Development Of Ex Vivo Gene Delivery By Myoblast Transplantation

A. Evidence for Myogenic Precursor Cells: Myoblasts

Skeletal muscle comprises approximately 10% of the total human body mass and is highly accessible to manipulation. A typical striated mature skeletal muscle cell, known as a myofiber, is large (1-40 mm in length and 10-50 microns wide), cylindrical, and multinucleated (as many as 100 nuclei per cell). Myofibers contain many nuclei because they are formed during development by the fusion of mononucleated precursor cells known as myoblasts. Myoblasts persist in mature muscle tissue as satellite cells, which can be viewed by electron microscopy as being ''wedged'' between the plasma membrane of the myofiber and the surrounding extracellular matrix (30) (Fig. 2). These cells can continue to fuse to neighboring myofibers in mature muscle, aiding in new muscle formation during regeneration following injury (31).

Interest in myoblasts as vehicles for gene delivery arose from studies of muscle cell biology and development. Studies of pattern formation in skeletal muscle showed that myoblasts are greatly influenced by extrinsic factors and become integrated into existing muscle fibers. Mammalian skeletal muscle is composed of a complex pattern of myofibers. Fiber types differ in their rate of contraction (fast and slow), determined in part by the ratio of fast and slow myosin heavy chain (MyHC) isoforms contained in each fiber (32). Although both fiber types occur in all skeletal muscles, the ratio of the 2 classes differs between muscles and even between different regions of a single muscle (33). Whereas lineage and myoblast-intrinsic

Figure 2 Satellite cell viewed by electron microscopy. Electron micrograph of a satellite cell in frog skeletal muscle, seen in longitudinal view. Extreme poles of the cell are indicated (sc). The arrow marks where the plasma membrane of the satellite cell juxtaposes that of the muscle fiber. (Reproduced from Ref. 30, p. 495, by copyright permission of The Rockefeller University Press.)

Figure 2 Satellite cell viewed by electron microscopy. Electron micrograph of a satellite cell in frog skeletal muscle, seen in longitudinal view. Extreme poles of the cell are indicated (sc). The arrow marks where the plasma membrane of the satellite cell juxtaposes that of the muscle fiber. (Reproduced from Ref. 30, p. 495, by copyright permission of The Rockefeller University Press.)

properties play a role in muscle fiber patterning, as shown by the finding that myoblasts expressing different slow isoforms are characteristic of different developmental stages (34), a number of experiments suggest that the environment is important in the generation and maintenance of that pattern (35). During early development of human limb muscle when multiple fiber types are forming, virtually all myoblasts, irrespective of the stage of development from which they are taken, give rise to clones expressing slow MyHC upon differentiation in culture. This is seen to be true even when myoblasts are taken from muscle at midgestation, when only 3% of fibers in vivo express slow MyHC. These results suggest that although culture conditions allow for slow MyHC expression in my-oblasts, such expression seems to be repressed by extrinsic factors in vivo.

Additional experiments using retroviruses as heritable markers of cell fate in vivo further solidified these findings. When retroviral vectors encoding the reporter gene lacZ were injected directly into muscles of postnatal rats, clusters of multiply labeled fibers arising from progeny of single satellite cells were observed (24). The basal lamina, a connective sheath surrounding each muscle fiber, did not appear to prevent migration of labeled myoblasts into multiple fibers. Moreover, rat myoblast clones were shown to contribute progeny to both slow and fast muscle fiber types in their vicinity in vivo (25). These results demonstrate that mammalian my-oblasts fuse randomly with all fiber types encountered, and adopt the pattern of myogenic gene expression characteristic of the host muscle fiber.

