|Home | About | Journals | Submit | Contact Us | Français|
Trans-splicing adeno-associated viral (tsAAV) vectors hold great promise for delivering large therapeutic genes. One potential application is in the treatment of Duchenne muscular dystrophy (DMD). In this case, it is necessary to transduce whole body muscle. We demonstrated body-wide AAV-9 tsAAV transduction in normal neonatal mice. However, it was not clear whether such an approach would work in diseased mice. In this study we delivered the AAV-9 alkaline phosphatase (AP) tsAAV vector (3×1012 vector genome particles per vector per mouse, tail vein injection) to 2-month-old mdx mice, the most widely used DMD model. Four months later, we observed widespread AP expression in the heart. It reached the same level as we have seen in normal neonatal puppy. Interestingly, myocardial transduction correlated with β-myosin heavy chain expression but not with LamR, the putative AAV-9 receptor. AP expression was also detected in various skeletal muscles but at levels much lower than in normal newborn mice. Despite the existing inflammatory milieu, we did not see any appreciable increase in CD4+ and CD8+ T cells and macrophages in striated muscles after systemic tsAAV infection. In summary, our results have paved the way for tsAAV-mediated gene therapy for Duchenne cardiomyopathy.
Adeno-associated virus (AAV) has been considered a promising viral vector for Duchenne muscular dystrophy (DMD) gene therapy (Athanasopoulos et al., 2004; Blankinship et al., 2006). DMD is a lethal, degenerative muscle disease caused by mutations in the dystrophin gene (Kunkel, 2005). The 12-kb dystrophin coding sequence has been a great challenge for the AAV vector, which has a 5-kb maximal packaging capacity. To overcome the size hurdle, investigators have developed synthetic micro- and minidystrophin genes. The microgenes carry only about 30% of the coding information and they are usually less than 4kb. The minigenes are larger than 6kb and contain about 50 to 60% of the coding sequence. Although a microgene can fit into a single AAV virion, it is less effective in ameliorating muscle disease (Harper et al., 2002; Lai et al., 2009). From the therapeutic standpoint, the minigene would be preferable to the microgene.
A series of dual-AAV vector systems have the capacity to carry the minigene. These include the overlapping, trans-splicing (ts) and hybrid vectors (Duan, 2006b; Ghosh and Duan, 2007; Ghosh et al., 2008). Among these, the rationally designed tsAAV vectors are particularly promising (Lai et al., 2005, 2006, 2008). After local injection into dystrophin-deficient mdx mice, the transduction efficiency of the tsAAV minidystrophin vectors reached that of a single AAV microdystrophin vector (Lai et al., 2005; Liu et al., 2005). Because all body muscles are affected in DMD, the next critical step will be to achieve body-wide transduction with the tsAAV vectors. As a proof of principle, we first tested systemic AAV-9 tsAAV transduction in normal newborn mice (Ghosh et al., 2007). Six weeks after intravenous delivery, we observed robust whole body muscle transduction. In skeletal and cardiac muscle, transduction efficiency reached approximately 80 and 50%, respectively (Ghosh et al., 2007). These levels would meet the need for DMD gene therapy (Chamberlain, 2002; Duan, 2006a).
Despite the encouraging results, it remains to be determined whether systemic tsAAV delivery is achievable in a disease model such as adult mdx mice. Compared with normal neonatal mice, there are several challenges. First, AAV-9 has not yet been tested in dystrophic muscles. Disease-associated cellular and/or biochemical changes may alter AAV-9 transduction and/or tsAAV reconstitution in mdx muscle, especially in the context of systemic delivery. Second, there is a 20-fold body weight difference between a newborn puppy and an adult mouse. Scaling up could represent an important barrier for systemic tsAAV transduction. Third, in our previous neonatal study, we observed transduction for only 6 weeks. A longer observation period would allow us to better evaluate the relative persistence of systemic tsAAV transduction.
