Trị liệu trong điều trị các bệnh cơ xương pdf

Gene Therapy for the Treatment of Musculoskeletal Diseases Christopher H. Evans, PhD, Steven C. Ghivizzani, PhD, James H. Herndon, MD, and Paul D. Robbins, PhD Abstract Gene therapy involves the transfer of genes to patients for therapeutic pur- poses. 1 This approach is intuitively obvious for the treatment of mende- lian disorders, butit also has wide ap- plication for diseases that lack a strong or simple genetic basis. In such instances, gene transfer is used as a biologic delivery system for thera- peutic RNAs or proteins encoded by the transgene. The definition of gene therapy can be expanded to include both the delivery of non-coding nu- cleic acids (eg, oligonucleotides), which have the ability to modify gene expression in the recipient cells, and the in situ repair of mutations through gene correction. 2 At a minimum, a successful gene therapy protocol must answer the fol- lowing questions: (1) Which gene or genes should be transferred? (2) Where should the therapeutic genes be transferred? (3) How can the trans- genes be transferred to the target cells? (4) How should the level and duration of transgene expression be regulated? (5) How can safety be en- sured? Gene Transfer Vectors, which can be viral or nonvi- ral, are vehicles that deliver genetic material into a living cell. To create vectors, wild-type viruses are genet- ically altered to eliminate virulence and, in most cases, their ability to rep- licate, while retaining infectivity. Vi- ral vectors being used in human clin- ical trials include oncoretrovirus (ie, retrovirus), adenovirus, adeno- associated virus (AAV), and herpes simplex virus. Lentivirus, another type of retrovirus, is also undergoing rapid development. The key charac- teristics of any viral vector include its host range, ability to infect nondivid- ing cells, packaging capacity, immu- nogenicity, titer, ease of manufacture, and safety, as well as whether it in- tegrates into the host genomic DNA. 3 Gene transfer using a viral vector is known as transduction. Nonviral vectors may be as sim- Dr. Evans is The Robert Lovett Professor of Or- thopaedic Surgery, Center for Molecular Ortho- paedics, Department of Orthopaedic Surgery, Har- vard Medical School, Boston, MA. Dr. Ghivizzani is Associate Professor, Department of Orthopaedic Surgery, University of Florida College of Medi- cine, Gainesville, FL. Dr. Herndon is The Wil- liam Harris Professor of Orthopaedic Surgery, Center for Molecular Orthopaedics, Department of Orthopaedic Sur gery, Harvard Medical School. Dr. Robbins is Professor, Department of Molec- ular Genetics and Biochemistry, University of Pittsburgh School of Medicine, Pittsburgh, PA. Reprint requests: Dr. Evans, Center for Molec- ular Orthopaedics, BLI-152, 221 Longwood Av- enue, Boston, MA 02115. Copyright 2005 by the American Academy of Orthopaedic Surgeons. Research into the orthopaedic applications of gene therapy has resulted in progress toward managing chronic and acute genetic and nongenetic disorders. Gene ther- apy for arthritis, the original focus of research, has progressed to the initiation of several phase I clinical trials. Preliminary findings support the application of gene therapy in the treatment of additional chronic conditions, including osteoporosis and aseptic loosening, as well as musculoskeletal tumors. The most rapid progress is likely to be in tissue repair because it requires neither long-term transgene expres- sion nor closely regulated levels of transgene expression. Moreover, healing prob- ably can be achieved with existing technology. In preclinical studies, genetically mod- ulated stimulation of bone healing has shown impressive results in repairing segmental defects in the long bones and cranium and in improving the success of spinal fu- sions. An increasing amount of evidence indicates that gene transfer can aid the repair of articular cartilage, menisci, intervertebral disks, ligaments, and tendons. These developments have the potential to transform many areas of musculoskeletal care, leading to treatments that are less invasive, more effective, and less expensive than existing modalities. J Am Acad Orthop Surg 2005;13:230-242 Orthopaedic Research Society Special Article 230 Journal of the American Academy of Orthopaedic Surgeons ple as naked, plasmid DNA. Trans- fer efficiency can be increased by com- bining the DNA with natural or synthetic polymers or by applying biophysical methods, such as elec- troporation. Nonviral gene transfer, known as transfection, is less expen- sive, safer, and simpler than transduc- tion, but it is considerably less effi- cient. 4 Regardless of the vector, genes may be transferred to their targets by in vivo or ex vivo strategies. For in vivo delivery, vector is intr oduced di- rectly into the recipient. During ex vivo delivery, cells are recovered, ge- netically manipulated outside the body, then returned to the recipient. Of the two, in vivo gene transfer is less expensive and technically sim- pler, but its use raises safety concerns because infectious or transfecting agents are introduced directly into the body. Moreover, many of these agents, particularly viral vectors, are antigenic, which may provoke im- mune problems and prevent repeat dosing. The major limitation of in vivo gene transfer is the inability of the vector selectively to target cells as- sociated with the tissue of interest. Ex vivo gene transfer is considered safer because transduced or transfect- ed cells—not vectors—are introduced into the body. Moreover, the geneti- cally modified cells can be exhaus- tively tested before reimplantation. Ex vivo transfer also facilitates the use of oncoretroviral vectors, which transduce only dividing cells, because many cells with low mitotic indices in vivo replicate readily in culture. Ex vivo gene transfer also helps address certain immune problems because vectors can be