Expansion of revertant fibers in dystrophic mdx muscles reflects activity of muscle precursor cells and serves as an index of muscle regeneration Free

Duchenne muscular dystrophy (DMD) is one of the commonest hereditary muscular diseases and is characterized by progressive muscle degeneration in 1 out of 3500 male live births. The gene responsible for DMD is located on the X-chromosome and was identified by positional cloning (Koenig et al., 1987; Hoffman et al., 1987). It is the largest and among the most complex of known genes, comprising 79 exons spanning more than 2.4 million base pairs (Koenig et al., 1987). The gene encodes at least 17 protein products of which the main muscle isoform is among the largest, consisting of a 3685 amino acid (427 kDa) single chain that can be divided into the following four domains; actin-binding N-terminal, rod, cysteine-rich and C-terminal (Bies et al., 1992; Nishio et al., 1994; Muntoni and Strong, 1989; Nudel et al., 1989; Monaco et al., 1986; Gorecki et al., 1992; Holder et al., 1996; Feener et al., 1989; Lidov et al., 1995; Byers et al., 1993; Austin et al., 1995). Dystrophin protein plays a central role in organizing a multiprotein complex at the sarcolemma, and linking cytoskeleton proteins to extracellular matrix proteins (Monaco et al., 1986; Holt et al., 1998; Rando, 2001). Structural and functional analyses have assigned most of the known functions of the protein to the two terminals and the cysteine-rich domain, whereas the rod domain, consisting of 24 repeated units and spanning about half the length of the protein, appears not to be essential to function. Most DMD mutations, be they deletions, duplications or point mutations, occur within the rod domain, but disrupt the reading frame and therefore prevent translation of the crucial C-terminal domain. This distinguishes DMD from the milder Becker muscular dystrophy (BMD), in which mutations occur in the same regions, but the mutated transcripts retain the reading frame, and are translated into truncated partially functional dystrophin. The mdx mouse, a homolog of human DMD (Bulfield et al., 1984), contains a nonsense point mutation in exon 23 (Sicinski et al., 1989), resulting in the lack of dystrophin expression, and cycles of muscle degeneration and regeneration.

An intriguing feature in muscles both of DMD patients and mdx mice is the presence of sporadic dystrophin-positive muscle fibers, so called `revertant fibers' (RFs) in an otherwise dystrophin-negative background (Hoffman et al., 1990; Nicholson, 1993; Sherratt et al., 1993). Our previous studies showed that revertant dystrophins in muscles of mdx mice involves massive loss of up to 30 exons, sometimes from two non-contiguous regions. Deletions of this size are rare in DMD or BMD, and double deletions are not reported. This is more consistent with aberrant splicing than genomic deletion as the mechanism for the restoration of dystrophin in RFs. This view is bolstered by our failure to detect deletions in myonuclei of RFs from the genomic regions coding for the missing part of dystrophin protein.

We also showed that RFs commonly form clonal clusters that increase in size from a mean of 1 short segment at birth to up to 100 fibers of more than 1 mm in length by 18 months of age. Our working hypothesis is that revertant events are initiated within a subset of muscle precursor cells that proliferate in response to muscle degeneration and participate in regeneration of new fibers. Expansion in the size and number of RFs is attributable mainly to their relative resilience to subsequent degeneration and would not be expected to be reflected in any parallel expansion of the revertant precursors, since these do not express dystrophin prior to fusion into fibers and can enjoy no selective advantage from their prospects of doing so. The low initial frequency of revertant events (around 0.1–0.01% as estimated at birth), together with this lack of selective advantage for revertant myogenic precursors in vivo or in tissue culture, is probably a major reason for the failure of our attempts, like those of others (Gussoni and Kunkel, personal communication), to establish clonal cultures of the myoblast precursors of RFs.

As an alternative, we have explored the relationship between expansion of RF clusters and myofiber degeneration and regeneration in vivo using several animal models in which degeneration or regeneration was experimentally modified. First, we examined three strains of `micro'-dystrophin-transgenic mdx mice – CS1, AX11 and M3 – generated by Sakamoto and colleagues to test the functionality of truncated (i.e. micro) dystrophins (Sakamoto et al., 2002; Yuasa et al., 1998). Expression of the transgenes protects the dystrophic muscles from degeneration with varying degrees of efficiency. The largest transgene, CS1, almost completely prevents muscle degeneration, whereas the smallest transgene, M3, has limited impact. In each case, we used an antibody against the missing rod region of dystrophin to identify `revertant' muscle fibers. We carried out similar studies of RFs in transgenic mdx mice over-expressing full-length and mini-utrophin, both of which have been shown to reduce dystrophic pathology markedly in limb and diaphragm muscle (Tinsley et al., 1996; Tinsley et al., 1998). Thus, these transgenics provide models that can be used to evaluate the effect of muscle degeneration and regeneration on the development and expansion of RFs. To eliminate regeneration specifically in degenerating mdx muscles, we applied local high-dose radiation, previously shown to block regeneration (Wakeford et al., 1991), but not degeneration in the mdx mouse (Pagel and Partridge, 1999), and compared the pattern of RF expansions in irradiated versus non-irradiated mdx muscles.