To address these clinically relevant issues, we administered AAV-9 alkaline phosphatase tsAAV vectors (tsAAV-AP) to 2-month-old female mdx mice by a single bolus tail vein injection and evaluated transgene expression 4 months later. The highest AP expression was observed in the heart, which reached the 50% transduction threshold required for dystrophic cardiomyopathy gene therapy. Surprisingly, limited skeletal muscle transduction was observed. Nevertheless, we did not see a strong cellular immune reaction in either the heart or skeletal muscle in tsAAV-infected mdx mice. These results raise the hope of exploring tsAAV-mediated minidystrophin therapy for Duchenne cardiomyopathy. Further optimization of the technology may pave the way for future body-wide treatment with minidystrophin tsAAV vectors.
All animal experiments were approved by the animal care and use committee at the University of Missouri (Columbia, MO) and were in accordance with National Institutes of Health (NIH, Bethesda, MD) guidelines. The original dystrophin-deficient mdx mice (C57BL/10ScSn-Dmdmdx/J) were purchased from Jackson Laboratory (Bar Harbor, ME). Experimental mice (2-month-old female) were obtained from a local breeding colony. All mice were housed in specific pathogen-free animal care facilities and kept under a 12-hr light (25lx)/12-hr dark cycle with free access to food and water.
The cis plasmids used for AAV packaging (including pcisAV.AP.Donor and pcisAV.AP.Acceptor) have been described previously (Ghosh et al., 2007). The reconstituted expression cassette expresses the human placental AP gene under the transcriptional regulation of the Rous sarcoma virus promoter and the simian virus 40 (SV40) polyadenylation signal. The AAV-9 packaging plasmid (pRep2/Cap9) was a gift from J. Wilson (University of Pennsylvania, Philadelphia, PA) (Gao et al., 2004). Recombinant AAV-9 vectors were generated by a triple-plasmid transfection protocol described previously (Bostick et al., 2007; Ghosh et al., 2007; Yue et al., 2008). Viral stocks were purified by three rounds of isopycnic CsCl2 ultracentrifugation followed by three changes of HEPES buffer at 4°C for 48hr. Viral titer and quality control were performed according to our previously published protocol (Bostick et al., 2007; Ghosh et al., 2007; Yue et al., 2008). The average viral titer was 1×1010 vector genome (VG) particles/μl for both the donor and acceptor vectors.
The study was performed in 2-month-old female mdx mice. The average body weight was 23.8±0.5g (n=10). We first determined the maximal tolerable fluid volume for tail vein injection. Up to 600μl of saline was well received in this age group of female mdx mice without noticeable adverse response (such as respiratory stress and death). Systemic delivery was then performed in conscious mice by a single bolus injection through the tail vein. Before injection, AV.AP.Donor and AV.AP.Acceptor vectors were thoroughly mixed by repeated pipetting and gentle shaking at 4°C overnight. A total of 600μl of tsAAV-AP vectors (including 300μl of AV.AP.Donor and 300μl of AV.AP.Acceptor) was delivered to each mouse (25μl/g body weight). This yielded a dose of 3×1012 VG particles/vector/mouse or 1.26×1011 VG particles/vector/g body weight. After injection, mice were returned to the cage for recovery. All tsAAV-injected mice survived the procedure.
Mice were euthanized 4 months after systemic tsAAV-AP infection. Various skeletal muscles (including the tibialis anterior, extensor digitorum longus, quadriceps, gastrocnemius, soleus muscles, diaphragm, and tongue) as well as several internal organs (the heart, lung, and liver) were harvested for AP staining/activity assays. In situ expression was determined by histochemical staining of 8-μm cryosections as described previously (Duan et al., 2000). AP activity in whole muscle lysate was measured with a StemTAG AP activity assay kit (Cell Biolabs, San Diego, CA) according to the manufacturer's instructions (Ghosh et al., 2007). In these assays, endogenous heat-labile AP activity was first inactivated by incubating samples at 65°C for 60min.
