Adenovirus-Mediated Dystrophin Minigene Transfer Improves Muscle Strength in Adult Dystrophic (MDX) Mice

L Yang1, H Lochmuller2,3, J Luo1, B Massie4, J Nalbantoglu2, G Karpati2 and BJ Petrof1 1Respiratory Division, Royal Victoria Hospital, and Meakins-Christie Laboratories, McGill University, Montreal; 2Neuromuscular Research Group, Montreal Neurological Institute, McGill University, Montreal, Quebec; 3Genzentrum and Friedrich-Baur-Institut, Munich, Germany; and 4Biotechnology Research Institute, National Research Council of Canada, Montreal, Quebec, Canada

Duchenne muscular dystrophy (DMD) and murine X-linked logy and muscle weakness. The main findings are as fol- muscular dystrophy (mdx) are both due to absence of the lows: (1) acute myofiber toxicity and gene transfer subsarcolemmal protein dystrophin. Recombinant adeno- efficiency are both AdV dose-dependent, such that the virus vectors (AdV) are considered a promising means for therapeutic margin of safety is fairly narrow; (2) immuno- delivering a functional dystrophin gene to muscle. How- suppressive therapy (FK506) prevents immune-mediated ever, the usefulness of AdV for this purpose is limited by elimination of dystrophin-positive fibers but not the dose- vector toxicity as well as immune-mediated elimination of dependent toxic effects; (3) at the optimal vector dosage infected fibers. In addition, studies to date of AdV-mediated and with effective immunosuppression, AdV-mediated dys- dystrophin gene transfer have either failed to examine trophin minigene transfer is capable of alleviating the loss effects on muscle strength or been performed in immuno- of force-generating capacity as well as histopathological logically immature neonatal animals with little baseline evidence of disease progression normally seen in adult abnormality of force-generating capacity. In the present mdx muscles over a 2-month period. These findings have study, AdV-mediated dystrophin gene transfer was perfor- important implications for the eventual application of AdV- med in adult mdx mice with pre-existent dystrophic patho- mediated dystrophin gene transfer in DMD patients.

Keywords: Duchenne muscular dystrophy; mdx diaphragm; viral toxicity; immunosuppression; muscle function

Introduction a promising means for delivering a functional dystrophin gene to muscle. Advantages of AdV as a vehicle for Patients with Duchenne muscular dystrophy (DMD) gen- therapeutic gene transfer in DMD include the ability to erally succumb to respiratory failure by the third decade infect nonreplicating muscle fibers, little apparent risk of of life due to disease involvement of the diaphragm and oncogenesis and the availability of replication-defective 1 other respiratory muscles. Although pharmacologic variants at high viral titers.8 However, a limitation to treatment with corticosteroids may delay disease pro- AdV-mediated gene transfer for the treatment of DMD is 2 gression, definitive therapy would require replacing the the transient nature of dystrophin expression in immuno- 3 missing gene product, dystrophin. Dystrophin is a large competent experimental animals, which is primarily due cytoskeletal protein that is normally present on the cyto- to a CD8+ cytotoxic T lymphocyte (CTL) response 3 plasmic aspect of the myofiber surface membrane. Cur- directed against the AdV-infected myofibers.9,10 In rent evidence suggests that myofibers lacking dystrophin addition, we have recently reported that AdV-mediated are abnormally susceptible to contraction-induced sarco- transfer of reporter genes (as opposed to the therapeutic 4,5 lemmal damage, which secondarily leads to muscle dystrophin gene) to diaphragm muscle caused significant 6,7 fiber dysfunction, necrosis and eventual replacement of impairment of force generation in normal as well as dys- the lost fibers by adipose and connective tissue. There- trophic (mdx) mice.11 Further studies employing fore, the muscle weakness found in DMD patients is due immunodeficient animal models12 have pointed to at to both myofiber loss and impaired contractile function least two possible etiologies for this reduction in force- 6 in the surviving myofiber population. Gene therapy generating capacity: (1) a nonantigen-specific myofiber approaches in DMD should ideally be capable of revers- toxicity of early onset; and (2) an antigen-specific ing and/or preventing these changes without entailing immune-based component of more delayed onset that is significant adverse effects on muscle function. dependent upon intact CD8+ T cell activity. These find- Recombinant adenovirus vectors (AdV) are considered ings raise the concern that toxic effects and immunologic responses to AdV administration could partially or even completely negate the therapeutic effects of dystrophin Correspondence: BJ Petrof, Royal Victoria Hospital, Room L408, 687 Pine gene replacement. Avenue West, Montreal, Quebec, Canada H3A 1A1 In the present study, we show for the first time that Received 3 July 1997; accepted 11 November 1997 AdV-mediated dystrophin minigene transfer can in fact Adenovirus-mediated dystrophin transfer improves muscle strength L Yang et al 370 alleviate the usual loss of muscle strength and histo- slight trend toward reduced force production in AdV- pathological changes found in dystrophin-deficient mdx Dys-injected diaphragms over the titer range of 1 × 1010 to muscles, provided that effective immunosuppression is 1 × 1011 particles per ml, but this did not reach statistical also instituted in order to prevent CTL-mediated elimin- significance. However, at an AdV-Dys inoculum dosage ation of dystrophin-expressing myofibers. We also dem- of 5 × 1011 particles per ml, there was a significant fall onstrate that the appropriate AdV inoculum dosage lies in maximal force-generating capacity as compared with within a fairly narrow therapeutic window, since the control. Importantly, there was no apparent increase in magnitude of both AdV toxicity and dystrophin muscle inflammation at the highest AdV titer despite the expression are shown to increase in a dose-dependent reduction in force-generating capacity. This is consistent manner. Finally, our data suggest that the degree of func- with the early, direct toxic effect of high-titer AdV infec- tional amelioration attainable in response to dystrophin tion on myofiber function previously noted after AdV- gene replacement may vary with the level of underlying mediated transfer of reporter genes.11,12 Based on these disease progression present at the time of AdV delivery, results, we selected an AdV-Dys titer of 1 × 1011 particles with initially more severely affected muscles potentially per ml as the optimal inoculum dosage representing the being less responsive to this form of therapy. best compromise between high-level dystrophin gene transfer efficiency on the one hand, and minimization of Results cytotoxic effects on mdx myofibers with attendant adverse consequences for contractile function on the other. Optimization of AdV-Dys inoculum dosage We first attempted to identify the ‘optimal’ AdV-Dys inoculum dosage, which was defined as the AdV titer Long-term dystrophin expression in adult mdx mice with achieving the highest level of dystrophin expression