Sie sind auf Seite 1von 10

Biomaterials 21 (2000) 431}440

Review: tissue engineering for regeneration of articular cartilage


Johnna S. Temeno! !, Antonios G. Mikos!,",*
!Department of Bioengineering, Rice University, 6100 Main, Houston, TX 77005-1892, USA
"Department of Chemical Engineering, Rice University, 6100 Main, Houston, TX 77005-1892, USA
Received 27 July 1999; accepted 30 September 1999

Abstract

Joint pain due to cartilage degeneration is a serious problem, a!ecting people of all ages. Although many techniques, often surgical,
are currently employed to treat this a%iction, none have had complete success. Recent advances in biology and materials science have
pushed tissue engineering to the forefront of new cartilage repair techniques. This review seeks to condense information for the
biomaterialist interested in developing materials for this application. Articular cartilage anatomy, types of injury, and current repair
methods are explained. The need for biomaterials, current commonly used materials for tissue-engineered cartilage, and consider-
ations in scale-up of cell}biomaterial constructs are summarized. ( 2000 Elsevier Science Ltd. All rights reserved.

Keywords: Cartilage repair; Cartilage regeneration; Cartilage tissue engineering; Cell transplantation; Bioreactors; Chondrogenesis

1. Introduction to compression and the ability to distribute loads that


cannot be fully replaced by any other tissue or device
Joint pain is a major cause of disability in middle-aged designed to date [5,6].
and older people [1]. Pain usually results from degener- This review will provide an overview of cartilage anat-
ation of the joint's cartilage due to primary osteoarthritis omy and types of injury, as well as focus on current
or from trauma causing loss of cartilage [1]. As many as strategies for cartilage repair in the knee joint, including
36 million Americans su!er from some form of arthritis promising new strategies involving tissue engineering.
[2] and cartilage damage from sports injury is also com- The knee, with two distinct articulating surfaces is a con-
mon. It has been shown that up to 16% of injuries to the dylar-type joint. Both bones, the femur and the tibia, that
knee cause intra-articular bleeding [3,4]. Since cartilage comprise the joint are covered with a layer of hyaline
shows very little tendency for self-repair, these injuries cartilage at the joint surface. Immediately below this is
are maintained for years and can eventually lead to the periosteal covering of the bone. The synovial mem-
further degeneration (secondary osteoarthritis) [4]. brane encircles the joint, thereby forming a barrier to
Given the debilitating nature of severe joint pain, retain the synovial #uid in the knee. This #uid provides
scientists and surgeons have tried for decades to repair or lubrication and nutrients for the cartilage since no blood
regenerate lost cartilage, but there has been little success vessels penetrate into the tissue from the subchondral
due to the complex properties of the tissue and its essen- bone [7].
tial function in the body. Hyaline cartilage, the type of Although only hyaline cartilage is found in the knee,
cartilage found in joints, provides stable movement with two other kinds, "brocartilage and elastic cartilage, are
less friction than any prosthetic replacement, and can seen in other areas of the body. All three are composed of
alter its properties in response to di!erences in loading. chondrocytes and extracellular matrix macromolecules.
Although appearing to be a simple, avascular matrix, Elastic cartilage forms the ear and nose and is character-
hyaline cartilage possesses properties such as resistance ized by the presence of elastin in the extracellular matrix
(ECM). Fibrocartilage has a higher proportion of col-
lagen in the ECM than hyaline cartilage and is found at
the ends of tendons and ligaments in apposition to bone.
* Correspondence address: Department of Bioengineering, Rice Uni-
versity, 6100 Main, MS 142, Houston, TX 77005-1892, USA. Tel.:
Hyaline cartilage has a white, glassy appearance, and
#713-285-5355; fax: #713-285-5353. unlike "brocartilage, shows no macroscopic evidence of
E-mail address: mikos@rice.edu (A.G. Mikos) "bers [8].

0142-9612/00/$ - see front matter ( 2000 Elsevier Science Ltd. All rights reserved.
PII: S 0 1 4 2 - 9 6 1 2 ( 9 9 ) 0 0 2 1 3 - 6
432 J.S. Temenow, A.G. Mikos / Biomaterials 21 (2000) 431}440

2. Composition and anatomy of articular cartilage another sugar. At least one component of the disaccharide
has a negatively charged sulfate or carboxylate group, so
2.1. Chondrocytes the GAGs tend to repel each other and other anions while
attracting cations and facilitating interaction with water.
In humans, chondrocytes represent only about 1% of Hyaluronic acid, chondroitin sulfate, keratan sulfate, der-
the volume of hyaline cartilage but are essential since it is matan sulfate and heparan sulfate are some of the GAGs
these cells that replace degraded matrix molecules to generally found in articular cartilage [5,10,11].
maintain the correct size and mechanical properties of There are both large aggregating monomers and
the tissue. Thus, microscopically, the cells' endoplasmic smaller proteoglycans present in articular cartilage. The
reticulum and Golgi apparatus are prominent. In addi- aggregating proteoglycans, or aggregans, are composed
tion, many contain lipid and glycogen stores and of monomers with keratan sulfate and chondroitin sul-
secretory vesicles. Some chondrocytes have cilia that fate GAGs attached to the protein core. In most aggre-
extend from the cell into the ECM and are believed to gan molecules, link proteins connect many (up to 300) of
play a role in sensing the mechanical environment of the these monomers to a hyaluronic acid chain. Aggregans
cell, since chondrocytes are known to modify matrix "ll most of the inter"brillar space of the ECM and are
properties in response to loading [5]. thought to be responsible for much of the resilience and
Chondrocytes originate from mesenchymal stem cells stress distribution in articular cartilage through their
(MSCs) found in the bone marrow in mature individuals. ability to attract water. There are no chemical bonds
During embryogenesis, the MSCs start to di!erentiate between the proteoglycans and collagen "bers; aggrega-
into chondrocytes and secrete a cartilaginous matrix. tion prevents di!usion of the proteoglycans out of the
During this time, the cells continue to divide. They pass matrix during joint loading [5,6,10].
through various lineage states and eventually the chon- The smaller proteoglycans include decorin, biglycan
drocytes in the central zone, located next to what will and "bromodulin. They have shorter protein cores and
soon be bone, enter the "nal stage of development and fewer GAG chains than their larger counterparts. Unlike
become hypertrophic chondrocytes, producing proteins aggregans, these molecules do not a!ect physical proper-
that are important in calci"cation of the matrix. Other ties of the tissue, but are thought to play a role in cell
chondrocytes (on the periphery) secrete collagen and function and organization of the collagen matrix [5].
matrix molecules in the right proportions to produce
hyaline cartilage. Mature articular chondrocytes, unable 2.2.3. Noncollagenous proteins
to proliferate, appear rounded and are completely en- In contrast to proteoglycans, glycoproteins have only
cased in matrix [5,9]. a small amount of oligosaccharide associated with the
protein core. These polypeptides help to stabilize the
2.2. Extracellular matrix ECM matrix and aid in chondrocyte}matrix interac-
tions. Both anchorin CII and cartilage oligomeric protein
2.2.1. Collagen anchor chondrocytes to the surrounding matrix. Other
Collagen types II, VI, IX, X, and XI are found in noncollagenous proteins commonly found in most tis-
articular cartilage, although type II accounts for 90}95% sues, such as "bronectin and tenascin, are also observed
of the collagen in the matrix. Type II collagen has a high in articular cartilage and are believed to perform similar
amount of bound carbohydrate groups, allowing more functions as the glycoproteins [5].
interaction with water than some other types. Types IX
and XI, along with type II, form "brils that interweave to 2.2.4. Tissue yuid
form a mesh. This organization provides tensile strength Tissue #uid is an essential part of hyaline cartilage,
as well as physically entrapping other macromolecules. comprising up to 80% of the wet weight of the tissue. In
Although the exact function of types IX and XI are addition to water, the #uid contains gases, metabolites and
unknown, type IX has been observed to bind super"cially a large amount of cations to balance the negatively
to the "bers and extending into the inter-"ber space to charged GAG's in the ECM. It is the exchange of this #uid
interact with other type IX molecules, possibly acting to with the synovial #uid that provides nutrients and oxygen
stabilize the mesh structure. Type X is found only near to the avascular cartilage. In addition, the entrapment of
areas of the matrix that are calci"ed [6,10]. this #uid though interaction with ECM components pro-
vides the tissue with its ability to resist compression and
2.2.2. Proteoglycans return to normal shape after deformation [5,10].
Proteoglycans are composed of about 95% polysac-
charide and about 5% protein. The protein core is asso- 2.3. Zones of organization
ciated with one or more varieties of glycosaminoglycan
(GAG) chains. GAG chains are unbranched polysaccha- Articular cartilage can be divided into four zones
rides made from disaccharides of an amino sugar and (super"cial, transitional, middle and calci"ed) based on
J.S. Temenow, A.G. Mikos / Biomaterials 21 (2000) 431}440 433