Transplantation studies further developed the notion of employing myoblasts for gene delivery, using allografts of minced muscle tissue (36) or pieces of intact muscle (37), and also by injection of muscle precursor cells (38-41). In these studies, isoforms of the enzyme glucose-6-phosphate iso-merase were employed as markers to differentiate contributions of donor and host myoblasts to myofibers. The detection of hybrid fibers expressing isoforms containing subunits de rived from donor- and host-provided evidence that grafts of muscle precursor cells could alter the genetic makeup of, and contribute muscle proteins to, mature myofibers.

These initial experiments were of importance in establishing that myoblasts can fuse with all muscle fiber types in their vicinity, becoming fully integrated into mature muscle tissue that has access to the circulation and is innervated. In addition, muscle precursor cells of one genotype that are injected into muscle tissue of another genotype fuse to form hybrid muscle fibers, where they are capable of expressing donor genes. These findings paved the way for later studies examining applications of myoblast transplantation for the correction of various diseases.

B. Methodology: Purification, Growth, and Transduction of Primary Myoblasts

The ease of isolating myoblasts from both mouse and human muscle, and purifying, growing, and transducing them in vitro is a major advantage of using myoblasts rather than other cell types for gene transfer. Primary myoblasts can be isolated from any mouse strain—including strains carrying genetic mutations or transgenic strains (42,43)—providing a broad array of genotypes either for study in tissue culture or for transplantation. Moreover, because myoblasts may be isolated from a specific donor for implantation into a syngeneic host, problems of immunoincompatibility are obviated. Although established myoblast cell lines, such as the C2 myogenic cell line, may also be used for implantation, these cells can proliferate and form tumors when implanted into mice (44). In contrast, despite their impressive capacity to proliferate in culture, primary muscle cells do not form tumors upon injection into mouse muscle (44), and in the case of human myoblasts exhibit Hayflick-like senescence but not transformation (45).

Myoblasts can be isolated from muscle tissues from individuals of all ages, although both the yield and number of doublings tend to be higher if obtained from younger donors. Primary cultures are derived from postnatal muscle using mechanical or enzymatic dissociation methods (44,46), and readily obtained from human biopsy or autopsy tissue (47). Because such cultures are composed of a mixture of cell types, the population of myoblasts must be further purified. For primary cultures isolated from mice, this is accomplished using cell culture conditions that favor myoblast growth at the expense of other cell types, such that a pure population of my-oblasts can be obtained within 2 weeks of normal growth (26,44). Human myoblasts can be purified by sorting in a fluorescence-activated cell sorter (FACS), employing fluorescent antibodies specific to the muscle surface antigen H31 or neural cell adhesion molecule (NCAM) (46). The cells that are isolated are capable of self-renewal and can undergo at least 40 cell doublings without differentiating (46). This implies that a kilogram of cells for transplantation use may be derived from a 5-mm3 biopsy. Mouse (48) andrat (49) primary cells have also been isolated by FACS using antibodies to a 7 integrin. Thus, a large population of rodent or human myoblasts may be easily obtained, purified, and expanded in cell culture.

Recombinant genes can be stably introduced into isolated and purified myoblasts using a number of methods, including lipofection, calcium-mediated transfection, or (more readily) by retroviral infection. Using conditions optimized for retroviral infection of myoblasts at high efficiency, 99% of primary myoblasts in culture are easily transduced without use of a selectable marker (50). This enables the creation of pure populations of primary myoblasts expressing a gene of interest, free of contamination by most nonexpressing cells. My-oblasts do not lose their ability to mature and differentiate by the process of being genetically altered.


Primary mouse myoblasts stably express recombinant genes following transduction. Myoblasts retrovirally transduced with the bacterial lacZ gene and injected into mouse skeletal muscle fuse with muscle fibers and express high levels of p-galactosidase (Fig. 3). The p-galactosidase expression can be observed in hybrid myofibers for at least 6 months (44). Other studies have shown that stable levels of recombinant proteins are produced for at least 10 months (51). Myoblasts that have been genetically engineered to express a recombinant gene may thus be used for stable delivery of that gene into the body (Fig. 4).