Immunofluorescence staining for dystrophin and utrophin was performed essentially as described previously (Yue et al., 2003; Lai et al., 2005). Dystrophin was detected with a mouse monoclonal antibody against the dystrophin C-terminal domain (Dys-2, diluted 1:100; clone Dy8/6C5, IgG1; Novocastra, Newcastle, UK). Utrophin was examined with a mouse monoclonal antibody against the utrophin N-terminal domain (VP-U579, diluted 1:20; clone DRP3/20C5, IgG1; Vector Laboratories, Burlingame, CA). Slow and fast skeletal muscle fiber and cardiac β-MHC were determined by immunofluorescence staining according to a previously published protocol (Lai et al., 2005). Slow twitch fiber (type I) was revealed with a monoclonal antibody against MHC-I (M8421, diluted 1:2000; clone NOQ7.5.4D, IgG1; Sigma-Aldrich, St. Louis, MO). Fast twitch fiber (type II) was revealed with a monoclonal antibody against MHC-II (VP-M665, diluted 1:20; clone WB-MHCf, IgG1; Vector Laboratories). Cardiac β-MHC was determined with the M8421 monoclonal antibody (Sigma-Aldrich) at a dilution of 1:4000. Embryonic MHC was detected with a monoclonal antibody against embryonic MHC (F1.652, diluted 1:250; IgG1; Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA). Capillaries were revealed by means of an alkaline phosphatase–metanil yellow staining protocol described previously (Yue et al., 2008). The AAV-9 receptor (laminin receptor) was detected by immunofluorescence staining with a rabbit polyclonal antibody (diluted 1:200; Abcam, Cambridge, MA) (Yue et al., 2008). A VECTASTAIN ABC kit (Vector Laboratories) was used for immunohistochemical detection of CD4+ and CD8+ T cells and macrophages. Macrophages were recognized with a rat anti-mouse F4/18 antibody (#RM2920, diluted 1:200; clone CI:A3-1, IgG2b; Caltag Laboratories, Burlingame, CA). CD4+ T cells were revealed with a rat anti-mouse CD4 (L3T4) antibody (cat. no. 553727, diluted 1:800; clone GK1.5, IgG2b; BD Biosciences, San Jose, CA). CD8+ T cells were revealed with a rat anti-mouse CD8 (Ly-2) antibody (cat. no. 553027, diluted 1:800; clone 53-6.7, IgG2b; BD Biosciences).
Data are presented as means±standard error of mean (SEM). Statistical analysis was performed with SPSS software (SPSS, Chicago, IL). Statistical significance was determined by one-way analysis of variance (ANOVA) followed by Bonferroni post hoc analysis. Differences were considered significant when p<0.05.
To evaluate intravascular tsAAV transduction in dystrophic muscle, we delivered tsAAV-AP vector at 6×1012 vector genome (VG) particles/mouse to 2-month-old mdx mice via tail vein injection (3×1012 VG particles/mouse of each vector, AV.AP.Donor and AV.AP.Acceptor). AP expression was examined 4 months after gene delivery. We observed minimal to moderate transduction in various skeletal muscles. It ranged from about 5 to 15% on tissue cross-sections (Fig. 1A). In whole muscle lysate, the AP activity ranged from 6 to 22μM/μg protein (Fig. 1B). This represents approximately one tenth to one fourth of what we have seen in tsAAV-AP-infected normal neonatal mouse skeletal muscle (Fig. 1B) (Ghosh et al., 2007).
To explore the potential mechanisms underlying the dramatic skeletal muscle transduction difference between adult mdx mice and newborn normal mice, we examined whether the absence of dystrophin limited systemic tsAAV transduction. In mdx skeletal muscle, there exists about 1 to 2% dystrophin-positive myofibers (Lu et al., 2000). These are called revertant fibers. Interestingly, we did not see a transduction preference for the revertant fibers (Fig. 1A, arrow in quadriceps panels).
We next examined fiber type preference. Although we observed a few AP-positive slow twitch myofibers (Fig. 2, arrows at bottom; Fig. 3B–F, arrowheads), the vast majority of AP-positive cells were fast twitch fibers (Fig. 3A). It is interesting to note that the only known AAV-9 receptor, LamR, was enriched in slow twitch myofibers (see Supplementary Fig. 1 at www.liebertonline.com/hum) (Akache et al., 2006; Yue et al., 2008). Besides slow and fast twitch myofibers, mdx skeletal muscle also contains a significant amount of embryonic/regenerating myofibers. Interestingly, these fibers were rarely transduced by the tsAAV vectors (Fig. 3).
Additional study on capillarity density suggests that systemic tsAAV transduction was not limited by capillary distribution (see Supplementary Fig. 2 at www.liebertonline.com/hum). The muscles (or the regions) that contain substantially more capillaries (presumably better blood perfusion) did not show high numbers of AP-positive cells (Supplementary Fig. 2).
