- Email: [email protected]
Gene Transfer Approaches to the Healing of Bone and Cartilage
doi:10.1006/mthe.2000.0663, available online at http://www.idealibrary.com on IDEAL
REVIEW
Gene Transfer Approaches to the Healing of Bone and Cartilage Jay R. Lieberman,1,* Steven C. Ghivizzani,2 and Christopher H. Evans2 1Department
of Orthopaedic Surgery, The David Geffen School of Medicine at UCLA, Los Angeles, California 90095, USA for Molecular Orthopaedics, Harvard Medical School, Boston, Massachusetts 02115, USA *To whom correspondence and reprint requests should be addressed. Fax: (310) 206-0063. E-mail: [email protected]. 2Center
Gene therapy is traditionally considered a treatment modality for genetic diseases and a limited number of complex, life-threatening disorders, such as cancer. However, gene therapy also holds promise as a novel means of treating more mundane conditions, including broken bones and damaged cartilages [1–3]. These injuries frequently present enormous clinical challenges to both orthopedic surgeons and rheumatologists, and can result in considerable human suffering and economic loss. Various gene products, particularly growth factors, show remarkable promise as agents that can improve the healing of bone, cartilage, and other connective tissues. Their clinical utility, however, is limited by delivery problems. The attraction of gene transfer approaches is the unique ability to deliver authentically processed gene products to precise anatomical locations at therapeutic levels for sustained periods of time. However, unlike the treatment of chronic disease, it is neither necessary nor desirable for transgene expression to persist beyond the few weeks or months needed to achieve healing. It is also unlikely that the level of transgene expression will need to be closely regulated. Moreover, achieving targeted delivery is not an issue, because the same orthopedic procedures that are already used when operating on bones and joints can be adapted for the purposes of targeted gene transfer. Thus, the applications proposed in this review represent several of the few examples in which gene therapy has a good chance of clinical success using existing technology.
BONE HEALING There is great interest in developing gene therapy to enhance bone repair because the potential for sustained protein production may enable the host to respond to an osteoinductive stimulus in a more robust fashion. Regional gene therapy is an attractive option to enhance bone formation and repair because the genes can be delivered to a specific anatomic site, and the duration of protein production can be determined by selection of a particular vector. In fact, gene therapy for bone repair may be easier to develop for human use than for other diseases because the requirements for protein production are more variable. In many cases, protein production will
MOLECULAR THERAPY Vol. 6, No. 2, August 2002 Copyright © The American Society of Gene Therapy
only be required until the bone has healed [3–5]. One exception to this would be the treatment of osteoporosis, which may require the development of systemic gene therapy. In the future, gene therapy may be one option in a comprehensive strategy for the tissue engineering of bone. Gene therapy will not be for all patients, as some may be best treated with an autologous bone graft and others, with recombinant proteins and/or bone graft substitutes. However, in some patients gene therapy may offer the best potential for healing [3–5]. Enhanced bone formation is often required to treat bone loss associated with trauma, revision total joint arthroplasty, pseudarthrosis of the spine, and tumor resection. Autologous bone grafts harvested from the pelvis are the current gold standard for treating bone loss problems. However, only a limited amount of autogenous bone graft is available, and there are problems with donor site morbidity [6]. In a large retrospective study, a 10% incidence of minor complications and a 5.8% incidence of major complications were associated with bone graft harvest [7]. These limitations have prompted increased interest in the development of alternative bone grafts, such as allografts and demineralized bone matrices. Allograft bone has limited osteoinductive potential, and there is also the possibility of transmission of viral disease. At this time, demineralized bone matrices seem to be most appropriately used as bone graft extenders rather than bone graft substitutes. These materials are osteoconductive and provide a scaffold to promote bone formation, but they have limited osteoinductive activity. Several osteoinductive growth factors that can enhance bone repair have been identified. In particular, the bone morphogenetic proteins (BMPs) have been shown to induce bone formation in a variety of different animal models. Both BMP-2 and BMP-7 (osteogenic protein-1) have been used to heal critical-sized defects in sheep, dogs, and nonhuman primates [8–10] and to induce the fusion of the spine in canines and nonhuman primates [11,12]. Although these preclinical models demonstrate the potential efficacy of different BMPs, these animal models do not truly simulate the clinical situations often associated with bone loss problems. These studies were carried out using young animals
141
REVIEW
doi:10.1006/mthe.2002.0663, available online at http://www.idealibrary.com on IDEAL
with well-vascularized bone and intact soft tissue. In contrast, the healing potential of the bone in a clinical setting may be quite limited because of compromised vascularity, limited bone stock, and abundant fibrous tissue. In addition, bone repair may be inhibited in diabetic, elderly, or nicotine-addicted patients [4]. Although recombinant BMP has been used successfully in a variety of animal studies, the success rate in humans is variable. In two clinical trials, large doses of BMP were required to induce adequate bone repair, suggesting that the mode of BMP delivery still requires further optimization [13,14]. It is not known what side effects are associated with such high doses of BMP. Therefore, there is concern that a single exposure to an exogenous growth factor may not induce an adequate osteogenic signal in many clinical situations. Gene therapy has the potential to provide more sustained protein release when necessary and to deliver protein in a more physiologic manner than recombinant proteins [3–4].
IN VIVO GENE TRANSFER STRATEGY In vivo gene delivery involves directly delivering the gene into a specific anatomic site by transducing local cells. The advantages of this strategy are that it is a relatively simple technique and has the potential for lower cost. The disadvantages are the difficulty in targeting specific cells for transduction and in achieving high transduction efficiency. When using an adenovirus vector there is the additional concern about the development of an immune response as a result of the viral particles injected directly into the anatomic site, which can inhibit transgene expression [15–17]. Both viral and nonviral delivery methods have successfully induced bone formation in several animal models. Baltzer et al. used a first-generation adenoviral vector containing the cDNA for BMP-2 to heal criticalsized femoral defects in New Zealand white rabbits [18]. In this experiment, healing of the bone defects was noted 7 weeks after viral injection. Ectopic bone formation has also been induced by the injection of a BMP-2containing adenovirus into the thigh muscles of nude rats. However, bone formation was inhibited in immune competent Sprague-Dawley rats. In addition, an injection of the BMP-2 adenovirus produced less bone when injected into the triceps of immune competent mice compared with the same adenovirus injected into the triceps of nude mice [19]. Okubu et al. noted increased bone formation after the direct injection of the BMP-2-containing adenovirus when the rats were also given cyclophosphamide. Although these experiments resulted in bone formation, they also suggest that the presence of an immune response against the adenoviral vectors limits the biological activity of the secreted protein [20]. To avoid potential problems with the adenoviral vectors, nonviral delivery methods are being developed.
142
Plasmid DNA encoding parathyroid hormone (PTH) 1-34 or BMP-4 placed in a gene-activated matrix (GAM) is one example of a nonviral gene therapy strategy that has been successfully used to promote osteogenesis in a rat bone-defect model. In this strategy, the plasmid PTH 1-34 or BMP-4 is incorporated into a collagen matrix and then implanted directly into a bone defect. Local fibroblasts in the area of the defect can then acquire the DNA and become local bioreactors for synthesis of the PTH or BMP-4 protein [21]. This strategy was successfully used to heal 5-mm defects in rats. A combination of both PTH 134 and BMP-4 showed enhanced healing (4 weeks versus 9 weeks) when compared with cDNA for either BMP-4 or PTH 1-34 alone. Bonadio et al. noted 6 weeks of in vivo protein production using a GAM containing PTH 1-34 cDNA to treat a canine tibial defect [22]. However, insufficient bone was produced to heal a critical-sized tibial defect in this canine model. These results suggest that either the local transfection efficiency is low or that PTH 1-34 has limited osteoinductive potential. Overall, the concept of using a GAM is attractive, and further study is necessary to determine the appropriate cDNA to use in this matrix.
