Adeno-associated virus vector-mediated gene transfer into dystrophin-deficient skeletal muscles evokes enhanced immune response against the transgene product

Gene Therapy (2002) 9, 1576–1588  2002 Nature Publishing Group All rights reserved 0969-7128/02 $25.00 www.nature.com/gt

RESEARCH ARTICLE

Adeno-associated virus vector-mediated gene transfer into dystrophin-deficient skeletal muscles evokes enhanced immune response against the transgene product K Yuasa1, M Sakamoto1, Y Miyagoe-Suzuki1, A Tanouchi1, H Yamamoto2, J Li3, JS Chamberlain4, X Xiao3 and S Takeda1 1

Department of Molecular Therapy, National Institute of Neuroscience, National Center of Neurology and Psychiatry, Tokyo, Japan; Department of Immunology, Graduate School of Pharmaceutical Sciences, Osaka University, Osaka, Japan; 3Department of Molecular Genetics and Biochemistry, Gene Therapy Center, and Duchenne Muscular Dystrophy Research Center, University of Pittsburgh, Pittsburgh, PA, USA; and 4Department of Neurology, University of Washington School of Medicine, Seattle, WA, USA 2

Duchenne muscular dystrophy (DMD) is an X-linked, lethal muscular disorder caused by a defect in the DMD gene. AAV vector-mediated micro-dystrophin cDNA transfer is an attractive approach to treatment of DMD. To establish effective gene transfer into skeletal muscle, we examined the transduction efficiency of an AAV vector in skeletal muscles of dystrophin-deficient mdx mice. When an AAV vector encoding the LacZ gene driven by a CMV promoter (AAVCMVLacZ) was introduced, ␤-galactosidase expression markedly decreased in mdx muscle 4 weeks after injection due to immune responses against the transgene product. We also injected AAV-CMVLacZ into skeletal muscles of mini-dystrophin-transgenic mdx mice (CVBA3’), which show ameliorated phenotypes without overt signs of muscle

degeneration. AAV vector administration, however, evoked substantial immune responses in CVBA3’ muscle. Importantly, AAV vector using muscle-specific MCK promoter also elicited responses in mdx muscle, but at a considerably later period. These results suggested that neo-antigens introduced by AAV vectors could evoke immune reactions in mdx muscle, since increased permeability allowed a leakage of neo-antigens from the dystrophindeficient sarcolemma of muscle fibers. However, resident antigen-presenting cells, such as myoblasts, myotubes and regenerating immature myofibers, might also play a role in the immune response. Gene Therapy (2002) 9, 1576–1588. doi:10.1038/sj.gt.3301829

Keywords: AAV vector; gene transfer; skeletal muscle; mdx mouse; immune response; micro-dystrophin

Introduction Duchenne muscular dystrophy (DMD) is an X-linked, lethal disorder of skeletal muscle caused by a defect in the dystrophin gene, which encodes a large sub-sarcolemmal cytoskeletal protein. The absence of dystrophin accompanied the loss of dystrophin–glycoprotein complex at the sarcolemma and results in progressive muscle weakness, cardiomyopathy and early death. DMD is characterized by a high incidence (one among 3500 boys) and a high frequency of de novo mutation1 and currently has no effective treatment. Several treatment modalities aimed at correcting the dystrophic phenotype have been researched using the dystrophin-deficient mdx mouse and the canine X-linked muscular dystrophy (CXMD), and tried on human DMD patients. These include the adenoviral vector-mediated introduction of functional dystrophin,2–6 naked DNA,7 Correspondence: S Takeda, Department of Molecular Therapy, National Institute of Neuroscience, National Center of Neurology and Psychiatry, 4-1-1 Ogawa-higashi, Kodaira, Tokyo 187-8502, Japan Received 22 January 2002; accepted 16 June 2002

and primary myoblast8 or stem cell9 transplantation. Alternatively, correction of the endogenous dystrophin gene was attempted using chimeric RNA/DNA oligonucleotides10,11 or antisense 2’-O-methyl oligoribonucleotides.12 The antibiotic gentamicin has been reported to increase read-through dystrophin protein over a premature stop codon,13 and clinical trials have been conducted recently.14 On the other hand, compensation for dystrophin lack by up-regulation of endogenous utrophin15 is another possible strategy. These innovative strategies have not been successful because of several difficulties, including host immune response and low efficiency. Adeno-associated virus (AAV) vector is a potential tool of gene therapy for treatment of genetic neuromuscular disorders. It is a non-pathogenic and replication-defective viral vector without any viral open reading frame that is able to infect effectively non-dividing cells, such as those of skeletal muscle16–18 and the central nervous system.19 It is widely believed that AAV vectors offer long-term expression of the transferred gene without immune response against the transferred gene product, and it has been demonstrated that intramuscular injection of AAV vectors expressing immunogenic proteins does not

AAV vector-mediated gene transfer into mdx muscle K Yuasa et al

stimulate humoral and cellular immune response to transgene products in immunocompetent mice.16–18,20 Thus, AAV vectors may be applicable to treatment of inherited neuromuscular disorders such as DMD. In fact, the AAV vector has been successfully introduced into ␦sarcoglycan-deficient hamsters (Bio 14.6)21,22 and ␥-sarcoglycan-deficient mice,23 animal models of limb girdle muscular dystrophies (LGMD). A recent report, however, showed lower levels of transgene expression together with substantial immune response to the therapeutic transgene product in ␥-sarcoglycan-null mice. Restriction of transgene expression to the muscle is important to avoid transgene expression in antigen-presenting cells (APCs).24 Although AAV vector-mediated gene transfer is an attractive option for treatment of DMD, the size of the foreign gene is limited to 4.7–4.9 kb of exogenous DNA. Therefore, full-length dystrophin (14 kb) and mini-dystrophin (6.4 kb) cDNAs are too large to be incorporated into AAV vector as the therapeutic gene.25,26 We previously reported that rod-truncated micro-dystrophin, which is encoded by 3.7 kb cDNA, could effectively accumulate at the sarcolemma and recover dystrophinassociated proteins after adenovirus-mediated gene transfer into mdx skeletal muscles.27 Later, Wang et al28 reported the effective transfer of two kinds of micro-dystrophin cDNAs (⬍4.2 kb) into mdx skeletal muscle using AAV vectors. The AAV vector treatment ameliorated dystrophic pathology in mdx muscle, demonstrating micro-dystrophin cDNA to be a good candidate for insertion into the AAV vector. To establish effective gene transfer into dystrophic skeletal muscle, we examined the transduction efficiency of the AAV vector in the skeletal muscles of mdx mice, using the ␤-galactosidase (␤-gal) gene as a reporter. AAV vector-mediated gene transfer elicited strong immune responses and resulted in poor and short-term expression of the transferred gene in mdx muscles. Furthermore, an AAV vector elicited immune responses in transgenic mdx muscle expressing mini-dystrophin, despite amelioration of dystrophic phenotypes. We also demonstrate that enhancement of immune responses was associated closely with membrane permeability. In addition, use of muscle-specific promoter could not prevent the activation of the immune system, although the activation was significantly delayed. Our results suggest the importance of transient immunosuppression before the transgene product accumulates in the affected muscle.

