We have quantitatively assessed the levels of dystrophin and dystrophin-associated protein expression in 17 patients with Becker muscular dystrophy grouped by deletion according to currently studied exon skipping models for Duchenne muscular dystrophy. Using two independent methods, we demonstrate that all the patients with Becker muscular dystrophy in our cohort have reduced dystrophin expression levels of at least 40% of controls. A previous study of families with X-linked dilated cardiomyopathy determined that dystrophin levels of 30% of control can be sufficient to avoid muscle weakness in these families, although these patients expressed lower levels of normal dystrophin (Neri et al., 2007
). Thus, our data indicate that the production of internally deleted dystrophin proteins by exon skipping should provide a substantial benefit to patients with Duchenne muscular dystrophy.
Overall, patients in the model 51 group had higher dystrophin and dystrophin-associated protein complex levels than patients in the other two groups with significant differences in dystrophin and nNOS expression. This lends further support to the ongoing exon 51 skipping clinical trials for Duchenne muscular dystrophy (Kinali et al., 2009
; Cirak et al., 2011
; Goemans et al., 2011
). Intragroup variability was lowest with the model MS group, which is the only group to be comprised entirely of patients harbouring the same deletion (exons 45–55). However, moderate variability was observed in dystrophin and dystrophin-associated protein complex levels between patients of the same deletion.
The two different deletions within the model 51 group have comparable dystrophin and dystrophin-associated protein complex expression levels as well as a mild clinical phenotype. There is a greater variability in the clinical phenotypes between the patients of the model 53 group; our results also indicate a possible benefit for the retention of the nNOS-binding domain in internally deleted dystrophin.
We observed variability in dystrophin protein levels between patients with the same deletion; this could stem from the differences of individual intronic breakpoints in patients, which may affect alternative splicing and/or translation efficiency. Other explanations for intra and intergroup variability could be the differential stability of the internally deleted dystrophins (Krieger et al., 2010
; Henderson et al., 2011
) and/or the endogenous splicing of other exons. For example, exon 44 is known to spontaneously skip in patients with deletions of the surrounding exons (van Vliet et al., 2008
). Although we have not measured this in our population, its occurrence could affect the levels of the dystrophin protein observed. Other genetic modifiers of dystrophin translation efficiency could also be involved in these discrepancies; for example a mutation in the promoter region of the gene encoding the extracellular matrix protein, osteopontin, identified osteopontin genotype as a genetic modifier of Duchenne muscular dystrophy severity (Pegoraro et al., 2011
). Additionally the microRNA, miR-31, was recently shown to repress dystrophin expression through binding to the 3′ untranslated region of dystrophin RNA (Cacchiarelli et al., 2011
The model 51 group contained three out of the four asymptomatic patients, while the model 53 group consisted of severe, mild and asymptomatic patients; thus there is no clear-cut genotype–phenotype correlation in this latter group of patients with Becker muscular dystrophy. We do, however, demonstrate for the first time that in Becker muscular dystrophy, higher expression levels of dystrophin, β-dystroglycan and nNOS are significantly associated with a milder skeletal muscle phenotype.
Our results suggest that the milder phenotypes in our cohort of patients with Becker muscular dystrophy are associated with higher expression levels of dystrophin and those dystrophin-associated protein complex members that bind directly to dystrophin. We found that β-dystroglycan and nNOS expression correlates with clinical severity, while α-sarcoglycan expression does not. This could be due to the direct binding of β-dystroglycan and nNOS to dystrophin; while α-sarcoglycan binds dystrophin indirectly (Ervasti, 2007
; Ozawa, 2010
). Our findings are further supported by mouse model data that demonstrate that the skipping of exons that preserve the nNOS-binding domain results in a more favourable clinical outcome (Lai et al., 2009
); this information should be considered when developing exon skipping strategies.
While it is generally established that the 45–55 deletion is associated with a favourable prognosis, this deletion is also associated with X-linked dilated cardiomyopathy (Beroud et al., 2007
; Nakamura et al., 2008
; Ferreiro et al., 2009
; Miyazaki et al., 2009
). In our study, the four patients with the 45–55 deletion were classified as mild and none had developed dilated cardiomyopathy, including one patient who died at age 76.