chosen that express no viral proteins in transduced cells. Thus, the cells returned to the patient synthesize no foreign antigens, there- by enabling both long-term gene ex- pression and repeat dosing. Finally, ex vivo methods enable more specif- ic targeting and, therefore, better con- trol of the transduced cells. The primary disadvantage of ex vivo delivery is the expense and com- plexity of harvesting cells and main- taining them in cell culture before transducing, testing, and returning them. Patients are exposed to the ad- ditional procedures involved in cell harvesting, and cell transplantation brings its own set of issues that are absent from in vivo delivery proto- cols. Approaches for obviating the disadvantages of ex vivo gene trans- fer are being explored. Preliminary evidence indicates that certain cells may be successfully allografted, thus acting as so-called universal donors. For example, der- mal fibroblasts expanded from a sin- gle donor provide all the cells used in living artificial skin grafts. The do- nor cells persist in the r ecipient’s skin for extended periods, possibly be- cause of the dense, collagenous, ex- tracellular matrix that surrounds them. Similarly, mesenchymal stem cells may possess immunosuppres- sive properties, thereby enabling their survival in allogeneic hosts. 5 If allo- geneic cells can indeed be used in this manner, batches of transduced, screened, and standardized universal donor cells could be established and injected into recipients on demand, increasing the ease of ex vivo gene de- livery. Ex vivo gene delivery also may be expedited by using cells that can be recovered, transduced, and returned to the patient in one sitting. Blood and bone marrow lend themselves to these abbreviated ex vivo delivery strategies. 6-8 Duration and Regulation of Transgene Expression The optimal duration and level of transgene expression is specific to each application. For example, some cancer applications may require a very large burst of transgene expres- sion for a limited period to kill tumor cells without causing subsequent damage to uninvolved tissues. In con- trast, successful treatment of many monogenic diseases (eg, hemophilia) requires prolonged expression at low to moderate levels. Modest levels of transgene expression for limited pe- riods may be appropriate for tissue repair (eg, cartilage or bone healing). Episodic conditions, such as rheuma- toid arthritis (RA), which is charac- terized by flares and remissions, might require persistent carriage of the transgene, with levels of expres- sion increased or reduced to match disease activity. Transient transgene expression is Dr. Evans or the department with which he is affiliated has received research or institutional support from NIH–National Institute for Arthritis Mus- culoskeletal and Skin Diseases; National Institute for Diabetes, Digestive and Kidney Diseases; the Orthopaedic Trauma Association; Orthogen; Valentis; Osiris; and TissueGene. Dr. Evans or the department with which he is affiliated has received royalties from Valentis and TissueGene. Dr. Evans or the department with which he is affiliated has stock or stock options held in Valentis, GenVec, and Orthogen. Dr. Evans or the department with which he is affiliated serves as a consultant to or is an employee of Valentis and TissueGene. Dr. Evans is on the Scientific Advisory Board of TissueGene and Orthogen. Dr. Ghivizzani or the department with which he is affiliated has received research or institutional support from NIH–National Institute for Arthritis, Musculoskeletal and Skin Diseases. Dr. Herndon or the department with which he is affiliated has stock or stock options held in Valentis. Dr. Robbins or the department with which he is affiliated has received research or institutional support from Valentis and TissueGene. Dr. Robbins or the department with which he is affiliated has received nonincome support (such as equipment or services), commercially derived honoraria, or other non- research–related funding (such as paid travel) from TissueGene and Orthogen. Dr. Robbins or the department with which he is affiliated has received royalties from TissueGene and Valentis. Dr. Robbins or the department with which he is affiliated has stock or stock options held in Valentis. Dr. Robbins or the department with which he is affiliated serves as a consultant to or is an employee of TissueGene and Orthogen. Christopher H. Evans, PhD, et al Vol 13, No 4, July/August 2005 231 more easily achieved than long-term expression. When prolonged trans- gene expression is required, it is nec- essary to introduce the genes into either long-lived cells or, if an inte- grating vector is used, cells whose progeny can continue to express the transgene. Skeletal muscle as well as brain and liver cells are examples of nondividing cells that could provide extended periods of transgene ex- pression. Stem cells are alternative targets because their capacity for self- renewal could ensure carriage of the transgene for extended periods, pos- sibly for life. Moreover, progeny cells could carry the transgene as they dif- ferentiate into mature cells, which otherwise might be difficult to target. It has been difficult to transduce and maintain transgene expression with- in hematopoietic 9 and mesenchy- mal 10 stem cells, but these technical limitations may soon be overcome. The immune system acts as a bar- rier to long-term gene expression when the vector used for gene trans- fer confers antigenicity on the re- cipient cells. For example, cells transduced with first-generation ad- enoviruses express certain residual adenoviral proteins. 