di!erences in matrix morphology and biochemistry. In now aligned perpendicular to the joint surface. The "bers
each zone, there are three distinct regions: the pericellular extend into the tidemark, a basophilic line of unknown
region, the territorial region and the interterritorial re- composition that indicates the beginning of calci"ed tis-
gion. The pericellular and territorial regions provide sue [5,10].
means for chondrocyte attachment to the ECM and
protection of the cells during loading. The pericellular 2.3.4. Calcixed cartilage zone
region contains almost exclusively noncollagenous bind- The thin calci"ed cartilage zone lies closest to the
ing proteins. Part of the territorial region is composed of subchondral bone and acts as a transition from soft
collagen "brils that follow the pericellular envelope but hyaline cartilage to bone. Signi"cant shear stress can be
may extend to encompass two or more chondrocytes. produced in this area due to the interface between the soft
Farther away from the cells, the territorial collagen is less cartilage and much sti!er bone [6]. The chondrocytes
aligned and "brils cross each other to form a basket here are smaller, with almost no endoplasmic reticulum.
structure around the chondrocyte(s). The interterritorial In some places, they seem to be completely surrounded
region is marked by an increase in "bril diameter and by calci"ed ECM, indicating that they have very little
orientation of "bers in a more parallel fashion. This metabolic activity [5,10].
region is primarily responsible for the mechanical prop-
erties of the tissue [5].
3. Articular cartilage injury
2.3.1. Superxcial zone
The super"cial zone is the thinnest zone and is made of There are three main types of cartilage injury: matrix
two distinct layers. An acellular sheet of predominantly disruption, partial thickness defects, and full thickness
collagen "bers (the lamina splendens) covers the joint. defects. Matrix disruption occurs from blunt trauma,
Deep to this, the second layer is composed of #attened such as dashboard injuries in automobile accidents. The
chondrocytes with their long axes parallel to the articular ECM is damaged, but if the injury is not extreme, the
surface. The ECM in this area has more collagen and less remaining viable chondrocytes will increase their syn-
proteoglycan than the other zones. There are also large thetic activity to repair the tissue. Partial thickness
amounts of "bronectin and water. This combination of defects demonstrate disruption of the cartilage surface
molecules imparts more tensile strength to this area of ("ssures, etc.) but this does not extend to the subchondral
the matrix, which could be useful in resisting shear from bone. Immediately following the injury, nearby cells be-
the articulating surfaces. The super"cial zone also is gin to proliferate, but for reasons that remain unclear,
important for the compressive strength of the tissue and cellular attempts to "ll the defect cease before it is re-
possibly in isolation of cartilage from the immune system paired. Full thickness defects arise from damage that
[5,10]. transverses the entire cartilage thickness and penetrates
the subchondral bone. In this case, the defect is "lled
with a "brin clot and a classic wound healing response
2.3.2. Transitional zone ensues. With this type of injury, unlike the others,
Occupying more volume than the super"cial zone, the there is access to a population of progenitor cells from
transitional zone includes chondrocytes that are spheri- the bone marrow which can migrate to "ll the defect
cal and contain synthetic organelles such as the endo- [13,14]. These cells usually cause replacement of the
plasmic reticulum, Golgi bodies and mitochondria. The "brin clot with tissue intermediate between hyaline
ECM in this area has larger collagen "brils, more pro- and "brocartilage. This tissue is usually less sti! and
teoglycan and less collagen and water than in the pre- more permeable than native cartilage, which could
vious zone. The interterritorial "brils are aligned contribute to its eventual degradation over a period of
obliquely or randomly to the articular surface, as op- months [14].
posed to parallel in the super"cial zone [5,10]. Several factors may explain why cartilage demon-
strates very little capability for self-repair after injury.
2.3.3. Middle (radial) zone Chondrocytes are not required to proliferate to maintain
The middle zone is usually the biggest and has the cartilage, as in some other tissues (i.e. skin), and mature
largest diameter collagen "brils, the most proteoglycan, chondrocytes have a relatively low metabolic activity [5].
and the least water. The cells are rounded, like in the Except in the case of full-thickness defects, there is no
transitional zone, but are stacked in columns perpendicu- direct access to progenitor cells (such as are found in
lar to the articulating surface. The territorial region of the bone marrow) and the resident chondrocytes may be
matrix encloses each column [5,10]. These cells show impeded by the ECM from migrating to "ll the defect
high synthetic activity*ten times that of super"cial zone [13,15]. In addition, the proteoglycans in the ECM can
chondrocytes [12]. Also, in this area, the orientation of prevent cell adhesion, further undermining the native
the interterritorial "bers changes again so that they are repair process [15,16].
434 J.S. Temenow, A.G. Mikos / Biomaterials 21 (2000) 431}440