Although not widely viewed as a secretory tissue, skeletal muscle is highly vascularized and recombinant proteins secreted from myoblasts readily gain access to circulation. Initial studies using C2C12 myoblasts genetically altered to express human growth hormone (hGH) first demonstrated this to be the case (52,53). hGH was chosen as the gene of interest because it has a very short half-life in mouse serum (4 min) (54), providing a stringent test for sustained production and secretion into the circulation over time. After injection of genetically engineered myoblasts into mouse muscle, stable physiological levels of hGH could be detected for at least 3 months (Fig. 5). These results showed that myoblasts, by fusing with preexisting multinucleated myofibers, can serve as vehicles for systemic delivery of recombinant proteins. Thus, skeletal muscle may be used as a factory for production of a range of secreted gene products for treatment of nonmuscle-related disorders. Because muscle is capable of carrying out posttranslational modifications normally performed by other tissues (e.g., gamma carboxylation essential for production of functional coagulation factors in the liver), such recombinant nonmuscle proteins are biologically active even when pro-ducedby muscle (55-58). Applications of myoblast-mediated gene delivery to treat diseases affecting both muscle and other tissues are discussed in the following section.


A. Muscular Dystrophies

The concept of applying myoblast transplantation to the treatment of disease was a natural outcome of the many studies

Figure 3 Incorporation of P-galactosidase-expressing myoblasts into skeletal muscle. Primary mouse myoblasts transduced with the reporter gene lacZ, encoding the bacterial p-galactosi-dase enzyme, were injected into mouse leg skeletal muscle, where they formed hybrid myofibers with host muscle. Injected muscles were isolated and frozen, and cryostat sections were prepared for histological analysis. (A) Hybrid myofibers producing p-galac-tosidase at the implantation sites can be seen as dark fibers after staining with the enzyme's substrate X-gal. (B) An adjacent section was stained with hematoxylin/eosin to show tissue architecture, and demonstrates that the hybrid fibers are of normal diameter and morphology and are an integral part of the muscle tissue. The centrally located nuclei that can be observed in (B) are indicative of myofibers that have undergone regeneration, and represent a normal response to a needle injection. The arrows denote corresponding regions in the two sections.

establishing myoblasts as potent vehicles for delivering donor genes into host muscle. The first approaches centered on the use of allografts of normal precursor cells to insert donor nuclei, containing a normal genome, into genetically abnormal muscle. Although not technically gene therapy (because donor myoblasts were not genetically engineered in any way), such ''cell therapy'' experiments were important in establishing the utility of myoblast-mediated gene delivery for the treatment of disease, and are the only studies involving myoblast transplantation that have translated into human clinical trials to date. The first disorder to which this therapeutic approach was applied was Duchenne muscular dystrophy (DMD), the most common of heritable human muscular dystrophies. DMD affects 1 in 3000 males and causes progressive muscle weakness beginning in childhood; patients with severe forms rarely survive past early adulthood. DMD is caused by mutations in the gene dystrophin (59), a large gene encoding a structural protein involved in anchoring skeletal myofibers to the extracellular matrix. By implanting myoblasts that contained normal copies of the dystrophin gene into dystrophin-deficient muscle, researchers hoped to rescue the genetic defect in humans as previously achieved in mdx mice, the mouse model of dis ease (41). In mdx mice, the implanted myoblasts were able to render host myofibers dystrophin positive while counteracting the characteristic cycle of fiber degeneration and regeneration characteristic of mdx muscle (41,60).

