Beruflich Dokumente
Kultur Dokumente
)
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
Modern Biopharmaceuticals
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
6.1 Introduction 152
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
8.1 Introduction 213
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
9.1 Introduction 230
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
11.1 Introduction 270
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
12.1 Introduction 284
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
13.1 Introduction 306
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
Therapy 537
6.2 Preclinical Studies 539
6.3 PEGylation and Protection from Inactivation 541
6.4 Toxicology 545
6.5 Clinical Trial 545
6.6 Summary and Conclusions 546
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
8.1 Introduction 569
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
2.1 Introduction 605
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
3.1 Introduction 622
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
3.7 Summary and Conclusions 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
5.1 Introduction 664
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
6.1 Introduction 695
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
6.9 Summary and Conclusion 701
7 Accelerating Diagnostic Product Development Process with Molecular Rational Design
and Directed Evolution 703
Harald Sobek, Rainer Schmuck, and Zhixin Shao
7.1 Introduction 704
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
1.1 Introduction 724
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
2.7 Summary and Conclusion 776
3 PER.C6
995
Yann Echelard, Harry M. Meade, and Carol A. Ziomek
11.1 Introduction 996
11.2 Recombinant Production of AT 998
11.3 Characterization of rhAT 1003
11.4 Preclinical Studies 1007
11.5 Clinical Trials with rhAT 1011
11.6 Conclusions 1016
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
13.1 Introduction 1033
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
14.1 Introduction 1045
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
15.1 Introduction 1064
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
15.7 Conclusion 1079
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
16.1 Introduction 1084
16.2 Expression Systems and Manufacturing Procedures 1085
16.3 Outsourcing and Contract Manufacturing 1089
16.4 Summary and Outlook 1100
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
1.1 Introduction 1107
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
2.1 Introduction 1147
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
3.5 Conclusions 1205
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
4.1 Introduction 1212
4.2 Imaging Modalities and Contrast Agents 1213
4.3 Molecular Imaging 1225
4.4 Molecular Imaging for Drug Discovery and Development 1230
4.5 Concluding Remarks 1239
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
5.6 Conclusions 1264
6 Ligand-based Targeting of Disease:
From Antibodies to Small Organic (Synthetic) Ligands 1271
Michela Silacci and Dario Neri
6.1 Introduction 1272
6.2 Ligands 1273
6.3 Classes of Diseases 1276
6.4 From a Ligand to a Product 1288
6.5 Concluding Remarks 1289
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
:
Paving the Way for Individualized Cancer Therapy 127
Thorsten S. Gutjahr and Carsten Reinhardt
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
7 MIDGE Vectors and dSLIM Immunomodulators:
DNA-based Molecules for Gene Therapeutic Strategies 183
Manuel Schmidt, Barbara Volz, and Burghardt Wittig
8 Nonprotein-coding RNAs and their Potential as Biopharmaceuticals 213
Maciej Szymanski, Jan Barciszewski and Volker A. Erdmann
9 Double-stranded Decoy Oligonucleotides as new Biopharmaceuticals 229
Andreas H. Wagner and Heiko E. von der Leyen
10 Rational siRNA Design for RNA Interference:
Optimizations for Therapeutic Use and Current Applications 243
Anastasia Khvorova, Queta Boese, and William S. Marshall
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
12 Myocardial Regeneration Strategies using Human Embryonic Stem Cells 283
Izhak Kehat, Oren Caspi, and Lior Gepstein
13 Gene and Cell-based Therapies for Cardiovascular Disease 305
Abeel A. Mangi
Contents VI
14 Spheramine
Therapy 537
6.2 Preclinical Studies 539
6.3 PEGylation and Protection from Inactivation 541
6.4 Toxicology 545
6.5 Clinical Trial 545
6.6 Summary and Conclusions 546
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
Contents VIII
8 Combinatorial RNA-based Therapies for HIV-1 569
Kevin V. Morris and John J. Rossi
8.1 Introduction 569
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
2 Learning from Viruses: High-throughput Cloning using the Gateway
System
to Transfer Genes without Restriction Enzymes 605
Jonathan D. Chesnut
2.1 Introduction 605
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
Contents IX
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
3.1 Introduction 622
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
3.7 Summary and Conclusions 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
5.1 Introduction 664
5.2 The Gene Discovery Tool Box or Dictyostelium Research 665
5.3 Production of Recombinant Proteins in D. discoideum 672
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
6.1 Introduction 695
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
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
6.9 Summary and Conclusion 701
7 Accelerating Diagnostic Product Development Process with Molecular Rational Design
and Directed Evolution 703
Harald Sobek, Rainer Schmuck, and Zhixin Shao
7.1 Introduction 704
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
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
3 PER.C6
995
Yann Echelard, Harry M. Meade, and Carol A. Ziomek
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
13 Advanced Expression of Biopharmaceuticals in Yeast at Industrial Scale:
The Insulin Success Story 1033
Asser Sloth Andersen and Ivan Diers
14 Baculovirus-based Production of Biopharmaceuticals using Insect Cell Culture
Processes 1045
Wilfried Weber and Martin Fussenegger
15 Robust and Cost-effective Cell-free Expression of Biopharmaceuticals:
Escherichia Coli and Wheat Embryo 1063
Luke Anthony Miles
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
Contents XII
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
2 Modern Antibody Technology: The Impact on Drug Development 1147
Simon Moroney and Andreas Plckthun
3 Molecular Characterization of Autoantibody Responses in Autoimmune Diseases:
Implications for Diagnosis and Understanding of Autoimmunity 1187
Constanze Breithaupt
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
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
6 Ligand-based Targeting of Disease:
From Antibodies to Small Organic (Synthetic) Ligands 1271
Michela Silacci and Dario Neri
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
Getting Insight Sense the Urgency for Early Diagnostics 1325
8 Development of Multi-marker-based Diagnostic Assays with the ProteinChip
System 1325
Andreas Wiesner
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 XIII
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
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
3 Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems 1419
Karen Lingnau, Christoph Klade, Michael Buschle, and Alexander von Gabain
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
5 Improvement of Intestinal Absorption of Peptide and Protein Biopharmaceuticals
by Various Approaches 1463
Akira Yamamoto
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
7 Liposomal In vivo Gene Delivery 1507
Shigeru Kawakami, Fumiyoshi Yamashita, and Mitsuru Hashida
8 Programmed Packaging:
A New Drug Delivery System and its Application to Gene Therapy 1521
Kentaro Kogure, Hidetaka Akita, Hiroyuki Kamiya, and Hideyoshi Harashima
Getting Beyond Rocket Science vs. Science Fiction 1537
9 Bionanotechnology and its Role to Improve Biopharmaceuticals 1537
Oliver Kayser
Contents XIV
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
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
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
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
5 The Regulatory Environment for Biopharmaceuticals in the EU 1669
Axel F. Wenzel and Carina E. A. Sonnega
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
News and Views Quo Vadis, Biopharmaceuticals? 1741
2 mondoBIOTECH: The Swiss biotech BOUTIQUE 1741
Dorian Bevec and Fabio Cavalli
3 G-CSF and Bioequivalence: The Emergence of Healthcare Economics 1755
James Harris, III
Contents XV
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
5 Engineering and Overproduction of Polyketide Natural Products 1803
Martha Lovato Tse and Chaitan Khoslat
Epilog 1833
More about the Editor 1835
Supplement CD-ROM 1837
Subject Index 1841
Contents XVI
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
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
2 The Role of Pharmacogenetics/Pharmacogenomics
in Drug Development and Regulatory Review: Current Status 49
Shiew-Mei Huang and Lawrence J. Lesko
V
Contents
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ISBN: 3-527-31184-X
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 71
Joerg Geistlinger and Peter Ahnert
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
5 The Development of Herceptin
:
Paving the Way for Individualized Cancer Therapy 127
Thorsten S. Gutjahr and Carsten Reinhardt
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
7 MIDGE Vectors and dSLIM Immunomodulators:
DNA-based Molecules for Gene Therapeutic Strategies 183
Manuel Schmidt, Barbara Volz, and Burghardt Wittig
8 Nonprotein-coding RNAs and their Potential as Biopharmaceuticals 213
Maciej Szymanski, Jan Barciszewski and Volker A. Erdmann
9 Double-stranded Decoy Oligonucleotides as new Biopharmaceuticals 229
Andreas H. Wagner and Heiko E. von der Leyen
10 Rational siRNA Design for RNA Interference:
Optimizations for Therapeutic Use and Current Applications 243
Anastasia Khvorova, Queta Boese, and William S. Marshall
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
12 Myocardial Regeneration Strategies using Human Embryonic Stem Cells 283
Izhak Kehat, Oren Caspi, and Lior Gepstein
13 Gene and Cell-based Therapies for Cardiovascular Disease 305
Abeel A. Mangi
Contents VI
14 Spheramine
System
to Transfer Genes without Restriction Enzymes 605
Jonathan D. Chesnut
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
4 Genetically Modified Mice in Medical and Pharmaceutical Research 649
Cord Brakebusch
5 An NIH Model Organism for Biopharmaceutical and Biomedical Research:
The Lower Eukaryote Dictyostelium discoideum 661
Thomas Winckler, Ilse Zndorf, and Theodor Dingermann
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
7 Accelerating Diagnostic Product Development Process with Molecular Rational Design
and Directed Evolution 703
Harald Sobek, Rainer Schmuck, and Zhixin Shao
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
1.1 Introduction 724
Contents VIII
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
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
2.7 Summary and Conclusion 776
3 PER.C6
995
Yann Echelard, Harry M. Meade, and Carol A. Ziomek
11.1 Introduction 996
11.2 Recombinant Production of AT 998
11.3 Characterization of rhAT 1003
11.4 Preclinical Studies 1007
11.5 Clinical Trials with rhAT 1011
11.6 Conclusions 1016
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
13.1 Introduction 1033
13.2 Design and Optimization of the Insulin Precursor Molecule 1036
13.3 Production of Insulin 1041
13.4 Conclusions and Future Aspects 1042
Contents XI
14 Baculovirus-based Production of Biopharmaceuticals using Insect Cell Culture
Processes 1045
Wilfried Weber and Martin Fussenegger
14.1 Introduction 1045
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
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
15.1 Introduction 1064
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
15.7 Conclusion 1079
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
16.1 Introduction 1084