The results showed clearly that expansion of RF clusters is dependent on muscle regeneration. The fact that reactivation of radiation-tolerant muscle precursor cells produce newly regenerated RFs suggests that revertant events occur initially within a minority of muscle precursor cells. The proliferation of these cells, as part of the regeneration process, leads to the expansion of RF clusters within degenerating muscles. The expansion of revertant clusters depicts the cumulative history of regeneration, thus providing a useful index for functional evaluation of therapies that counteract muscle degeneration.

RFs in muscles of micro-dystrophin-transgenic mice To investigate the effects of truncated dystrophin expression on the expansion of RF clusters, we examined micro-dystrophin-transgenic mice CS1, AX11 and M3 (Sakamoto et al., 2002), which contain four rod repeats (4.9 kb), three rod repeats (4.4 kb) and one rod repeat (3.9 kb), respectively (Fig. 1). Two antibodies, P7 and MANDRA1, were used to distinguish the revertant dystrophin from transgenic micro-dystrophin (Ellis et al., 1990). P7, a polyclonal antibody against epitopes encoded by exon 57 of the dystrophin gene, which is absent from the micro-dystrophin constructs, stains only RFs (Fig. 2). MANDRA1 is a monoclonal antibody against a C-terminal epitope present both in RFs and in micro-dystrophins, and therefore detects both (Fig. 2). All muscle fibers in the three transgenic mice were stained with MANDRA1 (data not shown). Only a small number of isolated fibers were stained with P7, and the pattern of distribution is consistent with that observed in muscles of young non-transgenic mdx mice. The P7-positive fibers were therefore counted as RFs.

To understand the special distribution of both the transgenic micro-dystrophins and revertant dystrophin, serial sections were immunostained with both P7 and MANDRA1 antibodies. All fibers that stained positively with P7 were also stained with MANDRA1. Indeed, with MANDRA1, the staining intensity in RFs was clearly stronger than in surrounding fibers expressing only micro-dystrophin (Fig. 2). This indicates that the RFs express both micro-dystrophin and revertant dystrophin, and that expression of micro-dystrophin does not suppress the expression of revertant dystrophin. Because no antibody specific to the micro-dystrophins is available, we cannot determine whether the levels of the transgenic dystrophins were lower in the revertant dystrophin-positive segments than in non-RFs.

We then compared the expression and clustering levels of RFs in the three transgenic mice at 5 and 10 weeks of age. In control mdx muscles, the total number of RFs and clusters both increased markedly between 5 and 10 weeks of age, a period of conspicuous muscle pathology (Fig. 2, Fig. 3A-C). By comparison, both the number of RFs and clusters were significantly fewer in AX11 and CS1 transgenic muscles at 5 weeks of age, and especially at 10 weeks of age (Fig. 2, Fig. 3A-C). The difference was most marked between control mdx mice and CS1 mice, which are transgenic for the longest isoform of the micro-dystrophin. Fewer RFs were detected in muscles of M3 when compared with muscles from control mdx, but the difference was not significant (Figs 2, 3). Interestingly, both the total number and the maximal number of RFs per cluster decreased with increasing age in both AX11 and CS1 transgenics (Fig. 3). Since CS1 transgenics experience no obvious muscle degeneration during this period, it is likely that decreased expression rather than loss of RFs is responsible, suggesting that events leading to the expression of revertant dystrophins are reversible in some fibers.

To test the hypothesis that muscle regeneration plays an important role in the process of RF expansion, we directly compared the numbers of RFs and clusters to the percentage of central nucleated fibers (CNF), a useful indicator of muscle regeneration before it reaches plateau (about 80% or higher CNF) in dystrophic muscles. The percentage of CNFs in CS1, AX11 and M3 transgenic mice, similar to that reported by Sakamoto et al. (Sakamoto et al., 2002), varied from 1.1%, to 23.7% to 53.4%, respectively, and was significantly lower than the 70% observed in mdx mice at 10 weeks of age (Fig. 4A). The increase in regeneration correlated well with the increase in numbers of RFs and clusters from CS1 to M3 transgenics and to mdx mice (Fig. 4B,C). This result further supports the idea that muscle regeneration is an essential process for the expansion of RFs.

