oralin
TM
administration on the three study days
was in randomized order. Each subject
was provided with the placebo device to
practice the administration technique as
taught by a video demonstration. The Ora-
lin
TM
device comprises a standard MDI de-
vice with certain proprietary modifications
to deliver the insulin into patients mouth.
A simple procedure was used for Oralin
TM
dosing: patients were asked to position the
device in their mouth and to spray the Ora-
lin
TM
formulation by depressing the device
once. Patients were asked not to exhale or
TM
1451
breathe for 5 s in order to retain the dose in
the mouth and not to expel the mist during
exhalation. Patients were asked to repeat the
procedure for the next dose, with each re-
ceiving the following treatments, in random
order, and 37 days apart:
Treatment 1: Bolus dose of Oralin
TM
,
five puffs at time t =0 min.
Treatment 2: Bolus dose, 10 puffs at
time t =0 min.
Treatment 3: Bolus dose, 20 puffs at
time t =0 min.
At different doses (5, 10 and 20 puffs),
compared to the subcutaneous injection,
Oralin
TM
had an earlier onset of action
(29.4611.01 versus 84.247.4 min), an
earlier peak (44.6111.44 versus 134
44 min), and a shorter duration of action
(67.5318.2 versus 282.891.8 min). The
maximal effect of Oralin
TM
(20 puffs) was
comparable with that of subcutaneous in-
sulin. A doseresponse relationship was
demonstrated by an increase in maximal
glucose infusion rates (GIR
max
: 0.800.69,
2.181.33, and 5.192.51 mg kg
1
min
1
)
and in AUC
0120
(238.11145.1, 298.21
223.7, and 415.01219.9) with five, 10,
and 20 puffs, respectively.
The time to achieve maximum serum
insulin concentration was shorter for Ora-
lin
TM
than for the subcutaneous injection
(T
max
: 25.669.9 versus 18192 min, re-
spectively, p<0.05). Maximum insulin
levels were comparable with those after
subcutaneous injection (51.521.8 versus
55.341.8 lU mL
1
; p=NS). The AUC
0
120
and maximum insulin levels (12.69.6,
17.411.2, and 55.341.8 lU mL
1
, re-
spectively, p<0.01) for five, 10 and 20 Ora-
lin
TM
puffs, proved the doseresponse rela-
tionship for the spray insulin (see Figs. 4.5
and 4.6).
TM
Fig. 4.5 Dose-ranging study in type 1 diabetes patients. Mean glucose
infusion rate (GIR; mg kg
1
min
1
) after subcutaneous injection
(0.1 unit kg
1
) and 5, 10, or 20 puffs of Oralin
TM
spray.
TM
1453
F
i
g
.
4
.
6
D
o
s
e
-
r
a
n
g
i
n
g
s
t
u
d
y
i
n
t
y
p
e
1
d
i
a
b
e
t
e
s
p
a
t
i
e
n
t
s
.
M
e
a
n
s
e
r
u
m
i
n
s
u
l
i
n
l
e
v
e
l
s
(
l
U
m
L
1
)
a
f
t
e
r
s
u
b
c
u
t
a
n
e
o
u
s
i
n
j
e
c
t
i
o
n
(
0
.
1
u
n
i
t
k
g
1
)
a
n
d
5
,
1
0
,
o
r
2
0
p
u
f
f
s
o
f
O
r
a
l
i
n
T
M
s
p
r
a
y
.
4.4.5
Oralin
TM
as Meal Insulin in Treatment
of Type 1 Diabetes
This study was designed to compare the
efficacy of Oralin
TM
spray in subjects with
type 1 diabetes implanted with a Minimed
CSII pump to control post-prandial glu-
cose levels.
This was an open-label, randomized,
comparative study in 11 patients (males
and females, mean age 36 years, and BMI
<28 kg m
2
) with type 1 diabetes stabilized
on CSII pump insulin. All subjects com-
pleted the following: one screening visit
and three visits to the CRC scheduled at 1
to 14 days apart at the Barbara Davis Cen-
ter for Diabetes (BDC), University of Col-
orado Hospital, Denver. Blood glucose and
insulin levels were measured for 4 h after
Oralin
TM
spray on one occasion, and on
two other occasions subjects were treated
with their usual bolus dose of Humalog
from their CSII pump, or with placebo
puffs with their CSII pump running at ba-
sal rate of 0.71.0 U h
1
. The evening be-
fore the study visit, patients were asked to
consume their normal evening meal along
with their prescribed bolus dose of insulin
via the CSII pump. All subjects were in-
structed not to consume any food or su-
gary drinks after 22:00 h. Patients were
also advised to avoid smoking and ingest-
ing alcohol during this time. The studies
were commenced with a FBG level (capil-
lary sample) in the range of 47 mmol L
1
(70130 mg dL
1
) as suggested by the
FDA. Each subject was provided with the
placebo device to practice the administra-
tion technique, as taught by video demon-
stration. A simple procedure was followed
to administer the dose of Oralin
TM
as out-
lined: patients were asked to position the
device in the mouth and to spray the Ora-
lin
TM
formulation by depressing the device
once. Patients were asked not to exhale or