16.2 Expression Systems and Manufacturing Procedures 1085
16.3 Outsourcing and Contract Manufacturing 1089
16.4 Summary and Outlook 1100
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
1.1 Introduction 1107
1.2 Making Monoclonal Antibodies 1109
1.3 Other Antibody Formats: Antibody Fragments 1113
Contents XII
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
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
2.1 Introduction 1147
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
3.5 Conclusions 1205
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
4.1 Introduction 1212
4.2 Imaging Modalities and Contrast Agents 1213
4.3 Molecular Imaging 1225
4.4 Molecular Imaging for Drug Discovery and Development 1230
4.5 Concluding Remarks 1239
Contents XIII
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
5.5 Pharmacodynamic Studies 1254
5.6 Conclusions 1264
6 Ligand-based Targeting of Disease:
From Antibodies to Small Organic (Synthetic) Ligands 1271
Michela Silacci and Dario Neri
6.1 Introduction 1272
6.2 Ligands 1273
6.3 Classes of Diseases 1276
6.4 From a Ligand to a Product 1288
6.5 Concluding Remarks 1289
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
:
Paving the Way for Individualized Cancer Therapy 127
Thorsten S. Gutjahr and Carsten Reinhardt
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
7 MIDGE Vectors and dSLIM Immunomodulators:
DNA-based Molecules for Gene Therapeutic Strategies 183
Manuel Schmidt, Barbara Volz, and Burghardt Wittig
8 Nonprotein-coding RNAs and their Potential as Biopharmaceuticals 213
Maciej Szymanski, Jan Barciszewski and Volker A. Erdmann
9 Double-stranded Decoy Oligonucleotides as new Biopharmaceuticals 229
Andreas H. Wagner and Heiko E. von der Leyen
10 Rational siRNA Design for RNA Interference:
Optimizations for Therapeutic Use and Current Applications 243
Anastasia Khvorova, Queta Boese, and William S. Marshall
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
12 Myocardial Regeneration Strategies using Human Embryonic Stem Cells 283
Izhak Kehat, Oren Caspi, and Lior Gepstein
13 Gene and Cell-based Therapies for Cardiovascular Disease 305
Abeel A. Mangi
Contents VI
14 Spheramine
System
to Transfer Genes without Restriction Enzymes 605
Jonathan D. Chesnut
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
4 Genetically Modified Mice in Medical and Pharmaceutical Research 649
Cord Brakebusch
5 An NIH Model Organism for Biopharmaceutical and Biomedical Research:
The Lower Eukaryote Dictyostelium discoideum 661
Thomas Winckler, Ilse Zndorf, and Theodor Dingermann
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
7 Accelerating Diagnostic Product Development Process with Molecular Rational Design
and Directed Evolution 703
Harald Sobek, Rainer Schmuck, and Zhixin Shao
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
Contents VIII
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
3 PER.C6
995
Yann Echelard, Harry M. Meade, and Carol A. Ziomek
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
Contents IX
13 Advanced Expression of Biopharmaceuticals in Yeast at Industrial Scale:
The Insulin Success Story 1033
Asser Sloth Andersen and Ivan Diers
14 Baculovirus-based Production of Biopharmaceuticals using Insect Cell Culture
Processes 1045
Wilfried Weber and Martin Fussenegger
15 Robust and Cost-effective Cell-free Expression of Biopharmaceuticals:
Escherichia Coli and Wheat Embryo 1063
Luke Anthony Miles
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
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
2 Modern Antibody Technology: The Impact on Drug Development 1147
Simon Moroney and Andreas Plckthun
3 Molecular Characterization of Autoantibody Responses in Autoimmune Diseases:
Implications for Diagnosis and Understanding of Autoimmunity 1187
Constanze Breithaupt
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
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
6 Ligand-based Targeting of Disease:
From Antibodies to Small Organic (Synthetic) Ligands 1271
Michela Silacci and Dario Neri
Contents X
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
Getting Insight Sense the Urgency for Early Diagnostics 1325
8 Development of Multi-marker-based Diagnostic Assays with the ProteinChip
System 1325
Andreas Wiesner
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
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
1.1 Introduction 1362
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
2.1 Introduction 1394
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
Contents XI
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
3.7 Conclusions 1440
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
4.1 Introduction 1446
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
4.6 Conclusions 1460
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
6.1 Introduction 1488
6.2 Poly(Ortho Esters) 1494
6.3 Preparation and Characterization of Microspheres 1496
6.4 In vivo Evaluation of Immune Responses 1500
6.5 Concluding Remarks 1503
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
7.6 Conclusions 1517
Contents XII
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
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
9.1 Introduction 1537
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
1.1 Introduction 1558
1.2 Quality Control 1559
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
1.7 Conclusions 1577
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
2.5 Conclusions 1599
Contents XIII
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
3.2 Phase I Metabolism Prediction: Computational Approaches 1608
3.3 Phase I Metabolism Prediction: In vitro Techniques 1613
3.4 Genotoxicity Prediction 1624
3.5 Conclusions 1634
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
4.1 Introduction 1638
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
5.1 Introduction 1673
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
5.9 Conclusions and Outlook 1701
Contents XIV
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
1.1 Introduction 1712
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
1.5 Conclusions and Outlook 1738
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
3.1 Introduction 1756
3.2 Biogenerics and Bioequivalence 1756
3.3 Summary and Outlook 1767
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
4.1 Introduction 1772
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
5.1 Introduction 1804
5.2 Polyketide Synthases 1806
Contents XV
5.3 Engineering PKSs to Produce Novel Polyketides 1815
5.4 Development of Scalable Production Processes 1820
5.5 Conclusions 1825
Epilog 1833
More about the Editor 1835
Supplement CD-ROM 1837
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
(Physiology or Medicine, 1986)
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
(Physiology or Medicine, 2003)
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
Modern Biopharmaceuticals
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
: 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
, BMS) or yt-
trium-90 ([
90
Y]ibritumomab tiuxetan; Zeva-
lin
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
[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
) 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
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
. 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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
Huntsville, AL 35806
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
Constanze Breithaupt
Structure Research
Max-Planck-Institute of Biochemistry
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
Center for Drug Evaluation and Research
Food and Drug Administration
29 Lincoln Drive
Bethesda, MD 20892
USA
Elisabeth Brundke
ProBioGen AG
Goethestrasse 54
13089 Berlin
Germany
Kevin W. Burton
Nektar Therapeutics
490 Discovery Drive
Huntsville, AL 35806
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
Department of Chemistry
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
Campus Vienna Biocenter 6
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
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
Shimon Gepstein
Department of Biology
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
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
Graduate School of Pharmaceutical
Sciences
Hokkaido University
Kita 12, Nishi 6
Kita-ku, Sapporo City
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
Graduate School of Pharmaceutical
Sciences
Kyoto University
Sakyo-ku, Kyoto 606-8501
Japan
Peter Hauff
Research Laboratories
Schering AG
Mllerstrasse 178
13342 Berlin
Germany
Kirsten Hegmans-Brouwer
Crucell Holland N.V.
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
Center for Drug Evaluation and Research
Food and Drug Administration
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
Astex Technology Ltd.
436 Cambridge Science Park
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
Graduate School of Pharmaceutical
Sciences
Hokkaido University, Kita 12, Nishi 6
Kita-ku, Sapporo City
Hokkaido 060-0812
Japan
Sung Keun Kang
Department of Theriogenology
and Biotechnology
College of Veterinary Medicine
Seoul National University
San 56-1 Shillim-Dong, Kwanak-Gu
Seoul 151-742
Korea
List of Contributors CXXIX
Joachim-Friedrich Kapp
Special Therapeutics
Schering AG
Mllerstrasse 178
13342 Berlin
Germany
Shigeru Kawakami
Department of Drug Delivery Research
Graduate School of Pharmaceutical
Sciences
Kyoto University
Sakyo-ku, Kyoto 606-8501
Japan
Oliver Kayser
Pharmaceutical Biology
University of Groningen
Antonius Deusinglaan 1
9713 AV Groningen
The Netherlands
Izhak Kehat
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
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
Campus Vienna Biocenter 6
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
Graduate School of Pharmaceutical
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
Department of Gene Therapy and
Corporate Center for Biologics
Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Fija Lagerwerf
Crucell Holland N.V.
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
Department of Theriogenology
and Biotechnology
College of Veterinary Medicine
Seoul National University
San 56-1 Shillim-Dong, Kwanak-Gu
Seoul 151-742
Korea
Elisabeth Lehmberg
Department of Gene Therapy
and Corporate Center for Biologics
Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Lawrence J. Lesko
Office of Clinical Pharmacology
and Biopharmaceutics
Center for Drug Evaluation and Research
Food and Drug Administration
29 Lincoln Drive
Bethesda, MD 20892
USA
Heiko E. von der Leyen
Avontec GmbH
Fraunhoferstrasse 15a
82152 Martinsried
Germany
Karen Lingnau
Intercell AG
Campus Vienna Biocenter 6
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
Department of Gene Therapy
and Corporate Center for Biologics
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
Department of Gene Therapy
and Corporate Center for Biologics
Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Harry M. Meade
Department of Research
and Development
GTC Biotherapeutics, Inc.
5 the Mountain Road
Framingham, MA 01701
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
College of Veterinary Medicine
Seoul National University
San 56-1 Shillim-Dong, Kwanak-Gu
Seoul 151-742
Korea
Luis Moroder
Bioorganic Chemistry
Max-Planck-Institute of Biochemistry
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
Department of Chemistry
and Applied Biosciences
Institute of Pharmaceutical Sciences
Swiss Federal Institute of Technology
Zurich (ETH)
Wolfgang-Pauli-Strasse 10
8093 Zrich
Switzerland
David J. Newman
Natural Products Branch
National Cancer Institute
1003 W 7th Street, Suite no. 206
Frederick, MD 21702
USA
Yasushi Ogawa
Department of Gene Therapy
and Corporate Center for Biologics
Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Barry R. OKeefe
Molecular Targets Development Program
Center for Cancer Research
National Cancer Institute
Frederick, MD 21702
USA
Nico Oosterhuis
DSM Biologics Company B.V.
Zuiderweg 72-2
Groningen 9744 AP
The Netherlands
Dirk-Jan Opstelten
Crucell Holland N.V.
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.