Clinical trials in which donor myoblasts taken from normal human muscle were introduced into DMD patients were initiated at multiple institutions (61-67). All these studies demonstrated that myoblast implantation into humans has no adverse effects. However, all but 1 group reported the disappointing finding that only a very small percentage of host myofibers resulted in normal dystrophin expression. At the protein level, these results could have been due to reversion or occasional expression by mutant host fibers of a truncated dystrophin detectable by antibodies. One group, however, provided definitive evidence that donor dystrophin transcripts were being synthesized by polymerase chain reaction (PCR) (62). Experiments combining fluorescent in situ hybridization (FISH) together with immunohistochemistry were recently conducted to examine the fate of individual myoblasts after implantation into muscles of DMD patients (68,69). This combination of techniques allowed the localization of both the dystrophin protein and the donor nuclei themselves, permitting more quantitative assessment of the efficiency of myoblast transfer. Findings from these studies showed that a large proportion of donor myoblasts successfully integrated into host myofibers in almost every subject; donor nuclei were interspersed with and aligned with host nuclei. Furthermore, these experiments demonstrated that increased dystrophin expression observed in recipient muscle was contributed by the donor nuclei and was not due to spontaneous reversion of the mutated dystrophin gene because the antibodies used were specific to the product for the deleted gene regions in the recipient. Moreover, the dystrophin produced by single nuclei spanned regions, including 20 to 30 nuclei. Why only a subset of transduced myofibers expressed dystrophin is still not understood. One hypothesis is that variables related to the DMD disease state itself, such as increased fibrosis with patient age, impaired myoblast access. An alternative hypothesis is that nuclei were not tran-scriptionally active in regions of fibers undergoing degeneration.

For treatment of muscular dystrophies by gene therapy, a large proportion of muscle fibers must be transduced to produce a beneficial outcome. Furthermore, myoblasts must be implanted into all muscles, some of which are difficult to access, such as the diaphragm and heart. Failure of these latter muscles is the cause of death in patients with DMD. These represent major challenges to myoblast-mediated gene delivery in treating inherited myopathies and suggest that a cell-based method may not be practical. Histochemical staining and enzymatic activity assays of muscle transplanted with p-galactosidase-expressing myoblasts show that the total number of labeled fibers and the total p-galactosidase activity is maximal at the implantation site, and decreases in parallel with increasing distance from the site (70). Although my-oblasts were believed to be able to migrate from the circulation to damaged muscle (71), this is certainly not a frequent event. Thus, to target a high percentage of myofibers in multiple

Figure 4 Muscle-mediated gene therapy. Muscle-mediated gene therapy by implantation of genetically engineered myoblasts allows for delivery of diverse therapeutic proteins, either directly to muscle or (as shown) to the systemic circulation. (Adapted with permission from Ref. 23, p. 1555. Copyright © 1995 Massachusetts Medical Society. All rights reserved.)

Figure 4 Muscle-mediated gene therapy. Muscle-mediated gene therapy by implantation of genetically engineered myoblasts allows for delivery of diverse therapeutic proteins, either directly to muscle or (as shown) to the systemic circulation. (Adapted with permission from Ref. 23, p. 1555. Copyright © 1995 Massachusetts Medical Society. All rights reserved.)

muscles of large organisms such as humans, delivery of viral vectors and naked DNA encoding either full-length or truncated dystrophin genes (72-74), or in the future the ubiquitous utrophin (75), may be most effective.

B. Lysosomal Storage Diseases and Serum Protein Deficiencies: Treatment by Secreted Circulating Recombinant Proteins

As described above, studies with hGH showed that myofibers efficiently secrete recombinant proteins that readily gain access to the circulation. Myoblast-mediated gene transfer has been further employed to express therapeutic proteins not normally made by muscle (51-53,55-58,76-79). Here we describe in further detail progress made in studies in which genes encoding ^-glucuronidase, clotting factor IX, and erythropoi-etin were transferred to muscle using myoblasts.

Lysosomal storage diseases are a subset of disorders that may be appropriate for muscle-mediated gene therapy (80).