Results ␤-Gal expression in mdx skeletal muscle after intramuscular injection of an AAV vector To investigate the transduction efficiency of AAV vectormediated gene transfer in skeletal muscles of dystrophindeficient mdx mice, we injected the AAV vector encoding ␤-galactosidase driven by a ubiquitous CMV promoter (AAV-CMVLacZ) into the tibialis anterior (TA) muscles of 5-week-old C57BL/10 (B10) and mdx mice. One, 2 and 4 weeks after injection, the expression of ␤-galactosidase (␤-gal) was evaluated by X-gal staining and ONPG (onitrophenyl ␤-D-galactopyranoside) assay. X-gal staining showed that the number of ␤-gal-positive fibers in AAV vector-injected muscles of control B10 mice increased

with time, and high levels of ␤-gal were expressed 4 weeks after injection (Figure 1a). In contrast, ␤-gal activity and the number of ␤-gal-positive fibers in AAV vector-injected mdx muscle was much lower than that in B10 mice, and had decreased significantly 4 weeks after injection (Figure 1a and b). Poor and reduced ␤-gal expression in the late stage in mdx muscle after AAV vector-mediated gene transfer may be due to the elimination of large numbers of ␤-galpositive fibers by active muscle degeneration–regeneration cycles. Muscle fibers of 5-week-old mdx mice generally degenerate at a high rate. To test this hypothesis, we injected AAV-CMVLacZ into the skeletal muscles of 12week-old mdx mice, when muscle degeneration has almost ceased. Importantly, ␤-gal expression in muscles injected at 12 weeks was much lower than that at 5 weeks in both mdx and B10 mice. The expression of transferred gene products showed a similar pattern: it increased for 2 weeks after injection and at 2 weeks there was little difference in ␤-gal activities between mdx and B10 skeletal muscles, but it reduced quickly in mdx mice at 4 weeks after injection, as seen in 5-week-old mdx mice (Figure 1c). This result supports our hypothesis that active degeneration of mdx muscle is largely responsible for poor expression of the transgene in the early stage of the AAV vector administration. Decreased expression of the LacZ gene products was also observed at 4 weeks after AAV vector administration when introduced into 12-week-old mdx muscles. This observation suggests that the remarkable reduction of ␤-gal expression at the later stage in AAV vector-injected mdx muscle was caused by mechanisms other than the dystrophic process.

1577

Histology of mdx skeletal muscle after AAV vectormediated gene transfer To investigate why ␤-gal expression decreased quickly in mdx muscle 4 weeks after AAV-CMVLacZ injection, we scrutinized the patterns of X-gal staining and histology of AAV vector-injected skeletal muscles of mdx and B10 mice. X-gal staining in AAV vector-injected mdx muscles showed a mosaic pattern at 4 weeks after injection (Figure 1a): a fiber strongly positive for X-gal staining often neighbors totally negative muscle fibers. Next, we analyzed the histology of the AAV vector-injected mdx muscle (Figure 2, column H&E and X-Gal). The histology of skeletal muscles of B10 mice seemed to be normal even 4 weeks after injection, as previously reported.16–18,20 In AAV vector-injected mdx muscle, clusters of infiltrating cells were often observed 1 and 2 weeks after injection, but such infiltrates are observed even in the non-injected muscle. Four weeks after injection, however, a large number of mononuclear cells appeared around the ␤-gal-positive fibers, and the cellular infiltration was more apparent in AAV vector-injected mdx muscle than in untreated mdx muscle. Further, X-gal staining in muscle fibers surrounded by mononuclear cells in mdx muscle tended to be intense both in the cytoplasm and at the sarcolemma (Figure 2b, 4 weeks), suggesting that these fibers were in the process of degeneration. We next immunostained AAV vector-injected mdx muscle using an antibody against mouse IgG (Figure 2, column IgG) because IgG uptake by degenerated fibers has been reported in dystrophic muscles.29 In mdx muscle 4 weeks after AAV vector injection, mouse IgG was detected weakly within some fibers, which corresponded Gene Therapy

AAV vector-mediated gene transfer into mdx muscle K Yuasa et al

1578

Figure 1 Expression of ␤-galactosidase (␤-gal) after AAV vector-mediated gene transfer into skeletal muscle of dystrophin-deficient mdx mouse. AAVCMVLacZ was injected into tibialis anterior (TA) muscles of C57BL/10 (B10) and mdx mice at 5 (a, b) and 12 weeks old (c). (a) Representative Xgal staining patterns of AAV vector-injected muscles 2 and 4 weeks after injection are shown (whole cross-section). Bar = 1 mm. (b, c) ␤-gal activities in AAV vector-injected B10 (white column) and mdx (black column) TA muscles 1, 2 and 4 weeks after injection were quantified by ONPG assay. Significant differences (∗P ⬍ 0.05; ∗∗P ⬍ 0.01; ∗∗∗P ⬍ 0.001) between B10 and mdx muscles were obtained on all occasions, except 2 weeks after injection into 12-week-old mice. Results are given as means and vertical bars show standard errors (s.e.). The number of mice examined is shown under each column bar.

to ␤-gal-positive fibers with abnormal X-gal staining. In addition, we detected intense staining of endogenous IgG in the extracellular space. In contrast, no fiber showed deposits of IgG in AAV vector-treated B10 muscle. These results indicated that cellular infiltration around ␤-galpositive fibers in mdx muscle 4 weeks after injection with AAV-CMVLacZ accelerated degeneration of these muscle fibers. Enhanced immune response in AAV vector-injected mdx skeletal muscle We then analyzed the cell markers of infiltrating cells in AAV vector-injected mdx muscle (Figure 3). CD11b-positive cells, such as macrophages, were detected in noninjected mdx muscle, but CD4- or CD8-positive T lymphocytes were not. CD4- and CD8-positive cells began to appear 2 weeks after injection, and the number of these Gene Therapy

cells increased gradually (data not shown). CD4- or CD8positive cells were mainly detected in clusters of mononuclear cells around ␤-gal-positive muscle fibers. The number of CD11b-positive cells also increased. Major histocompatibility complex (MHC) class I molecules H-2Db (Figure 3) and H-2Kb (data not shown) were also upregulated in both infiltration cells and AAV-injected muscle fibers of mdx mice. These immuno-competent cells were scarcely detected in AAV vector-injected B10 muscles (data not shown). These observations indicated that injection with the AAV-CMVLacZ vector gradually activated the immune system from 2 weeks after injection in mdx skeletal muscle. We next measured antibodies against the transgene product, ␤-gal, or the AAV particle in the sera of AAV vector-injected mdx mice (Figure 4). Anti-AAV IgG levels in sera of B10 and mdx mice were nearly identical, implying that both mdx and B10 mice