To determine whether the expression of micro-dystrophins specifically contributes to the suppression of revertant dystrophin in the transgenics, we examined RFs in both full-length and mini-utrophin-transgenic mdx mice in comparison with control mdx mice (Fig. 5). Although both transgenes diminish muscle pathology, the expression of full-length utrophin significantly reduced the number of centrally nucleated regenerating fibers in tibialis anterior (TA), extensor digitorum longus (EDL) and diaphragm muscles, especially in young animals (Tinsley et al., 1996; Tinsley et al., 1998). As shown in Fig. 5, the total number of RFs and of clusters was, as expected, significantly lower in the TA muscles of the utrophin-transgenic mice, most notably in the mice transgenic for full-length utrophin, than the numbers in control mdx muscles in all age groups from 1 month, to 6 months to >12 months of age. The size of RF cluster, as seen in micro-dystrophin-transgenic mice, was also greatly reduced in the utrophin-transgenic mice, with nearly all RFs appearing as isolated single fibers in mice transgenic for full-length utrophin, even in the 1-year-old animals (Fig. 5A). However, small clusters with more than five RFs were still frequently detected in the muscles of the same-aged mini-utrophin-transgenic mice. The degree of expansion of RFs in the two transgenics is therefore consistent with our previous observations that full-length utrophin expression significantly diminishes muscle degeneration and regeneration, and maintains the muscles of the transgenics at nearly full function despite the presence of some CNFs (not significantly different from the normal control C57Bl10 mice by the age of 5 weeks), whereas the mini-utrophin protects muscles much less effectively (with about 10% of CNFs by the age of 5 weeks) (Tinsley et al., 1996; Tinsley et al., 1998). This result thus suggests that it is the inhibition of muscle degeneration and regeneration per se rather than expression of a specific transgene that suppresses the expansion of RFs.

The importance of regeneration for the expansion of RFs was further supported by the examination of mdx muscles after irradiation. TA muscles of 3-week-old mdx mice were irradiated with 4, 10 or 18 Gy and examined at 1 month, 3 months and 1 year afterwards (Fig. 6). In agreement with our previous findings (Wakeford et al., 1991; Pagel and Partridge, 1999; Morgan et al., 2002), irradiation suppressed fiber regeneration in a dose-dependent manner, and 18 Gy irradiation abolished nearly all regeneration activity as demonstrated by the lack of fibers of small caliber and by the lack of central nucleation one month after irradiation (Fig. 6A). The number of RFs was inversely correlated with the dosage of irradiation. After irradiation with 18 Gy, nearly all RFs were singular and the number of RFs in the muscles at 1 and 3 months after the irradiation remained indistinguishable from those seen in the muscles at the time of irradiation (3 weeks) (Fig. 6B). The suppressive effect of irradiation on RF proliferation was most evident 1 year after irradiation (Fig. 6C). This reflected the continued growth in the number of RFs and the size of RF clusters in non-irradiated mdx muscles, whereas in the irradiated muscles such numbers remained at similar levels to those of mice at 3 weeks of age. These results provide compelling evidence that the expansion of RFs relies on muscle regeneration.

Our previous experiments have shown that most muscle precursor cells are radiation sensitive and that muscle regeneration in the mdx mouse is blocked by irradiation with 18 Gy. However, some muscle precursor cells, perhaps quiescent satellite cells, in irradiated muscles can be reactivated by severe injury mediated by treatment with notexin, a snake venom, and contribute to muscle regeneration (Heslop et al., 2000). We therefore investigated the association between activation of such precursor cells and generation of RFs and revertant clusters. The legs of mdx mice were irradiated with 18 Gy at 3 weeks of age followed by notexin injection into TA muscles 4 weeks later (Fig. 7A). As expected, the muscles, examined 1 week after injection, showed foci of regenerating fibers of small caliber in the notexin-treated areas. The total number of RFs in notexin-treated and pre-irradiation muscles was lower than that in the non-irradiated control muscles, but significantly higher than that in the irradiated muscles without notexin treatment (Fig. 7B,C). Most notable is the formation of clusters of dystrophin-positive fibers within the regenerating areas demonstrated by immunohistochemistry (Fig. 7B). All these fibers were small in caliber and centrally nucleated, and therefore were newly regenerated RFs. Although notexin treatment in non-irradiated muscle did not increase the total number of RFs significantly, it did create some clusters of newly regenerating RFs (Fig. 7B,C). These results show that at least some of the radiation-resistant muscle precursor cell populations are able to establish revertant phenotypes.