Department of Gene Therapy
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
Research Laboratories
Schering AG
Mllerstrasse 178
13342 Berlin
Germany
Yair Reisner
Department of Immunology
Weizmann Institute of Science
PO Box 26
Rehovot 76100
Israel
Elke Reissig
Special Therapeutics
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
Department of Gene Therapy and
Corporate Center for Biologics
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 Diagnostics GmbH
Roche Centralized Diagnostics
Nonnenwald 2
82377 Penzberg
Germany
List of Contributors CXXXV
Marjorie A. Shapiro
Offices of Biotechnology Products
New Drug Chemistry and Drug
Evaluation VI
Center for Drug Evaluation and Research
Food and Drug Administration
29 Lincoln Dr.
Bethesda, MD 20892
USA
Katrin Sichler
Quality Management Centralizied
Diagnostics
Roche Diagnostics GmbH
Sandhofer Strasse 116
68305 Mannheim
Germany
Michela Silacci
Institute of Pharmaceutical Sciences
Department of Chemistry and Applied
Biosciences
Swiss Federal Institute of Technology
Zurich (ETH)
Wolfgang-Pauli-Strasse 10
8093 Zrich
Switzerland
Harald Sobek
Global Biotechnology Operations
Roche Diagnostics GmbH
Roche Applied Science
Nonnenwald 2
82377 Penzberg
Germany
Carina E. A. Sonnega
Casa Oberti, U Pianu
20225 Muro, Corsica
France
Patrick G. Swann
Offices of Biotechnology Products
New Drug Chemistry and Drug
Evaluation VI
Center for Drug Evaluation and Research
Food and Drug Administration
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
Department of Gene Therapy
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
Huntsville, AL 35806
USA
Dijana Matak Vinkovic
Astex Technology Ltd.
436 Cambridge Science Park
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)
Swiss Federal Institute of Technology
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
Department of Gene Therapy
and Corporate Center for Biologics
Development
Berlex, Inc.
2600 Hilltop Drive
Richmond, CA 94806
USA
Sheena Whyte
Astex Technology Ltd.
436 Cambridge Science Park
Milton Road
Cambridge CB4 0QA
United Kingdom
Andreas Wiesner
Ciphergen Biosystems
Hannah-Vogt-Strasse 1
37085 Gttingen
Germany
List of Contributors CXXXVII
Barbara Wilcox
Offices of Biotechnology Products
New Drug Chemistry and Drug
Evaluation VI
Center for Drug Evaluation and Research
Food and Drug Administration
29 Lincoln Drive
Bethesda, MD 20892
USA
Pamela A. Williams
Astex Technology Ltd.
436 Cambridge Science Park
Milton Road
Cambridge CB4 0QA
United Kingdom
Thomas Winckler
Institute of Pharmaceutical Biology
Goethe-University of Frankfurt (Biocenter)
Marie-Curie-Strasse 9
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
Swiss Federal Institute of Technology
Lausanne (EPFL)
1015 Lausanne
Switzerland
Chris Yallop
Crucell Holland N.V.
Archimedesweg 4
2333 CN Leiden
The Netherlands
Akira Yamamoto
Department of Biopharmaceutics
Kyoto Pharmaceutical University
Misasagi Yamashina-ku, Kyoto, 607-8414
Japan
Fumiyoshi Yamashita
Department of Drug Delivery Research
Graduate School of Pharmaceutical
Sciences
Kyoto University
Sakyo-ku, Kyoto 606-8501
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
GTC Biotherapeutics, Inc.
Suite 410, 175 Crossing Blvd.
Framingham, MA 01702
USA
Ilse Zndorf
Institute of Pharmaceutical Biology
Goethe-University of Frankfurt (Biocenter)
Marie-Curie-Strasse 9
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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,
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 2
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-
3 Upstream and Downstream Processing 3
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-
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 4
3 Upstream and Downstream Processing 5
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.
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 6
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
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 8
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
6 Products Approved for Human Use 9
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-
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 10
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.
6 Products Approved for Human Use 11
Table 2 Recombinant tPA-based products thus far approved
Product Company Therapeutic
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-
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 12
6 Products Approved for Human Use 13
Table 3 Recombinant insulins/insulin analogs thus far approved
Product Company Therapeutic
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]
Novo Nordisk diabetes mellitus 2000 (EU)
Novolog (Insulin Aspart, short-acting
rhInsulin analog produced in
S. cerevisiae; see also NovoRapid)
Novo Nordisk diabetes mellitus 2001 (US)
Novolog mix 70/30 (contains insulin
Aspart, short-acting rhInsulin analog
produced in S. cerevisiae as one
ingredient; see also Novomix 30)
Novo Nordisk diabetes mellitus 2001 (US)
Actrapid/Velosulin/Monotard/Insula-
tard/Protaphane/Mixtard/Actraphane/
Ultratard (all contain rhInsulin pro-
duced in S. cerevisiae formulated as
short/intermediate/long-acting product)
Novo Nordisk diabetes mellitus 2002 (EU)
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
rhInsulin analog produced in
S. cerevisiae)
Novo Nordisk diabetes mellitus 2004 (EU)
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
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 14
Table 4 Additional recombinant hormones approved for general medical use
Product Company Therapeutic
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,
6 Products Approved for Human Use 15
Table 4 (continued)
Product Company Therapeutic
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-
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 16
Table 5 Recombinant hemopoietic growth factors approved for general medical use
Product Company Therapeutic
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;
produced in E. coli)
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)
6 Products Approved for Human Use 17
Table 6 Recombinant IFNs and ILs approved for general medical use
Product Company Therapeutic
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
produced in E. coli)
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
produced in E. coli)
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
produced in E. coli)
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;
produced in E. coli)
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
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 18
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,
containing rHBsAg produced
in S. cerevisiae as one component)
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,
containing rHBsAg produced
in S. cerevisiae as one component)
SmithKline
Beecham
immunization against
diphtheria, tetanus,
pertussis, polio,
hemophilus influenzae B
and hepatitis B
2000 (EU)
Infanrix-Penta (combination vaccine,
containing rHBsAg produced
in S. cerevisiae as one component)
SmithKline
Beecham
immunization against
diphtheria, tetanus,
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.
6 Products Approved for Human Use 19
Table 7 (continued)
Product Company Therapeutic indication Approved
Ambirix (combination vaccine,
containing rHBsAg produced
in S. cerevisiae as one component)
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)
immunization against
hepatitis A and B
1996
(EU, adult),
1997 (EU,
pediatric),
2001 (US)
Primavax (combination vaccine,
containing rHBsAg produced
in S. cerevisiae as one component)
Pasteur Merieux
MSD
immunization against
diphtheria, tetanus and
hepatitis B
1998 (EU)
Pediarix (combination vaccine,
containing rHBsAg produced
in S. cerevisiae as one component)
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
immunization against
hemophilus influenzae
type B and hepatitis B
1999 (EU)
Hexavac (combination vaccine,
containing rHBsAg produced
in S. cerevisiae as one component)
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
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 20
6 Products Approved for Human Use 21
Table 8 Monoclonal/engineered antibody-based products approved for general medical use
Product Company Therapeutic indication Approved
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)
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 22
Table 8 (continued)
Product Company Therapeutic indication Approved
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
6 Products Approved for Human Use 23
Table 9 Additional biopharmaceuticals approved for general medical use
Product Company Therapeutic
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
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 24
Table 9 (continued)
Product Company Therapeutic
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
Product Company Therapeutic indication Approved
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).
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 26
Table 10 (continued)
Product Company Therapeutic indication Approved
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)
Merial active immunization of
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)
Merial active immunization of
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.
8 Likely Future Directions 27
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-
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 28
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 Likely Future Directions 29
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
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 30
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
8 Likely Future Directions 31
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
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 32
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.
Current Status of Biopharmaceuticals: Approved Products and Trends in Approvals 34
Part I
Biopharmaceuticals Used in Molecular Medicine
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
1 Beginning to Understand the End of the Chromosome 38
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
1 Beginning to Understand the End of the Chromosome 40
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-
1 Beginning to Understand the End of the Chromosome 42
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
1 Beginning to Understand the End of the Chromosome 44
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
1 Beginning to Understand the End of the Chromosome 46
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.
1 Beginning to Understand the End of the Chromosome 48
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
Shiew-Mei Huang and Lawrence J. Lesko
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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-
2 The Role of Pharmacogenetics/Pharmacogenomics in Drug Development and Regulatory Review: Current Status 52
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
Drugdrug interactions
1999 1988 Astemizole
(Hismanal)
Antihistamine Torsades de Pointes
Drugdrug interactions
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
Drugdrug interactions
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
2 The Role of Pharmacogenetics/Pharmacogenomics in Drug Development and Regulatory Review: Current Status 54
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
2 The Role of Pharmacogenetics/Pharmacogenomics in Drug Development and Regulatory Review: Current Status 56
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
2 The Role of Pharmacogenetics/Pharmacogenomics in Drug Development and Regulatory Review: Current Status 58
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.
2 The Role of Pharmacogenetics/Pharmacogenomics in Drug Development and Regulatory Review: Current Status 60
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
2 The Role of Pharmacogenetics/Pharmacogenomics in Drug Development and Regulatory Review: Current Status 62
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-
2 The Role of Pharmacogenetics/Pharmacogenomics in Drug Development and Regulatory Review: Current Status 64
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
2 The Role of Pharmacogenetics/Pharmacogenomics in Drug Development and Regulatory Review: Current Status 66
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
FDA/PhRMA/JHU educational workshop,
September 1314, 2004, Rockville, MD. http://
www.fda.gov/cder/offices/ocpb/
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.
2 The Role of Pharmacogenetics/Pharmacogenomics in Drug Development and Regulatory Review: Current Status 68
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,
September 1314, 2004, Rockville, MD. http://
www.fda.gov/cder/offices/ocpb/
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://
www.fda.gov/cder/offices/ocpb/
workshops.htm
73 Ratain M. Special consideration of metabolic
biomarker: UGT1A1 presentation at the
FDA/PhRMA/JHU educational workshop,
September 1314, 2004, Rockville, MD. http://
www.fda.gov/cder/offices/ocpb/
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
2 The Role of Pharmacogenetics/Pharmacogenomics in Drug Development and Regulatory Review: Current Status 70
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ISBN: 3-527-31184-X
3
Large-scale Detection of Genetic Variation:
The Key to Personalized Medicine
Joerg Geistlinger and Peter Ahnert
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
TNF tumor necrosis factor
TPMT thiopurine methyltrans-
ferase
TSER thymidilate synthase
enhancer region
VNTR variable number of tandem
repeats
WNK4 (with-no-lysine) kinase 4
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 72
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
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 74
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
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 76
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 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 78
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
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 80
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
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 82
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
3.6 SNP Technologies 83
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
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 84
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-
3.6 SNP Technologies 85
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
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 86
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,
3.6 SNP Technologies 87
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-
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 88
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
3.7 Polydimensional SNP-Chips: The Array-On Technology 89
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
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 90
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-
3.7 Polydimensional SNP-Chips: The Array-On Technology 91
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
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 92
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-
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 94
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.