AAV vector-mediated gene transfer into mdx muscle K Yuasa et al

1579

Figure 2 Enhanced cellular infiltration and degeneration in AAV vector-injected mdx muscle. AAV-CMVLacZ was injected into TA muscles of female B10 and mdx mice at 5 weeks (left, H&E staining; middle, X-gal staining; right, anti-mouse IgG immunostaining). Inset in IgG column shows the non-injected muscle on the non-injected side. Bar = 200 ␮m. Male mice showed the same results as female mice (data not shown).

immediately developed humoral immunity against viral particles. Interestingly, the levels of IgG reacting to ␤-gal in mdx mice were higher than that in B10 mice. We also noticed that anti-␤-gal IgG in the sera of female mice tended to be higher than in the sera of male mice, at least in the mdx strain. Although IgG antibodies specific for ␤gal or AAV vector particles were detected in the sera of AAV-CMVLacZ-treated mdx mice, mouse IgG was detected at higher levels in the AAV vector-injected muscle than in the non-injected side (Figures 2b and 4). To investigate the specificity of the IgG antibody detected in AAV vector-injected mdx muscle, a binding assay was carried out using Cy3-labeled ␤-gal (Cy3-␤-gal), AAV vector particle (Cy3-AAV), and mouse albumin (Cy3albumin). Cy3-␤-gal bound strongly to the extracellular space and the cytoplasm of some muscle fibers in AAV vector-injected mdx muscle, where deposits of endogenous IgG were detected (Figure 5a). In contrast, no binding of Cy3-␤-gal was detected in control mdx muscle, and weak binding was detected in non-injected mdx muscle. To verify the specificity, we performed a competitive inhibition experiment with un-labeled ␤-gal and a pre-

adsorption assay of Cy3-␤-gal with anti-␤-gal antibodies derived from the sera of the mice. Competition for Cy3␤-gal showed that inhibition was dependent on the amount of excess of un-labeled ␤-gal (Figure 5b). The binding was inhibited when Cy3-␤-gal was pre-adsorbed with the serum of AAV-CMVLacZ-treated mdx mouse, but not with the serum of non-treated mdx mouse (Figure 5c). On the other hand, in binding assay using Cy3-AAV, no clear binding to the extracellular space or the cytoplasm was detected in AAV vector-injected and noninjected muscles of AAV vector-treated mdx mouse (Figure 5a). Binding of Cy3-AAV to the nucleus was not inhibited by addition of an excess of un-labeled AAV vector particle and pre-adsorption with the serum containing anti-AAV antibody, showing that the signals were nonspecific (data not shown). Furthermore, binding of Cy3albumin was not detected in the extracellular space or the cytoplasm in any muscle sample tested (Figure 5a). Thus, IgG antibody against ␤-gal accumulated significantly in AAV vector-injected muscle of AAV-CMVLacZtreated mdx mice. Then, to further confirm the increase of anti-␤-gal IgG in AAV vector-injected mdx muscle, Gene Therapy

AAV vector-mediated gene transfer into mdx muscle K Yuasa et al

1580

Figure 3 Infiltrating cells in AAV vector-injected mdx skeletal muscle at 4 weeks after injection. AAV-CMVLacZ was injected into TA muscles of 5week-old female mice. In non-injected mdx muscle, CD11b-positive cells were predominant. In the AAV vector-injected mdx muscle, numerous CD4and CD8-positive cells (arrows) were detected in clusters of infiltrating cells around ␤-gal-expressing fibers. Expression of MHC class I H-2Db was highly up-regulated both in mononuclear cells (arrows) and in muscle fibers (asterisks). Bar = 100 ␮m.

Figure 4 AAV vector-mediated gene transfer elicits more severe humoral immune response against the transgene products in mdx mice than in B10 mice. AAV-CMVLacZ was injected into TA muscles of 5-week-old mice. Sera were analyzed for the presence of antibodies against ␤-gal (a) or AAV particle (b) 4 weeks after injection, using ELISA.

Gene Therapy

AAV vector-mediated gene transfer into mdx muscle K Yuasa et al

1581

Figure 5 Deposition of mouse IgG and antibodies against ␤-gal, AAV particles or albumin in muscle. (a) Cryosections of mdx muscle from non-treated mice and injected and non-injected sides of AAV vector-treated mice were incubated with FITC-labeled anti-mouse IgG (IgG), Cy3-labeled ␤-gal (Cy3␤-gal), Cy3-labeled AAV particle (Cy3-AAV) or Cy3-labeled mouse albumin (Cy3-Albumin). (b, c) Binding specificity of Cy3-␤-gal to extracellular spaces was examined using competition assay (b) or pre-adsorption assay (c). In (b), cryosections were incubated with Cy3-␤-gal in the presence of 0, 1, 10, or 100-fold excess of un-labeled ␤-gal. In (c), Cy3-labeled ␤-gal was pre-adsorbed with serum from either non-treated or AAV vector-treated mdx mouse before incubation with cryosections. Bar = 200 ␮m. (d) Western blot analysis for anti-␤-gal IgG in muscle tissue is shown. Purified ␤-gal was separated on 10% SDS-PAGE gel and transferred on to a membrane. Membrane strips were incubated with the muscle extract, and then binding of mouse IgGs to ␤-gal was detected by anti-mouse IgG antibodies. CBB staining of purified ␤-gal (lane 1), Western blot detection of extracts from control mdx muscle (lane 2), non-injected muscle (lane 3), injected muscle (lane 4) from the AAV vector-treated mdx mouse, and a negative control without muscle extracts (Lane 5). The arrow indicates ␤-gal.

Western blot analysis was performed using muscle extracts (Figure 5d). Commercially available ␤-gal was separated on SDS-PAGE gel and blotted on to a PVDF membrane. Consistent with the results of the ELISA, IgG antibodies bound to ␤-gal were detected in both AAV vector-injected and non-injected muscles of AAV vectortreated mdx mouse, but not in untreated mdx muscle. Anti-␤-gal IgG was more abundant in AAV vectorinjected muscle than in non-injected mdx muscles of AAV vector-treated mdx mouse, showing selective accumulation of antibodies specific for ␤-gal in the injected mdx muscle.