11 These proteins are highly antigenic, and cells ex- pressing them are killed by the im- mune system. This difficult problem has been put to rest only with the ad- vent of “gutted” adenoviruses from which all viral coding sequences have been eliminated. Retroviral and AAV vectors do not express viral proteins in transduced cells. Nonviral vectors also avoid the expression of viral proteins. Howev- er, they may nevertheless activate the immune system because plasmid DNA used by most nonviral systems is grown in bacteria that, unlike eu- karyotic cells, do not methylate cy- tosine residues in DNA. The un- methylated dinucleotide sequence cytosine-guanosine (CpG) strongly activates cell-mediated immunity. 12 In the absence of problems with cell turnover or the immune system, long-term gene expression also can be curtailed at the level of the promoter. The strong viral promoters often fa- vored for gene therapy experiments may become turned off (silenced) in certain eukaryotic cells. As a result, there is interest in using constitutive eukaryotic promoters. In general, this strategy can be successful, but in many cases, the level of transgene ex- pression is considerably lower than that achieved with viral promoters. Manipulation of gene expression at the level of the promoter currently offers the best prospect of achieving regulated gene expression; there are two general approaches to regulating transgene expression in this way. One method consists of using an exoge- nous molecule to control the level of gene expression. Systems responsive to agents such as tetracycline, rapa- mycin, and RU486 are available. 13 An alternative strategy makes use of in- trinsic regulation, taking advantage of the natural responsiveness of many promoters to endogenous stimuli, such as inflammation. 14 These types of inducible systems ar e attractive be- cause they are self-regulating. How- ever, they raise safety concerns be- cause there is no easy way to control them, should that become medically necessary. Safety Several safety concerns are associat- ed with gene therapy, some more psy- chological than actual. Recombinant viruses used for gene transfer are de- rived from wild-type viruses that cause disease, thus raising tangible concerns regarding the use of viral vectors. For example, lentiviral vec- tors are derived from HIV; oncoret- roviral vectors are commonly derived from the Moloney murine leukemia virus; wild-type adenoviruses cause colds and flu; and herpes simplex vi- rus causes conditions such as cold sores and herpes. In contrast, AAV causes no known human disease. The viruses used for gene transfer have been altered and in principle are no longer virulent. Theoretically, how- ever, during the production of large batches of virus for clinical use or dur- ing transduction of the target cells, vi- ral vectors may undergo genetic re- arrangements that restore virulence. There is particular concern regarding the possible generation of replication- competent viruses, which not only would spread within the r ecipient but also could permit horizontal transfer to other individuals, with unknown consequences. The presence of replication-competent virus also in- creases the likelihood of germ-line gene transfer, another matter of concern. Considerable effort has been ex- pended in developing very sensitive assays for replication-competent vi- ruses; in fact, this is mandatory in hu- man clinical studies. With lentivirus- es, using vectors derived fr om equine or feline sources rather than from HIV may be safer. Although these nonhu- man lentiviruses do not normally cause disease in humans, the prop- erties of recombinant viruses engi- neered in the laboratory to transduce human cells may be different. Ironically, the first documented death as a result of gene transfer oc- curred with adenovirus, a vector con- sidered to be safe because it is non- integrating and, in its wild-type state, is associated with only mild respira- tory infections. 15 A large adenoviral load of approximately 10 14 particles was infused into the hepatic portal vein of a patient, which led to an uncon- trollable, systemic inflammatory re- action and death from respiratory fail- ure. The exact mechanism remains unclear, but a hypersensitivity reac- tion could occur with a high antigenic load. Moreover, infection of cells with adenoviruses activates the intracellular signaling machinery (mitogen-activated protein kinases and the transcription factor NF κB), which are involved with the induction of inflammatory cyto- kines that could trigger a massive, sys- temic inflammatory response. Gener- Gene Therapy for the Treatment of Musculoskeletal Diseases 232 Journal of the American Academy of Orthopaedic Surgeons alized reactions should not occur when smaller doses of adenovirus ar e locally applied. Insertional mutagenesis has al- ways been a theor etic possibility with retroviral vectors, but until recently, it had never been observed despite the widespread use of retroviral vec- tors in human trials. However, in 1999, a lymphoproliferative disorder resembling leukemia occurred in a child treated for X-linked severe com- bined immunodeficiency disease with retroviral gene transfer. 16 Two more children in the same study also developed leukemia, which resulted from insertion of the retrovirus near a known oncogene. Tw o of these three children subsequently died from the secondary leukemia. Several circum- stances conspired to make this clin- ical trial singularly vulnerable to this type of adverse event: the subjects lacked adaptive immunity, the retro- virus was targeted to hematopoietic stem cells, the transgene encoded one chain of a receptor common to sev- eral different growth factors, and the genetically modified cells had a se- lective, in vivo growth advantage over unmodified cells. Ironically, the protocol that produced the leukemia successfully treated the genetic dis- ease. Because childhood leukemia can be successfully treated in most cases and because X-linked severe com- bined immunodeficiency disease is lethal, permission has been given to treat additional patients with retro- viral gene transfer. However, the ep- isode has renewed concerns about insertional