EX VIVO GENE TRANSFER STRATEGIES In addition to providing an osteoinductive gene to a desired site, the advantage of an ex vivo approach is that it enables surgeons to select specific cells (that is, bone marrow cells, muscle cells, or stem cells) that can participate in osteoinduction. This may enhance the bone repair process as both autocrine and paracrine responses can be elicited with secretion of the desired gene product. In addition, ex vivo methods may be safer than in vivo techniques when working with an adenovirus because no viral particles or DNA complexes are injected into the body. In general, ex vivo techniques have a high efficiency of cell transduction. The major disadvantage of this strategy is the requirement of an extra harvesting step and the increased time and cost of the process. However, it is theoretically feasible to harvest cells from the patient, have a short period of infection and then reimplant the transduced cells at the appropriate anatomic site [23]. Because they are osteogenic, bone marrow cells are obvious candidates for ex vivo gene transfer in bone healing. Furthermore, bone marrow cell harvest is a well-established procedure and methods have been developed to produce a purified population of mesenchymal cells. These progenitor cells can be expanded over 1 billion-fold without losing their potency, and they have demonstrated the ability to induce sufficient bone formation to heal critical-sized defects in both rat and canine models [24–26]. Although these results are promising and have the possible advantage of avoiding the use of gene transfer techniques, it remains to be determined whether these cells alone would be able to heal a bone defect in a clinical situation. However, genetically manipulating these cells to MOLECULAR THERAPY Vol. 6, No. 2, August 2002 Copyright © The American Society of Gene Therapy
doi:10.1006/mthe.2002.0663, available online at http://www.idealibrary.com on IDEAL
REVIEW
produce an osteoinductive protein would further enhance the bone repair response. Lieberman et al. [5] used an ex vivo gene transfer approach to heal critical-size femoral defects in a rat model. Rat bone marrow cells were cultured, infected with a BMP-2-containing adenovirus, and then implanted in the femoral defect site. Histomor- FIG. 1. Radiographs of specimen made 2 months after the surgical procedure to test the efficacy of the ex vivo gene phometric analysis revealed transfer strategy with BMP-2-producing bone marrow cells that were created by adenoviral gene transfer. We used 20 a more robust pattern of mg of guanidine hydrochloride-extracted, demineralized bone matrix as a carrier in all defects. (A) rhBMP-2 (20 g, positive control). The healed defect is filled with lace-like trabecular bone. (B) BMP-2-producing bone marrow cells bone formation in femoral (5 ⫻ 106). Dense, coarse trabecular bone, remodeled to form a new cortex, was present. (C) -Galactosidase-prodefects treated with BMP-2- ducing rat bone marrow cells. (D) Rat bone marrow cells alone. (E) Demineralized bone matrix alone. Groups (C), (D), producing bone marrow and (E) were negative controls with minimal bone formation. cells as compared with defects treated with recombinant BMP-2 protein placed on a collagen sponge [5] (Fig. a fibroblast-like population of cells that can behave like 1). Although biomechanical testing of the healed femurs mesenchymal stem cells [29]. When these lipoaspirate did not show any differences between these groups with cells are grown in the appropriate medium, they can difrespect to energy-to-failure or torque-to-failure, this may ferentiate into bone, cartilage, fat, or muscle. Although the be due to a lack of sensitivity of the rat model. The more full osteogenic potential of these cells is still being anarobust bone formed by cells transduced with BMP-2 may lyzed, it is known that these stem cells can form bone in be the result of a more continuous or physiological release vivo when they are transduced with a BMP-2-containing of BMP-2 protein compared with the kinetics of BMP adenovirus and implanted into a SCID mouse muscle release from a collagen sponge. It is hypothesized that a pouch [30]. Further work with these types of cells is being more vigorous osteogenic activity is seen because both carried out (J.R.L.’s laboratory). paracrine and autocrine responses occur, that is, the transSPINE FUSION duced bone marrow cells can respond to the BMP-2 protein that they are secreting. Spinal fusion is one of the most commonly performed A variety of other cell types have been used to induce orthopedic surgical procedures in the United States, with bone formation. BMP-7-transfected rat skin fibroblasts over 980,000 performed every year. Approximately onehave been shown to heal critically-sized calvarial defects in third of these procedures requires a bone graft [31]. In Lewis rats. Skin fibroblasts are an attractive cellular delivgeneral, autogenous bone graft is a successful method for ery vehicle because they are easy to harvest and are readienhancing spine fusion, but non-union rates range from ly available in all patients [27]. There is one caveat when 5% to 35%. A variety of factors can influence the success using fibroblasts to deliver osteoinductive signals: fibrous of the spine fusion including mechanical instability of tissue can actually inhibit bone formation and is found in the spine, the stability of fixation, the quality of bone, fracture non-union sites. the health of surrounding soft tissue, the type of bone Genetically engineered muscle-derived cells also have graft used, and the concurrent use of medications and osteogenic potential [28]. Lee et al. demonstrated that drugs such as nicotine. Concerns regarding the osteoinmuscle cells transduced with a BMP-2-containing adenductive potential of various exogenous growth factors ovirus were able to heal calvarial defects in SCID mice. have lead investigators to explore the potential of gene These investigators also noted that a small percentage of therapy to induce fusion of the spine. muscle-derived cells implanted in calvarial defects actually Bone marrow cells transduced with a DNA plasmid condifferentiated into osteoblasts in vivo. Both muscle cells taining the cDNA for LMP-1 (LIM mineralization proteinand bone marrow cells have a similar advantage in that 1) were assessed in a single intertransverse process lumbar they not only deliver an osteoinductive signal, but also spine fusion model [32]. LMP-1 is a transcription factor contain osteogenic precursor cells. that can promote endogenous expression of BMPs. Stem cells harvested from fat also have potential for use Control animals were treated with a reverse copy of the in tissue engineering. Zuk et al. demonstrated that human cDNA that does not express LMP-1. Successful spine fusion adipose tissue obtained at the time of liposuction contains was noted in all 11 sites treated with the LMP-1 transduced MOLECULAR THERAPY Vol. 6, No. 2, August 2002 Copyright © The American Society of Gene Therapy
143
REVIEW
A
doi:10.1006/mthe.2002.0663, available online at http://www.idealibrary.com on IDEAL
B
FIG. 2. Responses of articular cartilage to injury. Partial-thickness injuries produce defects that fail to heal (A). Marrow enters full-thickness defects that penetrate the subchondral bone (B). Within the marrow are mesenchymal stem cells (MSCs) that synthesize a fibrocartilagenous repair tissue. Although this material fills the defect, it cannot withstand physiological loads for prolonged periods and eventually degenerates.
bone marrow cells. The spine fusion sites treated with control bone marrow cells alone did not fuse. Buffy-coat cells have also been engineered to promote bone formation using an LMP-1-containing adenovirus. Rabbits treated with these transduced peripheral bloodderived cells develop solid spine fusions [23]. Peripheral blood cells are appealing as a cellular gene delivery vehicle because these cells are easier to harvest than bone
144
marrow cells. However, it is unlikely that cells derived from peripheral blood have the same autocrine effects that are seen with transduced bone marrow cells. The results with LMP-1 in these preclinical models are impressive and could be adapted to treat other bone repair problems. However, no mice mutant or transgenic for LMP1 have been reported and little is known of the possible side effects of LMP-1 overexpression. Wang et al. used a recombinant adenovirus containing the BMP-2 cDNA to induce an intertransverse lumbar spinal fusion in rats [33]. Autologous bone marrow cells were harvested from Lewis rats, expanded in tissue culture, and then transduced with the BMP-2-containing adenovirus. These transduced cells were then loaded onto either guanidine-extracted demineralized bone matrix or collagen sponges and then implanted into the appropriate fusion site. The BMP-2-producing bone marrow cells induced spine fusion in all 15 spines that were treated. Bone marrow cells alone did not induce spine fusion. The results of these gene transfer preclinical studies are quite exciting and clearly demonstrate the adaptability of gene therapy for human use. The ability to deliver growth factors via gene therapy may lead to the development of less invasive operative techniques such as laparoscopic spine fusion. Such a development carries the potential for reduced operative morbidity, shortened time to wound healing, and diminished costs. The goal is to develop gene therapy as a part of an overall comprehensive tissue engineering strategy to enhance bone repair. Clearly gene therapy is not necessary for all patients, but it does have the potential to enable orthopedic surgeons to treat difficult bone repair problems that can not be handled successfully with our present technology.
ARTICULAR CARTILAGE REPAIR Cartilage is an avascular, aneural, alymphatic tissue that overlies the ends of the long bones and helps ensure the smooth, almost frictionless movement of joints. Approximately 95% of the tissue consists of a highly hydrated extracellular matrix with a unique composition and microarchitecture, enabling it to bear load effectively. This matrix is synthesized and maintained by the sparse population of articular chondrocytes embedded within it. The poor ability of cartilage to heal has been recognized since the time of Hippocrates. Indeed, it mounts almost no reparative response to traumatic injuries that fail to penetrate the subchondral bone underlying the cartilage. Such lesions are often painful and debilitating. They may also progress to osteoarthritis, a disease characterized by focal erosion of the articular cartilage. Deeper MOLECULAR THERAPY Vol. 6, No. 2, August 2002 Copyright © The American Society of Gene Therapy
doi:10.1006/mthe.2002.0663, available online at http://www.idealibrary.com on IDEAL
injuries that engage the subchondral bone are associated with the entry of blood and marrow contents into the lesion. Progenitor cells within these infiltrates form fibrocartilagenous repair tissue which, although inferior to native articular cartilage, can provide symptomatic improvement for extended periods of time before deteriorating because of defective load-bearing properties (Fig. 2). Cartilage injuries are common, symptomatic, and difficult to repair by surgical means alone. Therefore, there is great interest in improving outcomes through harnessing the biology of the system [34]. Among the more commonly used procedures is abrasion arthroplasty, in which the subchondral plate is deliberately violated to encourage the influx of marrow elements. Less commonly, autologous cartilage, harvested from the margins of the joint that do not bear load, is used as a source of autologous chondrocytes in partial-thickness defects. After expansion in culture, the cells are surgically transplanted to the injured site, where they effect a repair. This procedure is largely restricted to knee joints and remains controversial. Problems common to all methods of attempted cartilage repair are fusion of the repair cartilage with the surrounding endogenous undamaged cartilage and the formation of a uniform cartilaginous surface. The repair of full-thickness lesions is also challenged by the need for reconstitution of the underlying subchondral bone. Biological repair is further challenged by the need to recapitulate the exquisite molecular microarchitecture of cartilage. The mechanically demanding environment of articular cartilage is a harsh discriminator of inadequate repair tissue.