breathe for 5 s in order to retain the dose
in the mouth, without expelling the mist
during exhalation. Patients were asked to
repeat the same procedure to take the next
dose.
Vital signs (blood pressure, body weight,
pulse) were monitored during the study
period at 0 (baseline), 1, 2, and 4 h after
Oralin
TM
or CSII bolus dose or placebo
spray treatments. All subjects received the
following treatments in a completely ran-
domized fashion, at 314 days apart.
Treatment 1: regular bolus dose of insu-
lin via a CSII pump.
Treatment 2: Oralin
TM
spray (10 puffs)
administered in <15 s.
Treatment 3: CSII pump running (0.7
1.0 U h
1
) with 10 placebo puffs.
At 10 min after the Oralin
TM
dose, bolus
CSII or placebo puffs, subjects were asked
to consume 360 kcal of Boost Plus
liquid
meal. Blood samples for glucose and insu-
lin (free and total) were taken just prior to
the meal (30 and 0 min) and over the fol-
lowing 4-h period, at 15, 30, 45, 60, 90,
120, 180, and 240 min.
Post-prandial glucose levels were signifi-
cantly lowered with Oralin
TM
compared to
CSII pump injection treatment (1055 ver-
sus 1247 mg dL respectively at 30 min,
and 1426 versus 1869 mg dL
1
respec-
tively injection (p<0.003) at 60 min). The
rise in serum insulin levels were significant-
ly greater with Oralin
TM
than with subcuta-
neous injection (C
max
=986 lUmL
1
for
Oralin
TM
at 30 min versus 653 lU mL
1
CSII bolus at 62 min, p<0.001). The ab-
sorption and onset of Oralin
TM
action were
faster than the CSII bolus (202 versus
607 min). Oralin
TM
did not adversely af-
fect post-prandial glycemic control when
compared to subcutaneous insulin. There
was no statistical difference in the variability
of absorption of Oralin
TM
and subcutaneous
TM
injection, as estimated from individual data
of each treatment, and both treatments
were comparable in absorption characteris-
tics (Figs. 4.7 and 4.8).
4.4.6
Oral Spray Insulin in Treatment
of Type 2 Diabetes: Comparison of Efficacy of
Oralin
TM
and Subcutaneous Insulin Injection
The aim of this proof-of-concept study was
to introduce Oralin
TM
as meal insulin in
place of meal-time insulin injections in
the treatment of type 2 diabetes, and to
evaluate the efficacy, safety, of the new for-
mulation.
This was a randomized, single-dose,
two-way, crossover and comparative study
which involved 23 middle-aged subjects
(12 males, 11 females; age range 3570
years) with type 2 diabetes. Before ran-
domization, the two treatment groups
were similar in terms of baseline clinical
features, including lipid parameters. All
subjects were currently receiving multiple
daily injections to control their diabetes.
All subjects were counseled by a quali-
fied dietician and then monitored closely
TM
1455
Fig. 4.7 Meal study in type 1 diabetes patients. Mean post-prandial blood glucose ex-
cursions after Oralin
TM
spray or CSII bolus doses, then challenged on three different
occasions with 360 cal Boost + liquid meal.
Fig. 4.8 Meal study in type 1 diabetes patients. Mean insulin levels after Oralin
TM
spray
or CSII bolus doses, then challenged on three different occasions with 360 cal Boost +li-
quid meal.
for 68 weeks during the run-in period.
Insulin doses were adjusted to achieve
morning glucose levels in the range 72
144 mg dL
1
, as suggested by the FDA.
When the morning glucose levels were
within this range, the subjects were re-
screened for their general well-being and
randomized to the study.
Subjects received each of the following
treatments, in a random order (37 days
apart), following the blood sample at
0 minutes:
Treatment 1: subcutaneous injection
(0.1 unit kg
1
) Humalog with placebo
puffs (n = 4), at time 0 min.
Treatment 2: Oralin
TM
spray (100 units,
four puffs, 25 units per puff) equivalent
to 78 units subcutaneous insulin, at
time 0 min.
At 10 min after each treatment administra-
tion, subjects were given a standard break-
fast (360 cal, Ensure, or Boost liquid meal)
as specified by the FDA.
Blood samples for glucose, insulin and
C-peptide were taken just before dosing,
and during the next 4 h (i.e., 30, 0, +15,
+30, +60, +90, +120, +150, +180, +210,
and +240 min). Vital signs (blood pressure,
pulse rate) were monitored during each
study day as follows: 0 min (pre-drug)
+0.5, +1.0, +2.0, and +4.0 h after treatment
administration.
The primary efficacy parameter was post-
prandial glucose (PPG) control during the
4-h study, together with an increase in se-