mutagenesis, and the use of retroviral vectors for non–life- threatening diseases has been subject to renewed questioning. Although nonviral vectors involve fewer safety concerns than their vi- ral counterparts, they are not devoid of potential side effects. For example, DNA is inflammatory, and unmeth- ylated CpG dinucleotide sequences present in plasmids generated in bac- teria stimulate cell-mediated immu- nity. The inefficiencies of nonviral gene delivery often require the ad- ministration of very large amounts of DNA, thus increasing the chances of unwanted side effects. Despite the disproportionate amount of negative publicity attract- ed by these events, there have been only scattered reports of nonfatal side effects and three deaths among more than 4,000 individuals treated. Musculoskeletal Applications of Gene Therapy Interest in gene therapy for muscu- loskeletal applications began with re- search focused on gene delivery to synovium to treat arthritis. 17 Howev- er, the rich potential of gene therapy for other musculoskeletal indications was quickly appreciated and, by the time the first review was published in 1995, 1 most major applications of the technology had been foreseen. To facilitate communication and collab- oration between the growing num- bers of investigators in this area, sev- eral workshops on orthopaedic gene therapy have been held. 18-20 Although gene therapy was con- ceived of as a method for treating mendelian diseases, much attention is devoted to its use in nongenetic dis- orders. It is useful to divide the field of orthopaedic gene therapy into four main areas based on the genetics and chronicity of the target diseases be- cause each entails differ ent gene ther- apy approaches (Fig. 1). Of the four areas illustrated, gene therapy for or - thopaedic tumors has received very little experimental attention. Mendelian Diseases Considerable progress has been made in identifying the mutations Figure 1 Categories of orthopaedic disease amenable to gene therapy. CACP = campodactyly- arthropathy-coxa vara-pericarditis. Christopher H. Evans, PhD, et al Vol 13, No 4, July/August 2005 233 that lead to mendelian disorders of the musculoskeletal system. 21 With completion of the Human Genome Project and rapid advances in tech- nology, there is a reasonable prospect of determining the molecular basis for all of them within the next decade. Despite such progress, these diseas- es present considerable challenges to gene therapy. Many of them are rare, dominant negative disorders, which require suppression of mutant gene expression. Another problem is the developmental nature of many of these diseases. Thus, gene therapy may need to be administered at a very early developmental age, possibly in utero, before the musculoskeletal sys- tem becomes fully developed and dif- ficult to alter. As well as challenging the limits of gene therapy, such constraints also require sophisticated early diagnosis. Even when postnatal gene therapy is a reasonable option, the abundant ex- tracellular matrix present in many musculoskeletal tissues renders gene delivery inefficient. Finally, effective gene therapy of most genetic disor- ders probably requires transgene ex- pression for life and, in the case of dominant negative mutations, equal- ly long suppression of mutant alleles. Nevertheless, for most genetic diseas- es, the choice of transgene is obvious and, in many cases, the level of trans- gene expression does not need to be finely regulated. A gene therapy ap- proach may be optimal because these diseases currently are often difficult to treat and impossible to cure. Osteogenesis Imperfecta Osteogenesis imperfecta (OI) is caused by mutations in the genes en- coding the alpha chains of type I col- lagen. Type III OI is recessive; types I, II, and IV are dominant. In tissue culture, antisense RNA both inhibits expression of the mutated gene and reduces expression of the unmutat- ed gene. Ribozymes and small, inter- fering RNAs, however, achieve sub- stantial suppression of the mutant allele without influencing expression of the wild-type allele. 22,23 The oim mouse, which lacks the alpha-2 chain of type I collagen, serves as a useful experimental mod- el for recessive forms of human OI. Niyibizi et al 24 corrected the molec- ular defect in vitro by introducing a cDNA encoding the wild-type alpha-2 chain into fibroblasts derived from the oim mouse. They also cor- rected the molecular defect in vivo in a small patch of skin injected with an adenovirus vector carrying the wild- type gene. The current challenge is to develop techniques that permit the introduction of the therapeutic gene into a sufficient proportion of osteo- blasts to correct the disease and to maintain expression of the gene for the life of the animal. Ex vivo strat- egies using stem cells (eg, mesenchy- mal stem cells) seem to be promising for correcting genetic defects, not only in bones but also in other collagenous tissues affected by the disease. Lysosomal Storage Disorders Several lysosomal storage diseas- es have important orthopaedic se- quelae, and they appeal to gene ther- apists for several reasons. The genes whose mutations cause the diseases are cloned and well characterized, the diseases are recessive, and treatment with recombinant protein or by bone marrow transplant typically is suc- cessful. In addition, the level of gene expression does not need to be tight- ly regulated and, in many cases, the therapeutic gene may be expressed in any convenient tissue with access to the systemic circulation. 25 Gaucher’s disease is caused by mutations in the gene encoding the enzyme glucocerebrosidase. In a phase I clinical trial in which gene therapy was used to treat Gaucher’s disease, a retrovirus was used to transfer the glucocerebrosidase cDNA via ex vivo delivery into he- matopoietic stem cells. 