GENE THERAPY APPROACHES
TO
CARTILAGE REPAIR
Cartilage healing presents far greater challenges than bone healing to gene therapy. Unlike bone, cartilage lacks a robust, native healing response that could serve as the basis of a gene therapy. Furthermore, there is no clinical experience of cartilage repair using recombinant proteins upon which to draw. Nevertheless, gene therapy approaches to cartilage repair are encouraged by the ability of various gene products to enhance chondrogenesis [2]. Examples include growth factors [35] such as insulin-like growth factor-1 (IGF-1), transforming growth factor- (TGF-), fibroblast growth factors, and various members of the BMP family, as well as transcription factors such as SOX-9 [36], certain signaling molecules such as SMADs [37], and molecules that inhibit apoptosis such as BCL-2 [38]. Growth factors, however, are difficult to administer exogenously to sites of cartilage injury in a sustained and therapeutically useful fashion, whereas the intracellular mediators have the additional problem of needing to enter responsive cells. Gene delivery strategies promise to overcome these limitations. One issue to be addressed in implementing such strategies is the choice of target tissue. MOLECULAR THERAPY Vol. 6, No. 2, August 2002 Copyright © The American Society of Gene Therapy
REVIEW
Synovium is the most accessible tissue in the joint. It is readily transduced by a wide variety of different vectors and has already served as the site of gene transfer in a clinical study related to the gene therapy of arthritis [39]. In the context of cartilage repair, genes encoding secreted repair factors could be delivered to synovium, which would then serve as a local, intraarticular source of diffusible transgene products. The basis of this approach has been demonstrated in rabbit knee joints. Adenoviral delivery of IGF-1 or BMP-2 cDNAs to synovium increases matrix synthesis by articular cartilage in the same knee joint by about 50% [40]. However, delivery of genes to synovium can generate cartilage ectopically within the joint, and lead to the formation of osteophytes [41]. Moreover, transfer of a TGF- cDNA is deleterious, causing massive fibrosis of the joint capsule (unpublished data). However, the local, intracartilagenous expression of growth factors might exploit the chondrogenic properties of these molecules without provoking adverse extra-cartilagenous reactions. In addition, the effectiveness of all such potentially reparative factors may be improved by targeted delivery to cartilage. The dense matrix that surrounds chondrocytes impedes in vivo gene delivery to these cells by nearly all vectors that have been tested. Nevertheless, transfer of marker genes to chondrocytes in situ has been reported with liposomes incorporating the hemagglutinating virus of Japan (Sendai virus) [42] and with AAV [43]. However, cartilage repair requires the presence of cells, as well as matrix, within the lesion; gene delivery to synovium or preexisting cartilage does not by itself satisfy this requirement. As noted, autologous chondrocyte transplantation is already used clinically [44]. It is thus tempting to improve the quality of the repair tissue by ex vivo genetic modification of the chondrocytes while they are undergoing expansion in culture. These cells are readily transduced by several viral vectors, including retrovirus, adenovirus, AAV, and lentivirus, and can be transfected with certain nonviral formulations [45]. Adenoviral transfer of cDNAs encoding IGF-1, BMP-2, or TGF- to cultures of articular chondrocytes dramatically increases matrix production, even in the presence of interleukin-1 (IL-1), a powerful physiological inhibitor of this process [46,47]. Moreover, use of IGF1 has the additional advantage of maintaining the phenotypic stability of the cells [48]. There have been few experimental studies of the fate of genetically modified chondrocytes after ex vivo transfer to defects in articular cartilage, but the data question the long-term survival of these cells [49–51]. An alternative to the transfer of chondrocytes is to adapt tissue engineering methods to generate genetically modified cartilage, rather than cells in suspension, for transplant. Encouraging results have been obtained when the IGF-1 cDNA is introduced into chondrocytes by transfection, and cartilage is grown from these cells in a bioreactor [52]. Similar experiments, in which cDNAs encod-
145
REVIEW
doi:10.1006/mthe.2002.0663, available online at http://www.idealibrary.com on IDEAL
ing various BMPs were transfected into chondrocytes, identified BMP-7 as a particularly promising molecule for this purpose [53]. Using articular chondrocytes as vehicles for ex vivo gene therapy in cartilage repair has several drawbacks. Supplies of autologous tissue are severely limited, and the culture techniques used to expand the chondrocytes lead to phenotypic modulation, commonly referred to as de-differentiation. It is not clear that these cells, or indeed authentic chondrocytes (genetically modified or not), can generate a repair tissue with the appropriate biochemistry, microarchitecture, and biomechanical properties. If these criteria are not met, the tissue fails. Thus, there is considerable interest in using chondroprogenitor cells, whose greater plasticity may enable the efficient recapitulation of true articular cartilage by way of biologically authentic routes of chondrogenesis. Various tissues, including bone marrow, perichondrium and periosteum, and (perhaps) adipose tissue and skin, contain cells with the ability to differentiate into articular chondrocytes. The chondroprogenitor cells of bone marrow are thought to be mesenchymal stem cells (MSCs); whether these cells are responsible for the chondrogenic potential of other tissues remains an unanswered question. Under the appropriate conditions, which in vitro include the presence of TGF-, dexamethasone, and spherical shape, MSCs predictably differentiate into chondrocytes [54]. When transplanted into experimental lesions in the articular cartilages of rabbits, MSCs effect better repair than similar transplants of articular chondrocytes [55]. Whether the appropriate genetic modification of MSCs will enhance repair even further remains to be demonstrated experimentally [56]. The resistance of MSCs to efficient transduction by many vectors hinders such studies. Periosteal cells, however, have been retrovirally transduced with the BMP-7 cDNA with promising results [57]. Likewise, efficient transfection of perichondrial cells has been reported [58]. One disadvantage of these ex vivo approaches is the need for the appropriate scaffolds and matrices upon which to seed the cells before transplant or complex surgical procedures, as in the case of autologous cartilage transplantation. There is preliminary evidence that genetically modified chondrocytes and MSCs selectively adhere to articular cartilage [50,59], suggesting that simple intraarticular injection of the cells may serve as an adequate delivery method. In situ gene delivery techniques that obviate the need for ex vivo cell culture would bring many advantages. Along these lines, methods are being developed for introducing vectors into sites of full-thickness lesions such that chondrogenic cells become transduced within the lesion [60].
OTHER CARTILAGINOUS TISSUES Certain joints, such as the knee, contain additional cartilaginous structures known as menisci. These are also
146
frequently subject to symptomatic injuries and, like articular cartilage, the menisci have limited ability to heal spontaneously. Gene therapy approaches similar to those used to address healing of articular cartilage can be applied to the healing of meniscus. Both retroviral ex vivo and adenoviral in vivo transfer methods have been successfully applied for the transfer of marker genes to the menisci of rabbits and dogs [61]. Adenoviral transfer of TGF- cDNA to cultures of human and canine meniscal cells leads to very large increases in matrix synthesis [62], but this has not yet been attempted in an in vivo model of meniscal healing. The intervertebral disc is another important cartilaginous structure whose injury is a major clinical and social problem. It, too, lacks an efficient endogenous repair mechanism and is the subject of gene transfer approaches to healing [63]. Cells of the intervertebral disc are easily transduced by a variety of different vectors. The direct, intradiscal injection of first-generation adenoviral vectors leads to high levels of transgene expression for over a year in immunocompetent rabbits. The absence of immune surveillance within this dense, avascular, alymphatic tissue and the low mitotic index of disc cells presumably account for the remarkable longevity of transgene expression. Delivery of growth factor genes to disc cells leads to large increases in matrix synthesis, but it is not yet known whether this alone is sufficient to produce healing of discal lesions. Several different gene therapy strategies for cartilage repair are under investigation, but research remains at the developmental stage and, unlike the case with bone healing, human clinical trials are not yet on the horizon. In particular, it is still unclear which gene, or combination of genes, will provide the most effective repair, and the choice of vector is not apparent. The most appropriate cellular targets for gene transfer have not been identified, and there has been no published demonstration of efficacy in an animal model. Nevertheless, gene transfer approaches hold much promise as a means of enhancing the repair of cartilaginous tissues, especially because this remains a major clinical problem.