Activation of immune response in mini-dystrophintransgenic mdx mice To determine whether the dystrophic phenotype influences the extent of immune responses after AAV vectormediated gene transfer, we injected AAV-CMVLacZ vector into skeletal muscles of transgenic mdx mice carrying the mini-dystrophin gene (CVBA3’ mice). In skeletal muscles of CVBA3’ mice, the dystrophic phenotype was greatly ameliorated by the supplement of mini-dystrophin, and there is almost no muscle degeneration, which in centrally nucleated fibers are less than 1% at 5–9 weeks of age26 (also see Figure 6d, CVBA3’). When AAVGene Therapy

AAV vector-mediated gene transfer into mdx muscle K Yuasa et al

1582

Figure 6 AAV vector-mediated gene transfer into mini-dystrophin-transgenic mdx (CVBA3’) muscles. AAV-CMVLacZ was injected into the TA muscles of 5-week-old B10, mdx, and CVBA3’ mice. (a) ␤-gal-positive fibers were counted using whole cross-sections of AAV vector-injected B10 (white column), CVBA3’ (gray column), and mdx (black column) muscles 1, 2 and 4 weeks after injection. Results are given as means ± s.e. The number of mice examined is shown under each column. (b) Cellular infiltration (H&E), low ␤-gal expression (X-Gal), deposits of IgG (IgG), and deposits of antibodies against ␤-gal (Cy3-␤-gal) are noticeable in AAV vector-injected CVBA3’ muscles 4 weeks after injection. Inset shows the muscle on the noninjected side. (c) Immunohistochemistry for macrophages and dendritic cells (DCs) was done in skeletal muscles of B10, mdx, and CVBA3’ mice. CD11b, CD11c and CD205 were used as markers of macrophages (CD11b+/CD11c+) and DCs (CD11c+/CD205+). (d) Sarcolemmal integrity of CVBA3’ muscle (top, H&E staining; bottom, anti-mouse IgG immunostaining). (b, d, bar = 100 ␮m; c, bar = 50 ␮m).

CMVLacZ was introduced into TA muscles of 5-week-old CVBA3’ mice, high levels of the ␤-gal expression were observed as in B10 mice 1 and 2 weeks after injection (Figure 6a). Four weeks later, however, some mice showed a great loss of ␤-gal-positive fibers: the percentGene Therapy

age of positive fibers greatly varied among mice (6.4%, 9.7%, 13.3%, 19.7%, 60.1%, 77.9% and 83.5%). In AAV vector-injected CVBA3’ muscles with low expression of ␤gal, remarkable cellular infiltration, including CD4- and CD8-positive cells (data not shown), was observed

AAV vector-mediated gene transfer into mdx muscle K Yuasa et al

around the ␤-gal-positive fibers, which were stained abnormally with X-gal. Abundant IgG antibodies against ␤-gal were also detected in the extracellular space and within some ␤-gal-positive fibers (Figure 6b). In contrast, no or little cellular infiltration and negligible anti-␤-gal antibodies were detected in AAV vector-injected CVBA3’ muscles with high ␤-gal expression (data not shown). Thus, AAV vector elicited immune responses against the transgene product, even in the nearly normal muscles of CVBA3’ mice. It has been reported that immature dendritic cells (DCs) infected with AAV vector play an important role in activation of cytotoxic T lymphocytes (CTLs), leading to elimination of transduced cells.30 To assess the possibility that antigen-presenting cells (APCs) are involved in activation of immune responses against the transgene product, we detected macrophages (CD11b+/CD11c+) or DCs (CD11c+/CD205+) by immunohistochemistry to characterize the infiltrates in native skeletal muscles of mdx and CVBA3’ mice (Figure 6c). A large number of CD11b- and CD11c-positive cells were detected within and around degenerating fibers in mdx muscle, but CD11b- and CD11c-positive cells were much fewer in CVBA3’ muscle. In addition, strong staining for CD205 was detected in the T cell areas of the spleen and cortical epithelium of the thymus,31 but not in skeletal muscles of mdx and CVBA3’ mice. These observations indicate that CVBA3’ muscle is devoid of DCs at the time of the AAV vector injection. Dystrophin forms a complex with dystrophin-associated proteins (DAPs) linking the cytoskeleton to the extracellular matrix at the inner surface of the sarcolemma to protect the membrane from stress. The absence of dystrophin increases membrane permeability.26 One activator of the immune response in mdx and CVBA3’ mice, but not in wild-type B10 mice, might be the release of cellular proteins into the extracellular space, so we investigated the sarcolemmal integrity of mdx and CVBA3’ muscles using anti-mouse IgG staining. Depositions of IgG within fibers were observed in apparently necrotic fibers with phagocytosis in mdx and CVBA3’ muscles (Figure 6d). Furthermore, slight IgG uptake into some fibers was detected in both mdx and CVBA3’ muscles although these muscle fibers often lack signs of degeneration on H&Estained sections (Figure 6d). These observations indicate that membrane integrity of CVBA3’ muscle is not as complete as that of normal B10 muscle, supporting our hypothesis that release of cellular proteins might stimulate the humoral and cellular immune responses against the transgene products. Effects of immunosuppression of AAV vector-injected mdx mice using anti-CD4 antibody To test whether immunosuppression prolongs the expression of the transgene in AAV vector-mediated gene transfer into mdx skeletal muscle, we administered antibodies against CD4 (GK1.5) intraperitoneally to 5week-old mdx mice on day ⫺2, ⫺1, 0 and +1 of AAVCMVLacZ vector injection and then weekly until they were killed. ␤-Gal activities in anti-CD4 antibody-treated mdx muscles remained high 4 weeks after injection (Figure 7a). Furthermore, anti-CD4 treatment inhibited both deposition of IgG in the extracellular space and uptake into muscle fibers in AAV vector-injected mdx muscle (Figure 7b). These results indicate that immune

response markedly reduced transgene expression in AAV vector-mediated gene transfer into mdx skeletal muscle. Skeletal muscle in 5-week-old mdx mouse shows vigorous degeneration and regeneration. This might lead to poor expression of the transgene in AAV-injected mdx muscle because AAV-transduced cells can be replaced by regenerating fibers at a high rate. We also noticed that anti-CD4-treated mdx muscles showed milder dystrophic changes than non-treated age-matched mdx mice (data not shown).

1583

Gene transfer with muscle-specific MCK promoter into mdx skeletal muscle To clarify the mechanism of humoral and cellular immune responses in mdx mice, we next investigated whether ␤-gal expression driven by a muscle-specific MCK promoter could suppress the immune response or not. The depressed expression of highly immunogenic ␤gal in potent antigen-presenting cells, such as dendritic cells or macrophages, might greatly attenuate the immune response. Expression of ␤-gal continued to increase at least for 4 weeks in AAV-MCKLacZ-injected mdx muscles (Figure 8a), but it significantly decreased 8 weeks after injection. Four weeks after injection, cellular infiltration and deposition of IgG were not detected in AAV-MCKLacZ-injected mdx muscle (Figure 8b), whereas it was obviously detected in AAV-CMVLacZinjected mdx muscles.