immunosuppressive therapy while minimizing AdV-induced toxic effects on mdx Diaphragm and soleus muscles were injected with the muscle function. Accordingly, progressively increasing optimal AdV-Dys titer and studied after 2 months of con- dosages of AdV-Dys particles were injected into separate tinuous immunosuppressive therapy. Under FK506 treat- diaphragms of mdx mice at 7–8 weeks of age, while using ment, there was no evidence of AdV-induced inflam- concomitant immunosuppression with the drug FK50610 mation and no significant loss of dystrophin-positive to eliminate potential confounding immunologic fibers in the diaphragm between 5 and 60 days after AdV responses. Five days after AdV-Dys administration, injection (197 ± 63 and 183 ± 29, respectively; P = 0.82). analysis of the number of dystrophin-positive fibers in The dystrophin-positive fibers were found to be present AdV-injected muscles revealed a direct correlation in a somewhat nonuniform pattern of distribution between AdV-Dys titer and successful myofiber trans- throughout both muscles, as can be seen in Figure 3. duction, as depicted in Figure 1. The total number of dystrophin-positive myofibers at The effect of increasing AdV-Dys titer on maximal tet- 60 days after injection and their relative contribution to anic force production in adult mdx diaphragms is shown total muscle cross-sectional area in diaphragm as well as in Figure 2. As compared with contralateral control hemi- soleus are shown in Table 1. The absolute number of dys- diaphragms that did not receive AdV-Dys, there was a trophin-positive fibers was significantly higher in dia-

Figure 1 Effect of AdV-Dys particle dosage on myofiber transduction efficiency. There was a direct relationship between the vector particle dosage administered to the diaphragm and the number of dystrophin-positive fibers. Data were obtained 5 days after AdV-Dys injection, and represent mean values ± s.e. Vector dosage is expressed in terms of AdV-Dys particle titer, which corresponded to the following values for total particles injected: 1.5 × 109, 7.5 × 109, 1.5 × 1010 and 7.5 × 1010.*PϽ 0.05 versus contralateral control muscle. Adenovirus-mediated dystrophin transfer improves muscle strength L Yang et al 371

Figure 2 Acute myofiber toxicity secondary to AdV-Dys administration. Maximal force generation was examined over the same range of AdV-Dys dosages as shown in Figure 1. At an AdV-Dys inoculum titer of 5 × 1011 particles per ml (total particles injected, 7.5 × 1010), AdV-Dys-injected diaphragms demonstrated a significant decrease in maximal force-generating capacity as compared with control mdx diaphragms. Although there was also a slight trend toward reduced force production in AdV-Dys-injected diaphragms over the titer range of 1 × 1010 to 1 × 1011 particles per ml, this did not reach statistical significance. Data were obtained 5 days after AdV-Dys injection, and represent mean values ± s.e. *P Ͻ 0.05 versus contralateral control muscle. phragm than in the soleus muscle. However, due to the positive and dystrophin-negative myofibers within AdV- greater overall cross-sectional area of the excised dia- Dys-injected soleus muscles demonstrated significantly phragm bundles, the mean contribution of dystrophin- less central nucleation than the contralateral control. Fur- positive fibers to total myofiber cross-sectional area actu- thermore, there was no significant difference in the ally tended to be higher in the soleus (P = 0.06). prevalence of central nucleation between dystrophin- positive and dystrophin-negative myofiber populations Effect of AdV-Dys on dystrophic muscle histopathology within the AdV-Dys-injected soleus muscles. Whereas myonuclei are normally found in a peripheral location in healthy muscle fibers, centrally located nuclei are characteristic of mouse myofibers that have under- Ability of AdV-Dys to prevent further deterioration of gone previous episodes of necrosis and subsequent regen- contractile function in adult mdx muscles eration. Therefore, quantification of the prevalence of cen- Data for mdx soleus and diaphragm twitch kinetics 2 tral nucleation within mdx myofibers provides a useful months after AdV-Dys administration are provided in 13–15 index of the antecedent level of muscle necrosis. Since Table 3. As can be seen, twitch contraction time and half- AdV-Dys was administered to adult mdx animals that relaxation time were both unaffected by AdV-Dys treat- had already undergone cycles of necrosis, central ment in mdx soleus, as well as diaphragm muscles. nucleation was present within a substantial proportion of On the other hand, AdV-Dys administration had a sub- fibers at the time of gene transfer. In this regard, Figure stantial effect on the generation of maximal twitch force 4 demonstrates that both centrally and peripherally by mdx muscles, as shown in Figure 5. Note that the nucleated myofibers were effectively transduced at 5 days baseline (control) force-generating capacity of the dia- after AdV-Dys delivery to adult mdx muscle. phragm was on average 27% lower than in soleus The prevalence of central nucleation within dystro- muscles obtained from the same animals, a finding that phin-positive and dystrophin-negative myofibers at 2 is consistent with the known accelerated progression of months after AdV injection is shown in Table 2. In the 16 AdV-Dys-injected hemidiaphragm, the prevalence of cen- the dystrophic process in the mdx diaphragm. tral nucleation was significantly lower in dystrophin- Importantly, AdV-Dys administration was associated positive as compared with dystrophin-negative fibers, with a mean increase in maximal twitch force of approxi- which strongly suggests that the incidence of necrosis mately 33% in the mdx soleus muscle. In contrast, maxi- had been substantially reduced in dystrophin-expressing mal twitch force production in the mdx diaphragm was fibers over the previous 2-month period. Moreover, there not altered by AdV-Dys injection. Similar results were was no difference in the prevalence of central nucleation obtained for maximal tetanic force production as illus- between dystrophin-negative fibers in AdV-Dys-injected trated in Figure 6, where baseline force production by the hemidiaphragms and contralateral control hemidiaphragm was 49% lower than found in the soleus. Once diaphragms that had not received AdV-Dys. again, a significant improvement in force-generating Table 2 also indicates the effects of AdV-Dys adminis- capacity was observed after AdV-Dys administration for tration on mdx limb muscle histopathology. In contrast the soleus muscle only, which amounted to an increase to results obtained in the diaphragm, both dystrophin- of approximately 14%. Adenovirus-mediated dystrophin transfer improves muscle strength L Yang et al 372

Figure 3 Dystrophin immunohistochemistry of mdx mouse diaphragm (A) and soleus (B) at 2 months after AdV-mediated dystrophin minigene transfer. Animals were immunosuppressed with FK506 throughout this period. Dystrophin-positive myofibers are identified by positive sarcolemmal staining. Note that in some myofibers, dystrophin overexpression is also present as indicated by the presence of cytoplasmic staining. Magnification × 100.