rum insulin levels within the first 60 min
after dosing. There was a significant differ-
ence in glucose excursion at 30 and
60 min after a standard meal challenge, as
indicated by lower glucose levels after Ora-
lin
TM
treatment than after injection. The
30-min and 60-min PPG levels were signif-
icantly lowered by Oralin
TM
compared to in-
jection (1465 and 1847 mg dL
1
respec-
tively; 21% lower at 30 min; and 1926
mg dL
1
Oralin
TM
versus 2369 mg dL
1
injection: 19% lower at 60 min, p<0.003).
This difference had disappeared at 2 h and
at the end of the study period at 240 min,
with glucose levels almost identical and
there being no difference between the two
treatments. The rise in serum insulin level
was significantly higher (C
max
=986 lU
TM
Fig. 4.9 Meal study in type 2 diabetes patients. Mean post-prandial
blood glucose excursions after Oralin
TM
spray or subcutaneous injec-
tion bolus, then challenged with 360 cal Boost +plus liquid meal.
mL
1
for Oralin
TM
at 30 min versus
653 lU mL
1
for injection, 35% higher,
p<0.001). The absorption of Oralin
TM
through the buccal mucosa was significantly
faster when compared to subcutaneously in-
jected, rapid-acting insulin. Oralin
TM
did not
adversely affect PPG control when com-
pared to subcutaneous insulin injection, this
being attributed to the much more rapid ab-
sorption of Oralin
TM
through the buccal mu-
cosa (T
max
=305 min for Oralin
TM
versus
6010 min for injection (Humalog); C
max
=
986 lU mL
1
for Oralin
TM
at 30 min ver-
sus 653 lU mL
1
for injection). Reduc-
tions in C-peptide levels were also signifi-
cantly greater during the first 1 h of the
study after Oralin
TM
treatment, mainly
due to the much more rapid absorption
and onset of action of Oralin
TM
[21% de-
crease at 30 min and 1 h (1.380.21 ng
mL
1
for Oralin
TM
versus 1.750.38 ng
mL
1
for injection); p<0.001] when com-
pared to subcutaneously injected insulin
(Figs. 4.9 and 4.10). This difference had dis-
appeared at 2 h, and at the end of the study
period, as seen from the available data.
There was no statistically significant differ-
ence in the variability of absorption of
Oralin
TM
versus subcutaneous injection, as
estimated from the individual data of each
treatment, and both treatments were com-
parable to each other in terms of absorption
characteristics.
4.5
Phase II, Long-term Safety and Efficacy Study
4.5.1
Replacement of Failing Oral Hypoglycemic
Agents with Oralin
TM
; Improvement in PPG
and Overall Glycemic Control (HbA1c)
in Subjects with Type 2 Diabetes
This study was designed as a proof-of-con-
cept Phase II study to determine the safety
and efficacy of Oralin
TM
in place of glybur-
ide (a sulfonylurea) or insulin secretagogues
on a long-term basis (90 days or more) in
subjects with type 2 diabetes. The primary
hypothesis was that Oralin
TM
could be used
safely in combination with metformin to
help maintain or improve the eight-point
glucose profiles and the baseline HbA1c lev-
els at 90 days or more after treatment.
This was a single-blind, randomized,
parallel group study involving 50 subjects
(28 males, 22 females; age range 3570
Fig. 4.10 Meal study in type 2 diabetes patients. Mean insulin levels after Oralin
TM
spray
or subcutaneous injection bolus, then challenged with 360 cal Boost +liquid meal.
years) with type 2 diabetes which was sub-
optimally controlled on OHAs, as assessed
by measurement of the patients HbA
1c
levels. Before randomization, the two treat-
ment groups were similar in baseline clini-
cal features, including lipid parameters.
The inclusion criteria included: FBG levels
of 72200 mg dL
1
during the screening
process; absence of other clinical anoma-
lies except those derived from their meta-
bolic condition that were relatively minor;
physical examination without reasonably
major abnormalities; HbA
1c
811%, and
BMI <38 kg m
2
. Patients were excluded if
they showed low fasting plasma C-peptide
concentrations (<0.2 nmol L
1
), significant
ketonuria (more than trace amounts), evi-
dence of renal disease, plasma creatinine
>150 lmol L
1
, severe retinopathy (prolif-
erative or pre-proliferative), severe cardiac
disease, and other potentially life-threaten-
ing disease. All subjects were counseled by
a qualified dietician and provided with
guidance to control their diet; they were
also encouraged to perform physical exer-
cise regularly. Patients were asked to con-
tinue with their regular therapy (metfor-