GENERAL CONCLUSIONS There is a major, unmet clinical need for improved ways to heal bone and cartilage. Through its ability to deliver osteogenic and chondrogenic substances in a precise, economical, local, authentic, and sustained fashion, gene transfer is a particularly promising technology with which to address this need. Impressive preclinical studies have demonstrated the power of gene therapy in several different models of bone healing, and clinical trials would seem to be only a matter of time. In several important ways, cartilage healing is a more intractable biological problem than bone healing, and less headway has been made in adapting gene therapy to this end. However, considerable experimental progress has been made and future success is likely. MOLECULAR THERAPY Vol. 6, No. 2, August 2002 Copyright © The American Society of Gene Therapy
doi:10.1006/mthe.2002.0663, available online at http://www.idealibrary.com on IDEAL
ACKNOWLEDGMENT J.R.L. is supported by a grant from the National Institutes of Health (AR467890IAI)
REFERENCES 1. Evans, C. H., and Robbins, P. D. (1995). Possible orthopaedic applications of gene therapy. J. Bone Joint Surg. Am. 77: 1103–1114. 2. Evans, C. H., et al. (2000). Using gene therapy to protect and restore cartilage. Clin. Orthop. 379 (Suppl.): S214–s219. 3. Oakes, D. A., and Lieberman, J. R. (2000). Osteoinductive applications of regional gene therapy: ex vivo gene transfer. Clin. Orthop. 379 (Suppl.): S101–s112. 4. Scaduto, A. A., and Lieberman, J. R. (1999). Gene therapy for osteoinduction. Orthop. Clin. North Am. 30: 625–633. 5. Lieberman, J. R., et al. (1999). The effect of regional gene therapy with bone morphogenetic protein-2-producing bone-marrow cells on the repair of segmental femoral defects in rats. J. Bone Joint Surg. Am. 81: 905–917. 6. Colterjohn, N. R., and Bednar, D. A. (1997). Procurement of bone graft from the iliac crest. An operative approach with decreased morbidity. J. Bone Joint Surg. Am. 79: 756–759. 7. Arrington, E. D., et al. (1996). Complications of iliac crest bone graft harvesting. Clin. Orthop. 329: 300–309. 8. Gerhart, T. N., et al. (1993). Healing segmental femoral defects in sheep using recombinant human bone morphogenetic protein. Clin. Orthop. 293: 317–326. 9. Cook, S. D., et al. (1994). Recombinant human bone morphogenetic protein-7 induces healing in a canine long-bone segmental defect model. Clin. Orthop. 301: 302–312. 10. Cook, S. D., et al. (1995). Effect of recombinant human osteogenic protein-1 on healing of segmental defects in non-human primates. J. Bone Joint Surg. Am. 77: 734–750. 11. Sandhu, H. S., et al. (1995). Evaluation of rhBMP-2 with an OPLA carrier in a canine posterolateral (transverse process) spinal fusion model. Spine 20: 2669–2682. 12. Boden, S. D., et al. (1998). Laparoscopic anterior spinal arthrodesis with rhBMP-2 in a titanium interbody threaded cage. J. Spinal Disord. 11: 95–101. 13. Friedlaender, G. E., et al. (2001). Osteogenic protein-1 (bone morphogenetic protein7) in the treatment of tibial nonunions. J. Bone Joint Surg. Am. 83-A (Suppl. 1 (Pt. 2)): S151–S158. 14. Boden, S. D., et al. (2000). The use of rhBMP-2 in interbody fusion cages. Definitive evidence of osteoinduction in humans: a preliminary report. Spine 25: 376–381. 15. Crystal, R. G. (1995). Transfer of genes to humans: early lessons and obstacles to success. Science 270: 404–410. 16. Anderson, W. F. (1998). Human gene therapy. Nature 392 (Suppl.): 25–30. 17. Wilson, J. M. (1996). Adenoviruses as gene-delivery vehicles. N. Engl. J. Med. 334: 1185–1187. 18. Baltzer, A. W., et al. (2000). Genetic enhancement of fracture repair: healing of an experimental segmental defect by adenoviral transfer of the BMP-2 gene. Gene Ther. 7: 734–739. 19. Musgrave, D. S., et al. (1999). Adenovirus-mediated direct gene therapy with bone morphogenetic protein-2 produces bone. Bone 24: 541–547. 20. Okubo, Y., et al. (2000). Osteoinduction by bone morphogenetic protein-2 via adenoviral vector under transient immunosuppression. Biochem. Biophys. Res. Commun. 267: 382–387. 21. Fang, J., et al. (1996). Stimulation of new bone formation by direct transfer of osteogenic plasmid genes. Proc. Natl. Acad. Sci. USA 93: 5753–5758. 22. Bonadio, J., et al. (1999). Localized, direct plasmid gene delivery in vivo: prolonged therapy results in reproducible tissue regeneration. Nat. Med. 5: 753–759. 23. Viggeswarapu, M., et al. (2001). Adenoviral delivery of LIM mineralization protein-1 induces new-bone formation in vitro and in vivo. J. Bone Joint Surg. Am. 83-A: 364–376. 24. Bruder, S. P., Jaiswal, N., and Haynesworth, S. E. (1997). Growth kinetics, self-renewal, and the osteogenic potential of purified human mesenchymal stem cells during extensive subcultivation and following cryopreservation. J. Cell Biochem. 64: 278–294. 25. Bruder, S. P., et al. (1998). Bone regeneration by implantation of purified, cultureexpanded human mesenchymal stem cells. J. Orthop. Res. 16: 155–162. 26. Bruder, S. P., et al. (1998). The effect of implants loaded with autologous mesenchymal stem cells on the healing of canine segmental bone defects. J. Bone Joint Surg. Am. 80: 985–996. 27. Krebsbach, P. H., et al. (2000). Gene therapy-directed osteogenesis: BMP-7-transduced human fibroblasts form bone in vivo. Hum. Gene Ther. 11: 1201–1210. 28. Lee, J. Y., et al. (2001). Effect of bone morphogenetic protein-2-expressing musclederived cells on healing of critical-sized bone defects in mice. J. Bone Joint Surg. Am. 83-A: 1032–1039. 29. Zuk, P. A., et al. (2001). Multilineage cells from human adipose tissue: implications for cell-based therapies. Tissue Eng. 7: 211–228. 30. Dragoo, J., et al. (2001). Bone induction by genetically manipulated stem cells derived from fat. J. Bone Min. Res. 16 (suppl. 1): 5500. 31. Deutsche Banc, A. B. (2001). Estimates and Company Information.
MOLECULAR THERAPY Vol. 6, No. 2, August 2002 Copyright © The American Society of Gene Therapy
REVIEW
32. Boden, S. D., et al. (1998). Lumbar spine fusion by local gene therapy with a cDNA encoding a novel osteoinductive protein (LMP-1). Spine 23: 2486–2492. 33. Wang, J., et al. (2002). Intertransverse process fusion in Lewis rats using bone marrow cells infected with a BMP-2 containing adenovirus. Trans. Orthop. Res. Soc. 25: 327. 34. Athanasiou, K. A., et al. (2001). Basic science of articular cartilage repair. Clin. Sports Med. 20: 223–247. 35. O’Connor, W. J., et al. (2000). The use of growth factors in cartilage repair. Orthop. Clin. North Am. 31: 399–410. 36. Uusitalo, H., et al. (2001). Accelerated up-regulation of L-Sox5, Sox6, and Sox9 by BMP-2 gene transfer during murine fracture healing. J. Bone Miner. Res. 16: 1837–1845. 37. Fujii, M., et al. (1999). Roles of bone morphogenetic protein type I receptors and Smad proteins in osteoblast and chondroblast differentiation. Mol. Biol. Cell 10: 3801–3813. 38. Feng, L., et al. (1998). Evidence of a direct role for Bcl-2 in the regulation of articular chondrocyte apoptosis under the conditions of serum withdrawal and retinoic acid treatment. J. Cell. Biochem. 71: 302–309. 39. Evans, C. H., et al. (1996). Clinical trial to assess the safety, feasibility, and efficacy of transferring a potentially anti-arthritic cytokine gene to human joints with rheumatoid arthritis. Hum. Gene Ther. 7: 1261–1280. 40. Mi, Z., et al. (2000). Adenovirus-mediated gene transfer of insulin-like growth factor 1 stimulates proteoglycan synthesis in rabbit joints. Arthritis Rheum. 43: 2563–2570. 41. Gelse, K., et al. (2001). Fibroblast-mediated delivery of growth factor complementary DNA into mouse joints induces chondrogenesis but avoids the disadvantages of direct viral gene transfer. Arthritis Rheum. 44: 1943–1953. 42. Tomita, T., et al. (1997). In vivo direct gene transfer into articular cartilage by intraarticular injection mediated by HVJ (Sendai virus) and liposomes. Arthritis Rheum. 40: 901–906. 43. Arai, Y., et al. (2000). Gene delivery to human chondrocytes by an adeno associated virus vector. J. Rheumatol. 27: 979–982. 44. Brittberg, M., et al. (2001). Autologous chondrocytes used for articular cartilage repair: an update. Clin. Orthop. 391 (Suppl.): S337–S348. 45. Madry, H., and Trippel, S. B. (2000). Efficient lipid-mediated gene transfer to articular chondrocytes. Gene Ther. 7: 286–291. 46. Shuler, F. D., et al. (2000). Increased matrix synthesis following adenoviral transfer of a transforming growth factor 1 gene into articular chondrocytes. J. Orthop. Res. 18: 585–592. 47. Smith, P., et al. (2000). Genetic enhancement of matrix synthesis by articular chondrocytes: comparison of different growth factor genes in the presence and absence of interleukin-1. Arthritis Rheum. 43: 1156–1164. 48. Nixon, A. J., et al. (2000). Insulinlike growth factor-I gene therapy applications for cartilage repair. Clin. Orthop. 379 (suppl.): S201–S213. 49. Kang, R., et al. (1997). Ex vivo gene transfer to chondrocytes in full-thickness articular cartilage defects: a feasibility study. Osteoarth. Cartilage 5: 139–143. 50. Baragi, V. M., et al. (1997). Transplantation of adenovirally transduced allogeneic chondrocytes into articular cartilage defects in vivo. Osteoarth. Cartilage 5: 275–282. 51. Ostrander, R. V., et al. (2001). Donor cell fate in tissue engineering for articular cartilage repair. Clin. Orthop. 389: 228–237. 52. Madry, H., Zurakowski, D., and Trippel, S. B. (2001). Overexpression of human insulinlike growth factor-I promotes new tissue formation in an ex vivo model of articular chondrocyte transplantation. Gene Ther. 8: 1443–1449. 53. Kaps, C., et al. (2002). Bone morphogenetic proteins promote cartilage differentiation and protect engineered artificial cartilage from fibroblast invasion and destruction. Arthritis Rheum. 46: 149–162. 54. Johnstone, B., et al. (1998). In vitro chondrogenesis of bone marrow-derived mesenchymal progenitor cells. Exp. Cell Res. 238: 265–272. 55. Wakitani, S., et al. (1994). Mesenchymal cell-based repair of large, full-thickness defects of articular cartilage. J. Bone Joint Surg. Am. 76: 579–592. 56. Yoo, J. U., et al. (2000). Chondrogenitor cells and gene therapy. Clin. Orthop. 379 (Suppl.): S164–S170. 57. Mason, J. M., et al. (1998). Expression of human bone morphogenic protein 7 in primary rabbit periosteal cells: potential utility in gene therapy for osteochondral repair. Gene Ther. 5: 1098–1104. 58. Goomer, R. S., et al. (2000). Nonviral in vivo gene therapy for tissue engineering of articular cartilage and tendon repair. Clin. Orthop. 379 (Suppl.): S189–S200. 59. Doherty, P. J., et al. (1998). Resurfacing of articular cartilage explants with geneticallymodified human chondrocytes in vitro. Osteoarth. Cartilage 6: 153–159. 60. Palmer, G., et al. (2002). Direct gene transfer to osteochondral defects. Trans. Orthop. Res. Soc. 27: 292. 61. Goto, H., et al. (1999). Transfer of lacZ marker gene to the meniscus. J. Bone Joint Surg. Am. 81: 918–925. 62. Goto, H., et al. (2000). Gene therapy for meniscal injury: enhanced synthesis of proteoglycan and collagen by meniscal cells transduced with a TGF(1)gene. Osteoarth. Cartilage 8: 266–271. 63. Nishida, K., et al. (1999). Modulation of the biologic activity of the rabbit intervertebral disc by gene therapy: an in vivo study of adenovirus-mediated transfer of the human transforming growth factor  1 encoding gene. Spine 24: 2419–2425.