4. Treatments for articular cartilage defects chondrogenesis was observed using periosteal grafts than
perichondrial grafts [4]. Furthermore, there was up to
4.1. Symptom management 70% failure of these grafts at "ve years due to ossi"cation
of the perichondrial-derived tissue [3,18]. Although these
Due to the di$culties in natural repair of cartilage, procedures have proven to be e!ective for some patients,
clinicians have turned to other methods to return func- and remove the problem faced in other autografts of
tionality of the joint. If cartilage damage is severe, symp- tissue damage due to alteration of position in the joint,
toms are managed using resection or other surgical they do result in donor site morbidity and can require
techniques. Resection of articular cartilage is often fol- a second incision to obtain the graft tissue [4].
lowed by implantation of a prosthesis in the joint [4].
Relief of pain can also be provided by an osteotomy, 4.2.2. Cartilage regeneration
which removes part of one of the bones in an articulating In an e!ort to eliminate the need for a donor site, and
surface to decrease loading of the joint. This also gives because of concerns associated with allogeneic and auto-
the surgeon an opportunity to bring regions of the joint geneic implants, many attempts have been made to heal
with remaining cartilage in contact with each other or to or regenerate existing cartilage, rather than replace it.
correct misalignments due to previous degeneration [1]. The proposed techniques have focused on either enhanc-
These techniques are often employed after other methods ing the intrinsic regenerative properties of the tissue or
have failed [4]. transplanting extra chondrocytes to form more tissue.
Unfortunately, neither of these has been completely suc-
4.2. Cartilage restoration cessful, especially in older populations [1,4].

In patients with smaller lesions, attempts are made to 4.2.2.1. Regeneration enhancement. The most common
restore cartilage to the joint surface through a variety of treatment of cartilage degeneration is to penetrate the
methods. These include implanting replacement tissue subchondral bone, either by abrasion or drilling. This, in
grafts, or employing techniques that encourage the native essence, creates a full-thickness defect. A clot forms over
repair process [4]. the bone surface which, if left unloaded, can provide
a sca!old for migration of MSCs and their eventual
4.2.1. Tissue grafts di!erentiation into chondrocytes and osteocytes [19].
Cartilage can be replaced either with small plugs from However, results of this method are mixed: the tissue
a less weight bearing region of the joint (autografts) or formed ranges from no cartilage to "brocartilage to hya-
with a full allograft. Autografts are limited by the small line cartilage, depending on the patient. Even when hya-
amount of cartilage available in the body for transplanta- line-like cartilage is created, the mechanical properties
tion to other sites. Usually, this cartilage is taken from and durability are less than that of the original tissue.
the patella, the femoral condyle or the proximal "bula Due to the large variability in outcomes, the best that can
[1,4,17]. While this procedure does seem to decrease pain be concluded is that this method has some chance of
in many patients (70% good results at two to "ve years helping and little chance of harming the patient [1,3,4].
[11]), the e!ects of damage to the donor site must be Less invasive techniques to restore cartilage include
considered. There is also concern that tissue from a less treatment with lasers or electrical stimulation, decreased
weight bearing region will not be able to withstand the joint loading in conjunction with continuous passive
forces imparted at the joint surface [15]. Allografts have motion, or the use of pharmacological agents to stimu-
resulted in a 95% survival rate at "ve years, making them late chondrocytic activity (see [4] for a more complete
a useful treatment. However, fresh grafts have been seen review of each technique). Passive motion shows con-
to induce an immune response once implanted [1,4,17]. siderable regeneration for small defects (less than 3 mm),
Frozen grafts may reduce this response as well as provide and has been found to enhance other procedures for
more time for screening of the donor tissue for pathogens, cartilage healing [1,4,14]. Injections of various pharma-
but freezing can reduce tissue viability [15]. cologics, such as growth factors (transforming growth
Another option is the use of periosteal or perichondrial factor-b1, insulin-like growth factor-1 and bone mor-
grafts. For the former, periosteal tissue is taken from an phogenic proteins) and corticosteroids, have been attem-
adjacent area and placed in the cartilage defect with the pted as a means to increase chondrocyte proliferation
belief that the undi!erentiated periosteal cells will be and matrix deposition. However, corticosteroids have
induced to form chondrocytes in their new environment. produced ambiguous results and growth factors can also
This technique has been used for more than a decade in stimulate the formation of osteophytes that can be quite
the United States with somewhat encouraging results, painful. Therefore, more research should be done to
depending on exact placement of the periosteal tissue in determine which pharmacologics should be administered
the defect and other factors [4,8]. A similar procedure and in what doses before this treatment will be e!ec-
can be done with perichondrium from the rib, but more tive [4].
J.S. Temenow, A.G. Mikos / Biomaterials 21 (2000) 431}440 435