Our previous studies showed that most revertant dystrophins are able to relocalize nitric oxide synthase (nNOS) to sarcolemma of the fibers (Lu et al., 2000). By contrast, nNOS was not restored to the sarcolemma of muscle fibers in mice transgenic for micro-dystrophin (Fig. 8). We therefore examined RFs in the muscles of these transgenic mice to see if the expression of micro-dystrophins affects the localization of nNOS in RFs. Serial sections from all three transgenics showed that most RFs (positively stained with antibody P7) were also positively stained for nNOS. This suggests that most revertant dystrophins are functionally superior in this respect to the micro-dystrophins, and that interaction of nNOS with revertant dystrophins is not hampered by the presence of overexpressed micro-dystrophins.

Although it was first described 15 years ago (Hoffman et al., 1990; Sherratt et al., 1993), the phenomenon of revertant muscle fibers remains enigmatic. In particular, we do not have a firm idea of the mechanism behind the expression of dystrophin in these fibers. We have argued previously (Lu et al., 2000) that the revertant phenomenon is an epigenetic event that arises in individual satellite cells at around birth. Here, from the study of the pattern of appearance and expansion of RFs in the dystrophic mice, we derive some further support for this hypothesis.

One of the most prominent features of the RFs in the mdx mouse is their expansion with age, which appears clonal in that the pattern of epitope loss is variable between clusters but is conserved within each cluster (Lu et al., 2000). RFs first appear at around birth as short segments, some 10 μm in length, of sporadic single fibers, expanding over the following 18 months up to clusters with more than 100 fibers and spanning 1 mm or more of fiber length, although never attaining a sufficient proportion to ameliorate muscle pathology significantly. This expansion has been modeled as a consequence of cycles of muscle degeneration and regeneration in combination with a preferential survival of the fibers that contain dystrophin (Lu et al., 2000; Garcia et al., 1999). However, this presumption has never been tested and this was a major goal of this investigation. All data in our study point to a strong and direct dependence of RF cluster expansion on the intensity of muscle regeneration. In mdx mice carrying micro-dystrophin transgenes, the number of RFs is strongly correlated with other measures of the protection afforded against muscle damage by each individual transgene. Similarly, in the utrophin-transgenic mdx mice, where muscle degeneration and consequent regeneration is diminished, there was a clear relationship between the protective effect of the two versions of the transgenes and the increase in size of RF clusters. The dependence of RF expansion on muscle regeneration was further confirmed by the effects of irradiation on RFs. By varying the dose of radiation and by examining the target muscles at different time points after irradiation, we found that expansion of RFs was diminished by radiation in a dose-dependent manner and that, at the highest dose, RFs were, in effect, frozen at the cluster size they would have attained at the time of application of the radiation (Fig. 6). This was most clearly seen 1 year after irradiation with 18 Gy, when the disparity between irradiated and control muscles had become more conspicuous. In combination, these data clearly establish that muscle regeneration is an absolute requirement for expansion of RF clusters.

The dependence of RF expansion on regeneration argues strongly against the mechanism we had raised previously: namely that the expansion represents the progressive increase in territory of factors each of which determines a specific pattern of alternative splicing, and spreads by diffusion within each fiber and between adjacent fibers (Lu et al., 2000). This hypothesis would predict that revertant clusters would grow within the existing stable muscle fibers. However, the present findings that expansion of RFs does not occur in the absence of regeneration, even when degeneration continues (after irradiation), argue unequivocally against significant participation of a diffusible factor for the expansion of RFs within dystrophic muscles. The clonal individuality of epitope expression (Lu et al., 2000) and the dependence on muscle regeneration therefore indicate clearly the involvement of myogenic stem cells for the expansion of RFs.