147
REVIEW
Gene Transfer Approaches to the Healing of Bone and Cartilage Jay R. Lieberman,1,* Steven C. Ghivizzani,2 and Christopher H. Evans2 1Department
of Orthopaedic Surgery, The David Geffen School of Medicine at UCLA, Los Angeles, California 90095, USA for Molecular Orthopaedics, Harvard Medical School, Boston, Massachusetts 02115, USA *To whom correspondence and reprint requests should be addressed. Fax: (310) 206-0063. E-mail: [email protected]. 2Center
Gene therapy is traditionally considered a treatment modality for genetic diseases and a limited number of complex, life-threatening disorders, such as cancer. However, gene therapy also holds promise as a novel means of treating more mundane conditions, including broken bones and damaged cartilages [1–3]. These injuries frequently present enormous clinical challenges to both orthopedic surgeons and rheumatologists, and can result in considerable human suffering and economic loss. Various gene products, particularly growth factors, show remarkable promise as agents that can improve the healing of bone, cartilage, and other connective tissues. Their clinical utility, however, is limited by delivery problems. The attraction of gene transfer approaches is the unique ability to deliver authentically processed gene products to precise anatomical locations at therapeutic levels for sustained periods of time. However, unlike the treatment of chronic disease, it is neither necessary nor desirable for transgene expression to persist beyond the few weeks or months needed to achieve healing. It is also unlikely that the level of transgene expression will need to be closely regulated. Moreover, achieving targeted delivery is not an issue, because the same orthopedic procedures that are already used when operating on bones and joints can be adapted for the purposes of targeted gene transfer. Thus, the applications proposed in this review represent several of the few examples in which gene therapy has a good chance of clinical success using existing technology.
BONE HEALING There is great interest in developing gene therapy to enhance bone repair because the potential for sustained protein production may enable the host to respond to an osteoinductive stimulus in a more robust fashion. Regional gene therapy is an attractive option to enhance bone formation and repair because the genes can be delivered to a specific anatomic site, and the duration of protein production can be determined by selection of a particular vector. In fact, gene therapy for bone repair may be easier to develop for human use than for other diseases because the requirements for protein production are more variable. In many cases, protein production will
MOLECULAR THERAPY Vol. 6, No. 2, August 2002 Copyright © The American Society of Gene Therapy
only be required until the bone has healed [3–5]. One exception to this would be the treatment of osteoporosis, which may require the development of systemic gene therapy. In the future, gene therapy may be one option in a comprehensive strategy for the tissue engineering of bone. Gene therapy will not be for all patients, as some may be best treated with an autologous bone graft and others, with recombinant proteins and/or bone graft substitutes. However, in some patients gene therapy may offer the best potential for healing [3–5]. Enhanced bone formation is often required to treat bone loss associated with trauma, revision total joint arthroplasty, pseudarthrosis of the spine, and tumor resection. Autologous bone grafts harvested from the pelvis are the current gold standard for treating bone loss problems. However, only a limited amount of autogenous bone graft is available, and there are problems with donor site morbidity [6]. In a large retrospective study, a 10% incidence of minor complications and a 5.8% incidence of major complications were associated with bone graft harvest [7]. These limitations have prompted increased interest in the development of alternative bone grafts, such as allografts and demineralized bone matrices. Allograft bone has limited osteoinductive potential, and there is also the possibility of transmission of viral disease. At this time, demineralized bone matrices seem to be most appropriately used as bone graft extenders rather than bone graft substitutes. These materials are osteoconductive and provide a scaffold to promote bone formation, but they have limited osteoinductive activity. Several osteoinductive growth factors that can enhance bone repair have been identified. In particular, the bone morphogenetic proteins (BMPs) have been shown to induce bone formation in a variety of different animal models. Both BMP-2 and BMP-7 (osteogenic protein-1) have been used to heal critical-sized defects in sheep, dogs, and nonhuman primates [8–10] and to induce the fusion of the spine in canines and nonhuman primates [11,12]. Although these preclinical models demonstrate the potential efficacy of different BMPs, these animal models do not truly simulate the clinical situations often associated with bone loss problems. These studies were carried out using young animals
141
REVIEW
doi:10.1006/mthe.2002.0663, available online at http://www.idealibrary.com on IDEAL
with well-vascularized bone and intact soft tissue. In contrast, the healing potential of the bone in a clinical setting may be quite limited because of compromised vascularity, limited bone stock, and abundant fibrous tissue. In addition, bone repair may be inhibited in diabetic, elderly, or nicotine-addicted patients [4]. Although recombinant BMP has been used successfully in a variety of animal studies, the success rate in humans is variable. In two clinical trials, large doses of BMP were required to induce adequate bone repair, suggesting that the mode of BMP delivery still requires further optimization [13,14]. It is not known what side effects are associated with such high doses of BMP. Therefore, there is concern that a single exposure to an exogenous growth factor may not induce an adequate osteogenic signal in many clinical situations. Gene therapy has the potential to provide more sustained protein release when necessary and to deliver protein in a more physiologic manner than recombinant proteins [3–4].
IN VIVO GENE TRANSFER STRATEGY In vivo gene delivery involves directly delivering the gene into a specific anatomic site by transducing local cells. The advantages of this strategy are that it is a relatively simple technique and has the potential for lower cost. The disadvantages are the difficulty in targeting specific cells for transduction and in achieving high transduction efficiency. When using an adenovirus vector there is the additional concern about the development of an immune response as a result of the viral particles injected directly into the anatomic site, which can inhibit transgene expression [15–17]. Both viral and nonviral delivery methods have successfully induced bone formation in several animal models. Baltzer et al. used a first-generation adenoviral vector containing the cDNA for BMP-2 to heal criticalsized femoral defects in New Zealand white rabbits [18]. In this experiment, healing of the bone defects was noted 7 weeks after viral injection. Ectopic bone formation has also been induced by the injection of a BMP-2containing adenovirus into the thigh muscles of nude rats. However, bone formation was inhibited in immune competent Sprague-Dawley rats. In addition, an injection of the BMP-2 adenovirus produced less bone when injected into the triceps of immune competent mice compared with the same adenovirus injected into the triceps of nude mice [19]. Okubu et al. noted increased bone formation after the direct injection of the BMP-2-containing adenovirus when the rats were also given cyclophosphamide. Although these experiments resulted in bone formation, they also suggest that the presence of an immune response against the adenoviral vectors limits the biological activity of the secreted protein [20]. To avoid potential problems with the adenoviral vectors, nonviral delivery methods are being developed.
142
Plasmid DNA encoding parathyroid hormone (PTH) 1-34 or BMP-4 placed in a gene-activated matrix (GAM) is one example of a nonviral gene therapy strategy that has been successfully used to promote osteogenesis in a rat bone-defect model. In this strategy, the plasmid PTH 1-34 or BMP-4 is incorporated into a collagen matrix and then implanted directly into a bone defect. Local fibroblasts in the area of the defect can then acquire the DNA and become local bioreactors for synthesis of the PTH or BMP-4 protein [21]. This strategy was successfully used to heal 5-mm defects in rats. A combination of both PTH 134 and BMP-4 showed enhanced healing (4 weeks versus 9 weeks) when compared with cDNA for either BMP-4 or PTH 1-34 alone. Bonadio et al. noted 6 weeks of in vivo protein production using a GAM containing PTH 1-34 cDNA to treat a canine tibial defect [22]. However, insufficient bone was produced to heal a critical-sized tibial defect in this canine model. These results suggest that either the local transfection efficiency is low or that PTH 1-34 has limited osteoinductive potential. Overall, the concept of using a GAM is attractive, and further study is necessary to determine the appropriate cDNA to use in this matrix.