4.2.2.2. Cell transplantation Cells in scawolds. Alternatively, instead of tissue #aps,


Cell suspensions. Another option is transplantation of highly porous sca!olds may be used to maintain di!eren-
chondrocytes or undi!erentiated cells to restore lost tis- tiated cells in a given area. This design is optimal because
sue mass. An optimal approach would be to take a small it signi"cantly reduces donor site morbidity and, in addi-
biopsy of cells (progenitor or di!erentiated chondrocytes) tion to simply providing a boundary for retention of cells,
expand the number of cells, and then return them to the the sca!old also acts as a substrate to which the anchor-
defect once the proper mass has been generated. Im- age-dependent chondrocytes can adhere [28]. When cul-
plantation of individual chondrocytes has been re- tured two-dimensionally, such as in Petri dishes, these
searched since 1968, but the success rate has been less cells dedi!erentiate, assuming a more #attened appear-
than 40%, due to the problem of retaining the cells in the ance and producing Type I instead of Type II collagen.
defect for a length of time that would allow them to begin However, when grown in three dimensions, chondrocytes
to produce matrix [20,21]. maintain their di!erentiated phenotype and function. In
However, recently, a surgical method has been de- this way, sca!olds can encourage the proliferation of
veloped which uses a periosteal #ap sutured over the chondrocytes without sacri"cing important functions to
defect as a barrier beneath which cultured autologous dedi!erentiation [29]. Sca!olds have also been used
chondrocytes are injected. The cells remain in the defect without seeded chondrocytes as a technique for regenera-
and this technique has shown promise in human trials tion enhancement by encouraging cell migration [30,31],
conducted over the past 12 years in Sweden and the but these studies are not reviewed here. Please refer to the
United States. A small biopsy of autologous tissue is recent article by Hunziker [15] for further treatment of
obtained during arthroscopic evaluation of the joint and this subject.
is cultured for approximately three weeks, resulting in Various sca!old materials have been tested, including
a 10}12-fold increase in cell number. During the surgical both naturally derived and synthetic polymers. (For
procedure, the periosteum (cambium layer toward the more complete listings, see [11,14,15]). This review will
defect) is attached with sutures and "brin glue to healthy not discuss all of these materials, but rather focus on
cartilage near the lesion and the cultured cells are injec- those that have been most extensively used in tissue-
ted beneath it [18,22,23]. engineered constructs. Although recent work with hy-
In the initial study, 14 of 16 patients with defects in the aluronic acid-based carriers is promising [32,33], the
femoral condyle had good results after 2 yr, while healing natural polymer that has received the most attention is
in patellar defects was less encouraging (two of seven had collagen. In 1983, it was found that chondrocytes main-
good outcomes) [22]. Since that time, further studies tain di!erentiated phenotype and GAG production for
have corroborated these results. A US study begun in six weeks in collagen gels [34]. Since that time, re-
1995 reports graft failure in only 7% of 70 patients and searchers have continued to show encouraging results
indicates that this procedure may hold promise to repair from use of these sca!olds. A key advantage is that,
joints with multiple lesions [3]. In a long-term study, 30 because collagen can be recognized by cellular enzymes,
of 31 patients maintained good results over 5}10 yr in it can be remodeled and degraded to provide space for
Sweden [18]. the growing tissue [35]. Collagen matrices have also been
Despite these encouraging results, this procedure, found to have the proper molecular cues to stimulate new
although clinically available in the US and other collagen production by transplanted cells as compared
countries, should not be regarded as a panacea. As with other sca!old types [36]. Caplan et al. reported that
indicated above, the best results are found for healing when MSCs were seeded into collagen gels implanted in
lesions in the femoral condyle [3] and the need for a osteochondral defects in rabbits, embryogenesis was
periosteal #ap requires a donor site. Interestingly, recapitulated and both bone and hyaline cartilage were
although short-term (three months) studies with rabbits formed, although mechanical properties of the regen-
reported healing similar to that found in humans [24], erated tissue were signi"cantly less than normal values,
long-term results in canines have not been as promising. and there was some evidence of degeneration after 24
Although initially the procedure described above weeks [37,38].
caused cartilage regeneration in canine knees, this was Although preliminary results are promising for natu-
not permanent, and 12}18 months after surgery, the rally derived polymers, concerns about the feasibility of
cartilage began to degrade [25]. In another study [26], "nding large amounts of these materials for clinical ap-
at 12}18 months, no di!erence was observed between plications, as well as assurance of pathogen removal has
the healing of control defects, defects with addition of prompted other researchers to investigate the use of
a periosteal #ap alone, or with a #ap and autologous synthetic polymers. These materials can be easily mass-
chondrocytes. Because the canine cartilage repair produced and their properties can be tailored for speci"c
process seems very similar to that of the human [27], applications. This includes creating degradable polymers,
explanations for these discrepancies are yet to be which like collagen, allow room for tissue growth in
determined. the construct and eliminate the need for a second surgery
436 J.S. Temenow, A.G. Mikos / Biomaterials 21 (2000) 431}440

to remove the implant. There is also the possibility copolymers of ethylene and propylene oxide P(EO-co-
that growth factors or other pharmacologics could be PO), have also been tested as injectable matrices for
included in the sca!old to encourage certain cellular chondrocyte transplantation. Sims et al. reported that 12
responses, such as proliferation or di!erentiation. weeks after subcutaneous delivery of bovine chon-
However, few of these polymers have been FDA ap- drocytes in a PEO gel into nude mice, cartilage that was
proved for use in humans [28]. biochemically similar to natural bovine cartilage had
Much current research has focused on chondrocyte been produced [54]. In a comparative study, when
interaction with those biodegradable polymers that are seeded with autologous chondrocytes and injected in
FDA approved: poly(glycolic acid) (PGA), poly(L-lactic pigs, the P(EO-co-PO) copolymer was found to produce
acid) (PLLA) and their copolymer, poly(DL-lactic-co- more cartilage-like tissue than either PGA sca!olds or
glycolic acid) (PLGA) [36,39}46]. They are poly(a-hy- alginate hydrogels. In this case, however, it must be noted
droxy esters) and are degraded by hydrolysis. PLLA is that the chondrocytes were of aural rather than articular
more hydrophobic and less crystalline than PGA and origin [53].
degrades at a slower rate [28]. Firm hyaline-like cartilage Types of cells used. Certain polymer properties, such
has been observed after six weeks when undi!erentiated as degradation time, may be dictated by the type of cells
perichondrial cells were seeded onto PLLA meshes and to be used in the sca!olds. As evidenced in the above
implanted in the femoral condyles of rabbits [42,43]. examples, both fully di!erentiated chondrocytes and pro-
Similar cartilage morphology was found using PGA por- genitor cells have been employed to produce hyaline-like
ous non-woven sca!olds seeded with bovine chon- cartilage. Chondrocytes are very easy to obtain, but it has
drocytes and cultured in vitro for 12 weeks. In this case, been found that as the cells are cultured for longer peri-
mechanical properties such as compressive modulus and ods before implantation, the cartilage formed is increas-
aggregate modulus were similar to that of normal bovine ingly "brous in nature [29]. Also, according to Caplan
cartilage [44]. Both types of degradable polyester tended et al., chondrocytes implanted in osteochondral defects in
to increase proteoglycan synthesis compared to a col- rabbits did generate cartilage, but a subchondral plate
lagen sca!old [36]. Cell growth was approximately twice was only observed if undi!erentiated MSCs were used
as high initially (less than two months) on PGA than [20,38]. MSCs can be expanded many-fold with little
PLLA matrices when seeded with bovine chondrocytes, e!ect on the tissue that is eventually formed and have
but after six months, total cellularity was found to be been shown to di!erentiate into both bone and cartilage
similar. Initial di!erences were attributed to the fact that cells, making them prime candidates for transplantation
PLLA degrades much more slowly and so left less space in tissue-engineered constructs [38,55}57].
for cell proliferation [45]. PLLA has been found to be With either type of cell chosen, good sca!old seeding is
less toxic to human chondrocytes than PGA in studies required for the formation of functional cartilage. Seed-
maintaining a constant pH over 12 days [46]. ing density is important since it has been found that cells
A main disadvantage of all of these sca!old materials is seeded too sparsely result in incomplete "lling of the
that they require an operation for implantation. This has sca!old, leading to the possibility of "brous ingrowth
led to the development of polymers that can be injected which would adversely a!ect the properties of the tissue
along with cells and then cross-linked in situ to form generated [58,59]. In addition, surface tension can pre-
matrices. Several investigators are exploring the option vent liquid media with cells from entering some synthetic
of combining "brinogen and thrombin to form a degra- sca!olds, such as PGA and PLLA. Therefore, these
dable "brin mesh that can be used to support chon- meshes must be pre-wet with alcohol which is then dis-
drocytes [47}49]. When the cell}"brinogen}thrombin placed by media to obtain seeding deep within the con-
mixture was injected into defects in horses, more GAGs, struct [28]. Other techniques to improve seeding include
aggregan and type II collagen were present in the new dipping the sca!olds in poly(L-lysine) or type II collagen
tissue at eight months than in defects that were left to improve initial cell attachment [46]. If materials that
untreated [49]. Alginate and bovine articular chon- set in situ are used, in addition to good cell seeding, the
drocytes, injected into nude mice and crosslinked with cytotoxicity of the reagents and temperature change dur-
calcium, also formed cartilage, but studies longer than 12 ing setting must be minimized to promote cartilage
weeks have not been conducted to determine if there is formation in the construct [28,60].
any long-term tissue degradation [29,50,51]. Some forms
of alginate have been found to be immunogenic, as evi-
denced by increased lymphocyte number and presence of 5. Bioreactors for chondrocyte and cartilage tissue growth
anti-alginate antibodies [52]. A further concern about
the biocompatibility of the material is that, according to In the past ten years, there have been signi"cant ad-
Cao et al., there is more in#ammatory response to al- vances in growing cartilage with properties similar to
ginate than to injectable synthetic materials [53]. Syn- those found in vivo, but to transfer this technology to
thetic polymers, such as poly(ethylene oxide) (PEO) or a clinical setting, scale-up remains a large problem. Both
J.S. Temenow, A.G. Mikos / Biomaterials 21 (2000) 431}440 437