AdV-Dys efficacy in protecting adult mdx muscle from contraction-induced mechanical stress Table 1 Dystrophin expression 60 days after AdV-mediated It was previously demonstrated that dystrophin-deficient dystrophin gene transfer with immunosuppressive therapy myofibers exhibit an abnormally increased susceptibility to contraction-induced injury that is reflected by an accel- Dystrophin-positive fibers erated loss of isometric force-generating capacity following the imposition of high-stress eccentric contractions.5 Total % Contribution to Therefore, we also determined the effects of AdV-Dys number total muscle administration on the vulnerability of mdx soleus cross-sectional area muscles to this form of damage. As depicted in Figure 7, both AdV-Dys and control mdx muscles exhibited ± ± Diaphragm 183 29 11.7 1.3 reductions in force-generating capacity over the course of ± a ± Soleus 106 13 16.6 2.3 the eccentric contraction protocol. However, the absolute magnitude of maximal isometric force production a Ͻ P 0.05, diaphragm versus soleus. remained significantly greater in AdV-Dys than in control soleus muscles throughout the imposition of high- stress eccentric contractions. When expressed as a per- Adenovirus-mediated dystrophin transfer improves muscle strength L Yang et al 373

Figure 4 Dystrophin immunohistochemistry of mdx muscle at 5 days after AdV-mediated dystrophin minigene transfer. Note that a large proportion of the dystrophin-positive fibers are centrally nucleated, indicating that myofibers in adult mdx muscles which have previously undergone necrosis and regeneration are effectively transduced by the vector. Magnification × 200.

leads to profound muscle weakness and eventual respir- Table 2 Prevalence of central nucleation among dystrophin- negative and dystrophin-positive myofibers atory failure. In the present study, we employed recombinant adenoviruses to achieve expression of a truncated Control AdV-Dys human dystrophin minigene in mdx mouse limb and diaphragm muscles. Although previous work in transgenic Dys-Neg Dys-Pos mdx mice indicates that the truncated dystrophin protein is somewhat less effective than its full-length counter- 14,15 Diaphragm 60 ± 1.8% 57 ± 3.0% 27 ± 2.5%a part, normal muscle strength can none the less be Soleus 69 ± 2.0% 53 ± 2.4%a 51 ± 3.7%a attained in mdx mice expressing the dystrophin minigene provided that the uniformity and level of expression are 14 aP Ͻ 0.05 versus control. sufficiently high. The fact that expression of even the truncated dystrophin is able to significantly ameliorate disease progression in transgenic mdx mice14,15 is encour- aging for future attempts at gene therapy in DMD. How- Table 3 Effect of AdV-mediated dystrophin gene transfer on ever, it is important to recognize that the threshold level twitch contraction kinetics of dystrophin replacement needed to achieve a therapeutic effect may differ in the context of AdV-mediated Contraction Half-relaxation gene transfer for a number of reasons. First, direct tox- time (ms) time (ms) icity to myofibers may result from either exposure to AdV capsid proteins17,18 delivered in the initial inoculum Diaphragm or interference with normal cellular function by factors Control 39 ± 138± 2 AdV-Dys 39 ± 139± 3 such as low-grade de novo expression of adenoviral genes.19–21 Second, host immune responses directed Soleus ± ± against AdV-infected cells and attendant elimination of Control 41 2534 transduced myofibers can lead to adverse effects on AdV-Dys 43 ± 351± 4 force-generating capacity.11,12 Third, the use of germ cell correction of the genetic abnormality to prevent disease onset in transgenic animals differs substantially from the anticipated clinical scenario, which would involve treat- centage of its initial value rather than in absolute terms, ing individuals with established disease of varying force production at the end of the protocol did not differ degrees of severity. significantly between the two groups, with the average The results of the present study provide insight into final force being 77 and 79% of its initial value in the these issues. We have demonstrated that both gene trans- AdV-Dys and control muscles, respectively. fer efficiency and myofiber toxicity are titer-dependent. Thus, at the highest AdV-Dys titer examined, a marked Discussion dissociation was observed between two commonly accepted endpoints of therapeutic efficacy, ie the number The primary goal of gene therapy in DMD is to prevent of dystrophin-expressing fibers and maximal force pro- the progressive deterioration of contractile function that duction. As suggested from transgenic mdx mice,14 Adenovirus-mediated dystrophin transfer improves muscle strength L Yang et al 374

Figure 5 Maximal twitch force generation in adult mdx muscles 2 months after AdV-Dys administration. AdV-Dys administration was associated with a signiﬁcant (approximately 33%) increase in maximal twitch force in the mdx soleus muscle. Maximal twitch force production in the control mdx diaphragm was substantially lower than in the soleus, and was not altered by AdV-Dys injection. Data were obtained 60 days after AdV-Dys injection (titer: 1 × 1011 particles per ml), and animals were immunosuppressed with FK506 throughout this period. Values are means ± s.e. *P Ͻ 0.05 versus contralateral control muscle.