In contrast to skeletal muscle, we observed widespread transduction in the mdx heart (Fig. 4A). The highest AP expression was observed in the left ventricle, which reached about 50% of the cardiomyocytes, a level sufficient for treating Duchenne cardiomyopathy (Fig. 4B) (Yue et al., 2004; Duan, 2006a; Bostick et al., 2008). Overall, AP activity in the mdx heart reached that of the normal newborn mouse heart (Fig. 4C) (Ghosh et al., 2007). To understand the mosaic transduction pattern in the mdx heart, we performed AP staining and β-myosin heavy chain (MHC) immunofluorescence staining (Fig. 5A). Interestingly, there appeared to be a correlation between AP and β-MHC expression. More AP-positive cardiomyocytes were seen in areas that showed high levels of β-MHC expression (Fig. 5A). Next, we examined AAV-9 receptor LamR expression in the mdx heart. In contrast to skeletal muscle, LamR was highly expressed at the intercalated disks rather than inside cardiomyocytes (Fig. 5B). This was completely different from the intracellular expression pattern seen in AP-positive cardiomyocytes (Fig. 5B).
The inflammatory milieu in dystrophic muscle represents a great challenge for muscular dystrophic gene therapy. To determine whether systemic AAV-9 tsAAV-AP delivery provoked additional immune cell infiltration, we performed immunohistochemical staining with macrophage- and T cell-specific markers. In the absence of tsAAV-AP infection, macrophages were the predominant inflammatory cells in striated muscles (Fig. 6). There were also a few CD4+ and CD8+ T cells (Fig. 6). In tsAAV-AP-infected mice, we did not see an appreciable increase in immune cell infiltration in either heart or skeletal muscle (Fig. 6).
Besides striated muscles, we also detected a few AP-positive cells in the lung and liver (see Supplementary Fig. 3 at www.liebertonline.com/hum). In the lung, the transduced cells were mainly alveolar cells (Supplementary Fig. 3). Surprisingly, we detected an extraordinary amount of AAV genome in the liver (see Supplementary Fig. 4 at www.liebertonline.com/hum). For reasons yet unknown, we observed extremely high-level endogenous AP expression in the kidney of uninfected mdx mice (data not shown). This high background prevented us from reaching a conclusion about tsAAV transduction in the mdx kidney.
The ability to deliver a larger therapeutic expression cassette is the primary motivation in developing the tsAAV vectors. Theoretically, the tsAAV system offers an opportunity to capitalize on the therapeutic benefit of the minidystrophin gene for DMD treatment (Harper et al., 2002; Duan, 2006b; Lai et al., 2009). To achieve the long-term goal of tsAAV-mediated DMD gene therapy, we generated a pair of rationally designed minidystrophin tsAAV vectors (Lai et al., 2005, 2008). Local injection into mdx mice resulted in robust transduction comparable to that of a single AAV vector (Lai et al., 2005). However, the benefit of such an approach will be limited unless we can demonstrate efficient whole body muscle transduction in a dystrophic animal model. To build on our previous success in normal neonatal mice, in this work we tested the hypothesis that systemic tsAAV transduction is achievable in adult mdx mice.
After a single tail vein injection of the tsAAV-AP vectors, we observed variable, but in general low-level, transduction in various skeletal muscles (Fig. 1). However, robust tsAAV transduction was detected in the mdx heart (Fig. 4). Importantly, we did not see aggravated immune cell intrusion in tsAAV-transduced muscles (Fig. 6). Collectively, these results are encouraging but somewhat unexpected.
In our study with the same set of AAV-9 tsAAV-AP vectors, we delivered 3.5×1011 VG particles/vector/mouse to normal newborn mice (Ghosh et al., 2007). This equals 2.3×1011 VG particles/vector/g body weight. In the current study, we used a dose of 3×1012 VG particles/vector/mouse or 1.26×1011 VG particles/vector/g body weight. On the basis of the viral titer of 1×1010VG particles/μl and a tolerable volume of 600μl/mouse, this is the maximal dose we can deliver to a 2-month-old mdx mouse. Although the total viral dose was about 10-fold higher in adult mdx mice, the dose per gram body weight was reduced by about 50%. The reduction in per-unit dose seems to correlate with the reduced transduction in skeletal muscle. However, it cannot explain our findings in the heart.