methods to mass-produce the needed biomaterials, and ance of the di!erentiated chondrocytic phenotype
large-scale production of cells must be developed. Be- [64,66].
cause synthesis is very speci"c to the material chosen
(natural vs. synthetic), only large-scale expansion of cells 5.3. Perfusion culture
is reviewed here.
It is estimated that to resurface an entire joint While the above types of bioreactors show favorable
would require an engineered cartilage implant of the results, there is concern that the mechanical mixing may
size 5 cm in diameter (1}5 mm thick), while current induce unwanted shear gradients in the reactor vessel. To
laboratory techniques test constructs only approximately eliminate this possibility, microcarriers can be placed in
5 mm in diameter [61]. High-density cell cultures #uidized bed or air-lift reactors where mixing is achieved
can be used to prevent chondrocytic dedi!erentiation by #uid recirculation rather than impeller motion [64].
that often occurs in two-dimensional cultures [62,63], Another simple design to improve mass transfer while
but bioreactors present numerous advantages for eliminating mechanical mixing is the perfusion culture
growing large amounts of tissue quickly [64]. These system. Sca!olds are seeded with cells statically and
advantages include that bioreactors provide uniform subsequently hooked to a perfusion system that uses
mixing and precise control over mass transfer rates, a peristaltic pump to deliver a constant #ow of media to
facilitating maintenance of nutrient levels and pH. In the culture chamber containing the constructs. The me-
addition, shear stress can be easily monitored in many dia then exits through another port that empties into
types of reaction vessels [64]. It has been found that a waste vessel. Good matrix formation and mature chon-
shear stress can have a signi"cant impact on the morpho- drocytic phenotype was observed after two weeks in
logy and thus the mechanical properties of cartilage, so culture [67] and, after nine days, the cellularity in test
regulation of this parameter is essential for quality constructs was higher than static controls [68]. In
control of the product [61]. Using process control tech- a long-term study, tissue with histological and mechan-
niques, the #ow rates in the reactor can be modulated to ical characteristics similar to native tissue was evident
maintain important parameters (i.e. nutrient concentra- after 50 days of perfusion [69]. Similar experiments using
tion) over time as the construct becomes less permeable marrow stromal cells indicate that perfusion enhances
due to matrix deposition and increased chondrocyte growth and function over two weeks [70]. However,
number [64]. currently, this system is only useful for three-dimensional
expansion of cells in sca!olds.
5.1. Spinner yasks
5.4. Rotating-wall bioreactors
One of the simplest bioreactor designs is the spinner
#ask. Typically, sca!olds seeded with chondrocytes are A promising bioreactor design that can be used with
attached to needles hanging from the stopper of the #ask. either microcarriers or sca!olds without the concern of
Media is added to cover the sca!olds and mixing is mechanical mixing is the rotating-wall vessel reactor
maintained with a magnetic stir bar in the bottom of the originally designed by NASA to simulate the e!ects of
#ask. Media is changed every few days to insure high microgravity. The most common type is the slow turning
nutrient concentration. Studies by Freed et al. showed lateral vessel (STLV). The reactor is comprised of two
that constructs cultured in spinner #asks were larger and concentric cylinders. The stationary inner cylinder has
contained more cells after "ve weeks than those grown in a membrane that allows gas exchange while the outer
Petri dishes [65]. cylinder, made from a non-permeable material, rotates.
The space between the two cylinders is perfused continu-
5.2. Microcarriers ously with media. Pre-seeded polymer constructs or
chondrocytes and microcarriers are placed inside the
To culture individual cells (not seeded in sca!olds) in reactor and are kept in a state of constant free-fall by
a bioreactor, microcarrier suspensions can be used. Vari- adjusting the speed of rotation of the outer cylinder so
ous types of microcarrier beads, such as collagen or that the centrifugal force just balances the force of gravity
dextran, have been tested. These carriers are suspended and the #uid drag on the objects inside. Mixing occurs
in a reactor with continuously stirred media and an due to the small amount of unavoidable settling which
inoculum of chondrocytes is added. As in the spinner creates movement of the sca!old/microcarrier relative to
#asks, the media is exchanged every few days. Cell doubl- the media [61,65].
ing times were shown to be around two to three days in With microcarrier suspensions, the STLV provided an
these bioreactors as compared up to "ve days in Petri environment that encourages aggregation of chon-
dishes, suggesting that the increased mass transfer in the drocytes to form tissue. This may be a means to grow
bioreactor encouraged cell growth. In addition, the larger sections of cartilage that could be placed directly
microcarrier suspension appeared to enhance mainten- into a defect without the use of an implanted matrix to
438 J.S. Temenow, A.G. Mikos / Biomaterials 21 (2000) 431}440