Discussion The AAV vector is a powerful tool to introduce therapeutic genes into non-dividing cells, such as skeletal muscle fibers. It evokes little immune reaction and enables long-term expression of the transferred genes, although full-length dystrophin cDNA is too large to be accommodated in an AAV vector. Several micro-dystrophins, in which the rod domain is largely deleted, introduced into mdx mice,27 were successfully expressed at the sarcolemma together with the recovery of the dystrophin–glycoprotein complex that protects muscle fibers from degeneration.28 In order to establish an AAV vectormediated gene therapy for DMD, we investigated the effect of AAV vector-mediated gene transfer into dystrophic skeletal muscles from mdx mice. When an AAV vector (AAV-CMVLacZ) encoding the LacZ gene driven by CMV promoter was introduced into skeletal muscle, earlier transduction efficiency was significantly lower in mdx muscle than in control C57BL/10 muscle. Poor transduction efficiency in mdx muscle in the early stage after AAV vector injection could be due to elimination of AAV vector-transduced cells by the spontaneous degeneration–regeneration cycles that are characteristic of young mdx mice. Interestingly, continuous immunosuppression by anti-CD4 antibody in mdx mice resulted in efficient transduction of the transferred gene 2 weeks, as well as 4 weeks after injection of the AAV vector (Figure 7). This result leads us to the hypothesis that immunosuppression could limit the degeneration by inhibiting immunocompetent cells that attack dystrophic muscle fibers. In fact, prednisone therapy has been confirmed to be of value in enhancing muscle strength and function in DMD.32 The anti-inflammatory and immunosuppressive effect of the steroid suppresses the progression of muscular dystrophy. In this respect, Gene Therapy

AAV vector-mediated gene transfer into mdx muscle K Yuasa et al

1584

Figure 7 Effect of immunosuppression with anti-CD4 antibody on ␤-gal expression in AAV vector-injected mdx muscle. AAV-CMVLacZ was injected into TA muscles of 5-week-old mdx mice and a monoclonal anti-CD4 antibody (GK1.5) was administered as described in Materials and methods. (a) ␤-gal activity in AAV-CMVLacZ-injected muscles of B10 mice (white column), anti-CD4-treated mdx mice (gray column) and non-treated mdx mice (black column) at 2 and 4 weeks after injection. Results are given as means ± s.e. The number of mice examined is shown under each column. (b) ␤gal expression and humoral immune response in anti-CD4-treated and non-treated mdx muscles 4 weeks after injection. (top, X-gal staining, bar = 1 mm; bottom, anti-mouse IgG immunostaining, bar = 200 ␮m). Inset shows the muscle on the non-injected side.

transient immunosuppression could be useful not only for efficient transduction of the transgene in AAV vectormediated gene transfer into dystrophic skeletal muscle, but also for prevention of progressive muscular degeneration itself. The minimal immune response that the AAV vector elicits in the host enables long-term expression of the transferred gene. We observed stable long-term expression of the LacZ gene products in normal B10 muscle. These muscles showed almost no signs of immune reactions against ␤-gal. However, enhanced immune responses were observed in mdx skeletal muscles after AAV-CMVLacZ-mediated gene transfer. We detected a large number of CD4- or CD8-positive T lymphocytes and up-regulation of MHC class I in mononuclear cells surrounding ␤-gal-expressing muscle fibers 4 weeks after injection. At this stage, ␤-gal expression remarkably decreased, and the remaining ␤-gal-positive muscle fibers showed signs of ongoing degeneration, implying that antigen-specific MHC class I-restricted CTLs play a major role in elimination of ␤-gal-expressing muscle fibers. In addition, we detected high levels of antibodies specific for ␤-gal in the sera of AAV vector-treated mdx mice and dense deposits of IgG including antibodies specific to ␤gal in muscle tissues. The results of immunosupression in later stages by anti-CD4 antibody treatment also support the hypothesis that immune response has been concerned in the AAV-vector mediated gene transfer into mdx skeletal muscle. Several groups have previously demonstrated effective AAV vector-mediated gene transfer into skeletal muscles of mdx mice,28 ␦-sarcoglycandeficient hamsters (Bio 14.6)21,22 and ␥-sarcoglycandeficient mice.23 On the other hand, Cordier et al24 have shown that an AAV vector driven by CMV promoter elicits strong cellular and humoral immune responses to the transgene product after intramuscular injection into ␥sarcoglycan-null mice. Infection of potent antigen-presenting cells (APCs), such as immature dendritic cells, by AAV vector with a ubiquitous promoter results in production of neo-antigens in APCs. Infected APCs would in turn activate CTLs to a great extent and eliminate the transduced muscle Gene Therapy

cells.24,30 Indeed, the number of DCs and macrophages dramatically increased in degenerating and regenerating skeletal muscle after artificial crush injury.33 It has been also reported that mdx muscles contained 20 times more macrophage and seven times more DCs than normal muscles.34 We could not detect CD205-positive dendritic cells in mdx muscle, but the decrease in ␤-gal expression was considerably delayed, when AAV-MCKLacZ was introduced into mdx muscle (Figure 8). This result indicates that APCs play a critical role in developing immunity against the AAV vector infection. Importantly, enhanced immune responses were observed in later stages of AAV-MCKLacZ-injected mdx muscles, as well as AAV-CMVLacZ-injected muscles. This finding could be explained by neo-antigens that were continuously released from dystrophin-deficient myofibers due to spontaneous degeneration. However, there is another possibility that myotubes and immature myofibers presented ␤-gal as a neo-antigen to immuno-competent cells. Indeed, myogenic cell, such as myoblast, myotubes and immature myofibers have been reported to function as local APCs.35 Therefore, myoblasts and regenerating myofibers in dystrophic skeletal muscle might also considerably enhance the immune responses in mdx muscle. CVBA3’ is a transgenic mdx mouse line carrying the mini-dystrophin gene isolated from a mild Becker MD patient that presents very mild symptoms of muscular dystrophy.25,26 In fact, CVBA3’ skeletal muscle showed only a few degenerating fibers and infiltrating cells. When the AAV-CMVLacZ vector was introduced into CVBA3’ mice at the age of 5 weeks, the expression level of the transferred gene products was as high as in normal mice until 2 weeks after injection, without overt sign of immune responses. We noticed, however, a reduction of ␤-gal expression 4 weeks after AAV-CMVLacZ injection, together with the enhanced cellular infiltration and extensive deposits of antibodies specific for the transgene product. Why were the immune responses activated in CVBA3’ muscle in spite of their ameliorated phenotype? Cordier et al24 suggested that a minimum number of DCs would be required to capture the AAV vector and/or its transduced gene products in order to elicit T cell immun-

AAV vector-mediated gene transfer into mdx muscle K Yuasa et al

1585

Figure 8 ␤-gal expression in mdx skeletal muscle after injection with an AAV vector driven by muscle-specific MCK promoter. (a) ␤-gal activity in TA muscles of mdx mice at 1, 2, 4 and 8 weeks after injection with AAV-CMVLacZ (white column) and AAV-MCKLacZ (black column). The mice were injected with AAV vectors at 5 weeks of age. Results are given as means ± s.e. The number of mice examined is shown under each column. (b) Cellular infiltration (H&E), ␤-gal expression (X-Gal) and deposits of IgG (IgG) in AAV-MCKLacZ-injected mdx muscles and AAV-CMVLacZ-injected mdx muscles 4 weeks or 8 weeks after injection. Inset shows the muscle on the non-injected side. Bar = 100 ␮m.