min + glyburide). All subjects were moni-
tored closely for 68 weeks during the
run-in period. The oral medication doses
were adjusted to achieve FBG levels in the
range of 72144 mg dL
1
. When this FBG
range was attained, the subjects were re-
screened for their general well-being and
randomized to the study. Following the
initial briefing and training session, the
subjects were randomly divided into two
groups:
Group A: oral insulin + metformin
group;
Group B: (control group) receiving met-
formin + glyburide and placebo puffs.
Group A subjects were asked to take met-
formin (500 or 850 mg, t.i.d.) and oral in-
sulin spray (seven puffs, 70 units, t.i.d.) at
1015 min before every meal (breakfast,
lunch and dinner) and snack, and before
the bed-time if needed. Group B subjects
(controls) were asked to take their usual
dose of metformin (500850 mg, t.i.d.)
with glyburide (510 mg, b.i.d., or as
directed) and the placebo puffs (seven
puffs) at 1015 min before every meal and
snack, for 90 days. The subjects were
asked to monitor themselves as many
times as possible every day, in the morn-
ing, at lunch time, and before the bedtime,
and once a week for eight-point glucose
profile throughout the 90-day study period,
and to note the values in their diary as in-
structed. Subjects were instructed not to
consume alcohol and to avoid smoking
during the trial period. Each subject was
screened for his or her routine blood
chemistry and the baseline HbA
1c
levels
when he or she entered the study, and at
15, 30, 60, and 90 days during the study
period.
The chronic administration of Oralin
TM
before each meal reduced hyperglycemia
(average glucose level 182 mg dL
1
after
Oralin
TM
versus 214 mg dL
1
after glybur-
ide, p<0.008) and the HbA
1c
levels in
comparison with the sulfonylurea (20 mg
glyburide per day) during the study period.
The effect of Oralin
TM
became more evi-
dent at about 60 days as the insulin resis-
tance decreased and patients became more
sensitive to the oral insulin treatment.
This effect was continued throughout the
study period, at the end of which it be-
came more pronounced as the HbA
1c
lev-
els were reduced significantly (~1%;
p<0.001) when compared to the regular
treatment with the oral agents metformin
+ glyburide, where HbA
1c
levels remained
unchanged in most cases. No adverse
events such as burning sensations, red-
ness, peeling of mucosal linings, and taste
TM
F
i
g
.
4
.
1
1
R
e
p
l
a
c
e
m
e
n
t
o
f
f
a
i
l
i
n
g
o
r
a
l
h
y
p
o
g
l
y
c
e
m
i
c
a
g
e
n
t
s
w
i
t
h
O
r
a
l
i
n
T
M
.
L
o
n
g
-
t
e
r
m
s
a
f
e
t
y
a
n
d
e
f
f
i
c
a
c
y
s
t
u
d
y
i
n
s
u
b
j
e
c
t
s
w
i
t
h
t
y
p
e
2
d
i
a
b
e
t
e
s
.
N
o
t
e
t
h
e
i
m
p
r
o
v
e
m
e
n
t
i
n
o
v
e
r
a
l
l
g
l
y
c
e
m
i
c
c
o
n
t
r
o
l
(
a
s
i
n
d
i
c
a
t
e
d
b
y
H
b
A
1
c
l
e
v
e
l
s
)
.
changes were observed during the study
period. Liver function, hematological pa-
rameters and lipid levels remained un-
changed in both groups. This proof-of-con-
cept study indicated that the oral insulin
spray formulation could be used safely to
control blood glucose levels effectively in
subjects with type 2 diabetes, where oral
agents such as sulfonylureas or insulin se-
cretagogues had failed (Fig. 4.11).
4.6
Conclusions
Taken together, the results of the above-
mentioned studies suggest that Oralin
TM
may offer important advantages in the
treatment of diabetes, and may increase
patient compliance, due mainly to the
avoidance of needle injections associated
with subcutaneous insulin. The non-inva-
sive, buccal delivery of Oralin
TM
should
make exogenous insulin treatment more
straightforward, thus improving patient
health and diabetes control, and avoiding
complications of treatment. Oralin
TM
will
also permit a safer and more effective con-
trol of meal-related glucose levels.
References
1 UK Prospective Diabetes Study (UKPDS)
Group: Effect of intensive blood glucose con-
trol with metformin on complications in over-
weight patients with type 2 diabetes (UKPDS
34): UK Prospective Diabetes Study (UKPDS)
Group. Lancet 352:854865, 1998.
2 UK Prospective Diabetes Study: Intensive
blood-glucose control with sulphonylureas or
insulin compared with conventional treatment
and risk of complications in patients with type
2 diabetes (UKPDS 33). Lancet 352:837853,
1998.
3 Matthews DR, Cull CA, Stratton IM, Holman
RR, Turner RC: UKPDS 26: sulphonylurea
failure in non-insulin diabetic patients over six