retain the cells [71]. Using sca!olds, when grown in the an optimal polymer for use in the sca!old. Also, the type
STLV for "ve weeks, cell}polymer constructs showed of cells (di!erentiated or progenitor) that are best for this
more GAG production than those placed in spinner application has yet to be decided. It is the interaction of
#asks [65]. In further experiments, the STLV constructs these two parameters (material and cells), along with the
were found to have 68% as much GAG, 33% as much possible addition of growth factors and/or mechanical
collagen and similar cellularity to natural cartilage [72]. stimulation to in#uence cellular development, which will
However, while these constructs demonstrated better determine the ultimate success of the implant.
mechanical properties than those grown in other condi- As these issues are considered, long-term results are of
tions (static and modi"ed spinnner #ask), the equilibrium utmost concern: no current experimental strategy has
modulus, dynamic sti!ness and streaming potential were succeeded in consistently producing cartilage that shows
only 20}25% as high as fresh cartilage explants [73]. no signs of degeneration after one year. Structure at these
In order to improve the quality of cartilage produced, late time points is also extremely important. None of the
researchers have begun to determine optimal operating studies reviewed here have created a tissue that fully
parameters for the STLV bioreactor. When rates of oxy- regenerates the natural zonal organization of articular
gen exchange were varied, it was discovered that, al- cartilage, even after several months in vivo. A "nal hurdle
though cartilage is generally regarded as a hypoxic tissue, to full cartilage regeneration is the integration of any
higher rates of oxygen tension resulted in larger tissue newly formed tissue with the existing cartilage. Without
constructs from pre-seeded polymer sca!olds. These con- this integration, forces that occur at the interface could
structs also showed higher amounts of ECM synthesis have deleterious long-term e!ects on the replacement
than those grown at a lower oxygen concentration [74]. tissue.
However, the STLV may be replaced by a larger rotat- Despite considerable work that must be done to opti-
ing-wall vessel, the high aspect ratio vessel (HARV), to mize parameters for formation of tissue-engineered
create the large constructs needed for joint resurfacing. cartilage, because this tissue is not vascularized and in-
Because of the size and design di!erences, the HARV has corporates only one cell type, it is likely the next tissue
"ve times as much surface area for gas exchange per unit after skin to be successfully regenerated. Further devel-
of reactor volume as the STLV. This increase in surface opment of techniques for chondrocyte transplantation
area may be especially important given the previous would be an incredible advance in the "eld, providing
"ndings relating construct size to oxygen tension. This hope for many patients that must endure chronic joint
vessel has been used to culture constructs up to 10 cm in pain due to failure of conventional methods to restore
diameter [61]. cartilage function.

5.5. Further development


Acknowledgements
Many in vitro studies have shown that mechanical
forces greatly in#uence the development of cartilage (see The work on skeletal tissue engineering has been
[75] for a review of the e!ects of mechanical forces) [75]. supported by the National Institutes of Health (R29-
Recently, researchers have begun to combine the above AR42639, R01-AR44381, and R01-DE13031). J.S. Te-
culture systems with external mechanical stimuli to im- meno! acknowledges the "nancial support of a Whitaker
prove the quality of the cartilage produced. Carver et al. Foundation Graduate Fellowship.
recently described a perfusion system which applies inter-
mittent hydrostatic pressure to chondrocytes imbedded
in polymer sca!olds. The pressurized sca!olds demon- References
strated cartilage development with higher GAG content
[1] Buckwalter JA, Mankin HJ. Articular cartilage: degeneration and
after "ve weeks than in samples that were not exposed to osteoarthritis, repair, regeneration and transplantation. AAOS
pressure [76]. Inst Course Lect 1998;47:487}504.
[2] Arthritis Brochure. American Academy of Orthopaedic Surgeons
website, www.aaos.org.
6. Conclusions [3] Minas T. The role of cartilage repair techniques, including chon-
drocyte transplantation, in focal chondral knee damage. AAOS
Inst Course Lect 1999;48:629}43.
The recent work with bioreactors holds promise that [4] O'Driscoll SW. The healing and regeneration of articular carti-
large numbers of cells can be grown e$ciently and quick- lage. J Bone Jt Surg 1998;80-A:1795}812.
ly, pushing cell transplantation to the forefront of new [5] Buckwalter JA, Mankin HJ. Articular cartilage: tissue design and
solutions for the problem of articular cartilage injury. chondrocyte}matrix interactions. AAOS Inst Course Lect
1998;47:477}86.
However, many questions remain before tissue engineer- [6] Cohen NP, Foster RJ, Mow VC. Composition and dynamics of
ing techniques can be used to mass produce cartilage articular cartilage: structure, function, and maintaining healthy
replacements. Further studies must be conducted to "nd state. J Orthop Sports Phys Ther 1998;28:203}15.
J.S. Temenow, A.G. Mikos / Biomaterials 21 (2000) 431}440 439