26 Four patients were treated, and the trial is now closed. The mucopolysaccharidoses (MPSs), a group of lysosomal storage disor- ders in which various enzymes nec- essary for the breakdown of glycosami- noglycans are missing, may have associated skeletal abnormalities (ie, Hunter’s and Hurler’s syndromes). Currently in progress is a phase I pro- tocol for subjects with a mild form of Hunter’s syndrome (MPS II), in which the enzyme iduronate-2-sulfatase is defective. Fibrodysplasia Ossificans Progressiva Fibrodysplasia ossificans pr ogres- siva is characterized by the exagger- ated deposition of ectopic bone fol- lowing even mild trauma, and afflicted individuals are said to devel- op a second skeleton. Although the molecular basis for the disease is un- known, it is thought to reflect muta- tions that disturb bone morphoge- netic protein (BMP)-4 synthesis or signaling. In an interesting approach to the therapy of a genetic disease whose molecular lesion is unknown, investigators are evaluating the nog- gin gene, whose product antagoniz- es BMP-4–induced heterotopic ossi- fication. 27 Chronic Nonmendelian Diseases The goal of gene therapy in man- aging the chronic nonmendelian dis- eases is not to compensate for a ge- netic abnormality in the patient but to use gene transfer as a biologic de- livery method for therapeutic gene products. In the absence of a clear ge- netic basis for the disease, the choice of therapeutic transgene is not always obvious, and its selection relies on an understanding of the etiology and pathogenesis of the disorder in ques- tion. Achieving long-term transgene expression is a major challenge; even developing convenient methods of re- administration may be problematic. Nevertheless, continued research is necessary because many of the target diseases are common, are poorly treated by existing modalities, and are increasing in incidence as the popu- Gene Therapy for the Treatment of Musculoskeletal Diseases 234 Journal of the American Academy of Orthopaedic Surgeons lation ages. Most progress has been made in the treatment of arthritis, the first musculoskeletal disorder target- ed for gene therapy. Rheumatoid Arthritis Although RA is an autoimmune condition with significant pathology involving the joints, there are impor- tant extra-articular and systemic manifestations of the disease. Accord- ingly, attempts to treat RA with gene therapy in animal models have con- sisted of local gene delivery to joints, systemic delivery to various organs, and delivery to lymphocytes and antigen-presenting cells, which have the ability to migrate between differ- ent lymphoid tissues. 28 Genes encod- ing a variety of type 2 cytokines (par- ticularly interleukins [ILs]-4, -10, and -13), antagonists of IL-1 and tumor necrosis factor, and antiangiogenic proteins, have shown efficacy in an- imal models. Rather than attempting to modulate the natural disease pro- cess, other investigators have pro- duced genetic synovectomies by in- jecting joints with genes whose products cause apoptosis within the synovium. The advantage of this ap- proach is that it circumvents the need for long-term gene expression. The disadvantage is that the clinical re- sults may be no better than those achieved by conventional synovecto- my. Preclinical studies have estab- lished a convincing proof of princi- ple that justifies and has propelled the development of the four human gene therapy protocols for RA. 29 The first clinical protocol 30 select- ed an IL-1 blocker, the IL-1 receptor antagonist (IL-1Ra), 31 as the trans- gene. Using ex vivo delivery, a retro- virus was used to transfer the IL-1Ra cDNA to autologous synovial fibro- blasts obtained from nine postmeno- pausal women with advanced RA. Control cells were not genetically modified. In a double-blind fashion, genetically modified and contr ol cells were delivered by intra-articular in- jection to the 2nd-5th metacarpopha- langeal (MCP) joints of one hand of each subject. One week later, these MCP joints were recovered and ex- amined for evidence of transgene ex- pression (Fig. 2). This study was not designed to determine efficacy; how- ever, it confirmed that it is indeed pos- sible to transfer genes to human joints Figure 2 The sequence of events in a phase I clinical trial of gene therapy in nine postmeno- pausal women with advanced RA who failed pharmacologic control and required multiple joint surgeries, including replacement of MCP joints 2 through 5 on one hand. Monolayers of autologous synovial fibroblasts were expanded in culture (1) and divided into two pop- ulations, one of which was transduced with a retrovirus carrying human IL-1Ra transgene (2). After safety testing (3), in a double-blinded fashion, two of the recipients’ MCP joints 2-4 were injected with genetically modified cells, while the other two were injected with naïve control cells (4). Seven days later, the four MCP joints were surgically replaced (5), and re- covered tissues were analyzed for expression of the transferred IL-1Ra gene (6). (Adapted with permission from Evans CH, Ghivizzani SC, Robbins PD: Blocking cytokines with genes. J Leukoc Biol 1998;64:55-61.) Christopher H. Evans, PhD, et al Vol 13, No 4, July/August 2005 235 and to successfully express them intra-articularly (Fig. 3) in a manner that is safe and acceptable to pa- tients. 32 A similar phase I protocol using ex vivo, retroviral transfer of human IL- 1Ra cDNAto MCP joints is underway in Germany. 29 However, in that study, there is a gap of 1 month between the introduction and surgical removal of the transgene. So far, four individu- als have been treated, with r esults sim- ilar to those in the United States trial. A phase I protocol involving the direct, intra-articular injection of a re- combinant AAV vector began last year. This vector carries a cDNA en- coding a fusion protein composed of two tumor necrosis factor–soluble re- ceptors combined on an immuno- globulin molecule. In essence, this is a gene that encodes the anti- rheumatic drug etanercept. The only clinical trial of gene ther- apy in RA using nonviral gene deliv- ery employs the genetic synovecto- my approach. Joints are injected with DNA encoding herpes simplex thy- midine kinase. Cells expressing this gene become susceptible to ganciclo- vir and, because of a pronounced by- stander effect, there is widespread death of cells within the synovium. 