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 96
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.
3 Large-scale Detection of Genetic Variation: The Key to Personalized Medicine 98
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
Bonnie E. Gould Rothberg, Carol E. A. Pena, and Jonathan M. Rothberg
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
PCR polymerase chain reaction
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
4.1 Limitations in the Chronic Disease Drug Discovery Process 101
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
4 A Systems Biology Approach to Target Identification and Validation 102
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
4.1 Limitations in the Chronic Disease Drug Discovery Process 103
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).
4 A Systems Biology Approach to Target Identification and Validation 104
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
4.2 Creating the Pharmaceutically Tractable Genome 105
4 A Systems Biology Approach to Target Identification and Validation 106
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-
4.2 Creating the Pharmaceutically Tractable Genome 107
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
4 A Systems Biology Approach to Target Identification and Validation 108
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
4.2 Creating the Pharmaceutically Tractable Genome 109
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 A Systems Biology Approach to Target Identification and Validation 110
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
.
4 A Systems Biology Approach to Target Identification and Validation 112
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,
4.3 Integrated Systems Biology Approaches to Drug Target Validation for Specific Clinical Indications 113
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
4 A Systems Biology Approach to Target Identification and Validation 114
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).
"
4.3 Integrated Systems Biology Approaches to Drug Target Validation for Specific Clinical Indications 115
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-
4 A Systems Biology Approach to Target Identification and Validation 116
4.3 Integrated Systems Biology Approaches to Drug Target Validation for Specific Clinical Indications 117
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
4 A Systems Biology Approach to Target Identification and Validation 118
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-
4.3 Integrated Systems Biology Approaches to Drug Target Validation for Specific Clinical Indications 119
4 A Systems Biology Approach to Target Identification and Validation 120
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
4.3 Integrated Systems Biology Approaches to Drug Target Validation for Specific Clinical Indications 121
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
4 A Systems Biology Approach to Target Identification and Validation 122
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
25 J. Drews, Drug Discov. Today 2000, 5, 54753.
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.
30 J. Drews, Drug Discov. Today 2003, 8, 411420.
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.
4 A Systems Biology Approach to Target Identification and Validation 124
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
:
Paving the Way for Individualized Cancer Therapy
Thorsten S. Gutjahr and Carsten Reinhardt
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
mAb monoclonal antibody
MBC metastatic breast cancer
ORR overall response rate
OS overall survival
PCR polymerase chain reaction
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
5 The Development of Herceptin
immunotherapy in meta-
static breast cancer. Oncology 2001, 61 (suppl
2), 5866.
72 Tan-Chiu E, Piccart M. Moving forward with
Herceptin
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-
10.3 Current siRNA Design Considerations 257
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.
10 Rational siRNA Design for RNA Interference: Optimizations for Therapeutic Use and Current Applications 258
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-
10.4 Therapeutic Applications of RNAi 259
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-
10 Rational siRNA Design for RNA Interference: Optimizations for Therapeutic Use and Current Applications 260
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
10.4 Therapeutic Applications of RNAi 261
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-
10 Rational siRNA Design for RNA Interference: Optimizations for Therapeutic Use and Current Applications 262
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
10.4 Therapeutic Applications of RNAi 263
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
10 Rational siRNA Design for RNA Interference: Optimizations for Therapeutic Use and Current Applications 264
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.
10 Rational siRNA Design for RNA Interference: Optimizations for Therapeutic Use and Current Applications 266
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,
Proc. Natl. Acad. Sci. USA 1998, 95(22),
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,
Proc. Natl. Acad. Sci. USA 1999, 96(24),
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,
Proc. Natl. Acad. Sci. USA 2003, 100(5),
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.
10 Rational siRNA Design for RNA Interference: Optimizations for Therapeutic Use and Current Applications 268
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
Woo Suk Hwang, Byeong Chun Lee, Sung Keun Kang, and Shin Yong Moon
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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,
11.2 Human Somatic Cell Nuclear Transfer (SCNT) 271
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-
11 The First Cloned Human Embryo: An Unlimited Source of Stem Cells for Therapeutic Cloning 272
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
11.2 Human Somatic Cell Nuclear Transfer (SCNT) 273
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
11 The First Cloned Human Embryo: An Unlimited Source of Stem Cells for Therapeutic Cloning 274
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
11.2 Human Somatic Cell Nuclear Transfer (SCNT) 275
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
11 The First Cloned Human Embryo: An Unlimited Source of Stem Cells for Therapeutic Cloning 276
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-
11 The First Cloned Human Embryo: An Unlimited Source of Stem Cells for Therapeutic Cloning 278
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.
11 The First Cloned Human Embryo: An Unlimited Source of Stem Cells for Therapeutic Cloning 280
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
11 The First Cloned Human Embryo: An Unlimited Source of Stem Cells for Therapeutic Cloning 282
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
Izhak Kehat, Oren Caspi, and Lior Gepstein
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
HLA human leukocyte antigen
ICM inner cell mass
IVF in vitro fertilization
LIF leukemia inhibitory factor
MEA micro-electrode array
MEF mouse embryonic fibroblasts
MHC major histocompatibility com-
plex
Mef myocyte enhancer factor Mef
PET positron emission tomography
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-
12 Myocardial Regeneration Strategies using Human Embryonic Stem Cells 284
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,
12.1 Introduction 285
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).
12 Myocardial Regeneration Strategies using Human Embryonic Stem Cells 286
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-
12 Myocardial Regeneration Strategies using Human Embryonic Stem Cells 288
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-
12.3 Cardiomyocyte Differentiation of ES Cells 289
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-
12 Myocardial Regeneration Strategies using Human Embryonic Stem Cells 290
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-
12.3 Cardiomyocyte Differentiation of ES Cells 291
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
12 Myocardial Regeneration Strategies using Human Embryonic Stem Cells 292
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
12 Myocardial Regeneration Strategies using Human Embryonic Stem Cells 294
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 Myocardial Regeneration Strategies using Human Embryonic Stem Cells 296
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
12 Myocardial Regeneration Strategies using Human Embryonic Stem Cells 298
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
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
.
13.2 Gene Therapy as Novel Drug Delivery 309
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
13 Gene and Cell-based Therapies for Cardiovascular Disease 310
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
13.2 Gene Therapy as Novel Drug Delivery 311
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
13 Gene and Cell-based Therapies for Cardiovascular Disease 312
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
13.2 Gene Therapy as Novel Drug Delivery 313
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
13 Gene and Cell-based Therapies for Cardiovascular Disease 314
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
13.2 Gene Therapy as Novel Drug Delivery 315
Table 13.3 (continued)
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
13 Gene and Cell-based Therapies for Cardiovascular Disease 316
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-
13.2 Gene Therapy as Novel Drug Delivery 317
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 Gene and Cell-based Therapies for Cardiovascular Disease 318
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
13 Gene and Cell-based Therapies for Cardiovascular Disease 320
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.
13 Gene and Cell-based Therapies for Cardiovascular Disease 322
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).
13 Gene and Cell-based Therapies for Cardiovascular Disease 324
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
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
15 Applying Human Cells to Organogenesis and Transplantation 370
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.
15 Applying Human Cells to Organogenesis and Transplantation 372
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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-
1.1 Introduction 379
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
1 Mechanisms of Serine Proteinase Activation 380
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
1.2 Bacterial Activators of Host Zymogens 381
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 Mechanisms of Serine Proteinase Activation 382
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-
1.2 Bacterial Activators of Host Zymogens 383
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
1 Mechanisms of Serine Proteinase Activation 384
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
1.2 Bacterial Activators of Host Zymogens 385
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
1 Mechanisms of Serine Proteinase Activation 386
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-
1.2 Bacterial Activators of Host Zymogens 387
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
1 Mechanisms of Serine Proteinase Activation 388
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
1.3 Some Remarks on Nonproteolytic Activators 389
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.
1 Mechanisms of Serine Proteinase Activation 390
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
1 Mechanisms of Serine Proteinase Activation 392
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
2 Application of the Principle of Polyvalency to Protease Inhibition 396
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-
2.2 Thermodynamic Model of Bivalent Ligand Binding 397
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].
2 Application of the Principle of Polyvalency to Protease Inhibition 398
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
2.3 Homo- and Heterobivalent Inhibitors of the Yeast 20S Proteasome 399
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
2 Application of the Principle of Polyvalency to Protease Inhibition 400
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.
2.3 Homo- and Heterobivalent Inhibitors of the Yeast 20S Proteasome 401
Table 2.2 Inhibition of 20S proteasome by mono- and bivalent inhibitors (IC
50
, lM)
Inhibitor PGPH Trypsin-like Chymotrypsin-like
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
|
PEG-CO-Arg-Val-Arg-H
(9) >100 0.097 0.031
2 Application of the Principle of Polyvalency to Protease Inhibition 402
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-
2.3 Homo- and Heterobivalent Inhibitors of the Yeast 20S Proteasome 403
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
2 Application of the Principle of Polyvalency to Protease Inhibition 404
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)
Inhibitor PGPH Trypsin-like Chymotrypsin-like
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
2.4 Bivalent Inhibition of Mast Cell b-Tryptase 405
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-
2 Application of the Principle of Polyvalency to Protease Inhibition 406
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
|
PEG-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-
2.4 Bivalent Inhibition of Mast Cell b-Tryptase 407
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
2 Application of the Principle of Polyvalency to Protease Inhibition 408
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
2.4 Bivalent Inhibition of Mast Cell b-Tryptase 409
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
2 Application of the Principle of Polyvalency to Protease Inhibition 410
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
2.5 Heterobivalent Inhibition of Thrombin 411
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
2 Application of the Principle of Polyvalency to Protease Inhibition 412
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
2.5 Heterobivalent Inhibition of Thrombin 413
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.