Figure 6 Maximal tetanic force generation in adult mdx muscles 2 months after AdV-Dys administration. AdV-Dys administration was associated with a signiﬁcant (approximately 14%) increase in maximal tetanic force in the mdx soleus muscle, whereas force production in the mdx diaphragm was unaffected by AdV-Dys injection. Data were obtained 60 days after AdV-Dys injection (titer: 1 × 1011 particles per ml), and animals were immunosuppressed with FK506 throughout this period. Values are means ± s.e. *P Ͻ 0.05 versus contralateral control muscle.

complete prevention of muscle weakness is dependent both immune and nonimmune origin. In immunocom- upon reaching a certain absolute level of dystrophin gene petent mice receiving intratracheal instillation of AdV, expression distributed in a fairly uniform manner McCoy et al22 found that pulmonary inflammation throughout the muscle. However, our data suggest that occurred in a titer-dependent fashion irrespective of the AdV inoculum dosage required to achieve this level whether the viral genes were functionally intact or inacti- of dystrophin replacement in mature mdx animals may vated by ultraviolet (UV)-irradiation. Similar findings be very close to the threshold for acute AdV-induced have been reported in the brain.23 Such early inflam- myofiber toxicity. Therefore, the effective therapeutic mation is likely to be a response of the innate immune index (ratio of desired therapeutic effects/undesirable system to adenoviral capsid proteins per se, and would toxic effects at a given dose) for AdV-Dys injection of not require de novo expression of either adenoviral genes mdx muscle appears to be relatively low. or the transgene in order to be stimulated. Although we The mechanisms underlying the acute dose-related tox- were able to essentially eliminate this early inflammatory icity observed after AdV administration appear to be of response by employing FK506 immunosuppressive Adenovirus-mediated dystrophin transfer improves muscle strength L Yang et al 375

Figure 7 Susceptibility to mechanical stress after AdV-mediated dystrophin minigene transfer. Imposition of high-stress eccentric contractions led to equivalent relative reductions in force-generating capacity in AdV-Dys-injected and control mdx soleus muscles. However, the absolute magnitude of maximal isometric force production remained significantly greater in AdV-Dys than in control throughout the protocol. Data were obtained 60 days after AdV-Dys injection (titer: 1 × 1011 particles per ml), and animals were immunosuppressed with FK506 throughout this period. Values are means ± s.e. *P Ͻ 0.05 versus contralateral control muscle. therapy in the present study, adverse effects on myofiber cycles of myofiber necrosis, AdV-mediated dystrophin function at high AdV titer were nonetheless observed. minigene transfer is able to significantly ameliorate histo- This may be explained, at least in part, by direct toxic pathological changes, provided that adequate immuno- effects of adenoviral proteins that are synthesized de novo suppressive treatment is also given to allow the mainte- after AdV infection of various target organs, including nance of dystrophin expression. To the extent that the muscle.19–21 Thus, while deletion of the E1 region should prevalence of central nuclei within the mdx soleus mus- prevent major transactivation of adenoviral genes, so- cle at 8 weeks of age (ie the age of vector administration called ‘leaky’ expression of adenoviral proteins can in the present study) has been shown to be approxi- nevertheless occur,19–21 possibly due to the presence of mately 55%,13 our finding of a similar prevalence of cen- cellular E1-like factors.24 This problem is probably most tral nucleation in mdx soleus muscles 2 months after pronounced at a higher multiplicity of infection (MOI). AdV-Dys injection suggests that further episodes of In support of this idea, we have recently reported that myofiber necrosis may have been minimal to absent fol- when administered at high MOI, UV-irradiated AdV are lowing vector delivery. Additionally, an interesting less toxic to cultured myoblasts than transcriptionally observation was that dystrophin gene transfer to the active AdV particles.12 Furthermore, UV-irradiated AdV soleus appeared to exert a protective effect (as determ- also caused less force depression in vivo than transcrip- ined by the prevalence of central nucleation) not only on tionally active AdV administered to muscle at the equiv- those fibers demonstrating detectable dystrophin alent particle dosage.12 expression, but also on neighboring fibers that were dys- To date, AdV-mediated dystrophin gene transfer has trophin-negative in the same muscle section. This may be generally been performed in immunologically immature related at least in part to the fact that so-called dystro- (neonatal) mdx animals during the period that precedes phin-negative fibers could have actually been dystro- the onset of major myofiber necrosis and reduced force- phin-positive in adjacent areas along the length of the generating capacity.25–27 Under these conditions, the myofiber, particularly if the nuclear domains covered by immune response against AdV-infected cells is signifi- the AdV-Dys-containing myonuclei were relatively cantly less pronounced than in adult animals, and the restricted.28 These results are consistent with data longevity of dystrophin gene expression can thus be suf- obtained in certain mdx transgenic lines14,15 as well as ficient to reduce histological evidence of disease pro- female heterozygote mdx mice28 in which the prevalence gression several weeks later26 despite the absence of of muscle necrosis also appears to be quite low despite immunosuppressive therapy. However, in human a mosaic pattern of dystrophin expression. patients suffering from DMD, the diagnosis is usually Importantly, this study also presents the first demon- made at a much later stage due to the development of stration of increased isometric force-generating capacity symptomatic muscle weakness. Therefore, both the after AdV-mediated dystrophin minigene transfer. Poten- immunological