33 This approach obviates the necessity of long-term gene expression; more- over, readministration of the gene upon recurrence of symptoms should be possible. It r emains to be seen how the clinical results compare with those of conventional synovectomy. Osteoarthritis IL-1 also may be an important me- diator in osteoarthritis (OA). Three studies confirm the promise of IL-1Ra gene therapy in treating OA. 34 The first showed that retroviral, ex vivo deliv- ery of human IL-1Ra cDNA to the knee joints of dogs after transection of the anterior cr uciate ligament slowed car- tilage loss. 35 In a subsequent study, plasmid DNAencoding canine IL-1Ra delivered nonvirally to the knee joints of rabbits suppressed development of surgically induced OA. 36 Convincing data were reported from a series of experiments in which equine IL-1Ra cDNA was delivered to the joints of horses by direct, in vivo, adenoviral delivery. 37 Intra-articular expression of equine IL-1Ra inhibit- ed the development of experimental OA induced by the surgical creation of osteochondral fragments. In addi- tion to strongly protecting the artic- ular cartilage, this therapy r educed the lameness index of the horses, dem- onstrating improvement in both clin- ical and laboratory parameters. Given the late stage at which hu- man OA is typically diagnosed, ar- resting the progress of the disease with an anti-inflammatory and chon- droprotective gene, such as IL-1Ra, may be insufficient. More often, it may be necessary to r estore damaged cartilage, possibly using gene thera- py approaches. Such complications could be avoided if earlier diagnosis were possible and if gene transfer could be given prophylactically after injuries known to predispose to OA, such as rupture of the anterior cru- ciate ligament. In several ways, OA is well suited to local, intra-articular gene therapy. Unlike RA, it is not a systemic con- dition; rather, i t i s restricted to a small number of accessible joints with lim- ited extra-articular manifestations of disease. Moreover, there are few ef- fective pharmacologic treatments. A phase I human protocol for gene transfer in subjects with OA is under- going the review process. Aseptic Loosening Proteins that maintain or restore bone mass around prosthetic joint ar- throplasties or inhibit cellular reac- tions to wear debris may prevent or reverse aseptic loosening. Delivery of genes encoding such proteins has shown promise in relevant animal models. Using a murine air pouch model, Yang et al 38 showed that in Figure 3 Expression of IL-1Ra transgene in human rheumatoid synovium following ex vivo gene transfer. The human IL-1Ra cDNA was transferred to human rheumatoid MCP joints by the protocol described in Figure 2. Genetically modified synovia were recovered at the time of joint arthroplasty, and expression of the transgene was detected by in situ hybrid- ization. The image is pseudocolored to show mRNA (green) and synovium (red). Arrows indicate areas of particularly high transgene expression. (Courtesy of Simon C. Watkins, PhD, Pittsburgh, PA.) Gene Therapy for the Treatment of Musculoskeletal Diseases 236 Journal of the American Academy of Orthopaedic Surgeons vivo delivery of cDNAs encoding IL- 1Ra and IL-10 strongly reduced the inflammatory cellular r eaction to par- ticles of ultra-high-molecular-weight polyethylene or polymethylmeth- acrylate. In another study, when frag- ments of bone were introduced into the air pouch along with the wear de- bris, transfer of the osteoprotegerin (OPG) cDNA inhibited loss of bone matrix. 39 In a complementary series of stud- ies, titanium particles were implant- ed onto the calvarium in a murine model. 40 Using adenovirus and AAV vectors, the investigators found that genes encoding a bivalent soluble tu- mor necrosis factor receptor, OPG, or IL-10 were able to inhibit bone resorp- tion in response to the particles. 41 Pro- tection occurred whether the vector was delivered locally to the calvarial surface or systemically via intramus- cular injection. Osteoporosis Genes whose pr oducts retard bone loss or promote bone formation have potential for managing osteoporosis. In a murine ovariectomy model of os- teoporosis, the injection of adenovi- rus carrying human IL-1Ra cDNA transduced cells in marrow and the surrounding bone, leading to a dra- matic reduction in bone loss. 42 Al- though gene expression persisted for only 2 to 3 weeks, the protective ef- fects of gene transfer lasted for at least 5 weeks. In similar experiments, ad- enovirus carrying OPG cDNAwas in- jected intravenously, 43 a route of application that pr edominantly trans- duces the liver. It led to high circu- lating levels of OPG, which produced a prolonged anti-osteoporotic effect. Of particular note was the remarkable duration of OPG gene expression achieved in this study, which may re- flect the ability of OPG to interfere with immune responses involved in the clearance of adenovirally infect- ed cells. Similar data were subse- quently obtained using an AAV vec- tor. 44 Tissue Repair There are several advantages to us- ing gene therapy to heal musculo- skeletal tissues: long-term transgene expression is neither necessary nor desirable; in most cases, the level of transgene expression need not be un- realistically high or closely regulated; and it may be possible to achieve clin- ical success using existing technolo- gy. Moreover, there is a need for bet- ter ways to heal injuries to bone and soft tissues. Many of these injuries oc- cur in younger individuals as a result of sporting activities; when unsatis- factorily repaired, such injuries have a major accumulated impact on qual- ity of life. The ultimate role of gene transfer strategies in musculoskeletal tissue repair will depend on the suc- cess of competing technologies, par- ticularly those based on the use of re- combinant growth factors and tissue engineering. Bone Healing The ability of gene transfer to in- duce bone formation has been con- firmed by multiple independent lab- oratories using both ex vivo and in vivo strategies. 