2 Application of the Principle of Polyvalency to Protease Inhibition 414
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.
2 Application of the Principle of Polyvalency to Protease Inhibition 416
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
ATP adenosine triphosphate
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
Norbert Riedel and Friedrich Dorner
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
PCR polymerase chain reaction
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
liquid chromatography
S/D solvent detergent
S
+
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-
3.1 Introduction 421
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
3 A New Technology Standard for Safety and Efficacy in Factor VIII Replacement Therapy 422
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-
3.1 Introduction 423
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
3 A New Technology Standard for Safety and Efficacy in Factor VIII Replacement Therapy 424
3.1 Introduction 425
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
]
.
3 A New Technology Standard for Safety and Efficacy in Factor VIII Replacement Therapy 426
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.
3.1 Introduction 427
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-
3 A New Technology Standard for Safety and Efficacy in Factor VIII Replacement Therapy 428
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
3.2 Development of rFVIII 429
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-
3 A New Technology Standard for Safety and Efficacy in Factor VIII Replacement Therapy 430
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-
3.3 Production of rFVIII 431
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-
3 A New Technology Standard for Safety and Efficacy in Factor VIII Replacement Therapy 432
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:
3.4 Pathogen Safety 433
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
), 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
4 Biopharmaceutical Drugs from Natural Sources 452
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
, Bex-
xar
, 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
)
which was first approved in Germany in
1982, through later versions (improved
purification systems) such as Bioclate
in
4 Biopharmaceutical Drugs from Natural Sources 454
4.1 Biotechnologically Produced Proteins and Peptides as Approved Drugs 455
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
4 Biopharmaceutical Drugs from Natural Sources 456
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
4.1 Biotechnologically Produced Proteins and Peptides as Approved Drugs 457
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
4 Biopharmaceutical Drugs from Natural Sources 458
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
4.1 Biotechnologically Produced Proteins and Peptides as Approved Drugs 459
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
,
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
, an alpha-1 antitrypsin
from human plasma launched in the USA
in 2003, and which had an earlier version,
Prolastin
, 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 mor-
4 Biopharmaceutical Drugs from Natural Sources 460
4.1 Biotechnologically Produced Proteins and Peptides as Approved Drugs 461
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
4 Biopharmaceutical Drugs from Natural Sources 462
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
4.1 Biotechnologically Produced Proteins and Peptides as Approved Drugs 463
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
4 Biopharmaceutical Drugs from Natural Sources 464
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
4.1 Biotechnologically Produced Proteins and Peptides as Approved Drugs 465
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
).
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 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
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
4 Biopharmaceutical Drugs from Natural Sources 468
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
4.1 Biotechnologically Produced Proteins and Peptides as Approved Drugs 469
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
4 Biopharmaceutical Drugs from Natural Sources 470
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
4.1 Biotechnologically Produced Proteins and Peptides as Approved Drugs 471
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
4 Biopharmaceutical Drugs from Natural Sources 472
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
4.1 Biotechnologically Produced Proteins and Peptides as Approved Drugs 473
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
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
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
4 Biopharmaceutical Drugs from Natural Sources 474
4.1 Biotechnologically Produced Proteins and Peptides as Approved Drugs 475
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
4 Biopharmaceutical Drugs from Natural Sources 476
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
4.1 Biotechnologically Produced Proteins and Peptides as Approved Drugs 477
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
4 Biopharmaceutical Drugs from Natural Sources 478
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
4.1 Biotechnologically Produced Proteins and Peptides as Approved Drugs 479
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
4 Biopharmaceutical Drugs from Natural Sources 480
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
,
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
4 Biopharmaceutical Drugs from Natural Sources 482
[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-
4.2 Potential Agents from Non-mammalian Sources as Leads to Novel Therapies 483
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
4 Biopharmaceutical Drugs from Natural Sources 484
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-
4.2 Potential Agents from Non-mammalian Sources as Leads to Novel Therapies 485
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
4 Biopharmaceutical Drugs from Natural Sources 486
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-
4.2 Potential Agents from Non-mammalian Sources as Leads to Novel Therapies 487
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:
4 Biopharmaceutical Drugs from Natural Sources 488
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-
4 Biopharmaceutical Drugs from Natural Sources 490
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,
4 Biopharmaceutical Drugs from Natural Sources 492
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
4 Biopharmaceutical Drugs from Natural Sources 494
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.
4 Biopharmaceutical Drugs from Natural Sources 496
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ISBN: 3-527-31184-X
hEGF human epidermal growth factor
HSA human serum albumin
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-
5.2 Principles of Targeted In situ Radiotherapy of Malignancies 499
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
5 Biomolecules as Targeting Vehicles for In situ Radiotherapy of Malignancies 500
5.3 RIT of Non-Hodgkins B-Cell Lymphomas: The Pre-eminent Success Story 501
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
5 Biomolecules as Targeting Vehicles for In situ Radiotherapy of Malignancies 502
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
5.3 RIT of Non-Hodgkins B-Cell Lymphomas: The Pre-eminent Success Story 503
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
5 Biomolecules as Targeting Vehicles for In situ Radiotherapy of Malignancies 504
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
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 Biomolecules as Targeting Vehicles for In situ Radiotherapy of Malignancies 510
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
5.7 Pre-Targeting Strategies: Improving the Therapeutic Index of RIT 511
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.
5 Biomolecules as Targeting Vehicles for In situ Radiotherapy of Malignancies 512
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.
5.7 Pre-Targeting Strategies: Improving the Therapeutic Index of RIT 513
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-
5 Biomolecules as Targeting Vehicles for In situ Radiotherapy of Malignancies 514
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
) 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
, 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
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
ORF open reading frame
PCR polymerase chain reaction
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
; 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
3 Target Validation: An Important Early Step in the Development of Novel Biopharmaceuticals 640
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 Target Validation: An Important Early Step in the Development of Novel Biopharmaceuticals 642
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.
3 Target Validation: An Important Early Step in the Development of Novel Biopharmaceuticals 644
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
Summary and Conclusions
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
3.7 Summary and Conclusions 645
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-
3 Target Validation: An Important Early Step in the Development of Novel Biopharmaceuticals 646
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
LIF leukemia inhibitory factor
mRNA messenger RNA
neo neomycin
PAC phage artificial chromosomes
PCR polymerase chain reaction
siRNA small interfering RNA
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
4 Genetically Modified Mice in Medical and Pharmaceutical Research 650
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-
4.2 Generation of Genetically Modified Mice by Gene Targeting 651
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
4 Genetically Modified Mice in Medical and Pharmaceutical Research 652
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
4.2 Generation of Genetically Modified Mice by Gene Targeting 653
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
4 Genetically Modified Mice in Medical and Pharmaceutical Research 654
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
4.2 Generation of Genetically Modified Mice by Gene Targeting 655
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
4 Genetically Modified Mice in Medical and Pharmaceutical Research 656
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
4.2 Generation of Genetically Modified Mice by Gene Targeting 657
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.
4 Genetically Modified Mice in Medical and Pharmaceutical Research 658
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.
4 Genetically Modified Mice in Medical and Pharmaceutical Research 660
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
Thomas Winckler, Ilse Zndorf, and Theodor Dingermann
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
FSH follicle-stimulating hormone
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
ORF open reading frame
PsA cell surface glycoprotein A
PSF prestarvation factor
REMI restriction enzyme-mediated
integration
REP transactivator of Ddp2 replication
RNAi RNA interference
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
5 Dictyostelium discoideum: Biopharmaceutical and Biomedical Research with a Lower Eukaryote 664
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
modern biopharmaceuticals.
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-
5 Dictyostelium discoideum: Biopharmaceutical and Biomedical Research with a Lower Eukaryote 666
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
5.2 The Gene Discovery Tool Box for Dictyostelium Research 667
5 Dictyostelium discoideum: Biopharmaceutical and Biomedical Research with a Lower Eukaryote 668
5.2 The Gene Discovery Tool Box for Dictyostelium Research 669
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-
5 Dictyostelium discoideum: Biopharmaceutical and Biomedical Research with a Lower Eukaryote 670
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
5.2 The Gene Discovery Tool Box for Dictyostelium Research 671
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
5 Dictyostelium discoideum: Biopharmaceutical and Biomedical Research with a Lower Eukaryote 672
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 Production of Recombinant Proteins in D. discoideum 673
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-
5 Dictyostelium discoideum: Biopharmaceutical and Biomedical Research with a Lower Eukaryote 674
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
5.3 Production of Recombinant Proteins in D. discoideum 675
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
5 Dictyostelium discoideum: Biopharmaceutical and Biomedical Research with a Lower Eukaryote 676
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
5.3 Production of Recombinant Proteins in D. discoideum 677
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
Table 5.2 (continued)
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].
5 Dictyostelium discoideum: Biopharmaceutical and Biomedical Research with a Lower Eukaryote 678
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
5.3 Production of Recombinant Proteins in D. discoideum 679
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
5 Dictyostelium discoideum: Biopharmaceutical and Biomedical Research with a Lower Eukaryote 680
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
5.3 Production of Recombinant Proteins in D. discoideum 681
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-
5 Dictyostelium discoideum: Biopharmaceutical and Biomedical Research with a Lower Eukaryote 682
Fig. 5.5 Scanning electron micrographs of Dictyostelium discoi-
deum cells inside SIRAN
, 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 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 726
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
1.2 Vectors, Transfections, and Cell Line Generation 727
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 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 728
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 Vectors, Transfections, and Cell Line Generation 729
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
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 730
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-
1.3 Host Cell Engineering 731
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
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 732
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-
1.4 Gene Transfer and Gene Amplification in Mammalian Cells 733
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
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 734
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
1.4 Gene Transfer and Gene Amplification in Mammalian Cells 735
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
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 736
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
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
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 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 738
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-
1.5 Production Principles for Mammalian Cells: Anchorage-dependent Cultures and Suspension Cultures 739
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-
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 740
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
Part IV, Chapter 7).
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-
1.5 Production Principles for Mammalian Cells: Anchorage-dependent Cultures and Suspension Cultures 741
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
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 742
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
, 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
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 744
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-
1.7 Regulatory Issues 745
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-
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 746
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-
1.7 Regulatory Issues 747
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
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 748
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-
1.7 Regulatory Issues 749
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 750
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-
1.8 Concluding Remarks 751
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.