[7] Harty M. Anatomy and development of joints. In: Simon WH, [29] Rodriguez AM, Vacanti CA. Tissue engineering of cartilage. In:
editor. The human joint in health and disease. Philadelphia: Patrick Jr CW, Mikos AG, McIntire LV, editors. Frontiers in
U. Pennsylvania, 1978. p. 3}8. tissue engineering. New York: Elsevier Science, 1998. p. 400}11.
[8] Stockwell RA. Biology of cartilage cells. Cambridge: Cambridge [30] von Schroeder HP, Kwan M, Amiel D, Coutts RD. The use of
University Press, 1979. polylactic acid matrix and periosteal grafts for the reconstruction
[9] Caplan AI, Boyan BD. Endochondral bone formation: the lineage of rabbit knee articular defects. J Biomed Mater Res 1991;
cascade. In: Hall BK, editor. Bone. Boca Raton, FL: CRC Press, 25:329}39.
1994. p. 1}46. [31] Athanasiou K, Korvick D, Schenck R. Biodegradable implants
[10] Buckwalter JA. Articular cartilage. AAOS Inst Course Lect for the treatment of osteochondral defects in a goat model. Tissue
1983;32:349}70. Eng 1997;3:363}73.
[11] Wirth CJ, Rudert M. Techniques of cartilage growth enhance- [32] Solchaga LA, Dennis JE, Goldberg VM, Caplan AI. Hyaluronic
ment: a review of the literature. Arthroscopy 1996;12:300}8. acid-based polymers as cell carriers for tissue-engineered repair of
[12] Wong M, Wuethrich P, Eggli P, Hunziker E. Zone-speci"c cell bone and cartilage. J Orthop Res 1999;17:205}13.
biosynthetic activity in mature bovine articular cartilage: a new [33] Butnariu-Ephrat M, Robinson D, Mendes DG, Halperin N, Nevo
method using confocal microscopy stereology and quantitative Z. Resurfacing of goat articular cartilage by chondrocytes derived
audioradiography. J Orthop Res 1996;14:424}32. from bone marrow. Clin Orthop Rel Res 1996;330:234}43.
[13] Coutts RD, Sah RL, Amiel D. E!ect of growth factors on cartilage [34] Kimura T, Yasui N, Ohsawa S, Ono K. Chondrocytes embedded
repair. AAOS Inst Course Lect 1997;46:487}94. in collagen gels maintain cartilage phenotype during long-term
[14] Buckwalter JA. Articular cartilage: injuries and potential for heal- cultures. Clin Orthop Rel Res 1984;186:231}9.
ing. J Orthop Sports Phys Ther 1998;28:192}202. [35] Speer DP, Chvapil M, Volz RG, Holmes MD. Enhancement of
[15] Hunziker EB. Articular cartilage repair: are the intrinsic biolo- healing in osteochondral defects by collagen sponge implants.
gical constraints undermining this process insuperable? Os- Clin Orthop Rel Res 1979;144:326}35.
teoarthr Cartil 1999;7:15}28. [36] Grande DA, Halberstadt C, Naughton G, Schwartz R, Manji R.
[16] Hunziker EB, Kap"nger E. Removal of proteoglycans from the Evaluation of matrix sca!olds for tissue engineering of articular
surface of defects in articular cartilage transiently enhances cover- cartilage grafts. J Biomed Mat Res 1997;34:211}20.
age by repair cells. J Bone Jt Surg 1998;80-B:144}50. [37] Wakitani S, Goto T, Pineda SJ, Young RG, Mansour JM, Caplan
[17] Goldberg VM, Caplan AI. Biologic resurfacing of articular surfa- AI, Goldberg VM. Mesenchymal cell-based repair of large, full-
ces. AAOS Inst Course Lect 1999;48:623}7. thickness defects of articular cartilage. J Bone Jt Surg 1994;76A:
[18] Gillogly SD, Voight M, Blackburn T. Treatment of articular 579}92.
cartilage defects of the knee with autologous chondrocyte im- [38] Caplan AI, Elyaderani M, Mochizuki Y, Wakitani S, Goldberg
plantation. J Orthop Sports Phys Ther 1998;28:241}51. VM. Principles of cartilage repair and regeneration. Clin Orthop
[19] Shapiro F, Koide S, Glimcher MJ. Cell origin and di!erentiation Rel Res 1997;342:254}69.
in the repair of full-thickness defects of articular cartilage. J Bone [39] Vacanti CA, Kim W, Schloo B, Upton J, Vacanti JP. Joint
Jt Surg 1993;75-A:532}53. resurfacing with cartilage grown in situ from cell}polymer struc-
[20] Wakitani S, Goto T, Young RG, Mansour JM, Goldberg VM, tures. Am J Sports Med 1994;22:485}8.
Caplan AI. Repair of large full-thickness articular cartilage defects [40] Vacanti CA, Upton J. Tissue-engineered morphogenesis of carti-
with allograft articular chondrocytes embedded in a collagen gel. lage and bone by means of cell transplantation using synthetic
Tissue Eng 1998;4:429}44. biodegradable polymer matrices. Clin Plast Surg 1994;21:445}62.
[21] Aston JE, Bentley G. Repair of articular surfaces by allografts of [41] Freed LE, Grande DA, Lingbin Z, Emmanual J, Marquis JC,
articular and growth-plate cartilage. J Bone Jt Surg 1986;68-B: Langer R. Joint resurfacing using allograft chondrocytes and
29}35. synthetic biodegradable polymer sca!olds. J Biomed Mater Res
[22] Brittberg M, Lindahl A, Nilsson A, Ohlsson C, Isaksson O, 1994;28:891}9.
Peterson L. Treatment of deep cartilage defects in the knee with [42] Chu CR, Coutts RD, Yoshioka M, Harwood FL, Monosov AZ,
autologous chondrocyte transplantation. N Engl J Med 1994; Amiel D. Articular cartilage repair using allogeneic perichon-
331:889}95. drocyte-seeded biodegradable porous polylactic acid (PLA): a tis-
[23] McPherson JM, Tubo R, Barone L. Letter to the editor. Arthro- sue-engineering study. J Biomed Mater Res 1995;29:1147}54.
scopy 1997;13:541}3. [43] Chu CR, Monosov AZ, Amiel D. In situ assessment of cell
[24] Brittberg M, Lindahl A, Nilsson A, Ohlsson C. Cellular aspects viability within biodegradable polylactic acid polymer matrices.
on treatment of cartilage injuries. Agents Actions Suppl 1993;39: Biomaterials 1995;16:1381}4.
237}41. [44] Ma PX, Schloo B, Mooney D, Langer R. Development of bio-
[25] Breinan HA, Minas T, Barone L, Tubo R, Hsu HP, Shortkro! S, mechanical properties and morphogenesis of in vitro tissue engin-
Nehrer S, Sledge CB, Spector M. Histological evaluation of eered cartilage. J Biomed Mater Res 1995;29:1587}95.
the course of healing of canine articular cartilage defects treated [45] Freed LE, Marquis JC, Nohria A, Emmanual J, Mikos AG,
with cultured autologous chondrocytes. Tissue Eng 1998;4: Langer R. Neocartilage formation in vitro and in vivo using cells
101}14. cultured on synthetic biodegradable polymers. J Biomed Mater
[26] Breinan HA, Minas T, Hsu HP, Nehrer S, Sledge C, Spector M. Res 1993;27:11}23.
E!ect of cultured autologous chondrocytes on repair of [46] Sittinger M, Reitzel D, Dauner M, Hierlemann H, Hammer C,
chondral defects in a canine model. J Bone Jt Surg 1997;79-A: Kastenbauer E, Planck H, Burmester GR, Bujia J. Resorbable
1439}51. polyesters in cartilage engineering: a$nity and biocompatibility
[27] Shortkro! S, Barone L, Hsu HP, Wrenn C, Gagne T, Chi T, of polymer "ber structures to chondrocytes. J Biomed Mater Res
Breinan H, Minas T, Sledge CB, Tubo R, Spector M. Healing of 1996;33:57}63.
chondral and osteochondral defects in a canine model: the role of [47] Sims CD, Butler PEM, Cao YL, Casanova R, Randolph MA,
cultured chondrocytes in regeneration of articular cartilage. Bio- Black A, Vacanti CA, Yaremchuk MJ. Tissue engineered neocar-
materials 1996;17:147}54. tilage using plasma derived polymer substrates and chondrocytes.
[28] Thomson RC, Wake MC, Yaszemski MJ, Mikos AG. Biodegrad- Plast Reconstr Surg 1998;101:1580}5.
able polymer sca!olds to regenerate organs. Adv Polym Sci 1995; [48] Ting V, Sims CD, Brecht LE, McCarthy JG, Kasabian AK,
122:245}74. Connelly PR, Elissee! J, Gittes GK, Longaker MT. In vitro
440 J.S. Temenow, A.G. Mikos / Biomaterials 21 (2000) 431}440