45,46 In evaluating heal- ing, the model of choice has been a defect of critical size surgically cre- ated in the long bones or crania of ex- perimental animals. In the ex vivo ap- proach, adenovirus was used to deliver BMP-2 cDNAto bone marrow stromal cells in cell culture. 47 The ge- netically modified cells were seeded onto a collagenous scaffold and in- serted into defects of critical size in the femurs of rats. Unlike control de- fects, the genetically treated femurs healed within a few weeks. By his- tologic criteria, healing achieved by BMP-2 gene transfer was superior to that achieved with recombinant BMP-2 protein. One advantage of using marrow stromal cells is their ability not only to express the BMP-2 transgene but also to respond to it and form bone. Subsequent investigators have con- firmed this appr oach, using osteopro- genitor cells derived from perios- teum, muscle, and fat. 46 Success also has been reported with cells (eg, skin fibroblasts) with no obvious osteo- genic potential. Other transgenes, such as BMP-4, also are effective in these models, and it is assumed that additional genes encoding osteogen- ic proteins, such as BMP-6, -7, and -9, also will be successful. Because of the cost and complex- ity of ex vivo delivery methods, ther e have been attempts to heal osseous defects by in vivo delivery of genes to the lesion. One approach involves the use of matrices impregnated with DNA, known as gene-activated ma- trices (GAMs). Fang et al 48 healed segmental defects in rat bone by in- serting GAMs containing a cDNA en- coding the first 34 amino acids of par- athyroid hormone (PTH 1-34) and BMP-4. This group later confirmed stimulated bone formation in large osseous defects in dogs using GAMs carrying PTH 1-34 cDNA. 49 An alternative in vivo strategy in- volves the direct, intralesional injec- tion of vectors, such as adenovirus- carrying osteoinductive genes, in the absence of a matrix or scaffold. Baltzer et al 50 demonstrated the feasibility of this in a rabbit segmental defect mod- el (Fig. 4). Those findings have been repr oduced in rats. 46 Safety is of great- er concern when using in vivo gene delivery of adenovirus. However, in the studies of Baltzer et al, 51 transgene expression was almost entirely re- stricted to the site of administration, with only slight and temporary ex- pression in the liver. No expression occurred in other organs that were ex- amined. It is not known whether there will be immunologic constraints to the application or reapplication of adenoviral vectors in the healing of human bones. Although the data from the afore- mentioned studies are impressive, it remains to be seen whether the osteo- genic response in humans, especial- ly those who are older, diabetic, or traumatized, or who smoke, will be Christopher H. Evans, PhD, et al Vol 13, No 4, July/August 2005 237 as vigorous as that of the young, oth- erwise healthy rats and rabbits stud- ied. Spine Fusion Gene transfer strategies are being developed to impr ove the outcome of spinal fusions using the osteogenic factor LIM mineralization protein-1 (LMP-1). 52 Because it is an intracellu- lar protein, its delivery by gene trans- fer is particularly appr opriate. One of the oddities about LMP-1 is its re- markable potency. In fact, investiga- tors have the problem of needing to prevent excessively high levels of gene expression because under such circumstances, the efficiency of bone formation is reduced. Limited transgene expression has been achieved with plasmid DNA and by transducing cells with adeno- virus vectors for only a short period. The latest version of the application consisted of an abbreviated ex vivo gene delivery approach in which buffy coat cells were isolated from au- tologous blood intraoperatively, brief- ly incubated with the adenoviral vec- tor, placed on a collagen-ceramic composite carrier, and immediately inserted into the fusion site. In rab- bits with a single-level arthrodesis of the lumbar spine, this procedure re- sulted in full spinal fusion within 4 weeks; none of the contr ol rabbits un- derwent spinal fusion. 8 Genes encod- ing additional osteogenic genes, such as BMP-2, also show preliminary promise in experimental spinal fusion studies. 53 Articular Cartilage and Meniscus Several approaches to repairing cartilage using gene transfer are be- ing evaluated. 54 One approach is to use technologies developed for man- aging arthritis and delivered to syn- Figure 4 Healing of an osseous defect of critical size (1.3 cm) by in vivo gene transfer in the femurs of rabbits. An adenovirus was used to deliver human BMP-2 cDNA to the defects shown in panels A-D. Control defects (E-H) received a luciferase gene. Plain radiographs were taken immediately after surgery (A and E) and at 5 weeks (B and F), 7 weeks (C and G), and 12 weeks (D and H) postoperatively. (Re- produced with permission from Baltzer AW, Lattermann C, Whalen JD, et al: Genetic enhancement of fracture repair: Healing of an ex- perimental segmental defect by adenoviral transfer of the BMP-2 gene. Gene Ther 2000;7:734-739.) Gene Therapy for the Treatment of Musculoskeletal Diseases 238 Journal of the American Academy of Orthopaedic Surgeons ovium growth factor genes whose se- creted products diffuse to areas of damaged cartilage. Other methods in- volve ex vivo gene delivery using chondrocytes or chondroprogenitor cells as vehicles and in vivo delivery using vectors associated with matri- ces or autologous blood and bone marrow clots. Delivery of transgenes encoding insulin-like growth factor-1 (IGF-1) or BMP-2 to synovium increases matrix synthesis by chondrocytes in the ad- jacent cartilage. 