These recessive disorders are caused by detrimental buildup of lysosomal enzyme substrates within affected tissues due to a single missing or dysfunctional lysosomal enzyme. Because these enzymes are marked with a specific targeting signal (mannose 6-phosphate) (81), missing lysosomal enzymes manufactured by muscle and delivered to the serum can be internalized by distant tissues and appropriately transported to lysosomes via mannose 6-phosphate receptors. By implanting into muscle genetically engineered primary myoblasts encoding ^-glucuronidase, a lysosomal enzyme, 1 group was able to demonstrate in vivo expression of the recombinant protein in adult p-glucuronidase-deficient mice (77). Production and secretion of the missing lysosomal enzyme by muscle led to correction of phenotypic abnormalities in the liver and spleen of treated animals.

A second disorder well suited to myoblast-mediated gene therapy is hemophilia B. Hemophilia B is a blood-clotting disease caused by a deficiency of a protein, clotting factor IX. Because conventional protein replacement therapies face drawbacks, including the necessity for frequently repeated

Figure 5 Systemic delivery of human growth hormone. A population of C2C12 myoblasts retrovirally transduced with hGH were implanted into hind limbs of 24 syngeneic mice, and serum hGH levels were monitored by radioimmunoassay of tail blood. Greater than 90% of the implanted cells expressed and secreted hGH as determined by clonal analysis in culture. Each point represents the mean ± SD for 4 to 24 mice; the dashed line shows the mean ± SD for serum samples taken from five uninjected control mice. Expression of hGH by implanted myoblasts persisted for at least 85 days in vivo. (Reprinted with permission from Ref. 53. Copyright © 1991 American Association for the Advancement of Science.)

Figure 5 Systemic delivery of human growth hormone. A population of C2C12 myoblasts retrovirally transduced with hGH were implanted into hind limbs of 24 syngeneic mice, and serum hGH levels were monitored by radioimmunoassay of tail blood. Greater than 90% of the implanted cells expressed and secreted hGH as determined by clonal analysis in culture. Each point represents the mean ± SD for 4 to 24 mice; the dashed line shows the mean ± SD for serum samples taken from five uninjected control mice. Expression of hGH by implanted myoblasts persisted for at least 85 days in vivo. (Reprinted with permission from Ref. 53. Copyright © 1991 American Association for the Advancement of Science.)

treatments and the risk of contaminating blood-borne pathogens in plasma-derived factors, gene therapy may provide a safer and more convenient alternative (82). Initial studies in which C2C12 myoblasts were transduced with a gene encoding human factor IX and implanted into immunocompetent mice led to a peak expression of recombinant protein (1 ^g/ mL) at day 12, and subsequent decline back to basal levels thereafter (55). The drop in human factor IX expression was shown to be due to production of specific antibodies targeted against the protein in wild-type mice. Other experiments (56-58) demonstrated that primary myoblasts engineered to constitutively express factor IX led to stable, low-level production of the protein in immunodeficient nude or SCID mice for many months. Another study (51) achieved stable production of human factor IX at therapeutic levels in SCID mice, using a promoter with muscle creatine enhancers to drive high levels of muscle-specific expression, for at least 8 months. Of importance, recombinant factor IX manufactured in muscle undergoes the gamma carboxylation required for functional activity of the protein (55,56). This finding demonstrates that muscle cells have efficient mechanisms for posttranslational modifications normally carried out by other tissue types such as liver. Moreover, the problems with immunogenicity are likely to affect only a percentage of hemophiliacs, as not all are null mutations but have some, albeit reduced, level of factor IX (83). Until recently, only dog models were available, however, now a mouse model that lacks factor IX has been created by homologous recombination (84), which should facilitate future preclinical gene therapy studies.

A third class of disorders for which myoblast-mediated expression of recombinant proteins into the circulation may be beneficial is in the treatment of erythropoietin (Epo)-re-sponsive anemias. Recombinant Epo replacement therapy has been employed for successful treatment of anemia associated with end-stage renal disease (85) and is being tested as a therapy for a broad array of other anemias (86). Epo is a mammalian hormone that controls the production of erythro-cytes, hemoglobin-carrying cells that deliver oxygen to tissues of the body (87). Anemic patients can currently be treated by repeated administration of recombinant Epo; such treatments, however, require frequent hospital visits by patients and are costly. Thus, Epo delivery by gene therapy could provide patients with long-term delivery of the protein, eliminating the need for multiple treatments. However, as with many gene therapies, regulated expression is desirable for Epo, because dosage must be tailored to the particular application and to the individual patient.