ity. Therefore, the number of inflammatory cells present at the time of the AAV injection would be critical in mobilizing sufficient APCs.24,30 We, however, observed very little infiltration of macrophages in CVBA3’ muscle, and immunostaining with CD205 detected no potent DCs. In addition, proliferating myoblasts and regenerating myofibers were rarely observed in CVBA3’ muscles. These observations suggested that immune responses in CVBA3’ muscle would not be directly related to the transgene expression in resident APCs including myogenic cells. A remaining possibility is that impaired membrane integrity of CVBA3’ led to immune reaction against AAV vector-infected muscle fibers. We saw some fibers with high membrane permeability in CVBA3’ muscle, although membrane integrity of the muscle was greatly improved over the original mdx muscle. Intracellular proteins could be leaking from dystrophic fibers into the extracellular space, as indicated by the increased serum

creatine kinase level in mdx and CVBA3’ mice.26 The transgene product released from AAV vector-transduced muscle cells into circulation might be trapped by APCs and stimulate both cellular and humoral immune systems. In mdx muscles, neo-antigens on APCs, in particular immature myofibers themselves, might stimulate CD4+ helper or CD8+ cytotoxic T cells. We need to further clarify the molecular mechanisms which underlie the unexpected immune response observed in CVBA3’ muscle. We clearly demonstrated that immune suppression improved the efficiency of the gene expression after AAV vector-mediated gene transfer, and our results suggest the importance of transient immunosuppression before the transgene product accumulates in the affected muscle. Nevertheless, we should independently consider how to minimize immune response for future approaches to DMD treatment. First of all, a muscle-specific promoter Gene Therapy

AAV vector-mediated gene transfer into mdx muscle K Yuasa et al

1586

should be used to suppress the transgene expression in AAV vector-transduced APCs, as suggested by this paper and others.24,34 Further, a fully functional micro-dystrophin would effectively protect muscle fibers from degeneration and reduce the release of neo-antigens or self-antigens into the extracellular space. A truncated micro-utrophin may be an alternative molecule for gene therapy of DMD. Utrophin has been shown to ameliorate dystrophic phenotypes when expressed exogenously15 or endogenously36 in dystrophin-deficient muscle. We may avoid undesirable activation of immune response by use of micro-utrophin instead of micro-dystrophin because the utrophin gene is intact in DMD patients.

Materials and methods Production of AAV vectors We used the AAV type2 vector named AAV-CMVLacZ37 and AAV-MCKLacZ, which contain the LacZ gene under the control of the cytomegalovirus (CMV) promoter and the truncated muscle creatine kinase (MCK) promoter, respectively. The truncated MCK promoter was shown to drive muscle-specific expression of the reporter gene38 and constructed by PCR from the mouse genomic DNA. Briefly, 206 bp of the enhancer region (from ⫺1256 to ⫺1050 relative to the transcriptional start site, Accession No. M21390) and 358 bp of the proximal regulatory region (from ⫺358 to +7) were amplified separately by PCR reactions, subcloned into pCR2.1 (Invitrogen, Groningen, The Netherlands), sequenced and connected directly, based on the report described.38 The recombinant viral vector stocks were produced by cotransfection methods as already described.39 The AAV virions were purified twice by CsCl density gradient centrifugation, dialyzed against sterile phosphate-buffered saline (PBS) and titrated by a quantitative DNA dot blot assay, according to the previously published protocol.16,39 Titers were routinely in the range of 1 × 1013 to 2 × 1013 viral genomes per ml. Administration of AAV vectors to murine skeletal muscles Normal C57BL/10 (B10), C57BL/10 mdx and mini-dystrophin-transgenic mdx (CVBA3’)26 mice were injected with AAV-CMVLacZ or AAV-MCKLacZ vectors (5–8 × 1011 viral genomes) in 50 ␮l of PBS into the right tibialis anterior (TA) muscles at 5 or 12 weeks of age. To suppress the immune response, we treated other AAVCMVLacZ-injected mdx mice with rat monoclonal antibody against mouse CD4 (GK1.5). The anti-CD4 antibody was administered intraperitoneally at a dose of 100 ␮g per animal on days ⫺2, ⫺1, 0 and +1 at AAV-CMVLacZ injection, and then weekly until they were killed. Histological analysis The AAV vector-injected mdx and normal B10 mice were killed by cervical dislocation 1, 2, 4 and 8 weeks after the AAV vector injection. The AAV vector-injected and contralateral non-injected TA muscles were isolated and frozen in liquid nitrogen-cooled isopentane. Cryosections (10 ␮m) from a third of the muscles were stained with hematoxylin/eosin (H&E). They were also stained with 5-bromo-4-chloro-3-indolyl-␤-D-galactopyranoside (Xgal; Wako Pure Chemical Industries, Osaka, Japan), as Gene Therapy

described elsewhere.40 Photographs were taken with a DAS Mikroskop LEITZ DMRB (Leica, Wetzlar, Germany), using a digital still camera system HC-2500 (Fujifilm, Tokyo, Japan). We counted the number of ␤galactosidase (␤-gal)-positive fibers in whole crosssections of the AAV vector-injected TA muscles of at least three mice at each time point. Measurement of ␤-galactosidase activity A cell extract was prepared from frozen muscle and used to measure ␤-gal activity by ONPG assay, as described elsewhere,40 with some modifications. Approximately 40 cryosections (each 10 ␮m thick) were suspended in 100 ␮l of sample buffer (250 mM Tris-HCl (pH 7.4), 2 mM MgCl2, 0.1 % Triton X-100) and then subjected to three cycles of freezing and thawing. After centrifugation at 18 000 g for 10 min, the supernatant was collected as the cell extract. For normalization of activity, we estimated the protein concentrations using the Bradford Coomassie Brilliant Blue binding assay (BioRad, Hercules, CA, USA). The cell extract was assayed in 1.0 ml of reaction buffer (60 mM Na2HPO4, 40 mM NaH2PO4, 10 mM KCl, 1 mM MgCl2) containing 0.8 mg/ml ONPG (Sigma, St Louis, MO, USA) and 4 ␮l/ml ␤-mercaptoethanol. The solutions were then incubated at 37°C until they turned yellow, the reactions were halted by adding 0.5 ml of 1 M Na2CO3, and their optical densities were measured at 420 nm. The results were expressed as ␤-gal units, in comparison with a standard enzymatic plot obtained from purified E. coli ␤-galactosidase (Roche Diagnostic, Mannheim, Germany). Immunohistochemistry For immunohistochemical study, we prepared transverse cryosections (6 ␮m) from AAV vector-injected and noninjected TA muscles. Sections from each muscle were placed on the same slide, then dried and fixed in acetone for 10 minutes at ⫺20°C. We used the following antibodies: biotin-conjugated rat monoclonal antibodies against mouse CD4 (L3T4; clone H129.19, PharMingen, San Diego, CA, USA) or against mouse CD8a (Ly-2; clone 53-6.7, PharMingen); rat monoclonal antibodies against mouse CD11b (Mac-1, Cedarlane, Ontario, Canada) and against mouse CD205 (DEC-205; clone NLDC-145, Serotec, Raleigh, NC, USA); hamster monoclonal antibody against mouse CD11c (clone N418), biotin-conjugated mouse monoclonal antibodies against mouse major histocompatibility complex (MHC) class I alloantigens H-2Db (clone 28-14-8, PharMingen) or H-2Kb (clone AF6-88.5, PharMingen); a fluorescein-conjugated F(ab’)2 fragment of goat antibody against mouse IgG (H&L) (Leinco Technologies, St Louis, MO, USA). The primary antibodies against CD11b, CD11c and CD205 were detected using biotinylated rabbit anti-rat immunoglobulins (Dako, Glostrup, Denmark) or biotinylated goat anti-hamster IgG (Cedarlane). The biotin molecules were detected using the avidin-biotinylated peroxidase supplied in the VECTASTAIN ABC kit (Vector Laboratories, Burlingame, CA, USA). The immune complex labeled with horseradish peroxidase was visualized with diaminobenzidine (DAB) (Vector Laboratories), then counterstained with methyl green, and photographed with a DAS Mikroskop Leitz DMRB using a digital still camera system HC-2500. The signals from fluorescein-conjugated antibody against mouse IgG were recorded photographically under the