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 752
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.
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 754
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.
Proc. Natl. Acad. Sci. USA 83:54145418.
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.
Proc. Natl. Acad. Sci. USA 75:55535556.
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.
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 756
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.
Proc. Natl. Acad. Sci. USA 92:72977301.
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
1 Manufacture of Recombinant Biopharmaceutical Proteins by Cultivated Mammalian Cells in Bioreactors 758
DNA transfected into mammalian cells.
Proc. Natl. Acad. Sci. USA 80:30153019.
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
BHK baby hamster kidney
CHO Chinese hamster ovary
CMV cytomegalovirus
DHFR dihydrofolate reductase
GS glutamine synthetase
IRES internal ribosome entry site
LCR locus control region
MSX methionine sulfoximine
MTX methotrexate
PCR polymerase chain reaction
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
of Biopharmaceuticals
Thomas Rose, Karsten Winkler, Elisabeth Brundke, Ingo Jordan and Volker Sandig
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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-
2.2 The Cell Line of Choice 763
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-
2 Alternative Strategies and New Cell Lines for High-level Production of Biopharmaceuticals 764
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 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
3.4 Fed-batch Process Development 789
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
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
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
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
4.2 Cell Line Construction and Selection 815
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-
4 Use of the Glutamine Synthetase (GS) Expression System 816
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-
4.2 Cell Line Construction and Selection 817
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
4 Use of the Glutamine Synthetase (GS) Expression System 818
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
4.4 Cell Engineering to Increase Productivity 819
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
4 Use of the Glutamine Synthetase (GS) Expression System 820
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
4.4 Cell Engineering to Increase Productivity 821
. . . 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 Use of the Glutamine Synthetase (GS) Expression System 822
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
4.6 Process Development 823
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 Use of the Glutamine Synthetase (GS) Expression System 824
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
4.6 Process Development 825
4 Use of the Glutamine Synthetase (GS) Expression System 826
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
4.6 Process Development 827
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
4 Use of the Glutamine Synthetase (GS) Expression System 828
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.
4.6 Process Development 829
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.
4 Use of the Glutamine Synthetase (GS) Expression System 830
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.
4 Use of the Glutamine Synthetase (GS) Expression System 832
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
HSA human serum albumin
IND investigational new drug
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-
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 834
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-
5.1 Introduction 835
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
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 836
5.1 Introduction 837
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
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 838
Table 5.1 (continued)
Production
strategy
Material used
for extraction
Host
organism
Institutions Product examples Reference
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
5.1 Introduction 839
Table 5.1 (continued)
Production
strategy
Material used
for extraction
Host
organism
Institutions Product examples Reference
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
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 840
Table 5.1 (continued)
Production
strategy
Material used
for extraction
Host
organism
Institutions Product examples Reference
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
5.1 Introduction 841
Table 5.1 (continued)
Production
strategy
Material used
for extraction
Host
organism
Institutions Product examples Reference
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,
Monoclonal 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
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 842
Table 5.1 (continued)
Production
strategy
Material used
for extraction
Host
organism
Institutions Product examples Reference
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
5.1 Introduction 843
Table 5.1 (continued)
Production
strategy
Material used
for extraction
Host
organism
Institutions Product examples Reference
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
Monoclonal antibodies,
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
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 844
Table 5.1 (continued)
Production
strategy
Material used
for extraction
Host
organism
Institutions Product examples Reference
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,
Monoclonal antibodies
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
of Biopharmaceuticals
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
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 846
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
5.2 Advantages and Disadvantages of Transgenic Systems for the Production of Biopharmaceuticals 847
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
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 848
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-
5.2 Advantages and Disadvantages of Transgenic Systems for the Production of Biopharmaceuticals 849
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
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 850
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-
5.2 Advantages and Disadvantages of Transgenic Systems for the Production of Biopharmaceuticals 851
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
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 852
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
.
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
) 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
(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
5.3 Commercial Biopharmaceuticals with Human Clinical Experience 855
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
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
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
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
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 860
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
5.3 Commercial Biopharmaceuticals with Human Clinical Experience 861
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
and
Humira
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
is ex-
pected to eliminate the decay-causing bac-
teria in 2 years.
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
5.3 Commercial Biopharmaceuticals with Human Clinical Experience 865
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-
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 866
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-
5.3 Commercial Biopharmaceuticals with Human Clinical Experience 867
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
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 868
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
)
[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
)
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.
5.3 Commercial Biopharmaceuticals with Human Clinical Experience 871
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-
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 872
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-
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 874
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
Part IV, Chapter 7).
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
5 Biopharmaceuticals Derived from Transgenic Plants and Animals 876
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
. 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
), 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
10.2 Drought and Salt Tolerance 973
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
10 Producing Biopharmaceuticals in the Desert 974
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
), 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-
10 Producing Biopharmaceuticals in the Desert 978
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
10.2 Drought and Salt Tolerance 979
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
10 Producing Biopharmaceuticals in the Desert 980
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-
10.3 High-temperature Stress 981
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
10 Producing Biopharmaceuticals in the Desert 982
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.
10.3 High-temperature Stress 983
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
10 Producing Biopharmaceuticals in the Desert 984
10.3 High-temperature Stress 985
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
10 Producing Biopharmaceuticals in the Desert 986
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
10.3 High-temperature Stress 987
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.
10.3.2.3 Transgenics Made for Altered
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
10 Producing Biopharmaceuticals in the Desert 988
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.
10.3.2.4 Transgenics Made for Altered
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.
10.3.2.5 Transgenics Made for Altered
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
to cause serious salinization of more than
50% of all arable land by the year 2050.
Conventional breeding programs for
raising abiotic stress-tolerant genotypes
10.4 Conclusions and Perspectives 989
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-
10 Producing Biopharmaceuticals in the Desert 990
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.
10 Producing Biopharmaceuticals in the Desert 992
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.
10 Producing Biopharmaceuticals in the Desert 994
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
CABG coronary artery bypass grafting
CD circular dichroism
CHO Chinese hamster ovary
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
LPS lipopolysaccharide
nvCJD new-variant Creutzfeld-Jakob
disease
PCR polymerase chain reaction
995
11
The First Biopharmaceutical from Transgenic Animals: ATryn
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.
11.1 Introduction 997
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
11 The First Biopharmaceutical from Transgenic Animals: ATryn
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-
11.2 Recombinant Production of AT 999
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
11 The First Biopharmaceutical from Transgenic Animals: ATryn
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
11.2 Recombinant Production of AT 1001
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
11 The First Biopharmaceutical from Transgenic Animals: ATryn
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-
11.3 Characterization of rhAT 1003
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
(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-
11 The First Biopharmaceutical from Transgenic Animals: ATryn
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
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
11.4 Preclinical Studies 1007
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-
11 The First Biopharmaceutical from Transgenic Animals: ATryn
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-
11.4 Preclinical Studies 1009
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-
11 The First Biopharmaceutical from Transgenic Animals: ATryn
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
11.5 Clinical Trials with rhAT 1011
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
11 The First Biopharmaceutical from Transgenic Animals: ATryn
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
11.5 Clinical Trials with rhAT 1013
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
11 The First Biopharmaceutical from Transgenic Animals: ATryn
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%.
11.5 Clinical Trials with rhAT 1015
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-
11 The First Biopharmaceutical from Transgenic Animals: ATryn
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.
11 The First Biopharmaceutical from Transgenic Animals: ATryn
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.
11 The First Biopharmaceutical from Transgenic Animals: ATryn
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
GMP good manufacturing practice
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
12.1 The Escherichia coli Expression Platform 1023
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
12 Producing Modern Biopharmaceuticals 1024
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
12.1 The Escherichia coli Expression Platform 1025
12 Producing Modern Biopharmaceuticals 1026
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.
12.2 The Saccharomyces cerevisiae Expression Platform 1027
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-
12 Producing Modern Biopharmaceuticals 1028
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
12.3 The HKB11 Expression Platform 1029
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
12 Producing Modern Biopharmaceuticals 1030
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.
12 Producing Modern Biopharmaceuticals 1032
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
ER endoplasmic reticulum
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
Asser Sloth Andersen and Ivan Diers
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
13.1 Introduction 1035
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-
13 Advanced Expression of Biopharmaceuticals in Yeast at Industrial Scale: The Insulin Success Story 1036
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
13.2 Design and Optimization of the Insulin Precursor Molecule 1037
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
13 Advanced Expression of Biopharmaceuticals in Yeast at Industrial Scale: The Insulin Success Story 1038
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
13.2 Design and Optimization of the Insulin Precursor Molecule 1039
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-
13 Advanced Expression of Biopharmaceuticals in Yeast at Industrial Scale: The Insulin Success Story 1040
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
13.3 Production of Insulin 1041
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).
13 Advanced Expression of Biopharmaceuticals in Yeast at Industrial Scale: The Insulin Success Story 1042
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).
13 Advanced Expression of Biopharmaceuticals in Yeast at Industrial Scale: The Insulin Success Story 1044
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
ORF open reading frame
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
Wilfried Weber and Martin Fussenegger
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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]).
14.4 Insect Cell Glycosylation and Glycoengineering 1047
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 Baculovirus-based Production of Biopharmaceuticals using Insect Cell Culture Processes 1048
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
14 Baculovirus-based Production of Biopharmaceuticals using Insect Cell Culture Processes 1050
14.7 Generic Protocol of Optimized Protein Production 1051
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-
14 Baculovirus-based Production of Biopharmaceuticals using Insect Cell Culture Processes 1052
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.
14 Baculovirus-based Production of Biopharmaceuticals using Insect Cell Culture Processes 1054
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].
14.8 Case study: Rapid Optimization of Expression Conditions and Large-scale Production of a (BTK) 1055
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.
14 Baculovirus-based Production of Biopharmaceuticals using Insect Cell Culture Processes 1056
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.
14.8 Case study: Rapid Optimization of Expression Conditions and Large-scale Production of a (BTK) 1057
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
14 Baculovirus-based Production of Biopharmaceuticals using Insect Cell Culture Processes 1058
Table 14.6 (continued)
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.
14.9 Conclusion 1059
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
14 Baculovirus-based Production of Biopharmaceuticals using Insect Cell Culture Processes 1060
Table 14.7 (continued)
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.