context and state of the underlying tial mechanisms underlying the improvement in force muscle disease differ substantially from the situation production include a reduction in the amount of fibrosis, encountered in neonatal animals. Here we show for the as well as an inhibition of myofiber loss. With regard to first time that even in muscles of immunologically the latter consideration in particular, AdV-mediated dys- mature mdx animals that have undergone previous trophin minigene transfer has previously been demon- Adenovirus-mediated dystrophin transfer improves muscle strength L Yang et al 376 strated to be capable of preventing myofiber loss in mdx or significant loss of dystrophin-positive fibers at muscles.9,26 In the present study we have reported data 2 months after AdV injection indicates that FK506 obtained after only 8 weeks of dystrophin gene replace- immunosuppressive therapy was effective in preventing ment in adult animals, but it is likely that differences CD8+ CTL-mediated elimination of transduced myofib- between AdV-Dys-injected and nontreated muscles ers, as we have previously reported.10 It is therefore would become increasingly pronounced over time, unlikely that antigen-specific adaptive immune responses assuming that the level of dystrophin expression were against AdV-infected myofibers9,12 significantly inter- adequately maintained. fered with the therapeutic efficacy of dystrophin gene At 15 to 16 weeks of age, maximal isometric force pro- transfer in this study. Whereas the dystrophin minigene duction by the nontreated mdx diaphragm was already replacement threshold for restoration of essentially nor- reduced to 40–50% of the normal value.11,12,29 Therefore, mal muscle strength in transgenic mdx mice appears to the lack of improved muscle strength in the mdx dia- lie at approximately 20% normal levels,14 successful AdV- phragm after AdV-Dys administration may be indicative mediated dystrophin replacement in approximately 17% of the response to AdV-mediated dystrophin gene of adult mdx myofibers produced less than complete cor- replacement in muscles with a high level of pre-existent rection of force-generating capacity. However, because dystrophic pathology. If these findings are extrapolated neither vector toxicity nor pre-existent disease are present to human DMD patients, this would imply that AdV- in the transgenic mouse model, the latter constitutes a mediated dystrophin gene transfer needs to be instituted best case scenario that does not accurately represent the very early in the course of the disease in order to achieve anticipated situation for clinical use of AdV-mediated the desired therapeutic effect. Other intrinsic differences gene transfer in DMD patients. In addition, given that a in the general characteristics of the two muscles (eg absol- fairly even distribution of dystrophin expression ute workload, pattern of contraction, fiber-type compo- throughout the muscle cross-sectional area appears to be sition, etc) could also conceivably contribute to differen- at least as important as the absolute dystrophin level in tial responses to AdV-Dys administration. In addition, determining complete functional correction in transgenic since the relative proportion of dystrophin-positive fibers animals,14 it is likely that the limited diffusion of viral tended to be lower in the diaphragm than in the soleus, particles within AdV-injected muscles results in reduced this may also account for the reduced therapeutic efficacy therapeutic efficacy. Unfortunately, attaining higher dys- observed in the former muscle. Finally, it should be trophin expression, both in terms of absolute quantity pointed out that because the optimal titer was initially and uniformity of expression, would likely require the established for the diaphragm and then applied to both use of higher AdV dosages that lie within the myotoxic muscles, our data could actually underestimate the bene- range. Therefore, although we have shown that mdx fit achievable in the mdx soleus muscle after AdV-Dys muscle function can be improved with the approach out- administration if the optimal titers for the soleus and dia- lined here, a more complete prevention of dystrophic phragm were to differ. manifestations would appear to require further modifi- In this study, the baseline improvement in force pro- cation of both the vector and delivery modalities. duction observed in AdV-Dys-injected soleus muscles Our findings suggest a number of potential thera- was maintained throughout a series of high-stress eccen- peutic strategies that have important implications for tric contractions, whereas the rate of decline was not sig- future application of AdV-mediated gene transfer to the nificantly altered. These data are consistent with those treatment of DMD. It should be possible to reduce recently reported by Decrouy et al,30 who also found no myofiber toxicity associated with adenoviral gene significant alteration in the decline of force-generating expression by creating further deletions of the viral gen- capacity caused by eccentric contractions in adult mdx ome within the vector. For example, the E4 region of the mouse diaphragms studied 3 weeks after plasmid- adenoviral genome encodes gene products with a num- mediated dystrophin minigene transfer. On the other ber of functions including blockage of both host cell pro- hand, in neonatal mdx gastrocnemius muscles injected tein synthesis32 and transcriptional activation of cellular with AdV-Dys and studied 14 weeks later,27 the percent- growth proteins by the p53 tumor suppressor.33 Apoten- age decline in force-generating capacity after eccentric tial role for E4 genes in target organ toxicity after AdV contractions was significantly blunted. The results of administration is suggested by a recent report that the these different studies further emphasize the potential early hepatotoxicity seen after E1-deleted AdV delivery variability in physiological responses to dystrophin gene to liver is reduced by using a doubly (E1 + E4)-deleted transfer in relation to factors such as age, level of disease vector.34 