In fact, a large percentage of patients with Becker muscular dystrophy develop dilated cardiomyopathy (Melacini et al., 1993
; Bushby et al., 2003
) and mutations that disrupt the phasing of the helical spectrin repeats encoded by exons 45–49 are thought to lead to an earlier-onset of dilated cardiomyopathy as a result of altered dystrophin structure (Kaspar et al., 2009
). The majority of the internally deleted dystrophin proteins studied within our cohort additionally remove the proline-rich hinge 3 region encoded by exons 50 and 51; the hinge 3 region is thought to be helical but does not necessarily form a stable tertiary structure (Bhasin et al., 2005
). One study designed nano-constructs of dystrophin based on in-frame Becker muscular dystrophy and therapeutic deletions including exons 45–51 (Krieger et al., 2010
); our cohort included three patients with this deletion in the model 51 group. The secondary structure prediction for the abnormal ‘linker’ created by the joining of exons 44 to 52 revealed that this region folds into a stable repeat domain that maintains its helicity (Krieger et al., 2010
). These findings and our results provide further support and guidance for exon skipping within the rod domain. However, a recent study has demonstrated that the removal of hinge 2 through to spectrin-like repeat 19 causes a significant loss in protein stability (Henderson et al., 2011
); thus caution is needed when removing large segments of the rod domain such as with multi-exon skipping models. Multi-exon skipping is also technically very challenging; only very minimal skipping has been detected at low frequency in preliminary cell culture work with a cocktail of antisense oligonucleotides for skipping of exons 45–55 (van Vliet et al., 2008
It has been hypothesized that patients with Becker muscular dystrophy with rod domain deletions suffer muscle cramps following exercise due to the mislocalization of nNOS from the sarcolemma with subsequent disruption of blood flow during exercise (Thomas et al., 1998
; Sander et al., 2000
; Kobayashi et al., 2008
). Our quantitative analysis of 17 patients with Becker muscular dystrophy does not completely support this hypothesis. Firstly, several patients with deleted nNOS-binding domains and/or low sarcolemmal expression levels do not suffer cramps and perhaps more convincingly, Becker muscular dystrophy Patients 5, 7 and 8 report cramps even though their deletions do not span the nNOS-binding domain. In fact, Patients 5 and 8 have near normal nNOS expression levels. The aetiology of cramps in Becker muscular dystrophy is complex and not necessarily or exclusively related to blood flow, but to physical exercise. Thus, mild individuals with preserved strength may have frequent cramps because they endure more physical activity, on the other hand growth and development may also play a role. A limitation of our conclusion is the relatively young age of a number of the patients, and we cannot rule out the possibility that some patients might develop cramps later on in life. Moreover, the recruitment of nNOS to the sarcolemma might also be affected by structural abnormalities located outside the nNOS-binding domain. Indeed, a complex correlation between sarcolemmal nNOS localization and muscle activity, oxidative stress and calcium signalling has been recently demonstrated (Pietri-Rouxel et al., 2010
; Finanger Hedderick et al., 2011
), suggesting the presence of multiple independent regulators of nNOS localization in muscle.
In summary, we demonstrate that all patients with the genotypes we have studied in this report had Becker muscular dystrophy and not Duchenne muscular dystrophy, in keeping with the reading frame hypothesis. While this not surprising, concerns that a significant number of patients with these genotypes could have Duchenne muscular dystrophy or severe Becker muscular dystrophy have been raised, although the source of the information relied on old studies when multiplex ligation-dependent probe amplification was not available and the end-points of the deletions were not systematically assessed (Yokota et al., 2009
). In fact, a more recent analysis of the same Leiden database used by Yokota et al. (2009)
has demonstrated that all patients with in-frame 45–51 or 50–51 deletions have a mild Becker muscular dystrophy phenotype (Helderman-van den Enden et al., 2010
). Our data are further supported by two other reports; firstly Beroud et al. (2007)
reported on the mild or asymptomatic phenotypes of 15 patients with a 45–55 in-frame deletion and secondly a study that focused on dilated cardiomyopathy in Becker muscular dystrophy reported the mild Becker muscular dystrophy phenotype of 24 patients with in-frame deletions that fit into one of the three models used in our study (Kaspar et al., 2009
We show that patients with Becker muscular dystrophy who express the same internally deleted dystrophin as could be induced by exon skipping therapies are mostly associated with mild phenotypes and express dystrophin at a high enough level (at least 40% of control) to provide a functional benefit to patients with Duchenne muscular dystrophy. We report that the asymptomatic Becker muscular dystrophy phenotype is associated with significantly higher dystrophin, β-dystroglycan and nNOS expression levels than the mild phenotype and highlight exon 51 as an optimal target exon for removal. This information is encouraging, as the dystrophin levels obtained in at least some of the patients recruited into the recently completed systemic clinical trials had levels of 15% (Goemans et al., 2011
) and 18% (Cirak et al., 2011
) of normal levels; the latter representing ~45% efficacy during a short 12 week study, suggesting that a significant clinical benefit from these dystrophins is a realistic possibility.