Studies of muscle-mediated delivery of Epo by gene therapy appear promising. Epo-secreting primary or C2 myoblasts have been introduced bilaterally into skeletal muscles of mice (77,78). Implantation of engineered cells led to an elevated hematocrit for 3 months, a direct measure of Epo production. At 3 months posttransplantation, implanted myoblasts were observed to have fused and fully differentiated into myofibers (78). Moreover, in an animal model of renal failure in which anemia is induced by nephrectomy of immunocompromised nude mice, injection into muscle of C2 myoblasts secreting human Epo led to reversal of the anemic phenotype (79). Levels of recombinant serum Epo measured by enzyme-linked immunosorbent assay (ELISA) remained elevated for the 2 months during which the animals were assessed following myoblast implantation. These studies lend credibility to using myoblast-mediated expression of recombinant Epo as a viable treatment for anemias.

Thus, myoblast-mediated gene transfer appears to be well suited for expression of recombinant proteins to the circulation. Unlike applications aimed at treating inherited myopathies, not all fibers need to be transduced with the gene of interest to achieve a therapeutic effect. Indeed, such therapies can be highly localized to a particular region of a single muscle. For a variety of disorders where patients may benefit from delivery of a recombinant gene product to the bloodstream, myoblast-mediated gene transfer to muscle tissue appears to be a promising treatment method. Stable, long-term expression of physiological levels can be achieved with therapeutic effects, and because there is no immune response, repeated administration of genetically engineered myoblasts is possible, unlike AAV or adenoviral gene delivery.

C. Vascular Insufficiencies and Cancer

Since the 1990s, a great deal has been learned about growth factors that induce angiogenesis, the sprouting of new blood vessels from preexisting vessels. There has been much interest in the use of angiogenic factors to stimulate new vessels to grow as a treatment for maladies including stroke, peripheral arterial disease, and myocardial infarction. As the genes that encode these proteins have been cloned, the concept of thera peutic angiogenesis has moved quickly into the realm of gene therapy and clinical trials are already underway as low levels provide therapeutic effects. Factors produced by genetically engineered myoblasts are continuously produced by contrast with injection of pure proteins, naked DNA, and viral vectors, and may be advantageous.

The angiogenic factor that has received the most attention to date is vascular endothelial growth factor (VEGF), a potent mitogen that was isolated by virtue of its ability to stimulate growth of endothelial cells and to increase permeability in vascular endothelium (hence its other designation, vascular permeability factor) (88-92). VEGF plays an important role in the induction of angiogenesis by tumors (93), and in the angiogenic response of normal tissue to decreased oxygen availability. VEGF is also known to serve as a critical signal during the initial embryonic development of the vasculature by a process known as vasculogenesis, or the de novo growth of blood vessels from precursor cells (94,95). In this case, VEGF induces endothelial cell migration via specific receptors. Therefore, VEGF is a crucial regulator of both modes of growth and development of the vasculature pre- and postna-tally.

Because of the potential clinical benefits of stimulating new blood vessel growth, much effort in recent years has been invested in the delivery of VEGF to tissues that are insufficiently vascularized. Injection of VEGF protein has resulted in angiogenic sprouting of vessels in muscle that was partially deprived of blood and oxygen, and therefore ischemic (96-98). However, presumably because of vascular permea-bilizing and/or vasodilating properties, bolus injections of the protein have been reported to be deleterious, causing hypotension (99,100). As a result, recent investigations have assessed the feasibility of localized delivery of VEGF by gene transfer using plasmid DNA injection or adenoviral vectors. Both of these delivery methods lead to transient production of the recombinant protein, and to angiogenic sprouting from preexisting vessels in matrigel in vitro (101,102), in adipose tissues in vivo (103), as well as in ischemic skeletal or cardiac muscle (104-107). These results have led to clinical trials of VEGF gene delivery for ischemic heart and limb diseases (104,108,109).