Age and disease status are two major differences between the current and previous studies. Although age-dependent differences have been shown in the aorta, retina, liver, and kidney after systemic AAV-9 transduction, age seems to have a minimal impact on striated muscle transduction (Bostick et al., 2007). The primary cellular defect in DMD is the loss of sarcolemmal dystrophin. To determine whether dystrophin deficiency per se affects systemic tsAAV transduction, we examined AP expression in dystrophin-positive revertant myofibers (Fig. 1). We did not see preferential transduction in these myofibers (Fig. 1). This result suggests that the absence of dystrophin is not the limiting factor in systemic tsAAV transduction in mdx mice. Efficient cardiac transduction in the mdx heart further strengthened this conclusion (Fig. 4).
We also failed to see a correlation between systemic tsAAV transduction and regenerating (embryonic) or necrotic myofibers (Figs. 1 and and3).3). Similar to normal mouse muscle, fast twitch myofibers were also favored in mdx muscle (Figs. 2 and and3)3) (Bostick et al., 2007). Additional studies on the LamR AAV-9 receptor and capillary density suggest that these factors were also unlikely to be responsible for the poor skeletal muscle transduction (Supplementary Figs. 1 and 2). We currently believe that the low transduction efficiency in mdx skeletal muscle may likely reflect the viral dose used in the study. In support of this notion, it has been shown that at the dose of 1.8×1012 VG particles/mouse, a tail vein-injected AAV-9 LacZ vector completely transduced both the heart and skeletal muscle in normal mice. However, at the dose of 1×1011 VG particle/mouse, efficient transduction was observed only in the heart and there was barely any transduction in skeletal muscle (Inagaki et al., 2006).
It has been established that mosaic expression in less than 20% myofiber cannot halt skeletal muscle disease in DMD (Phelps et al., 1995). Our findings reveal an unexpected challenge in AAV-9 tsAAV-mediated minidystrophin gene therapy. Considering the fact that we have already used the maximal dose that an adult mdx mouse can tolerate, novel strategies must be developed to further enhance skeletal muscle transduction efficiency. In this regard, several strategies (such as tyrosine mutant modification and in vivo AAV capsid evolution) hold great promise in generating more potent AAV vectors (Zhong et al., 2008; Yang et al., 2009).
Several interesting observations were seen in the heart. We have shown that AAV-9 preferentially transduces the fast twitch fibers (type II MHC) in skeletal muscle. The slow twitch fibers (type I MHC) were minimally transduced (Fig. 3) (Bostick et al., 2007). Cardiac β-MHC and skeletal muscle type I MHC (slow twitch fibers) share the same contractile properties. Surprisingly, β-MHC-positive cardiomyocytes were preferentially transduced in the mdx heart (Fig. 5A). A better understanding of this paradoxical phenomenon awaits future studies.
AAV-9 has been shown to efficiently transduce adult mouse liver after systemic gene delivery in normal mice (Inagaki et al., 2006; Bostick et al., 2007). Surprisingly, we detected only minimal liver transduction in adult mdx mice despite the abundant amount of viral genome in the liver (Supplementary Figs. 3 and 4). Future studies are needed to clarify the underlying mechanism(s) of the aberrant tsAAV transduction pattern in mdx mouse liver.
The most striking finding of our study is the robust transduction of tsAAV vector in adult mdx heart. A single AAV-9 vector has been shown to preferentially transduce the heart in normal mice (Inagaki et al., 2006; Pacak et al., 2006; Bostick et al., 2007). The tsAAV-9 vector has also been shown to result in robust cardiac transduction in normal neonatal mice (Ghosh et al., 2007). The outstanding myocardial transduction and the apparent lack of provocative cellular immune response seen in adult mdx mice suggest that AAV-9 tsAAV may hold great promise for treating Duchenne cardiomyopathy. This will be particularly relevant to the gene therapy of X-linked dilated cardiomyopathy in which dystrophin is selectively lost in the heart but not in skeletal muscle.
This work was supported by grants from the National Institutes of Health (AR-49419; D.D.) and the Muscular Dystrophy Association (D.D.).
No competing financial interests exist.