years: UK Prospective Diabetes Study (UKPDS)
Group. Diabet Med 15:297303, 1998.
4 Turner RC, Cull CA, Frighi V, Holman RR:
Glycemic control with diet, sulfonylurea, met-
formin or insulin in patients with type 2 dia-
betes mellitus: progressive requirement for
multiple therapies (UKPDS 49): UK Prospec-
tive Diabetes Study (UKPDS) Group. JAMA
281:20052012, 1999.
5 Tuomilento J, Lindstrom J, Eriksson JG, Valle
TT, Hamalainen H, Lianne-Parikka P, Keina-
nen-Kiukaanniami S, Laakae M, Louheranta A,
Rastas M, Salminen V, Uusitupa M: Prevention
of type 2 diabetes mellitus by changes in life-
style among subjects with impaired glucose
tolerance. N Engl J Med 344:13431350, 2001.
6 Gjessing HJ, Reinhold B, Pedersen O: The ef-
fect of chronic hyperglycaemia on the islet B-
cell responsiveness in newly diagnosed type 2
diabetes. Diabet Med 9:601604, 1992.
7 Horton ES: Role and management of exercise
in diabetes mellitus. Diabetes Care 11:201211,
1988.
8 Wing RR, Venditti E, Jakicic JM, Polley BA,
Lang W: Lifestyle intervention in overweight
individuals with a family history of diabetes.
Diabetes Care 21:350359, 1998.
9 Diabetes Prevention Program Research
Group: Reduction in the incidence of type 2
diabetes with lifestyle intervention or metfor-
min. N Engl J Med 346:393403, 2002.
10 Winocour PH: Effective diabetes care: a need for
realistic targets. Br Med J 324:15771580, 2002.
11 Inzucchi SE: Oral antihyperglycemic therapy
for type 2 diabetes: scientific review. JAMA
287:360372, 2002.
12 Peacock I, Tattersall RB: The difficult choice
of treatment for poorly controlled maturity on-
set diabetes: tablets or insulin? Br Med J
288:19561959, 1984.
13 Taylor R: Insulin for the non-insulin depen-
dent. Br Med J 296:1015, 1988.
14 Birkeland KI, Rishaug U, Hanssen KF, Vaaler
S: NIDDM: a rapid progressive disease: re-
sults from a long-term, randomised, compara-
tive study of insulin or sulphonylurea treat-
ment. Diabetologia 39:16291633, 1996.
15 Moellma ED, Snoek FJ, Ader HJ, Heine RJ,
van der Ploeg HM: Insulin treated diabetes
patients with fear of self-injecting or fear of
TM
self testing: psychological comorbidity and
general well-being. J Psychosom Res 51:655
672, 2001.
16 Jaber LA, Nowak SN, Slaughter RR: Insulin-
metformin combination therapy in obese pa-
tients with type 2 diabetes. J Clin Pharmacol
25:8994, 2002.
17 de Grauw WJC, Van de Lisdonk EH, van Ger-
wen WHEM, van den Hoogen HJM, van Weel
C: Insulin therapy in poorly controlled type 2
diabetic patients: does it affect quality of life?
Br J Gen Pract 51:527532, 2001.
18 Guevara-Aguirre J, Guevara M, Saavedra J,
Modi P: Beneficial Effects of Addition of Oral
Spray Insulin (Oralin
TM
) on Insulin Secretion
and Metabolic Control in Subjects with Type 2
Diabetes Mellitus Suboptimally Controlled on
Oral Hypoglycemic Agents. Diabetes Technol-
ogy & Therapeutics 6:1: 18, 2004.
Mihic M, Modi P: Oral spray insulin in treat-
ment of type 2 diabetes: a comparison of effi-
cacy of the oral spray insulin (Oralin
TM
) with
subcutaneous (SC) insulin injection, a proof
of concept study: Diabetes/Metabolism Re-
search and Reviews; 10.1002/dmrr.477, 2004.
20 Levin P, Yutzy P, Chez N, Modi, P: Improved
Post-prandial Glucose Control with Oralin
TM
at Breakfast, Lunch and Dinnertime. Diabetes:
A Journal of the American Diabetes Associations
50(Suppl. 2):A124, 2001 (Abstract).
20 Raz I, Kidron M, Wohlgernter J, Modi P: Time
Action Profile of Oralin
TM
in Comparison with
s.c. Injected Insulin in Type 1 Diabetic Patients
Under Euglycemic Clamp Technique. Diabetes:
A Journal of the American Diabetes Associations
52(Suppl. 1):A107, 2003 (Abstract).
Moncayo P, Benitez E, Modi P: Dose Ranging
Study of Oralin
TM
in Healthy Subjects. Dia-
betes: A Journal of the American Diabetes Asso-
ciations 52(Suppl. 1):A104, 2003 (Abstract).
References 1461
Abstract
Although peptide and protein drugs are an
increasingly important class of therapeutic
agents, their oral bioavailability is gener-
ally poor as they are poorly absorbed and
easily degraded by proteolytic enzymes in
the gastrointestinal tract [1]. For the sys-
temic delivery of peptide and protein
drugs, parenteral administration is cur-
rently required in order to achieve their
therapeutic activities. However, these ad-
ministration routes are poorly accepted by
patients, and may cause allergic reactions.
Thus, alternative routes such as nasal [2],
buccal [3], pulmonary [4], rectal [5], vaginal