Furthermore, AdV that are deleted of all viral progression at the time of therapeutic gene delivery, and coding sequences have recently been developed,35,36 the duration as well as magnitude of successful dystro- which have the additional advantage of being able to phin gene replacement. accommodate the more efficacious full-length dystro- Immunosuppressive agents such as corticosteroids2 phin gene. The availability of vectors with a higher and cyclosporine31 have been reported to have a transient therapeutic index, in turn, could allow for higher AdV beneficial effect on muscle strength in DMD in the dosages to be employed, as well as repeated vector absence of dystrophin gene replacement. Therefore, an administration (under immunosuppression) into the important methodological aspect of this study was the same muscle. Alternatively, an effective method for vas- use of contralateral muscles from the same mice as a con- cular administration of AdV9,37 could potentially permit trol, which permitted determination of the independent a lower viral particle exposure per infected myofiber effects of AdV-Dys injection. In addition, potential while at the same time allowing more widespread dis- between-animal variations in the degree of immune semination of AdV-Dys within the muscle, but this modulation as well as disease progression were also con- awaits further development. trolled for in this fashion. The lack of either inflammation In summary, AdV-mediated dystrophin minigene Adenovirus-mediated dystrophin transfer improves muscle strength L Yang et al 377 transfer is capable of alleviating the usual progression of mice.10 Immunosuppression was begun on the day before muscle weakness and histopathological changes in adult AdV-Dys injection and continued until the animals were dystrophic animals with pre-existent muscle pathology. killed. The study was approved by the institutional ani- In order to achieve this beneficial effect, however, it is mal ethics committee. necessary: (1) to establish initially the optimal vector dos- Measurement of isometric contractile properties age representing the best compromise between maximal Mice were first anesthetized to achieve a loss of deep gene transfer efficiency and minimization of direct toxic pain reflexes. Soleus muscles were then carefully effects; and (2) to provide effective immunosuppression removed in random order and tested to determine in vitro to prevent T cell-mediated elimination of AdV-infected isometric contractile properties as described below in myofibers. Although further efforts to improve vector detail. After completing contractility measurements in design and delivery methods are needed, these findings the soleus muscles, mice were killed by anesthetic over- provide optimism for the eventual applicability of this dose and the diaphragm was removed en bloc with its treatment modality in DMD patients. intact rib cage as previously described.5,29 Contractility measurements were then made in random order on Materials and methods muscle strips obtained from each side of the diaphragm. For both soleus and diaphragm, the freshly dissected Preparation of recombinant adenoviruses muscles were briefly transferred to chilled Ringer’s sol- Adenoviral recombinants (AdV-Dys) containing the 6.3 ution (composition: 119 mm NaCl, 4.7 mm KCl, 2.5 mm 38 CaCl2, 1.2 mm KH2PO4, 1.2 mm MgSO4,20mm NaHCO3 kb dystrophin minigene were constructed using E1/E3- ␮ deleted replication-defective serotype 5 human adeno- and 6 md-tubocurarine chloride) perfused with 95% virus as previously outlined in detail,9 where the dystro- O2:5%CO2 (pH 7.4). The muscles were then mounted phin cDNA was driven by human cytomegalovirus vertically in a jacketed tissue bath chamber filled with continuously perfused Ringer’s solution maintained at (CMV) promoter/enhancer elements inserted into the E1 ° region. The absence of E1-containing replication- 25 C, and a thermoequilibration period of 10 min was competent AdV was confirmed using a sensitive PCR observed before initiating contractile measurements. One screening assay as previously described.39 AdV-Dys titers end of each muscle was securely anchored to a platform (expressed as virus particles per ml) were determined in near the base of the chamber, while the opposite tendon a spectrophotometer at 260 nm, and the ratio of total was tied to the lever arm of a force transducer/length servomotor system (model 300B dual mode; Cambridge virus particles to infectious virus particles (100:1) was 5,29 ascertained by plaque assay as previously described.9 Technology, Watertown, MA, USA). The latter was mounted on a mobile micrometer stage (Newport Instru- Animals and surgical procedures ments, Toronto, Canada) to allow incremental adjust- ments of muscle length. Electrical field stimulation was Dystrophin-deficient mdx mice were purchased from the induced via platinum plate electrodes placed into the Jackson Laboratory (Bar Harbor, ME, USA). Mdx mice bath on both sides of the muscle. Supramaximal stimuli (50–60 days old) received injections of AdV-Dys (diluted with a monophasic pulse duration of 2 ms were delivered in saline) into the hemidiaphragm and soleus hindlimb using a computer-controlled electrical stimulator (model muscle on one side, whereas the same muscle on the con- S44; Grass Instruments, Quincy, MA, USA) connected in tralateral side was injected with saline alone to serve as series to a power amplifier (model 6824A; Hewlett Pack- a within-animal control. The mdx diaphragm and soleus ard, Palo Alto, CA, USA). Muscle force was displayed on were selected in order to study the response to AdV-Dys a storage oscilloscope (Tektronix, Beaverton, OR, USA), in muscles with relatively severe and mild pre-existing and the data were simultaneously acquired to computer functional impairment, respectively.16 Briefly, the mice (Labdat/Anadat software; RHT-InfoData, Montreal, were anesthetized with ketamine (130 mg/kg) and xyla- Canada) via an analog-to-digital converter at a sampling zine (20 mg/kg) by intramuscular (i.m.) injection. A lap- rate of 1000 Hz for later analysis. arotomy was performed