In principle, translatable transcripts might be generated by either further mutations that restore an open-reading frame, or by skipping of frame-disrupting exon(s) during splicing of the transcript. The clonal individuality of the RF cluster favors the mutation hypothesis (Klein et al., 1992; Wallgren-Pettersson et al., 1993), but the high frequency of revertant events and of clusters in which exons are lost from two non-adjacent parts of the gene encoding dystrophin argue against this view (Lu et al., 2000). A random mutational process would have been expected to generate at least some large multi-fiber domains at birth, arising from mutations occurring early during prenatal myogenesis. We found no such clusters at birth. Furthermore, we could not demonstrate deletion of the genomic sequence corresponding to the missing epitopes in the protein sequence within large RF clusters. Results from our present studies now provide more-compelling evidence against the mechanism of superimposed mutations (Fig. 3) (Crawford et al., 2001), leaving an epigenetic mechanism as the most plausible explanation (Wilton et al., 1997; Lu et al., 2000). All transgenic mdx mice showed similar numbers of RFs before the onset of muscle degeneration and nearly all of these RFs were singular. Furthermore, no expansion of RFs was observed in any transgenic mdx mouse despite a significant muscle growth before the age of 5 weeks. Instead, the number of RFs and the intensity of dystrophin signals decrease with age in mini-dystrophin- and micro-dystrophin-transgenic mdx mice, suggesting that at least some of the revertant phenotypes are not stably maintained and could still be subject to modulation. Because no counterpart of such a phenomenon has been described in other biological systems, the mechanism is still open to speculation, but seems likely to involve alteration in splicing. The dystrophin gene and expansion of RFs could be a unique model for exploring mechanisms in the evolution of alternative splicing.

Skeletal muscle has considerable regenerative capacity both in response to injury and in the recurrent dystrophic process. However, it is difficult to measure regenerative activity in vivo. The clonal expansion of RFs provides a unique tool for assessing the regenerative capacity of the muscle. Before the onset of muscle degeneration in mdx mice, dystrophin in nearly all RFs spans short segments of single fibers, corresponding in size to a single nuclear domain. The fact that these domains subsequently expand along the fiber and to neighboring fibers is, as we show here, best explained by the presence of an adjacent myogenic precursor harboring the revertant event. One year later, RF clusters have grown up to 100 fibers across and over 1 mm in length (Lu et al., 2000). From counts of myonuclear density, we estimate that a cluster of such size contains some thousands of myonuclei. The clonal nature of a revertant cluster implies that most of these thousands of myonuclei are the offspring of the single myogenic cell that was present at birth. The presence of myogenic cells with such mobility, and sufficient capacity to produce a large number of fibers in adult muscles in vivo, fits well with the evidence from our recent myofiber implantation experiments (Collins et al., 2005). When isolated fibers were transplanted into host muscles, we demonstrated that a single myofiber bearing on average fewer than seven satellite cells can produce a cluster of some 300 fibers, resembling in many respects the clusters of RFs in mdx mice. Similarly, we have demonstrated previously that the majority of satellite cells lose proliferative capability after irradiation with 18 Gy, but that a tiny subset of radiation-resistant myogenic precursors, constituting probably in the region of 1% of the radiation-sensitive population, retains the ability to restitute the satellite cell population partially (Heslop et al., 2000). The appearance of RF clusters in notexin-treated pre-irradiated muscles suggests that some of the revertant myogenic cells belong to this radiation-resistant precursor population. Expansion of RFs in mdx mice therefore provides potentially useful models for studies of myogenic cells.

The link between regeneration and expansion of revertant clusters allows us to use the degree of expansion as a cumulative index of past pathological activity in dystrophic muscles. At present, the most commonly used histological indicator of disease progression in the mdx mouse is the progressive increase in the proportion of centrally nucleated muscle fibers. However, this index rises very rapidly after the onset of the dystrophy and reaches a plateau of ∼80% in about 6 weeks, becoming insensitive to any continuation of pathology beyond this point in the dystrophic mdx mouse, which is the most widely used model for testing of new therapies. By contrast, the increase in size of RF clusters continues into the second year (Lu et al., 2000), confirming that the disease remains active into the later life of this animal model. This should, in principle, make it useful as a means of evaluating the beneficial effects of any treatment designed to combat muscle fiber degeneration. This view is vindicated in the present work by the clear concordance between the effects of the micro-dystrophin and utrophin transgenes on the overall measures of pathological change and on the suppression of RF cluster expansion. Thus, the degree of expansion of the revertant clusters in the mdx mice is by far the best available histological indicator for assessment of long-term functional value by therapeutic transgenes or other treatments that are designed to combat muscle degeneration. It has the additional advantage that it might be applicable in the dystrophic dog, in which central nucleation is not a persistent feature of the dystrophic pathology.