[6], conjunctival [7], and transdermal [8]
are under investigation for peptide and
protein delivery. Among these routes, the
oral route is the most common and conve-
nient for the administration of these
drugs. The intestinal absorption of peptide
and protein drugs is poor due to extensive
degradation by peptidases and digestive
enzymes, together with poor membrane
permeability characteristics. Thus, a variety
of strategies were examined to improve in-
testinal absorption of these drugs, and
these are outlined in this chapter. First,
the effects of absorption enhancers and
protease inhibitors on intestinal absorption
are detailed. The effect of chemical modifi-
cation (acylation) on intestinal absorption
of peptide and protein drugs, including in-
sulin and tetragastrin, is also examined.
The colon-specific delivery of insulin using
chitosan capsules is also described.
Abbreviations
ACTH adrenocorticotropic hormone
BL-9 polyoxyethylene-9-lauryl ether
CD circular dichroism
CF carboxyfluorescein
CLd degradation clearance
CLp permeation clearance
D% decrement of plasma glucose
concentration %
DM diethyl maleate
EB Evans blue
ECT [Asu
1.7
]-eel calcitonin
EDTA ethylenediaminetetra-acetic
acid
FD-4 fluorescein isothiocyanate-
labeled dextran with an aver-
age molecular weight of 4000
LA linoleic acid
LM n-lauryl-b-D-maltopyranoside
MM mixed micelle
NaCap sodium caprate
1463
5
Improvement of Intestinal Absorption of Peptide
and Protein Biopharmaceuticals by Various Approaches
Akira Yamamoto
ISBN: 3-527-31184-X
NaDC sodium deoxycholate
NaGC sodium glycocholate
NaSal sodium salicylate
NaTC sodium taurocholate
NO nitric oxide
PA% pharmacological availability %
Phe-Gly phenylalanyl-glycine
pMZ-azide p-methoxybenzoxycarbonyl
azide
STI soybean trypsin inhibitor
TFA trifluoroacetic acid
TG tetragastrin
YAGFM (D-Ala
2
)Met-enkephalinamide
5.1
Improvement of Peptide
and Protein Absorption
5.1.1
Use of Absorption Enhancers
Extensive studies have been conducted on
the intestinal absorption of peptides and
proteins, especially insulin. However, in
the absence of an absorption-promoting
adjuvant, the intestinal absorption of these
biopharmaceuticals is much less than after
intramuscular, intravenous, or subcuta-
neous administration. Incomplete absorp-
tion is probably due to a combination of
poor membrane permeability and metabo-
lism at the absorption site [1]. Thus, a
number of absorption enhancers have
been utilized for improving intestinal ab-
sorption of larger polypeptides and pro-
teins [916] (see Part VI, Chapters 1 and
3). Examples of the intestinal absorption
of peptides and proteins with various ab-
sorption enhancers are listed in Table 5.1.
As shown in the table, many absorption
enhancers have been utilized to enhance
the absorption of insulin, calcitonin, leu-
prolide, and interferon. Moreover, these
enhancers were adopted not only for the
gastrointestinal tract but also for other al-
ternative routes such as nasal, buccal, ocu-
lar, pulmonary, vaginal, and rectal routes.
5 Improvement of Intestinal Absorption of Peptide and Protein Biopharmaceuticals by Various Approaches 1464
Table 5.1 Enhancement of intestinal absorption of peptides/proteins by absorption enhancers
Peptides/proteins Absorption promoters Animal(s)
Insulin Various surfactants
Bile acids, Phospholipid
Enamine derivatives
Sodium salicylate
Sodium 5-methoxysalicylate
Rabbit
Rabbit, rat, dog
Dog
Gastrin Sodium 5-methoxysalicylate Rat
Pentagastrin
Lysozyme Enamine derivatives Rabbit
Heparin
(Asu
1.7
)-eel calcitonin Enamine derivatives Rat
Sodium salicylate
Human epidermal growth factor Sodium caprate Rat
Interferon (human fibroblast interferon) Mixed micelle (linoleic acid, HCO60) Rat
Des-enkephalin-c-endorphin Medium-chain glyceride
Na
2
-EDTA
Rat
Low molecular-weight heparin, buserelin Chitosan derivatives Rat
Insulin Labrasol Rat
There are many factors affecting the effec-
tiveness of absorption enhancers, includ-
ing the physico-chemical characteristics of
the drugs, the administration site of ab-
sorption enhancers, and species differ-
ences in the effectiveness of absorption en-
hancers. In this section, we describe those
factors that can regulate the effectiveness
of various absorption enhancers.
5.1.1.1 Effect of Absorption Enhancers on
the Intestinal Absorption of Peptides
Among the peptides and proteins detailed
in Table 5.1, insulin is probably the most