EX VIVO GENE TRANSFER STRATEGIES In addition to providing an osteoinductive gene to a desired site, the advantage of an ex vivo approach is that it enables surgeons to select specific cells (that is, bone marrow cells, muscle cells, or stem cells) that can participate in osteoinduction. This may enhance the bone repair process as both autocrine and paracrine responses can be elicited with secretion of the desired gene product. In addition, ex vivo methods may be safer than in vivo techniques when working with an adenovirus because no viral particles or DNA complexes are injected into the body. In general, ex vivo techniques have a high efficiency of cell transduction. The major disadvantage of this strategy is the requirement of an extra harvesting step and the increased time and cost of the process. However, it is theoretically feasible to harvest cells from the patient, have a short period of infection and then reimplant the transduced cells at the appropriate anatomic site [23]. Because they are osteogenic, bone marrow cells are obvious candidates for ex vivo gene transfer in bone healing. Furthermore, bone marrow cell harvest is a well-established procedure and methods have been developed to produce a purified population of mesenchymal cells. These progenitor cells can be expanded over 1 billion-fold without losing their potency, and they have demonstrated the ability to induce sufficient bone formation to heal critical-sized defects in both rat and canine models [24–26]. Although these results are promising and have the possible advantage of avoiding the use of gene transfer techniques, it remains to be determined whether these cells alone would be able to heal a bone defect in a clinical situation. However, genetically manipulating these cells to MOLECULAR THERAPY Vol. 6, No. 2, August 2002 Copyright © The American Society of Gene Therapy
doi:10.1006/mthe.2002.0663, available online at http://www.idealibrary.com on IDEAL
REVIEW
produce an osteoinductive protein would further enhance the bone repair response. Lieberman et al. [5] used an ex vivo gene transfer approach to heal critical-size femoral defects in a rat model. Rat bone marrow cells were cultured, infected with a BMP-2-containing adenovirus, and then implanted in the femoral defect site. Histomor- FIG. 1. Radiographs of specimen made 2 months after the surgical procedure to test the efficacy of the ex vivo gene phometric analysis revealed transfer strategy with BMP-2-producing bone marrow cells that were created by adenoviral gene transfer. We used 20 a more robust pattern of mg of guanidine hydrochloride-extracted, demineralized bone matrix as a carrier in all defects. (A) rhBMP-2 (20 g, positive control). The healed defect is filled with lace-like trabecular bone. (B) BMP-2-producing bone marrow cells bone formation in femoral (5 ⫻ 106). Dense, coarse trabecular bone, remodeled to form a new cortex, was present. (C) -Galactosidase-prodefects treated with BMP-2- ducing rat bone marrow cells. (D) Rat bone marrow cells alone. (E) Demineralized bone matrix alone. Groups (C), (D), producing bone marrow and (E) were negative controls with minimal bone formation. cells as compared with defects treated with recombinant BMP-2 protein placed on a collagen sponge [5] (Fig. a fibroblast-like population of cells that can behave like 1). Although biomechanical testing of the healed femurs mesenchymal stem cells [29]. When these lipoaspirate did not show any differences between these groups with cells are grown in the appropriate medium, they can difrespect to energy-to-failure or torque-to-failure, this may ferentiate into bone, cartilage, fat, or muscle. Although the be due to a lack of sensitivity of the rat model. The more full osteogenic potential of these cells is still being anarobust bone formed by cells transduced with BMP-2 may lyzed, it is known that these stem cells can form bone in be the result of a more continuous or physiological release vivo when they are transduced with a BMP-2-containing of BMP-2 protein compared with the kinetics of BMP adenovirus and implanted into a SCID mouse muscle release from a collagen sponge. It is hypothesized that a pouch [30]. Further work with these types of cells is being more vigorous osteogenic activity is seen because both carried out (J.R.L.’s laboratory). paracrine and autocrine responses occur, that is, the transSPINE FUSION duced bone marrow cells can respond to the BMP-2 protein that they are secreting. Spinal fusion is one of the most commonly performed A variety of other cell types have been used to induce orthopedic surgical procedures in the United States, with bone formation. BMP-7-transfected rat skin fibroblasts over 980,000 performed every year. Approximately onehave been shown to heal critically-sized calvarial defects in third of these procedures requires a bone graft [31]. In Lewis rats. Skin fibroblasts are an attractive cellular delivgeneral, autogenous bone graft is a successful method for ery vehicle because they are easy to harvest and are readienhancing spine fusion, but non-union rates range from ly available in all patients [27]. There is one caveat when 5% to 35%. A variety of factors can influence the success using fibroblasts to deliver osteoinductive signals: fibrous of the spine fusion including mechanical instability of tissue can actually inhibit bone formation and is found in the spine, the stability of fixation, the quality of bone, fracture non-union sites. the health of surrounding soft tissue, the type of bone Genetically engineered muscle-derived cells also have graft used, and the concurrent use of medications and osteogenic potential [28]. Lee et al. demonstrated that drugs such as nicotine. Concerns regarding the osteoinmuscle cells transduced with a BMP-2-containing adenductive potential of various exogenous growth factors ovirus were able to heal calvarial defects in SCID mice. have lead investigators to explore the potential of gene These investigators also noted that a small percentage of therapy to induce fusion of the spine. muscle-derived cells implanted in calvarial defects actually Bone marrow cells transduced with a DNA plasmid condifferentiated into osteoblasts in vivo. Both muscle cells taining the cDNA for LMP-1 (LIM mineralization proteinand bone marrow cells have a similar advantage in that 1) were assessed in a single intertransverse process lumbar they not only deliver an osteoinductive signal, but also spine fusion model [32]. LMP-1 is a transcription factor contain osteogenic precursor cells. that can promote endogenous expression of BMPs. Stem cells harvested from fat also have potential for use Control animals were treated with a reverse copy of the in tissue engineering. Zuk et al. demonstrated that human cDNA that does not express LMP-1. Successful spine fusion adipose tissue obtained at the time of liposuction contains was noted in all 11 sites treated with the LMP-1 transduced MOLECULAR THERAPY Vol. 6, No. 2, August 2002 Copyright © The American Society of Gene Therapy
143
REVIEW
A
doi:10.1006/mthe.2002.0663, available online at http://www.idealibrary.com on IDEAL
B
FIG. 2. Responses of articular cartilage to injury. Partial-thickness injuries produce defects that fail to heal (A). Marrow enters full-thickness defects that penetrate the subchondral bone (B). Within the marrow are mesenchymal stem cells (MSCs) that synthesize a fibrocartilagenous repair tissue. Although this material fills the defect, it cannot withstand physiological loads for prolonged periods and eventually degenerates.
bone marrow cells. The spine fusion sites treated with control bone marrow cells alone did not fuse. Buffy-coat cells have also been engineered to promote bone formation using an LMP-1-containing adenovirus. Rabbits treated with these transduced peripheral bloodderived cells develop solid spine fusions [23]. Peripheral blood cells are appealing as a cellular gene delivery vehicle because these cells are easier to harvest than bone
144
marrow cells. However, it is unlikely that cells derived from peripheral blood have the same autocrine effects that are seen with transduced bone marrow cells. The results with LMP-1 in these preclinical models are impressive and could be adapted to treat other bone repair problems. However, no mice mutant or transgenic for LMP1 have been reported and little is known of the possible side effects of LMP-1 overexpression. Wang et al. used a recombinant adenovirus containing the BMP-2 cDNA to induce an intertransverse lumbar spinal fusion in rats [33]. Autologous bone marrow cells were harvested from Lewis rats, expanded in tissue culture, and then transduced with the BMP-2-containing adenovirus. These transduced cells were then loaded onto either guanidine-extracted demineralized bone matrix or collagen sponges and then implanted into the appropriate fusion site. The BMP-2-producing bone marrow cells induced spine fusion in all 15 spines that were treated. Bone marrow cells alone did not induce spine fusion. The results of these gene transfer preclinical studies are quite exciting and clearly demonstrate the adaptability of gene therapy for human use. The ability to deliver growth factors via gene therapy may lead to the development of less invasive operative techniques such as laparoscopic spine fusion. Such a development carries the potential for reduced operative morbidity, shortened time to wound healing, and diminished costs. The goal is to develop gene therapy as a part of an overall comprehensive tissue engineering strategy to enhance bone repair. Clearly gene therapy is not necessary for all patients, but it does have the potential to enable orthopedic surgeons to treat difficult bone repair problems that can not be handled successfully with our present technology.
ARTICULAR CARTILAGE REPAIR Cartilage is an avascular, aneural, alymphatic tissue that overlies the ends of the long bones and helps ensure the smooth, almost frictionless movement of joints. Approximately 95% of the tissue consists of a highly hydrated extracellular matrix with a unique composition and microarchitecture, enabling it to bear load effectively. This matrix is synthesized and maintained by the sparse population of articular chondrocytes embedded within it. The poor ability of cartilage to heal has been recognized since the time of Hippocrates. Indeed, it mounts almost no reparative response to traumatic injuries that fail to penetrate the subchondral bone underlying the cartilage. Such lesions are often painful and debilitating. They may also progress to osteoarthritis, a disease characterized by focal erosion of the articular cartilage. Deeper MOLECULAR THERAPY Vol. 6, No. 2, August 2002 Copyright © The American Society of Gene Therapy
doi:10.1006/mthe.2002.0663, available online at http://www.idealibrary.com on IDEAL
injuries that engage the subchondral bone are associated with the entry of blood and marrow contents into the lesion. Progenitor cells within these infiltrates form fibrocartilagenous repair tissue which, although inferior to native articular cartilage, can provide symptomatic improvement for extended periods of time before deteriorating because of defective load-bearing properties (Fig. 2). Cartilage injuries are common, symptomatic, and difficult to repair by surgical means alone. Therefore, there is great interest in improving outcomes through harnessing the biology of the system [34]. Among the more commonly used procedures is abrasion arthroplasty, in which the subchondral plate is deliberately violated to encourage the influx of marrow elements. Less commonly, autologous cartilage, harvested from the margins of the joint that do not bear load, is used as a source of autologous chondrocytes in partial-thickness defects. After expansion in culture, the cells are surgically transplanted to the injured site, where they effect a repair. This procedure is largely restricted to knee joints and remains controversial. Problems common to all methods of attempted cartilage repair are fusion of the repair cartilage with the surrounding endogenous undamaged cartilage and the formation of a uniform cartilaginous surface. The repair of full-thickness lesions is also challenged by the need for reconstitution of the underlying subchondral bone. Biological repair is further challenged by the need to recapitulate the exquisite molecular microarchitecture of cartilage. The mechanically demanding environment of articular cartilage is a harsh discriminator of inadequate repair tissue.