It is also noteworthy that the majority of RFs were able to localize nNOS to the fiber membrane, whereas none of the micro-dystrophins nor the utrophin exhibited this function, supporting previous findings that the structure of part of the rod sequence is implicated in the association with nNOS (Wells et al., 2003). Because the RF number diminished in the background of the most effective micro-dystrophins, it would appear that nNOS function is not crucial to myofiber survival if other functions of dystrophin are adequately accomplished.

In summary, our studies show that initial formation of RFs in mdx mice is independent of dystrophic phenotype and that alternative splicing in myogenic stem cells is likely to be responsible for the production of revertant dystrophin and clonal expansion of RFs. The expansion of RFs is closely related to muscle regeneration and therefore the quantity and distribution of RFs is a useful indicator of the long-term functionality of therapeutic interventions. The study of revertant dystrophin and expansion of RFs in the mdx mouse could provide a unique model to understand the mechanisms for the regulation of alternative splicing.

This study used mdx mice, mdx mice transgenic for CS1, AX11 and M3 micro-dystrophin transgenes, mdx mice transgenic for full-length and mini-utrophin transgenes, and C57/BL10 and C57/BL6 mice as controls (Sakamoto et al., 2002; Vainzof et al., 1995; Tinsley et al., 1996; Tinsley et al., 1998).

The monoclonal antibody MANDRA1 was used to detect both RFs and micro-dystrophins (1:100). The polyclonal antibody against neuronal nitric oxide synthase (nNOS) (Santa Cruz) was also used (1:20).

P7 rabbit polyclonal antibody was raised against a polypeptide within exon 57 of the gene encoding dystrophin. P7 antibody detects only RFs and not the micro-dystrophins (1:1000). The anti-utrophin polyclonal antibody (G3; 1:25) was used for immunohistochemistry (Tinsley et al., 1998).

Acetone-fixed cryosections (6 μm) were incubated with primary antibody at 4°C Probes) was used as the secondary antibody (1:200). For immunofluorescent overnight. Alexa Fluor 488- or 594-labeled goat anti-rabbit IgG (H+L) (Molecular with a biotin-conjugated anti-rabbit IgG (1:200; Dako) for 60 minutes at room temperature, and treated with an avidin-biotin-horseradish peroxidase complex (Dako) for 30 minutes and visualized by 3,3′-diaminobenzidine tetrahydrochloride containing 0.015% hydrogen peroxide.

Muscle fibers were regarded as dystrophin positive only when more than half the membrane circumference was stained in cross-sections. RFs adjacent to each other were characterized as a single cluster. For closer comparison of RFs in mice of different groups, serial sections (6 μm thick) of TE muscle of every 100 μm were stained with antibodies. The maximal number of RFs, the number of revertant clusters, and the number of RFs in each cluster were counted and compared.

mdx mice were irradiated at the age of 3 weeks. Animals were anesthetized with subcutaneous injection of 50 μl of hypnorm (Janssen; fentanyl citrate, final concentration 0.79 mg/ml; fluanisone, final concentration 2.5 mg/ml) and hypnovel (Roche; midazolam, final concentration 1.25 mg/ml). Animals were restrained during irradiation, the legs were irradiated with 4, 10 and 18 Gy delivered by an IBL 637 cell irradiator as described previously (Gross et al., 1999). Non-irradiated contralateral legs and age-matched mdx were used as controls. At each of a series of time points, 1, 4, 12 weeks and 1 year after irradiation, mice were euthanized and muscles examined. For notexin treatment, mice were anesthetized as above and the TA muscles of their right legs that had been pre-irradiated with 18 Gy 4 weeks before were injured by injection of 10 μl of 10 μg/ml notexin (Heslop et al., 2000). The injected mice were killed 7 days after the notexin injection and its TA muscles taken for histopathology and immunohistochemistry. Three to eight samples were examined for each treatment regime and untreated controls.

The data between samples were compared using Student's t test. P<0.05 was considered statistically significant.

This work was supported by grants from the Medical Research Council, the Muscular Dystrophy Group of Great Britain, the Leopold Muller Foundation (UK), Grant-in-Aid for Scientific Research (KAKENHI), Grant-in-Aid for Japan Society of the Promotion of Science (JSPS) Fellows (Japan), Aktion Benni & Co EV (Germany) and the Neuromuscular/ALS Center (Carolinas Medical Center, Charlotte, NC). We thank G. E. Morris (MRIC Biochemistry Group, The North East Wales Institute, Wrexham, UK) for supplying the monoclonal antibodies. T.Y. is a research fellow of JSPS.