often studied protein with respect to in-
testinal absorption. Nishihata et al. found
that sodium salicylate and 5-methoxysalicy-
late both increased the rectal absorption of
insulin [5]. The absorption-promoting ef-
fect of sodium 5-methoxysalicylate was
also studied in rats with respect to rectal
delivery of pentagastrin and gastrin [17].
Rectal bioavailability was quantitated by di-
rect comparison of pharmacological effect
with intravenous dose response. Co-ad-
ministration of the absorption adjuvant
greatly enhanced the rectal bioavailability
of the model peptides. The bioavailability
of pentagastrin and gastrin in the absence
of absorption enhancer was 64 and 0, re-
spectively, while the bioavailability of these
peptides increased to 3310 and 187
with the adjuvant.
The effect of various absorption enhanc-
ers on insulin transport across the rectal
membrane of albino rabbits was examined
by an in vitro Ussing chamber method
[18]. Insulin was unable to cross the rectal
mucosa without absorption enhancers, but
its transport was improved in their pres-
ence. Among these enhancers, Na glyco-
cholate (NaGC) was more effective than
Na taurocholate (NaTC), but less effective
than Na deoxycholate (NaDC) and poly-
oxyethylene-9-lauryl (BL-9) in enhancing
rectal transport of insulin. The transport
of YAGFM ([D-Ala
2
]Met-enkephalinamide)
was also enhanced by the addition of 1%
NaGC. Increasing the NaGC concentra-
tions further increased rectal insulin trans-
port. Although EDTA at 0.01% and 0.1%
did not affect rectal transport of insulin, it
did augment the penetration enhancement
effect of 1% NaGC.
We also studied the transport of insulin
across colonic membranes in the presence
of various absorption enhancers, again
using the Ussing chamber method [19].
Insulin transport was enhanced by the ad-
dition of NaDC, EDTA, n-lauryl-b-D-malto-
pyranoside (LM) and Na caprate (NaCap)
(Fig. 5.1), but not by other enhancers.
The mechanisms whereby peptides and
protein absorption was improved by ab-
sorption enhancers were examined from
various aspects. These mechanisms involve
an increase in membrane fluidity, expan-
sion of the dimension of the intercellular
space, solubilization of the mucosal mem-
brane, increase in water flux, and reduc-
tion of the viscosity of the mucus layer ad-
hering to all mucosal surfaces [20].
Furthermore, for peptides and proteins, in-
hibition of peptidase activity is an impor-
tant factor to improve absorption [21].
Thus, in this chapter we will introduce the
use of protease inhibitors to improve the
stability and absorption of peptide and pro-
tein biopharmaceuticals in the gut.
5.1.2
Efficacy and Safety of Absorption Enhancers
As indicated above, a large number of ab-
sorption enhancers including surfactants,
bile salts, chelating agents, and fatty acids
have been used to enhance the intestinal
absorption of antibiotics and macromole-
cules [9, 22]. When these absorption en-
hancers are applied in practical use, it is
essential that they do not affect the mem-
brane integrity of the epithelium. Some of
these adjuvants cause membrane damage
and irritate the intestinal mucosal mem-
brane; consequently, it is necessary to de-
velop effective and non-toxic enhancers for
selective, practical use. Based on this view-
point, we examined the correlation be-
tween the effectiveness and toxicity of vari-
ous absorption enhancers in the intestine.
In these studies, phenol red which is
poorly absorbed and stable in the gastroin-
testinal tract was chosen as a model polar
drug, and a variety of absorption enhancers
was compared in a single experimental sys-
tem in order to rank them in terms of their
absorption-promoting ability [23]. The en-
hancers used were NaGC, NaTC, NaDC,
ethylenediaminetetra acetic acid (EDTA),
sodium salicylate (NaSal), NaCap, diethyl
maleate (DM), LM, and linoleic acid (LA)-
HCO60 mixed micelle (MM) at a concentra-
tion of 20 mM. A simultaneous evaluation
was made of local intestinal damage by
measuring the release of protein and phos-
pholipid as biological markers. In the small
intestine, NaDC, EDTA and LM were the
most effective absorption enhancers,
though NaDC and EDTA caused significant
release of protein and phospholipids. By
contrast, LM did not damage the small in-
testinal membrane. NaTC enhanced phenol
red absorption from the small intestine,
which resulted in little or no protein and
phospholipid release levels.