14 Baculovirus-based Production of Biopharmaceuticals using Insect Cell Culture Processes 1062
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
ATP adenosine triphosphate
CAT chloramphenicol acetyl-
transferase
CDP cytosine diphosphate
CTP cytidine triphosphate
DHFR dihydrofolate reductase
dNTP desoxynucleotide triphosphate
DTT dithiothreitol
EF elongation factor
GSH reduced glutathione
GSSG oxidized glutathiones
HEPES 2-[4-(2-hydroxyethyl)-1-pipera-
zinyl] ethanesulfonic acid
HTS high-throughput screening
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ISBN: 3-527-31184-X
PA plasminogen activator
PCR polymerase chain reaction
PDI protein disulfide isomerase
PEG polyethylene glycol
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
UTR untranslated region
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
15.1 Introduction 1065
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 Robust and Cost-effective Cell-free Expression of Biopharmaceuticals: Escherichia Coli and Wheat Embryo 1066
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.
15.2 Transcription 1067
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-
15 Robust and Cost-effective Cell-free Expression of Biopharmaceuticals: Escherichia Coli and Wheat Embryo 1068
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
15.3 Translational 1069
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
Initiation Factor 2 (IF2) His-tagged 40 lg mL
1
Initiation Factor 3 (IF3) His-tagged 15 lg mL
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-
15 Robust and Cost-effective Cell-free Expression of Biopharmaceuticals: Escherichia Coli and Wheat Embryo 1070
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.
15.3 Translational 1071
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
Magnesium acetate 14 mM
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
Magnesium acetate 5 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.
15 Robust and Cost-effective Cell-free Expression of Biopharmaceuticals: Escherichia Coli and Wheat Embryo 1072
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-
15.4 Treatment of Extracts for Synthesis of Disulfide-bonded Proteins 1073
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
15 Robust and Cost-effective Cell-free Expression of Biopharmaceuticals: Escherichia Coli and Wheat Embryo 1074
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
15.6 Reaction Conditions 1075
(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-
15 Robust and Cost-effective Cell-free Expression of Biopharmaceuticals: Escherichia Coli and Wheat Embryo 1076
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-
15.6 Reaction Conditions 1077
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-
15 Robust and Cost-effective Cell-free Expression of Biopharmaceuticals: Escherichia Coli and Wheat Embryo 1078
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
15.7 Conclusion 1079
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.
7 Kim, D.-M. and Swartz, J. R. (2001) Biotechnol.
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.
15 Robust and Cost-effective Cell-free Expression of Biopharmaceuticals: Escherichia Coli and Wheat Embryo 1080
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.
16 Kim, D.-M. and Swartz, J. R. (2004) Biotechnol.
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ISBN: 3-527-31184-X
16
Contract Manufacturing of Biopharmaceuticals
Including Antibodies or Antibody Fragments
J. Carsten Hempel and Philipp N. Hess
When Success Raises its Ugly Head
Outsourcing to Uncork the Capacity Bottleneck
Abbreviations
ATF alternating tangential filtration
CHO Chinese hamster ovary
CMC chemistry manufacturing and
control
CMO contract manufacturing organiza-
tion
CRO contract research organisation
DSP downstream processing
FDA Food and Drug Administration
GMP good manufacturing practice
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.
16.2 Expression Systems and Manufacturing Procedures 1085
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.
16 Contract Manufacturing of Biopharmaceuticals Including Antibodies or Antibody Fragments 1086
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
(infliximab)
sales >100 Mill$/year
Centocor, Inc. (subsidiary of Johnson &
Johnson)
August 1998
Herceptin
(trastuzumab)
sales >100 Mill$/year
Genentech, Inc./Roche September 1998
Mylotarg
TM
(gemtuzumab ozogamicin) Celltech Pharmaceuticals and Wyeth May 2000
Campath
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
1 Thirty Years of Monoclonal Antibodies: A Long Way to Pharmaceutical and Commercial Success 1112
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
1.3 Other Antibody Formats: Antibody Fragments 1113
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.)
1 Thirty Years of Monoclonal Antibodies: A Long Way to Pharmaceutical and Commercial Success 1114
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 Other Antibody Formats: Antibody Fragments 1115
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.
1 Thirty Years of Monoclonal Antibodies: A Long Way to Pharmaceutical and Commercial Success 1116
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
1.5 From Initial Failure to Success: Getting the Target Right 1117
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
(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
, 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
, 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
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
, 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-
1.7 Drug Targeting: The Next Generation in Cancer Treatment 1125
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]
(see Part IV, Chapter 16).
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 Thirty Years of Monoclonal Antibodies: A Long Way to Pharmaceutical and Commercial Success 1126
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
1.9 Routine Manufacture of MAbs 1127
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-
1 Thirty Years of Monoclonal Antibodies: A Long Way to Pharmaceutical and Commercial Success 1128
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.
1.9 Routine Manufacture of MAbs 1129
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
1 Thirty Years of Monoclonal Antibodies: A Long Way to Pharmaceutical and Commercial Success 1130
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-
1.9 Routine Manufacture of MAbs 1131
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
1 Thirty Years of Monoclonal Antibodies: A Long Way to Pharmaceutical and Commercial Success 1132
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
1.10 Glycosylation and Other Post-translational Modifications 1133
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
1 Thirty Years of Monoclonal Antibodies: A Long Way to Pharmaceutical and Commercial Success 1134
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
(see Part IV, Chapter 11).
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-
1.11 Emerging Issues in MAb Production 1135
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
1 Thirty Years of Monoclonal Antibodies: A Long Way to Pharmaceutical and Commercial Success 1136
in any case [204, 205]. Cytotoxic antitumor
MAbs such as Herceptin
. 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.
1 Thirty Years of Monoclonal Antibodies: A Long Way to Pharmaceutical and Commercial Success 1138
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
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
%
2 Modern Antibody Technology: The Impact on Drug Development 1152
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-
2 Modern Antibody Technology: The Impact on Drug Development 1154
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
2 Modern Antibody Technology: The Impact on Drug Development 1156
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-
2 Modern Antibody Technology: The Impact on Drug Development 1158
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-
2 Modern Antibody Technology: The Impact on Drug Development 1160
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 Modern Antibody Technology: The Impact on Drug Development 1162
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-
2.4 Reaching the Target: The Importance of Specificity, Affinity and Format 1163
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
2 Modern Antibody Technology: The Impact on Drug Development 1164
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),
2.4 Reaching the Target: The Importance of Specificity, Affinity and Format 1165
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
2 Modern Antibody Technology: The Impact on Drug Development 1166
(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
2.4 Reaching the Target: The Importance of Specificity, Affinity and Format 1167
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.
2 Modern Antibody Technology: The Impact on Drug Development 1168
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.
2.5 Exerting an Effect at the Target 1169
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-
2 Modern Antibody Technology: The Impact on Drug Development 1170
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
2.5 Exerting an Effect at the Target 1171
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-
2 Modern Antibody Technology: The Impact on Drug Development 1172
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
macrophage colony-stimulating factor
(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
2.5 Exerting an Effect at the Target 1173
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-
2 Modern Antibody Technology: The Impact on Drug Development 1174
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,
2.6 Antibodies in their Natural Habitat: Infectious Diseases 1175
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
2 Modern Antibody Technology: The Impact on Drug Development 1176
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
2.8 Future Directions and Concluding Statements 1177
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.
2 Modern Antibody Technology: The Impact on Drug Development 1178
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.
Methods Mol. Biol. 2002, 178, 137.
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.
2 Modern Antibody Technology: The Impact on Drug Development 1180
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.
J. Immunol. Methods 1999, 231, 249260.
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.
2 Modern Antibody Technology: The Impact on Drug Development 1182
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
2 Modern Antibody Technology: The Impact on Drug Development 1184
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.
2 Modern Antibody Technology: The Impact on Drug Development 1186
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
Constanze Breithaupt
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
3.1 Autoantibodies in Autoimmune Diseases 1189
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
3 Molecular Characterization of Autoantibody Responses in Autoimmune Diseases 1190
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
3.2 Autoantibody Epitopes 1191
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
3 Molecular Characterization of Autoantibody Responses in Autoimmune Diseases 1192
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-
3.2 Autoantibody Epitopes 1193
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
3 Molecular Characterization of Autoantibody Responses in Autoimmune Diseases 1194
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-
3.3 Visualization of Epitopes 1195
3 Molecular Characterization of Autoantibody Responses in Autoimmune Diseases 1196
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 Visualization of Epitopes 1197
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
3 Molecular Characterization of Autoantibody Responses in Autoimmune Diseases 1198
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.
3.4 Structural Characterization of AutoantibodyAutoantigen Complexes 1199
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
3 Molecular Characterization of Autoantibody Responses in Autoimmune Diseases 1200
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
3.4 Structural Characterization of AutoantibodyAutoantigen Complexes 1201
(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
3 Molecular Characterization of Autoantibody Responses in Autoimmune Diseases 1202
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
3.4 Structural Characterization of AutoantibodyAutoantigen Complexes 1203
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
3 Molecular Characterization of Autoantibody Responses in Autoimmune Diseases 1204
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.
3.5 Conclusions 1205
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-
3 Molecular Characterization of Autoantibody Responses in Autoimmune Diseases 1206
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.
Natl. Acad. Sci. USA 93, 712.
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.
Proc. Natl. Acad. Sci. USA 96, 1448714492.
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.
Natl. Acad. Sci. USA 87, 98489852.
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.
3.5 Conclusions 1207
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-
3 Molecular Characterization of Autoantibody Responses in Autoimmune Diseases 1208
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.