AAV vector-mediated gene transfer into mdx muscle K Yuasa et al

same conditions using the confocal laser scanning microscope system FLUOVIEW (Olympus Optical, Tokyo, Japan). Measurement of humoral immune response The titer of the antibody to the transgene product (␤-gal) or to the AAV vector particle was measured by ELISA. A SUMILON ELISA plate H (Sumitomo Bakelite, Tokyo, Japan) was coated overnight at 4°C with ␤-gal (0.2 ␮g/well; Roche Diagnostic) or the AAV vector particle (2 × 109 genomes/well) in 100 ␮l of PBS. The microtiter plate was washed and blocked overnight at 4°C with blocking buffer (Block Ace; Dainippon Pharmaceutical, Osaka, Japan). After washing, serial dilutions of mouse serum were applied to the wells and incubated 2 h at room temperature. Reactivity to ␤-gal and the AAV vector particle was determined by incubation with a 1/5000 dilution of horseradish peroxidase-conjugated goat antimouse IgG (H+L) (BioRad, Hercules, CA, USA) for 1 h at room temperature, followed by an extensive wash with PBS-T (0.05% Tween 20 in PBS). The Peroxidase Substrate System T (Sumitomo Bakelite) consisting of 3, 3’, 5, 5’tetramethylbenzidine and H2O2 was used for the color reaction and absorbance was read at 450 nm. Binding assay using the fluorescence-conjugated protein One milligram of ␤-gal (Roche), 0.11 mg of recombinant AAV-CMVLacZ viron (5.2 × 1012 genomes), and 1 mg of mouse albumin (Sigma) were labeled with Cy3 fluorescence dye, using a FluoroLink-Ab Cy3 labeling kit (Amersham Pharmacia Biotech, Buckinghamshire, UK) according to the supplemental protocol. After the labeling reaction, approximately 1 ml of the labeled ␤-gal and albumin was separated from non-conjugated dye by the supplied column, and stored in PBS containing 0.1% sodium azide at ⫺80°C. One milliliter of the labeled AAV vector viron was separated from non-conjugated dye by extensive dialysis against PBS, and stored at ⫺80°C. To detect the antibody against each antigen, the muscle section was incubated with Cy3-labeled proteins. Briefly, transverse cryosections (6 ␮m) from AAV vector-injected and non-injected TA muscles were fixed for 2 h at 4°C in 4% paraformaldehyde. The sections were blocked with 5% goat serum in PBS, and then incubated with Cy3-labeled protein (1/200 dilution for ␤-gal and AAV vector, 1/80 for albumin) for 1 h at room temperature. For preadsorption of Cy3-labeled protein, a dilution was incubated for 1 h with an equal volume of serum from either a non-injected or an AAV vector-injected mdx mouse (4 weeks after injection), before reaction with the cryosections. After extensive washing, binding of Cy3-labeled protein on muscle sections was detected by a confocal laser scanning microscope system FLUOVIEW (Olympus). Western blot analysis for detection of antibodies in muscle tissue For detection of anti-␤-gal IgG antibody in the muscles of non-injected and AAV vector-injected mice, an alternative Western blot analysis was carried out, as described elsewhere18 with some modifications. To prepare muscle extracts, 100 cryosections (each 10 ␮m thick) of noninjected or AAV vector-injected muscle were suspended in 200 ␮l of PBS with 0.1% Triton X-100 and subjected to three cycles of freezing and thawing. After centrifugation,

the supernatant was removed as the muscle extract and the protein concentrations estimated. Aliquots (5 ␮g/lane) of purified ␤-gal protein (Roche) were separated on 10% SDS-PAGE and electrotransferred on to a PVDF membrane (Immobilon; Millipore, New Bedford, MA, USA). Individual lanes were cut and blocked for 2 h at room temperature with Block Ace (Dainippon Pharmaceutical). The blots were then incubated overnight at 4°C in 30 ␮g/ml of the muscle extracts. IgG antibodies bound to ␤-gal were detected by horseradish peroxidase-conjugated goat anti-mouse IgG (1/2000, BioRad) and ECL Western blotting detection reagents (Amersham Pharmacia Biotech, Uppsala, Sweden). The blots were also stained with Coomassie Brilliant Blue R-250.

1587

Statistical analysis In this report, data are expressed as means ± s.e.m. Student’s t tests were used to evaluate the difference between two groups.

Acknowledgements We are grateful to Dr M Yoshida for providing mdx mice. We also thank colleagues in our laboratory for useful discussion and suggestions on this work. This work is supported by Grants-in-Aid for Scientific Research for Center of Excellence, Research on Nervous and Mental Disorders (10B-1, 13B-1) Health Sciences Research Grants for Research on the Human Genome and Gene Therapy (H10-genome-015, H13-genome-001), for Research on Brain Science (H12-Brain-028) from the Ministry of Heath, Labor and Welfare, Grant-in-Aids for Scientific Research (10557065, 11470153, 11170264 and 14657158) from the Ministry of Education, Culture, Sports, Science and Technology, and a Research Grant from Human Frontier Science Project.