to reveal the abdominal surface After adjusting each muscle to optimal length (Lo, the of the diaphragm and the soleus was also surgically length at which maximal twitch force is achieved), five exposed to permit direct visualization of the muscle. twitch stimulations were recorded and the mean value Using the aid of a dissecting microscope, the diaphragm was used to determine the following: twitch contraction and soleus muscles were directly injected with AdV-Dys, time (time to peak force), twitch half-relaxation time as previously described in detail.11 The injectate volume (time for force to decrease to one-half of its peak value), was roughly proportional to the muscle mass injected, and maximal isometric twitch force. Maximal isometric amounting to 150 ␮l and 15 ␮l for hemidiaphragm and tetanic force was then measured by stimulating the soleus, respectively. Animals recovered well from the muscle at 80 Hz for 1 s, allowing a clear plateau in force surgery and demonstrated no untoward effects. The to be attained. The muscle was removed from the bath mice were killed at either 5 or 60 days after AdV-Dys and Lo was directly measured under a dissecting micro- administration. scope with microcalipers accurate to 0.1 mm. Total mus- In order to prevent immune-mediated elimination of cle strip cross-sectional area was determined by dividing AdV-infected myofibers and associated adverse effects on muscle weight by its length and tissue density (1.06 muscle function,11,12 all mice were injected with FK506 (5 g/cm3). This allowed specific force (force/cross-sectional mg/kg/day subcutaneously), a potent inhibitor of acti- area) to be calculated, which was expressed as N/cm2. vated T-lymphocyte proliferation.10 This immunosuppressive regimen was selected based upon our previous Response to high-stress lengthening contractions demonstration of sustained high-level dystrophin Dystrophin is postulated to play an important role in expression 2 months after AdV-Dys delivery to mdx helping to protect myofibers from potentially damaging Adenovirus-mediated dystrophin transfer improves muscle strength L Yang et al 378 stresses associated with muscle contraction.4,5 Therefore, was supported by grants of the Deutsche Forschungsge- after allowing a 10-min recovery period following the meinschaft and Sander-Siftung, Germany. L Yang was above measurements of isometric contractile properties, the recipient of a post-doctoral fellowship from the Can- AdV-Dys-injected and control soleus muscles were also adian Lung Association. J Nalbantoglu is a Research compared to determine their ability to withstand high- Scholar of the Fonds de la Recherche en Sante du Quebec. stress eccentric (lengthening) contractions. The protocol We thank J Bourdon, N Matusiewicz, N Chughtai and has been described in detail previously.5 Briefly, the S Prescott for expert technical assistance. soleus muscle was supramaximally stimulated at 80 Hz for 700 ms; the muscle was held at Lo during the initial 500 ms, whereas a step-change in length amounting to a 10% Lo increase was imposed by the computer-controlled References servomotor lever arm during the final 200 ms of stimulation. A total of five such stimulations was performed, 1 Smith PEM et al. Practical problems in the respiratory care of patients with muscular dystrophy. New Engl J Med 1987; 316: each being separated by a 2-min recovery period at Lo. 1197–1204. The decline in maximal isometric force (obtained from 2 Khan MA. Corticosteroid therapy in Duchenne muscular dys- the first 500 ms of muscle stimulation at Lo) during suc- trophy. J Neurol Sci 1993; 120: 8–14. cessive contractions was then recorded and used as an 3 Hoffman EP, Brown RHJ, Kunkel LM. Dystrophin: the protein index of contraction-induced myofiber injury.5,27 product of the Duchenne muscular dystrophy locus. Cell 1987; 51: 919–928. Dystrophin immunostaining and histopathology 4 Weller B, Karpati G, Carpenter S. Dystrophin-deficient mdx After completing the muscle function studies, the soleus muscle fibers are preferentially vulnerable to necrosis induced and diaphragm were embedded in mounting medium by experimental lengthening contractions. J Neurol Sci 1990; 100: 9–13. and snap-frozen in isopentane pre-cooled with liquid N2. ␮ 5 Petrof BJ et al. Dystrophin protects the sarcolemma from stresses Transverse sections (6 m thick) were obtained in a cryo- developed during muscle contraction. Proc Natl Acad Sci USA stat and then fixed on slides in 1% acetone. Immunohisto- 1993; 90: 3710–3714. chemical procedures were carried out to detect dystro- 6 Fink RHA, Stephenson DG, Williams DA. Physiological proper- phin expression using a polyclonal antidystrophin (C- ties of skinned fibres from normal and dystrophic (Duchenne) terminus) primary antibody and biotinylated secondary human muscle activated by Ca2+ and Sr2+. J Physiol 1990; 420: antibody with subsequent visualization by peroxidase 337–353. staining, as previously outlined in detail.9,10 Muscle sec- 7 Williams DA, Head SI, Lynch GS, Stephenson DG. Contractile tions were also counterstained with hematoxylin and properties of skinned muscle fibres from young and adult normal and dystrophic (mdx) mice. J Physiol 1993; 460: 51–67. eosin (H & E) to allow determination of the prevalence of 8 Kozarsky KF, Wilson JM. Gene therapy: adenovirus vectors. centrally nucleated myofibers, an indicator of the degree Curr Opin Genet Dev 1993; 3: 499–503. of previous necrosis and regeneration in mouse 9 Acsadi G et al. Dystrophin expression in muscles of mdx mice 13–15 muscles. after adenovirus-mediated in vivo gene transfer. Hum Gene Ther Microscopically visualized sections were photo- 1996; 7: 129–140. graphed using a video camera and the image was stored 10 Lochmuller H et al. Transient immunosuppression by FK506 on a Macintosh computer. Analysis of the number and permits a sustained high-level dystrophin expression after aden- cross-sectional area of dystrophin-positive and dystro- ovirus-mediated dystrophin minigene transfer to skeletal phin-negative myofibers on the entire muscle cross-sec- muscles of adult dystrophic (mdx) mice. Gene Therapy 1996; 3: tion (averaging approximately 1500 and 700 myofibers 706–716. 