H088

Hochgeladen von

Dokumentinformationen

Originalbeschreibung:

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

H088

Hochgeladen von

Copyright:

Verfügbare Formate

Jrg Knblein (Ed.

: A Cell Therapeutic Approach to Parkinsons Disease 325

as Treatment for Cancer 537

Cells for the Manufacture of Biopharmaceutical Proteins 779

System at a Glance 1329

: A Cell Therapeutic Approach to Parkinsons Disease 325

as Treatment for Cancer 537

Cells for the Manufacture of Biopharmaceutical Proteins 779

: A Cell Therapeutic Approach to Parkinsons Disease 325

as Treatment for Cancer 537

Cells for the Manufacture of Biopharmaceutical Proteins 779

System at a Glance 1329

: A Cell Therapeutic Approach to Parkinsons Disease 325

as Treatment for Cancer 537

Cells for the Manufacture of Biopharmaceutical Proteins 779

experiment (turning on Lucifer-

was initially approved by the European

, Schering AG) has reinvigorated the

, Merck), a chemotherapeutic agent

. The principle here is

R&D. The Gateway recombinational clon-

was generated by immortaliza-

from Novo Nor-

library also incor-

system from Cipher-

in pulmonary arterial hyperten-

, an oral spray insulin. Pan-

has completed phase III clinical

(the human vasoactive intestinal

(trastuzumab), retained the high af-

: Paving the Way for Individualized Cancer Therapy 128

: Paving the Way for Individualized Cancer Therapy 130

: Paving the Way for Individualized Cancer Therapy 132

: Paving the Way for Individualized Cancer Therapy 134

: Paving the Way for Individualized Cancer Therapy 136

: Paving the Way for Individualized Cancer Therapy 138

: Paving the Way for Individualized Cancer Therapy 140

: Paving the Way for Individualized Cancer Therapy 142

: Paving the Way for Individualized Cancer Therapy 144

: Paving the Way for Individualized Cancer Therapy 146

in the adjuvant setting. Oncology

: Paving the Way for Individualized Cancer Therapy 148

: Paving the Way for Individualized Cancer Therapy 150

). Once a clone that sta-

: A Cell Therapeutic Approach to Parkinsons Disease

: A Cell Therapeutic Approach to Parkinsons Disease 326

: A Cell Therapeutic Approach to Parkinsons Disease 328

: A Cell Therapeutic Approach to Parkinsons Disease 330

: A Cell Therapeutic Approach to Parkinsons Disease 332

: A Cell Therapeutic Approach to Parkinsons Disease 334

: A Cell Therapeutic Approach to Parkinsons Disease 336

: A Cell Therapeutic Approach to Parkinsons Disease 338

: A Cell Therapeutic Approach to Parkinsons Disease 340

: A Cell Therapeutic Approach to Parkinsons Disease 342

: A Cell Therapeutic Approach to Parkinsons Disease 344

: A Cell Therapeutic Approach to Parkinsons Disease 346

: A Cell Therapeutic Approach to Parkinsons Disease 348

: A Cell Therapeutic Approach to Parkinsons Disease 350

focus assay murine sarcoma

focus assay (xenotropic retro-

Antihemophilic factor (Recombi-

database. Hence, there will be dis-

) from 1986 used as an immuno-