References 1 Emery AEH. Duchenne Muscular Dystrophy 2nd edn. Oxford University Press: Oxford, 1993. 2 Rogot T et al. Efficient adenovirus-mediated transfer of a human minidystrophin gene to skeletal muscle of mdx mice. Nature 1993; 361: 647–650. 3 Acsadi G et al. Dystrophin expression in muscles of mdx mice after adenovirus-mediated in vivo gene transfer. Hum Gene Ther 1996; 7: 129–140. 4 Kochanek S et al. A new adenoviral vector: replacement of all viral coding sequences with 28 kb of DNA independently expressing both full-length dystrophin and ␤-galactosidase. Proc Natl Acad Sci USA 1996; 93: 5731–5736. 5 Tuner G, Duncley MG, Dickson G. Gene therapy of Duchenne muscular dystrophy. In: Brown SC, Lucy JA (eds). Dystrophin: Gene, Protein and Cell Biology. Cambridge University Press: Cambridge, 1997, pp 274–309. 6 Chen H-H et al. DNA from both high-capacity and first generation adenoviral vectors remains intact in skeletal muscle. Hum Gene Ther 1999; 10: 365–373. 7 Acsadi G et al. Human dystrophin expression in mdx mice after intramuscular injection of DNA constructs. Nature 1991; 352: 815–818. 8 Gussoni E, Blau HM, Kunkel LM. The fate of individual myoblasts after transplantation into muscles of DMD patients. Nature Med 1997; 3: 970–977. 9 Gussoni E et al. Dystrophin expression in mdx mouse restored by stem cell transplantation. Nature 1999; 40: 390–394. Gene Therapy

AAV vector-mediated gene transfer into mdx muscle K Yuasa et al

1588

10 Rando TA, Disatnik MH, Zhou LZ. Rescue of dystrophin expression in mdx mouse muscle by RNA/DNA oligonucleotides. Proc Natl Acad Sci USA 2000; 97: 5363–5368. 11 Bartlett RJ et al. In vivo targeted repair of a point mutation in the canine dystrophin gene by a chimeric RNA/DNA oligonucleotide. Nat Biotechnol 2000; 18: 615–622. 12 Dunckley MG et al. Modification of splicing in dystrophin gene in cultured Mdx muscle cells by antisense oligonucleotides. Hum Mol Genet 1998; 7: 1083–1090. 13 Barton-Davis ER et al. Aminoglycoside antibiotics restore dystrophin function to skeletal muscles of mdx mice. J Clin Invest 1999; 104: 375–381. 14 Wagner KR et al. Gentamicin treatment of Duchenne and Becker muscular dystrophy due to nonsense mutations. Ann Neurol 2001; 49: 706–711. 15 Yamamoto K et al. Immune response to adenovirus-delivered antigens upregulates utrophin and results in mitigation of muscle pathology in mdx mice. Hum Gene Ther 2000; 11: 669–680. 16 Xiao X, Li J, Samulski RJ. Efficient long-term gene transfer into muscle tissue of immunocompetent mice by adeno-associated virus vector. J Virol 1996; 70: 8098–8108. 17 Kessler PD et al. Gene delivery to skeletal muscle results in sustained expression and systemic delivery of a therapeutic protein. Proc Natl Acad Sci USA 1996; 93: 14082–14087. 18 Fisher KJ et al. Recombinant adeno-associated virus for muscle directed gene therapy. Nat Med 1997; 3: 306–312. 19 Kaplitt MG et al. Long-term gene expression and phenotypic correction using adeno-associated virus vectors in mammalian brain. Nat Genet 1994; 8: 148–153. 20 Jooss K, Yang Y, Fisher KJ, Wilson JM. Transduction of dendritic cells by DNA viral vectors directs the immune response to transgene products in muscle cells. J Virol 1998; 72: 4212–4223. 21 Greelish JP et al. Stable restoration of the sarcoglycan complex in dystrophic muscle perfused with histamine and a recombinant adeno-associated viral vector. Nat Med 1999; 5: 439–443. 22 Li J et al. rAAV vector-mediated sarcoglycan gene transfer in a hamster model for limb girdle muscular dystrophy. Gene Therapy 1999; 6: 74–82. 23 Cordier L et al. Rescue of skeletal muscles of ␥-sarcoglycandeficient mice with adeno-associated virus-mediated gene transfer. Mol Ther 2000; 1: 119–129. 24 Cordier L et al. Muscle-specific promoters may be necessary for adeno-associated virus-mediated gene transfer in the treatment of muscular dystrophies. Hum Gene Ther 2001; 12: 205–215. 25 England SB et al. Very mild muscular dystrophy associated with the deletion of 46% of dystrophin. Nature 1990; 343: 180–182.

Gene Therapy

26 Phelps SF et al. Expression of full-length and truncated dystrophin mini-gene in transgenic mdx mice. Hum Gene Ther 1995; 4: 1251–1258. 27 Yuasa K et al. Effective restoration of dystrophin-associated proteins in vivo by adenovirus-mediated transfer of truncated dystrophin cDNAs. FEBS Lett 1998; 425: 329–336. 28 Wang B, Li J, Xiao X. Adeno-associated virus vector carrying human minidystrophin genes effectively ameliorates muscular dystrophy in mdx mouse model. Proc Natl Acad Sci USA 2000; 97: 13714–13719. 29 Straub V, Rafael JA, Chamberlain JS, Campbell KP. Animal models for muscular dystrophy show different patterns of sarcolemmal disruption. J Cell Biol 1997; 139: 375–385. 30 Zhang Y, Chirmule N, Gao G-P, Wilson J. CD40 ligand-dependent activation of cytotoxic T lymphocytes by adeno-associated virus vectors in vivo: role of immature dendritic cells. J Virol 2000; 74: 8003–8010. 31 Witmer-Pack MD et al. Tissue distribution of the DEC-205 protein that is detected by the monoclonal antibody NLDC-145. Cell Immunol 1995; 163: 157–162. 32 Khan MA. Corticosteroid therapy in Duchenne muscular dystrophy. J Neurol Sci 1993; 120: 8–14. 33 Pimorady-Esfahani A, Grounds MD, McMenamin PG. Macrophages and dendritic cells in normal and regenerating murine skeletal muscle. Muscle Nerve 1997; 20: 158–166. 34 Hartigan-O’Connor D et al. Immune evasion by muscle-specific gene expression in dystrophic muscle. Mol Ther 2001; 4: 525–533. 35 Hohlfeld R, Engel AG. The immunobiology of muscle. Immunol Today 1994; 15: 269–274. 36 Tinsley JM et al. Amelioration of the dystrophic phenotype of mdx mice using a truncated utrophin transgene. Nature 1996; 28: 349–353. 37 Xiao X, Li J, Samulski RJ. Production of high-titer recombinant adeno-associated virus vectors in the absence of helper adenovirus. J Virol 1998; 72: 2224–2232. 38 Shield MA et al. E-box sites and a proximal regulatory region of the muscle creatine kinase gene differentially regulate expression in diverse skeletal muscles and cardiac muscle of transgenic mice. Mol Cell Biol 1996; 16: 5058–5068. 39 Snyder R, Xiao X, Samulski RJ. Production of recombinant adeno-associated viral vectors. In: Dracopoli N et al (ed.). Current Protocols in Human Genetics. John Wiley: New York, 1996, pp 1–2. 40 Ishii A et al. Effective adenovirus-mediated gene expression in adult murine skeletal muscle. Muscle Nerve 1999; 22: 592–599.