prefabrication of human cartilage shapes using "brin glue and Chick W, editors. Principles of tissue engineering. New York:
human chondrocytes. Ann Plast Surg 1998;40:413}21. R.G. Landes Co., 1997. p. 151}65.
[49] Hendrickson DA, Nixon AJ, Grande DA, Todhunter RJ, Minor [62] Mackay AM, Beck SC, Murphy JM, Barry FP, Chichester CO,
RM, Erb H, Lust G. Chondrocyte}"brin matrix transplants for Pittenger MF. Chondrogenic di!erentiation of cultured human
resurfacing extensive articular cartilage defects. J Orthop Res mesenchymal stem cells from marrow. Tissue Eng 1998;4:415}28.
1994;12:485}97. [63] Fedewa MM, Oegema Jr TR, Schwartz MH, MacLeod A, Lewis
[50] Paige KT, Cima LG, Yaremchuk MJ, Schloo BL, Vacanti JP, JL. Chondrocytes in culture produce a mechanically functional
Vacanti CA. De novo cartilage generation using calcium al- tissue. J Orthop Res 1998;16:227}36.
ginate}chondrocyte constructs. Plast Reconstr Surg 1996;97: [64] Freed LE, Vunjak-Novakovic G, Langer R. Cultivation of
168}80. cell}polymer cartilage implants in bioreactors. J Cell Biochem
[51] Paige KT, Cima LG, Yaremchuk MJ, Vacanti JP, Vacanti CA. 1993;51:257}64.
Injectable cartilage. Plast Reconstr Surg 1995;96:1390}400. [65] Freed LE, Vunjak-Novakovic G. Microgravity tissue engineering.
[52] Kulseng B, Skjak-Braek G, Ryan L, Andersson A, King A, Fax- In Vitro Cell Dev Biol*Animal 1997;33:381}5.
vaag A, Espevik T. Transplantation of alginate microcapsules. [66] Frondoza C, Sohrabi A, Hungerford D. Human chondrocytes
Transplantation 1999;67:978}84. proliferate and produce matrix components in microcarrier sus-
[53] Cao Y, Rodriguez A, Vacanti M, Ibarra C, Arevalo C, Vacanti C. pension culture. Biomaterials 1996;17:879}88.
Comparative study of the use of poly(glycolic acid), calcium [67] Sittinger M, Bujia J, Minuth WW, Hammer C, Burmester GR.
alginate and pluronics in the engineering of autologous porcine Engineering of cartilage tissue using bioresorbable polymer car-
cartilage. J Biomater Sci Polym Edn 1998;9:475}87. riers in perfusion culture. Biomaterials 1994;15:451}6.
[54] Sims CD, Butler PEM, Casanova R, Lee BT, Randolph MA, Lee [68] Davisson TH, Wu FJ, Jain D, Sah RL, Ratcli!e AR. E!ect of
WPA, Vacanti CA, Yaremchuk MJ. Injectable cartilage using perfusion on the growth of tissue engineered cartilage. In: 45th
polyethylene oxide polymer substrates. Plast Reconstr Surg Annual Meeting. Anaheim, CA: Orthopaedic Research Society,
1996;98:843}50. 1999.
[55] Dennis JE, Merriam A, Awadalla A, Yoo JU, Johnstone B, [69] Bujia J, Sittinger M, Minuth WW, Hammer C, Burmester G,
Caplan AI. A quadripotential mesenchymal progenitor cell iso- Kastenbauer E. Engineering of cartilage tissue using bioresorb-
lated from the marrow of an adult mouse. J Bone Miner Res able polymer #eeces and perfusion culture. Acta Otolaryngol
1999;14:700}9. 1995;115:307}10.
[56] Pittenger MF, Mackay AM, Beck SC, Jaiswal RK, Douglas R, [70] Glowacki J, Mizuno S, Greenberger JS. Perfusion enhances func-
Mosca JD, Moorman MA, Simonetti DW, Craig S, Marshak DR. tion of bone marrow stromal cells in three-dimensional culture.
Multilineage potential of adult human mesenchymal stem cells. Cell Transplant 1998;7:319}26.
Science 1999;284:143}7. [71] Duke PJ, Daane EL, Montufar-Solis D. Studies of chondrogen-
[57] Grande DA, Southerland SS, Manji R, Pate DW, Schwartz RE, esis in rotating systems. J Cell Biochem 1993;51:274}82.
Lucas PA. Repair of articular cartilage defects using mesenchymal [72] Freed LE, Hollander AP, Martin I, Barry JR, Langer R, Vunjak-
stem cells. Tissue Eng 1995;1:345}53. Novakovic G. Chondrogenesis in a cell}polymer}bioreactor sys-
[58] Puelacher WC, Kim SW, Vacanti JP, Schloo B, Mooney D, tem. Exp Cell Res 1998;240:58}65.
Vacanti CA. Tissue-engineered growth of cartilage: the [73] Vunjak-Novakovic G, Martin I, Obradovic B, Treppo S, Grod-
e!ect of varying the concentration of chondrocytes seeded onto zinsky AJ, Langer R, Freed LE. Bioreactor cultivation conditions
synthetic polymer matrices. Int J Oral Maxillofac Surg 1994;23: modulate the composition and mechanical properties of tissue-
49}53. engineered cartilage. J Orthop Res 1999;17:130}8.
[59] Vacanti CA, Vacanti JP. Bone and cartilage reconstruction. In: [74] Obradovic B, Carrier RL, Vunjak-Novakovic G, Freed LE. Gas
Lanza R, Langer R, Chick W, editors. Principles of tissue engi- exchange is essential for bioreactor cultivation of tissue engineer-
neering. New York: R.G. Landes Co., 1997. p. 619}31. ed cartilage. Biotechnol Bioeng 1999;63:197}205.
[60] Yaszemski MJ, Payne RG, Hayes WC, Langer R, Mikos AG. [75] Heath CA, Magari SR. Mini-review: mechanical factors a!ecting
In vitro degradation of a poly(propylene fumarate)-based com- cartilage regeneration in vitro. Biotechnol Bioeng 1996;50:430}7.
posite material. Biomaterials 1996;17:2127}30. [76] Carver SE, Heath CA. Semi-continuous perfusion system for
[61] Freed LE, Vunjak-Novakovic G. Tissue culture bioreactors: delivering intermittent physiological pressure to regenerating car-
chondrogenesis as a model system. In: Lanza R, Langer R, tilage. Tissue Eng 1999;5:1}11.

Das könnte Ihnen auch gefallen