GENE THERAPY APPROACHES
TO
CARTILAGE REPAIR
Cartilage healing presents far greater challenges than bone healing to gene therapy. Unlike bone, cartilage lacks a robust, native healing response that could serve as the basis of a gene therapy. Furthermore, there is no clinical experience of cartilage repair using recombinant proteins upon which to draw. Nevertheless, gene therapy approaches to cartilage repair are encouraged by the ability of various gene products to enhance chondrogenesis [2]. Examples include growth factors [35] such as insulin-like growth factor-1 (IGF-1), transforming growth factor- (TGF-), fibroblast growth factors, and various members of the BMP family, as well as transcription factors such as SOX-9 [36], certain signaling molecules such as SMADs [37], and molecules that inhibit apoptosis such as BCL-2 [38]. Growth factors, however, are difficult to administer exogenously to sites of cartilage injury in a sustained and therapeutically useful fashion, whereas the intracellular mediators have the additional problem of needing to enter responsive cells. Gene delivery strategies promise to overcome these limitations. One issue to be addressed in implementing such strategies is the choice of target tissue. MOLECULAR THERAPY Vol. 6, No. 2, August 2002 Copyright © The American Society of Gene Therapy
REVIEW
Synovium is the most accessible tissue in the joint. It is readily transduced by a wide variety of different vectors and has already served as the site of gene transfer in a clinical study related to the gene therapy of arthritis [39]. In the context of cartilage repair, genes encoding secreted repair factors could be delivered to synovium, which would then serve as a local, intraarticular source of diffusible transgene products. The basis of this approach has been demonstrated in rabbit knee joints. Adenoviral delivery of IGF-1 or BMP-2 cDNAs to synovium increases matrix synthesis by articular cartilage in the same knee joint by about 50% [40]. However, delivery of genes to synovium can generate cartilage ectopically within the joint, and lead to the formation of osteophytes [41]. Moreover, transfer of a TGF- cDNA is deleterious, causing massive fibrosis of the joint capsule (unpublished data). However, the local, intracartilagenous expression of growth factors might exploit the chondrogenic properties of these molecules without provoking adverse extra-cartilagenous reactions. In addition, the effectiveness of all such potentially reparative factors may be improved by targeted delivery to cartilage. The dense matrix that surrounds chondrocytes impedes in vivo gene delivery to these cells by nearly all vectors that have been tested. Nevertheless, transfer of marker genes to chondrocytes in situ has been reported with liposomes incorporating the hemagglutinating virus of Japan (Sendai virus) [42] and with AAV [43]. However, cartilage repair requires the presence of cells, as well as matrix, within the lesion; gene delivery to synovium or preexisting cartilage does not by itself satisfy this requirement. As noted, autologous chondrocyte transplantation is already used clinically [44]. It is thus tempting to improve the quality of the repair tissue by ex vivo genetic modification of the chondrocytes while they are undergoing expansion in culture. These cells are readily transduced by several viral vectors, including retrovirus, adenovirus, AAV, and lentivirus, and can be transfected with certain nonviral formulations [45]. Adenoviral transfer of cDNAs encoding IGF-1, BMP-2, or TGF- to cultures of articular chondrocytes dramatically increases matrix production, even in the presence of interleukin-1 (IL-1), a powerful physiological inhibitor of this process [46,47]. Moreover, use of IGF1 has the additional advantage of maintaining the phenotypic stability of the cells [48]. There have been few experimental studies of the fate of genetically modified chondrocytes after ex vivo transfer to defects in articular cartilage, but the data question the long-term survival of these cells [49–51]. An alternative to the transfer of chondrocytes is to adapt tissue engineering methods to generate genetically modified cartilage, rather than cells in suspension, for transplant. Encouraging results have been obtained when the IGF-1 cDNA is introduced into chondrocytes by transfection, and cartilage is grown from these cells in a bioreactor [52]. Similar experiments, in which cDNAs encod-
145
REVIEW
doi:10.1006/mthe.2002.0663, available online at http://www.idealibrary.com on IDEAL
ing various BMPs were transfected into chondrocytes, identified BMP-7 as a particularly promising molecule for this purpose [53]. Using articular chondrocytes as vehicles for ex vivo gene therapy in cartilage repair has several drawbacks. Supplies of autologous tissue are severely limited, and the culture techniques used to expand the chondrocytes lead to phenotypic modulation, commonly referred to as de-differentiation. It is not clear that these cells, or indeed authentic chondrocytes (genetically modified or not), can generate a repair tissue with the appropriate biochemistry, microarchitecture, and biomechanical properties. If these criteria are not met, the tissue fails. Thus, there is considerable interest in using chondroprogenitor cells, whose greater plasticity may enable the efficient recapitulation of true articular cartilage by way of biologically authentic routes of chondrogenesis. Various tissues, including bone marrow, perichondrium and periosteum, and (perhaps) adipose tissue and skin, contain cells with the ability to differentiate into articular chondrocytes. The chondroprogenitor cells of bone marrow are thought to be mesenchymal stem cells (MSCs); whether these cells are responsible for the chondrogenic potential of other tissues remains an unanswered question. Under the appropriate conditions, which in vitro include the presence of TGF-, dexamethasone, and spherical shape, MSCs predictably differentiate into chondrocytes [54]. When transplanted into experimental lesions in the articular cartilages of rabbits, MSCs effect better repair than similar transplants of articular chondrocytes [55]. Whether the appropriate genetic modification of MSCs will enhance repair even further remains to be demonstrated experimentally [56]. The resistance of MSCs to efficient transduction by many vectors hinders such studies. Periosteal cells, however, have been retrovirally transduced with the BMP-7 cDNA with promising results [57]. Likewise, efficient transfection of perichondrial cells has been reported [58]. One disadvantage of these ex vivo approaches is the need for the appropriate scaffolds and matrices upon which to seed the cells before transplant or complex surgical procedures, as in the case of autologous cartilage transplantation. There is preliminary evidence that genetically modified chondrocytes and MSCs selectively adhere to articular cartilage [50,59], suggesting that simple intraarticular injection of the cells may serve as an adequate delivery method. In situ gene delivery techniques that obviate the need for ex vivo cell culture would bring many advantages. Along these lines, methods are being developed for introducing vectors into sites of full-thickness lesions such that chondrogenic cells become transduced within the lesion [60].
OTHER CARTILAGINOUS TISSUES Certain joints, such as the knee, contain additional cartilaginous structures known as menisci. These are also
146
frequently subject to symptomatic injuries and, like articular cartilage, the menisci have limited ability to heal spontaneously. Gene therapy approaches similar to those used to address healing of articular cartilage can be applied to the healing of meniscus. Both retroviral ex vivo and adenoviral in vivo transfer methods have been successfully applied for the transfer of marker genes to the menisci of rabbits and dogs [61]. Adenoviral transfer of TGF- cDNA to cultures of human and canine meniscal cells leads to very large increases in matrix synthesis [62], but this has not yet been attempted in an in vivo model of meniscal healing. The intervertebral disc is another important cartilaginous structure whose injury is a major clinical and social problem. It, too, lacks an efficient endogenous repair mechanism and is the subject of gene transfer approaches to healing [63]. Cells of the intervertebral disc are easily transduced by a variety of different vectors. The direct, intradiscal injection of first-generation adenoviral vectors leads to high levels of transgene expression for over a year in immunocompetent rabbits. The absence of immune surveillance within this dense, avascular, alymphatic tissue and the low mitotic index of disc cells presumably account for the remarkable longevity of transgene expression. Delivery of growth factor genes to disc cells leads to large increases in matrix synthesis, but it is not yet known whether this alone is sufficient to produce healing of discal lesions. Several different gene therapy strategies for cartilage repair are under investigation, but research remains at the developmental stage and, unlike the case with bone healing, human clinical trials are not yet on the horizon. In particular, it is still unclear which gene, or combination of genes, will provide the most effective repair, and the choice of vector is not apparent. The most appropriate cellular targets for gene transfer have not been identified, and there has been no published demonstration of efficacy in an animal model. Nevertheless, gene transfer approaches hold much promise as a means of enhancing the repair of cartilaginous tissues, especially because this remains a major clinical problem.