3.5 Conclusions 1209
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
CEA carcinoembryonic antigen
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
MRI magnetic resonance imaging
NIH National Institute of Health
NIR near infrared
NMR nuclear magnetic resonance
PET positron emission tomogra-
phy
1211
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ISBN: 3-527-31184-X
4
Molecular Imaging and Applications for Pharmaceutical R&D
Joke G. Orsel and Tobias Schaeffter
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
VEGF vascular endothelial growth
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 Molecular Imaging and Applications for Pharmaceutical R&D 1212
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-
4.2 Imaging Modalities and Contrast Agents 1213
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-
4 Molecular Imaging and Applications for Pharmaceutical R&D 1214
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
4.2 Imaging Modalities and Contrast Agents 1215
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
4 Molecular Imaging and Applications for Pharmaceutical R&D 1216
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
:
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-
4 Molecular Imaging and Applications for Pharmaceutical R&D 1226
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 Molecular Imaging 1227
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-
4 Molecular Imaging and Applications for Pharmaceutical R&D 1228
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-
4.3 Molecular Imaging 1229
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-
4 Molecular Imaging and Applications for Pharmaceutical R&D 1230
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-
4.4 Molecular Imaging for Drug Discovery and Development 1231
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
4 Molecular Imaging and Applications for Pharmaceutical R&D 1232
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
4.4 Molecular Imaging for Drug Discovery and Development 1233
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-
4 Molecular Imaging and Applications for Pharmaceutical R&D 1234
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
4.4 Molecular Imaging for Drug Discovery and Development 1235
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
4 Molecular Imaging and Applications for Pharmaceutical R&D 1236
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
4.4 Molecular Imaging for Drug Discovery and Development 1237
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
4 Molecular Imaging and Applications for Pharmaceutical R&D 1238
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
modern biopharmaceuticals.
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.
4.5 Concluding Remarks 1239
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.,
Proc. Natl. Acad. Sci. USA 2000, 97, 2767
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.
4 Molecular Imaging and Applications for Pharmaceutical R&D 1240
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
ATP adenosine triphosphate
ATSM diacetyl-bis(N4-methylthiosemi-
carbazone)
BrdU bromodeoxy uridine
CDK cyclin-dependent kinase
CEA carcinoembryonic antigen
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
FGF fibroblast growth factor
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
PEG polyethylene glycol
PET positron emission tomography
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
5.1 What is Positron Emission Tomography? 1245
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
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].
8 Development of Multi-marker-based Diagnostic Assays with the ProteinChip
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
8.2 Proteins are Best Choice Again 1327
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
8 Development of Multi-marker-based Diagnostic Assays with the ProteinChip
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
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
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-
8 Development of Multi-marker-based Diagnostic Assays with the ProteinChip
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
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).
8.6 The Pattern Track
TM
Process: From Biomarker Discovery to Assay Development 1335
8 Development of Multi-marker-based Diagnostic Assays with the ProteinChip
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
8.7 Protein Variants as Disease Markers 1337
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-
8 Development of Multi-marker-based Diagnostic Assays with the ProteinChip
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.
8 Development of Multi-marker-based Diagnostic Assays with the ProteinChip
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
Jrg Ingo Baumbach, Wolfgang Vautz, Vera Ruzsanyi, and Lutz Freitag
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
9.2 Material and Methods: IMS 1345
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.
9 Early Detection of Lung Cancer: Metabolic Profiling of Human Breath 1346
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
9.3 Results and Discussion 1347
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
9 Early Detection of Lung Cancer: Metabolic Profiling of Human Breath 1348
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]).
9.4 Clinical Study 1349
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-
9 Early Detection of Lung Cancer: Metabolic Profiling of Human Breath 1350
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-
9.4 Clinical Study 1351
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
9 Early Detection of Lung Cancer: Metabolic Profiling of Human Breath 1352
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.
9.4 Clinical Study 1353
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-
9 Early Detection of Lung Cancer: Metabolic Profiling of Human Breath 1354
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-
9.5 Conclusions 1355
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.
9 Early Detection of Lung Cancer: Metabolic Profiling of Human Breath 1356
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.
9 Early Detection of Lung Cancer: Metabolic Profiling of Human Breath 1358
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
Gesine E. Hildebrand and Stephan Harnisch
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
BSA bovine serum albumin
DDS drug delivery systems
G-CSF granulocyte colony-stimulating
factor
GRAS generally regarded as safe
HGR human growth hormone
HIV human immunodeficiency virus
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
TNF tumor necrosis factor
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
1 Advanced Drug Delivery Systems for Biopharmaceuticals 1362
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
of Biopharmaceuticals
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
1.2 Challenges for the Administration of Biopharmaceuticals 1363
1 Advanced Drug Delivery Systems for Biopharmaceuticals 1364
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.
1.2 Challenges for the Administration of Biopharmaceuticals 1365
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 Advanced Drug Delivery Systems for Biopharmaceuticals 1366
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
1.3 Drug Delivery Strategies 1367
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
1 Advanced Drug Delivery Systems for Biopharmaceuticals 1368
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.
1.3 Drug Delivery Strategies 1369
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).
1 Advanced Drug Delivery Systems for Biopharmaceuticals 1370
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.
1.3 Drug Delivery Strategies 1371
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-
1 Advanced Drug Delivery Systems for Biopharmaceuticals 1372
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
1.3 Drug Delivery Strategies 1373
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
1 Advanced Drug Delivery Systems for Biopharmaceuticals 1374
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
1.3 Drug Delivery Strategies 1375
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
1 Advanced Drug Delivery Systems for Biopharmaceuticals 1376
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
1.3 Drug Delivery Strategies 1377
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-
1 Advanced Drug Delivery Systems for Biopharmaceuticals 1378
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-
1.3 Drug Delivery Strategies 1379
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
1 Advanced Drug Delivery Systems for Biopharmaceuticals 1380
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-
1.3 Drug Delivery Strategies 1381
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,
1 Advanced Drug Delivery Systems for Biopharmaceuticals 1382
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]).
1.3 Drug Delivery Strategies 1383
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-
1 Advanced Drug Delivery Systems for Biopharmaceuticals 1384
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
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
2.2 The Polymer 1395
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
2 Poly(ethylene) Glycol Conjugates of Biopharmaceuticals in Drug Delivery 1396
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
2.4 PEG Reagents and Conjugation 1397
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].
2 Poly(ethylene) Glycol Conjugates of Biopharmaceuticals in Drug Delivery 1398
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
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.
2.4 PEG Reagents and Conjugation 1399
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.
2 Poly(ethylene) Glycol Conjugates of Biopharmaceuticals in Drug Delivery 1400
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
versus
PEG-Intron
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
2 Poly(ethylene) Glycol Conjugates of Biopharmaceuticals in Drug Delivery 1402
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.
2.5 Biopharmaceutical Conjugates 1403
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
(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-
2 Poly(ethylene) Glycol Conjugates of Biopharmaceuticals in Drug Delivery 1404
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-
2.5 Biopharmaceutical Conjugates 1405
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
(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-
2 Poly(ethylene) Glycol Conjugates of Biopharmaceuticals in Drug Delivery 1406
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
2.6 PEGylation of Peptides 1407
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
2 Poly(ethylene) Glycol Conjugates of Biopharmaceuticals in Drug Delivery 1408
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
2.7 Formulations of PEGylated Biopharmaceuticals 1409
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
2 Poly(ethylene) Glycol Conjugates of Biopharmaceuticals in Drug Delivery 1410
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) 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-
2.8 Analysis of PEG-conjugates 1411
Scheme 1
2 Poly(ethylene) Glycol Conjugates of Biopharmaceuticals in Drug Delivery 1412
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-
2.8 Analysis of PEG-conjugates 1413
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-
2 Poly(ethylene) Glycol Conjugates of Biopharmaceuticals in Drug Delivery 1414
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,
J. Immunol. Methods 81, 245257, 1985.
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
) 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.
2 Poly(ethylene) Glycol Conjugates of Biopharmaceuticals in Drug Delivery 1418
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
Karen Lingnau, Christoph Klade, Michael Buschle, and Alexander von Gabain
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
HIV human immunodeficiency virus
HSV herpes simplex virus
IFA incomplete Freunds adjuvant
IFN interferon
IL interleukin
LDH lactate dehydrogenase
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
3 Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems 1420
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.
3.1 Vaccines and their Importance in the Fight against Human Diseases 1421
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-
3 Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems 1422
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 Adjuvants: An Overview 1423
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
3 Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems 1424
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
3.2 Adjuvants: An Overview 1425
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
3 Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems 1426
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
3.3 Cationic Peptides as Novel Vaccine Adjuvants 1427
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
3 Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems 1428
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
3.3 Cationic Peptides as Novel Vaccine Adjuvants 1429
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-
3 Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems 1430
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
3.3 Cationic Peptides as Novel Vaccine Adjuvants 1431
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-
3 Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems 1432
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
3.4 Cationic Antimicrobial Peptides (CAMP) as Novel Adjuvants 1433
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-
3 Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems 1434
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
3.4 Cationic Antimicrobial Peptides (CAMP) as Novel Adjuvants 1435
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
3 Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems 1436
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
3.5 Cationic Peptide Delivery Systems in Combination with Other Adjuvants 1437
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
3 Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems 1438
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.
3.5 Cationic Peptide Delivery Systems in Combination with Other Adjuvants 1439
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
3 Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems 1440
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.
Proc. Natl. Acad. Sci. USA 97, 1300313008.
19 Fuchs, S. M. and Raines, R. T. (2004) Pathway
for polyarginine entry into mammalian cells.
Biochemistry 43, 24382444.
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
3 Novel Vaccine Adjuvants Based on Cationic Peptide Delivery Systems 1442
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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
using a Novel RapidMist
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
4 The Evolving Role of Oralin
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.
4.3 The Benefits of Oralin
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
4 The Evolving Role of Oralin
TM
(Oral Spray Insulin) in the Treatment of Diabetes 1448
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
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
4 The Evolving Role of Oralin
TM
(Oral Spray Insulin) in the Treatment of Diabetes 1454
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
4.4 The Preparation and Pharmaceutical Properties of Oralin
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
4 The Evolving Role of Oralin
TM
(Oral Spray Insulin) in the Treatment of Diabetes 1456
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
4.5 Phase II, Long-term Safety and Efficacy Study 1457
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
4 The Evolving Role of Oralin
TM
(Oral Spray Insulin) in the Treatment of Diabetes 1458
4.5 Phase II, Long-term Safety and Efficacy Study 1459
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
4 The Evolving Role of Oralin
TM
(Oral Spray Insulin) in the Treatment of Diabetes 1460
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.
19 Guevara-Aguirre J, Guevara M, Saavedra J,
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).
21 Guevara-Aguirre J, Guevara M, Saavedra J,
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
AUC area under the curve
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
FITC fluorescein isothiocyanate
LA linoleic acid
LDH lactate dehydrogenase
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
Modern Biopharmaceuticals. Edited by J. Knblein
Copyright 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
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-
5.1 Improvement of Peptide and Protein Absorption 1465
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.