11 Petrof BJ et al. Efficiency and functional consequences of adeno- for diaphragm and soleus, respectively) was performed virus-mediated in vivo gene transfer to normal and dystrophic using the public domain program NIH Image (version (mdx) mouse diaphragm. Am J Resp Cell Mol Biol 1995; 13: 1.49). The proportion of dystrophin-positive and dystro- 508–517. phin-negative myofibers exhibiting central nucleation 12 Petrof BJ et al. Impairment of force generation after adenovirus- was similarly determined. mediated gene transfer to muscle is alleviated by adenoviral gene inactivation and host CD8+ T cell deficiency. Hum Gene Statistical analysis Ther 1996; 7: 1813–1826. To assess the response to different AdV-Dys titers, a one- 13 Karpati G, Carpenter S, Prescott S. Small-caliber skeletal muscle way analysis of variance was employed along with post- fibers do not suffer necrosis in mdx mouse dystrophy. Muscle Nerve 1988; 11: 795–803. hoc application of Fisher’s least significant difference test. 14 Phelps SF et al. Expression of full-length and truncated dystro- Differences between AdV-Dys and contralateral control phin mini-genes in transgenic mdx mice. Hum Mol Genet 1995; muscles were determined using Student’s two-tailed t 4: 1251–1258. test for dependent samples. All values are means ± s.e. 15 Wells DJ et al. Expression of human full-length and minidystro- Statistical significance was defined as P Ͻ 0.05. phin in transgenic mdx mice: implications for gene therapy of Duchenne muscular dystrophy. Hum Mol Genet 1995; 4: 1245– 1250. Acknowledgements 16 Stedman HH et al. The mdx mouse diaphragm reproduces the This investigation was supported by grants from the degenerative changes of Duchenne muscular dystrophy. Nature Muscular Dystrophy Association of Canada, the Medical 1991; 352: 536–539. 17 Moscovici C, Maisel J. A toxin-like material separable from type Research Council of Canada, the National Research 5 adenovirus particles. Virology 1958; 6: 770–771. Council of Canada, the Muscular Dystrophy Association, 18 Seth P et al. Evidence that the penton base of adenovirus is Inc (USA) and the Association Pulmonaire du Quebec. BJ involved in potentiation of toxicity of pseudomonas exotoxin Petrof is the recipient of a Clinician-Scientist Award from conjugated to epidermal growth factor. Mol Cell Biol 1984; 4: the Medical Research Council of Canada. H Lochmuller 1528–1533. Adenovirus-mediated dystrophin transfer improves muscle strength L Yang et al 379 19 Yang Y et al. Cellular immunity to viral antigens limits E1- and contractile function in the dystrophic (MDX) mouse dia- deleted adenoviruses for gene therapy. Proc Natl Acad Sci USA phragm. Am J Physiol 1993; 265: C834–C841. 1994; 91: 4407–4411. 30 Decrouy A et al. Mini-dystrophin gene transfer in mdx4cv dia- 20 Yang Y et al. Inactivation of E2a in recombinant adenoviruses phragm muscle fibers increases sarcolemmal stability. Gene Ther- improves the prospect for gene therapy in cystic fibrosis. Nat apy 1997; 4: 401–408. Genet 1994; 7: 362–369. 31 Sharma KR, Mynhier MA, Miller RG. Cyclosporine increases 21 Dai Y et al. Cellular and humoral immune responses to adenovi- muscular force generation in Duchenne muscular dystrophy. ral vectors containing factor IX gene: tolerization of factor IX Neurology 1993; 43: 527–532. and vector antigens allows for long-term expression. Proc Natl 32 Halbert DN, Cutt JR, Shenk T. Adenovirus early region 4 enco- Acad Sci USA 1995; 92: 1401–1405. des functions required for efficient DNA replication, late gene 22 McCoy RD et al. Pulmonary inflammation induced by incom- expression, and host cell shutoff. J Virol 1985; 56: 250–257. plete or inactivated adenoviral particles. Hum Gene Ther 1995; 6: 33 Dobner T, Horikoshi N, Rubenwolf S, Shenk T. Blockage by 1533–1560. adenovirus E4orf6 of transcriptional activation by the p53 tumor 23 Brynes AP, Rusby JE, Wood MJA, Charlton HM. Adenovirus suppressor. Science 1996; 272: 1470–1473. gene transfer causes inflammation in the brain. Neuroscience 34 Gao G-P, Yang Y, Wilson JM. Biology of adenovirus vectors with 1995; 66: 1015–1024. E1 and E4 deletions for liver-directed gene therapy. J Virol 1996; 24 Spergel JM et al. NF-IL6, a member of the C/EBP family, regu- 70: 8934–8943. lates E1A-responsive promoters in the absence of E1A. J Virol 35 Kochanek S et al. A new adenoviral vector: replacement of all 1992; 66: 1021–1030. viral coding sequences with 28 kb of DNA independently 25 Ragot T et al. Efficient adenovirus-mediated transfer of a human expressing both full-length dystrophin and ␤-galactosidase. Proc minidystrophin gene to skeletal muscle of mdx mice. Nature Natl Acad Sci USA 1996; 93: 5731–5736. 1993; 361: 647–650. 36 Haecker SE et al. In vivo expression of full-length human dystro- 26 Vincent N et al. Long-term correction of mouse dystrophic phin from adenoviral vectors deleted of all viral genes. Hum degeneration by adenovirus-mediated transfer of a minidystro- Gene Ther 1996; 7: 1907–1914. phin gene. Nat Genet 1993; 5: 130–134. 37 Huard J et al. The route of administration is a major determinant 27 Deconinck N et al. Functional protection of dystrophic mouse of the transduction efficiency of rat tissues by adenoviral recom- (mdx) muscles after adenovirus-mediated transfer of a dystro- binants. Gene Therapy 1995; 2: 107–115. phin minigene. Proc Natl Acad Sci USA 1996; 93: 3570–3574. 38 England SB et al. Very mild muscular dystrophy associated with 28 Weller B et al. Inhibition of myosatellite cell proliferation by the deletion of 46% of dystrophin. Nature 1990; 343: 180–182. gamma irradiation does not prevent the age-related increase of 39 Lochmuller H et al. Emergence of early region 1-containing rep- the number of dystrophin-positive fibers in soleus muscles of lication-competent adenovirus in stocks of replication-defective mdx female heterozygote mice. Am J Pathol 1991; 138: 1497–1502. adenovirus recombinants during multiple passsages in 293 cells. 29 Petrof BJ et al. Adaptations in myosin heavy chain expression Hum Gene Ther 1994; 5: 1485–1491.