55 However, delivery of a transforming growth factor-β (TGF-β) gene in this way is deleteri- ous, causing massive fibrosis, ectop- ic cartilage formation, and osteo- phytes. 56 Moreover, this approach to gene therapy does not, by itself, in- crease the cellularity of the lesion. However, i t might be a useful adjunct to cell-based repair methods or mi- crofracture. Using marker genes, it has been es- tablished that genetically modified chondrocytes and periosteal cells can be implanted into cartilaginous le- sions, where they continue to express the transgene for up to several weeks. Transfer of cDNAs encoding BMP-2, BMP-7, IGF-1, or TGF-β dramatical- ly increases matrix synthesis by cul- tures of chondrocytes, even in the presence of IL-1. 57 In an equine mod- el of cartilage repair by chondrocyte transplantation, the introduction of a BMP-7 cDNA into the transplanted chondrocytes accelerated repair. 58 BMP-7 also promotes the chondro- genic differentiation of precursor cells derived from periosteum. The im- plantation of periosteal cells trans- duced with BMP-7 or sonic hedgehog cDNAs enhances repair of osteochon- dral defects in rabbits. 59 To avoid the complexities of ex vivo gene delivery, there have been attempts to deliver genes directly to full-thickness lesions in cartilage. In one approach, adenoviral vectors are associated with a collagen-gly- cosaminoglycan matrix that is in- serted into the defect. Alternatively, the vectors are mixed with autolo- gous blood or bone marrow during clotting. The resulting “gene plug” can be press-fit into lesions in artic- ular cartilage. 6 As an alternative to implanting vectors or genetically modified cells into lesions, the genetically modified cells can be allowed to develop into mature tissue before implantation. This approach combines gene thera- py with tissue engineering. Prelimi- nary success has been reported with chondrocytes that have been trans- fected with IGF-1 cDNA, seeded onto scaffolds, and incubated in a bioreac- tor. 60 Many of the principles for repair- ing articular cartilage can be extend- ed to the repair of meniscal lesions. Marker genes have been successful- ly expressed in experimental menis- cal lesions by ex vivo and gene plug approaches. 7 Using a tissue engineer- ing approach, genetically modified meniscal cells have been seeded onto a matrix and implanted into nude mice, where the cells develop into me- niscal tissue. 61 Intervertebral Disk Using strategies similar to those employed in the repair of articular cartilage, investigators are develop- ing methods of introducing genes into cells of intervertebral disks to prevent or reverse disk degenera- tion. 62 One interesting and unex- pected finding is the remarkably prolonged duration of transgene ex- pression that follows the intradiskal injection of recombinant adenoviral vectors. This duration appears to re- flect the immunologically protected environment of the disk and the nondividing state of its cells. It should be a major asset to the fur- ther development of this approach to therapy. Introduction of growth factor genes into disk cells elevates the synthesis of matrix macromole- cules, but whether this protects or heals disks in vivo has not yet been evaluated in animal models. Ligament and Tendon Cells recovered from ligaments and tendons are readily transduced by a variety of viral and nonviral vec- tors, and gene transfer can be accom- plished by ex vivo and in vivo strat- egies. 63 Delivery of cDNAs encoding growth factors promotes cell division and the deposition of extracellular matrix in vitro, 64 but it is not yet known whether this accelerates heal- ing in vivo. BMP-12 and -13 proteins are of particular interest because they pro- mote the differ entiation of mesenchy- mal stem cells into tissue with the ap- pearance of ligament and tendon. Intramuscular injection of an adeno- virus encoding BMP-12 leads to the formation of ectopic ligamentous tis- sue. When this vector is injected into chicken tendon cells, there is an in- crease in the synthesis of type I col- lagen. In a complete tendon lacera- tion chicken model, BMP-12 gene transfer doubled the tensile strength and stiffness of the repaired ten- dons. 65 Another strategy for improving the healing of ligaments and tendons is to reduce the synthesis of decorin. This small proteoglycan is an attrac- tive target because it limits the diam- eter of collagen fibrils and also acts as an antagonist of TGF-β. Blocking decorin production has been evalu- ated as a means to improve healing of the medial collateral ligament in a rabbit model. Inhibiting decorin ex- pression with antisense RNA in- creased the average diameter of the collagen fibers within the repair tis- sue and improved the mechanical properties of the ligament. 66 Summary Gene therapy offers a broad range of potential applications for treating musculoskeletal conditions in all specialty areas. 67 In particular, gene transfer offers novel therapeutic ap- proaches to all six focus areas iden- Christopher H. Evans, PhD, et al Vol 13, No 4, July/August 2005 239 . treatment of mende- lian disorders, butit also has wide ap- plication for diseases that lack a strong or simple genetic basis. In such instances, gene transfer is used as a biologic delivery system. cy- tosine residues in DNA. The un- methylated dinucleotide sequence cytosine-guanosine (CpG) strongly activates cell-mediated immunity. 12 In the absence of problems with cell turnover or the. immune system, long-term gene expression also can be curtailed at the level of the promoter. The strong viral promoters often fa- vored for gene therapy experiments may become turned off (silenced)