The effects of long-term stable production of VEGF were recently investigated using the myoblast-mediated gene transfer techniques described above (110). This resulted in many unexpected findings. Myoblasts were transduced with a retro-virus carrying a murine cDNA encoding the heparin-binding VEGF164 and injected into the muscles of SCID mice. A physiological response to VEGF was observed in every mouse that received VEGF-producing cells. At day 11 postimplantation, mice appeared outwardly normal and no differences were observed between VEGF and control muscle upon dissection. However, histological analysis of frozen muscle sections revealed that the implantation sites of VEGF-expressing my-oblasts, but not control myoblasts, were invariably associated with regions of infiltrating mononuclear cells, identified by fluorescent antibody staining to multiple markers as endothe-lial cells and macrophages. By days 44-47, 100% of legs injected with VEGF myoblasts contained large hemangiomas composed of vascular channels and pools of blood, whereas control legs appeared normal (Fig. 6). These results have been reproduced in the heart following myoblast implantation into myocardium (111). However, implantation into muscle of clonal populations of myoblasts that express lower levels of VEGF results in an increased number of vessels without associated hemangioma formation (C. Ozawa et al., submitted). Interestingly, if the implanted cells are a mixture of myoblast clones expressing different levels of VEGF, due to genomic differences in retroviral insertion, the induced vasculature ranges from seemingly normal capillaries to hemangioma-like vascular sacs. This reveals a surprisingly localized response to VEGF on a microscopic scale and shows the importance of not exceeding threshold levels of VEGF, not only on a tissuewide level, but also on a per cell level. These studies demonstrate that myoblast-mediated VEGF gene delivery is extremely potent and provide evidence that this single growth factor can lead to a cascade of events resulting in the formation of complex tissues of multiple cell types. These results also show for the first time that exogenous VEGF expression at high levels or long duration can have deleterious effects, a factor of importance as clinical trials of VEGF gene delivery by plasmid DNA injection or adenoviral-mediated delivery are underway. Moreover, because myoblast implantation affords higher expression levels of longer duration than other gene transfer techniques, a physiological response to VEGF was observed in nonischemic muscle for the first time. Thus, the dose and duration of VEGF expression appear critical in determining a range of effects.

These results point both to the potency of myoblast-me-diated gene transfer, and the necessity of regulation of recombinant gene expression for gene therapy applications. In the case of myoblast-mediated delivery of VEGF, too much of a good thing clearly can lead to adverse and unwanted effects. Gene therapy has usually been plagued by insufficient levels of the protein of interest. However, these VEGF results illustrate that current methods of gene delivery can be limited by a lack of ability to control gene expression. Both the ability to increase expression levels if an insufficient amount of a recombinant protein is being produced, and the option to intentionally reduce or cease expression, are likely to be necessary for the health of the patient in many cases.

In addition to studies aimed at triggering the growth of new blood vessels, other experiments are currently geared toward preventing blood vessel development in special circumstances. Because tumor growth and metastasis require persistent new blood vessel growth (112,113), therapies targeted at blocking this growth could lead to an arrest of tumor development. One of the most promising avenues for preventing an-giogenesis in tumors may lie in the utilization of recently discovered antiangiogenic agents.

Early in the 20th century, it was first noticed that primary tumors are able to suppress the growth of a second tumor inoculum (114). Resistance to secondary tumor challenge was shown to be inversely proportional to the size of the second tumor inoculum, and directly proportional to the size of the

Injected teg Uninjected leg

0 0

Post a comment