GENERAL CONCLUSIONS There is a major, unmet clinical need for improved ways to heal bone and cartilage. Through its ability to deliver osteogenic and chondrogenic substances in a precise, economical, local, authentic, and sustained fashion, gene transfer is a particularly promising technology with which to address this need. Impressive preclinical studies have demonstrated the power of gene therapy in several different models of bone healing, and clinical trials would seem to be only a matter of time. In several important ways, cartilage healing is a more intractable biological problem than bone healing, and less headway has been made in adapting gene therapy to this end. However, considerable experimental progress has been made and future success is likely. MOLECULAR THERAPY Vol. 6, No. 2, August 2002 Copyright © The American Society of Gene Therapy
doi:10.1006/mthe.2002.0663, available online at http://www.idealibrary.com on IDEAL
ACKNOWLEDGMENT J.R.L. is supported by a grant from the National Institutes of Health (AR467890IAI)
REFERENCES 1. Evans, C. H., and Robbins, P. D. (1995). Possible orthopaedic applications of gene therapy. J. Bone Joint Surg. Am. 77: 1103–1114. 2. Evans, C. H., et al. (2000). Using gene therapy to protect and restore cartilage. Clin. Orthop. 379 (Suppl.): S214–s219. 3. Oakes, D. A., and Lieberman, J. R. (2000). Osteoinductive applications of regional gene therapy: ex vivo gene transfer. Clin. Orthop. 379 (Suppl.): S101–s112. 4. Scaduto, A. A., and Lieberman, J. R. (1999). Gene therapy for osteoinduction. Orthop. Clin. North Am. 30: 625–633. 5. Lieberman, J. R., et al. (1999). The effect of regional gene therapy with bone morphogenetic protein-2-producing bone-marrow cells on the repair of segmental femoral defects in rats. J. Bone Joint Surg. Am. 81: 905–917. 6. Colterjohn, N. R., and Bednar, D. A. (1997). Procurement of bone graft from the iliac crest. An operative approach with decreased morbidity. J. Bone Joint Surg. Am. 79: 756–759. 7. Arrington, E. D., et al. (1996). Complications of iliac crest bone graft harvesting. Clin. Orthop. 329: 300–309. 8. Gerhart, T. N., et al. (1993). Healing segmental femoral defects in sheep using recombinant human bone morphogenetic protein. Clin. Orthop. 293: 317–326. 9. Cook, S. D., et al. (1994). Recombinant human bone morphogenetic protein-7 induces healing in a canine long-bone segmental defect model. Clin. Orthop. 301: 302–312. 10. Cook, S. D., et al. (1995). Effect of recombinant human osteogenic protein-1 on healing of segmental defects in non-human primates. J. Bone Joint Surg. Am. 77: 734–750. 11. Sandhu, H. S., et al. (1995). Evaluation of rhBMP-2 with an OPLA carrier in a canine posterolateral (transverse process) spinal fusion model. Spine 20: 2669–2682. 12. Boden, S. D., et al. (1998). Laparoscopic anterior spinal arthrodesis with rhBMP-2 in a titanium interbody threaded cage. J. Spinal Disord. 11: 95–101. 13. Friedlaender, G. E., et al. (2001). Osteogenic protein-1 (bone morphogenetic protein7) in the treatment of tibial nonunions. J. Bone Joint Surg. Am. 83-A (Suppl. 1 (Pt. 2)): S151–S158. 14. Boden, S. D., et al. (2000). The use of rhBMP-2 in interbody fusion cages. Definitive evidence of osteoinduction in humans: a preliminary report. Spine 25: 376–381. 15. Crystal, R. G. (1995). Transfer of genes to humans: early lessons and obstacles to success. Science 270: 404–410. 16. Anderson, W. F. (1998). Human gene therapy. Nature 392 (Suppl.): 25–30. 17. Wilson, J. M. (1996). Adenoviruses as gene-delivery vehicles. N. Engl. J. Med. 334: 1185–1187. 18. Baltzer, A. W., et al. (2000). Genetic enhancement of fracture repair: healing of an experimental segmental defect by adenoviral transfer of the BMP-2 gene. Gene Ther. 7: 734–739. 19. Musgrave, D. S., et al. (1999). Adenovirus-mediated direct gene therapy with bone morphogenetic protein-2 produces bone. Bone 24: 541–547. 20. Okubo, Y., et al. (2000). Osteoinduction by bone morphogenetic protein-2 via adenoviral vector under transient immunosuppression. Biochem. Biophys. Res. Commun. 267: 382–387. 21. Fang, J., et al. (1996). Stimulation of new bone formation by direct transfer of osteogenic plasmid genes. Proc. Natl. Acad. Sci. USA 93: 5753–5758. 22. Bonadio, J., et al. (1999). Localized, direct plasmid gene delivery in vivo: prolonged therapy results in reproducible tissue regeneration. Nat. Med. 5: 753–759. 23. Viggeswarapu, M., et al. (2001). Adenoviral delivery of LIM mineralization protein-1 induces new-bone formation in vitro and in vivo. J. Bone Joint Surg. Am. 83-A: 364–376. 24. Bruder, S. P., Jaiswal, N., and Haynesworth, S. E. (1997). Growth kinetics, self-renewal, and the osteogenic potential of purified human mesenchymal stem cells during extensive subcultivation and following cryopreservation. J. Cell Biochem. 64: 278–294. 25. Bruder, S. P., et al. (1998). Bone regeneration by implantation of purified, cultureexpanded human mesenchymal stem cells. J. Orthop. Res. 16: 155–162. 26. Bruder, S. P., et al. (1998). The effect of implants loaded with autologous mesenchymal stem cells on the healing of canine segmental bone defects. J. Bone Joint Surg. Am. 80: 985–996. 27. Krebsbach, P. H., et al. (2000). Gene therapy-directed osteogenesis: BMP-7-transduced human fibroblasts form bone in vivo. Hum. Gene Ther. 11: 1201–1210. 28. Lee, J. Y., et al. (2001). Effect of bone morphogenetic protein-2-expressing musclederived cells on healing of critical-sized bone defects in mice. J. Bone Joint Surg. Am. 83-A: 1032–1039. 29. Zuk, P. A., et al. (2001). Multilineage cells from human adipose tissue: implications for cell-based therapies. Tissue Eng. 7: 211–228. 30. Dragoo, J., et al. (2001). Bone induction by genetically manipulated stem cells derived from fat. J. Bone Min. Res. 16 (suppl. 1): 5500. 31. Deutsche Banc, A. B. (2001). Estimates and Company Information.
MOLECULAR THERAPY Vol. 6, No. 2, August 2002 Copyright © The American Society of Gene Therapy
REVIEW
32. Boden, S. D., et al. (1998). Lumbar spine fusion by local gene therapy with a cDNA encoding a novel osteoinductive protein (LMP-1). Spine 23: 2486–2492. 33. Wang, J., et al. (2002). Intertransverse process fusion in Lewis rats using bone marrow cells infected with a BMP-2 containing adenovirus. Trans. Orthop. Res. Soc. 25: 327. 34. Athanasiou, K. A., et al. (2001). Basic science of articular cartilage repair. Clin. Sports Med. 20: 223–247. 35. O’Connor, W. J., et al. (2000). The use of growth factors in cartilage repair. Orthop. Clin. North Am. 31: 399–410. 36. Uusitalo, H., et al. (2001). Accelerated up-regulation of L-Sox5, Sox6, and Sox9 by BMP-2 gene transfer during murine fracture healing. J. Bone Miner. Res. 16: 1837–1845. 37. Fujii, M., et al. (1999). Roles of bone morphogenetic protein type I receptors and Smad proteins in osteoblast and chondroblast differentiation. Mol. Biol. Cell 10: 3801–3813. 38. Feng, L., et al. (1998). Evidence of a direct role for Bcl-2 in the regulation of articular chondrocyte apoptosis under the conditions of serum withdrawal and retinoic acid treatment. J. Cell. Biochem. 71: 302–309. 39. Evans, C. H., et al. (1996). Clinical trial to assess the safety, feasibility, and efficacy of transferring a potentially anti-arthritic cytokine gene to human joints with rheumatoid arthritis. Hum. Gene Ther. 7: 1261–1280. 40. Mi, Z., et al. (2000). Adenovirus-mediated gene transfer of insulin-like growth factor 1 stimulates proteoglycan synthesis in rabbit joints. Arthritis Rheum. 43: 2563–2570. 41. Gelse, K., et al. (2001). Fibroblast-mediated delivery of growth factor complementary DNA into mouse joints induces chondrogenesis but avoids the disadvantages of direct viral gene transfer. Arthritis Rheum. 44: 1943–1953. 42. Tomita, T., et al. (1997). In vivo direct gene transfer into articular cartilage by intraarticular injection mediated by HVJ (Sendai virus) and liposomes. Arthritis Rheum. 40: 901–906. 43. Arai, Y., et al. (2000). Gene delivery to human chondrocytes by an adeno associated virus vector. J. Rheumatol. 27: 979–982. 44. Brittberg, M., et al. (2001). Autologous chondrocytes used for articular cartilage repair: an update. Clin. Orthop. 391 (Suppl.): S337–S348. 45. Madry, H., and Trippel, S. B. (2000). Efficient lipid-mediated gene transfer to articular chondrocytes. Gene Ther. 7: 286–291. 46. Shuler, F. D., et al. (2000). Increased matrix synthesis following adenoviral transfer of a transforming growth factor 1 gene into articular chondrocytes. J. Orthop. Res. 18: 585–592. 47. Smith, P., et al. (2000). Genetic enhancement of matrix synthesis by articular chondrocytes: comparison of different growth factor genes in the presence and absence of interleukin-1. Arthritis Rheum. 43: 1156–1164. 48. Nixon, A. J., et al. (2000). Insulinlike growth factor-I gene therapy applications for cartilage repair. Clin. Orthop. 379 (suppl.): S201–S213. 49. Kang, R., et al. (1997). Ex vivo gene transfer to chondrocytes in full-thickness articular cartilage defects: a feasibility study. Osteoarth. Cartilage 5: 139–143. 50. Baragi, V. M., et al. (1997). Transplantation of adenovirally transduced allogeneic chondrocytes into articular cartilage defects in vivo. Osteoarth. Cartilage 5: 275–282. 51. Ostrander, R. V., et al. (2001). Donor cell fate in tissue engineering for articular cartilage repair. Clin. Orthop. 389: 228–237. 52. Madry, H., Zurakowski, D., and Trippel, S. B. (2001). Overexpression of human insulinlike growth factor-I promotes new tissue formation in an ex vivo model of articular chondrocyte transplantation. Gene Ther. 8: 1443–1449. 53. Kaps, C., et al. (2002). Bone morphogenetic proteins promote cartilage differentiation and protect engineered artificial cartilage from fibroblast invasion and destruction. Arthritis Rheum. 46: 149–162. 54. Johnstone, B., et al. (1998). In vitro chondrogenesis of bone marrow-derived mesenchymal progenitor cells. Exp. Cell Res. 238: 265–272. 55. Wakitani, S., et al. (1994). Mesenchymal cell-based repair of large, full-thickness defects of articular cartilage. J. Bone Joint Surg. Am. 76: 579–592. 56. Yoo, J. U., et al. (2000). Chondrogenitor cells and gene therapy. Clin. Orthop. 379 (Suppl.): S164–S170. 57. Mason, J. M., et al. (1998). Expression of human bone morphogenic protein 7 in primary rabbit periosteal cells: potential utility in gene therapy for osteochondral repair. Gene Ther. 5: 1098–1104. 58. Goomer, R. S., et al. (2000). Nonviral in vivo gene therapy for tissue engineering of articular cartilage and tendon repair. Clin. Orthop. 379 (Suppl.): S189–S200. 59. Doherty, P. J., et al. (1998). Resurfacing of articular cartilage explants with geneticallymodified human chondrocytes in vitro. Osteoarth. Cartilage 6: 153–159. 60. Palmer, G., et al. (2002). Direct gene transfer to osteochondral defects. Trans. Orthop. Res. Soc. 27: 292. 61. Goto, H., et al. (1999). Transfer of lacZ marker gene to the meniscus. J. Bone Joint Surg. Am. 81: 918–925. 62. Goto, H., et al. (2000). Gene therapy for meniscal injury: enhanced synthesis of proteoglycan and collagen by meniscal cells transduced with a TGF(1)gene. Osteoarth. Cartilage 8: 266–271. 63. Nishida, K., et al. (1999). Modulation of the biologic activity of the rabbit intervertebral disc by gene therapy: an in vivo study of adenovirus-mediated transfer of the human transforming growth factor  1 encoding gene. Spine 24: 2419–2425.
147