DHPR α1S subunit controls skeletal muscle mass and morphogenesis

Abstract

The α1S subunit has a dual function in skeletal muscle: it forms the L-type Ca²⁺ channel in T-tubules and is the voltage sensor of excitation–contraction coupling at the level of triads. It has been proposed that L-type Ca²⁺ channels might also be voltage-gated sensors linked to transcriptional activity controlling differentiation. By using the U7-exon skipping strategy, we have achieved long-lasting downregulation of α1S in adult skeletal muscle. Treated muscles underwent massive atrophy while still displaying significant amounts of α1S in the tubular system and being not paralysed. This atrophy implicated the autophagy pathway, which was triggered by neuronal nitric oxide synthase redistribution, activation of FoxO3A, upregulation of autophagy-related genes and autophagosome formation. Subcellular investigations showed that this atrophy was correlated with the disappearance of a minor fraction of α1S located throughout the sarcolemma. Our results reveal for the first time that this sarcolemmal fraction could have a role in a signalling pathway determining muscle anabolic or catabolic state and might act as a molecular sensor of muscle activity.

Introduction

Skeletal muscle growth is dependent on innervation and exercise. Prolonged periods of muscle inactivity because of bed rest, hindlimb unloading, immobilization, microgravity or denervation result in a significant muscle atrophy (Borisov et al, 2001; Adams et al, 2003; Dedkov et al, 2003; De-Doncker et al, 2005). Direct electrical stimulations can, to a large extent, substitute for innervation and preserve or even restore almost normal properties in denervated muscle fibres (Kern et al, 2004; Squecco et al, 2008). Nevertheless, mechanisms implicated in muscle homeostasis under such circumstances are poorly understood. These molecular mechanisms are likely triggered by action potentials or prompted by the contractile function itself. In neurons, L-type Ca²⁺ channels have been shown to act as voltage-gated sensors linked to transcriptional activity controlling differentiation (Wheeler et al, 2008). In skeletal muscle, L-type Ca²⁺ channels are made of five subunits (α1S, α2, β, γ, δ). The α1S, known as the dihydropyridine receptor (DHPR), is a four-repeat transmembrane protein, which contains the basic functional elements of the channel, including the Ca²⁺ selective pore and the S4 ‘voltage-sensing' transmembrane segment (Ahern et al, 2001). It is predominantly located in T-tubules, although a minor fraction of DHPR has been found on the sarcolemma (Jorgensen et al, 1989). α1S is also located in triads where it interacts with the type 1 ryanodine sensitive Ca²⁺ release channel (RyR1) and forms the voltage sensor of the excitation–contraction (EC) coupling.

We hypothesize that subtle modification of α1S contents in adult muscle could allow to decipher unexpected functions for this multi-functional protein. Indeed, complete knockout of the α1S subunit during development is lethal as observed in the mdg/mdg dysgenic mouse model (Pai, 1965; Banker, 1977; Chaudhari, 1992). To create a viable model, which would overcome this problem, we have chosen a strategy for long-lasting RNA interference based on the use of U7 snRNA chimaeras. The U7 system has been successfully used in several paradigms for rescuing mutated mRNAs by either prompting exon skipping or exon inclusion to restore protein synthesis (Goyenvalle et al, 2004; Marquis et al, 2007). Here, we used this system to decline amounts of α1S subunit by forced skipping of an exon affecting its mRNA reading frame. To guarantee long-lasting effects, the U7 snRNA chimaeras were delivered intramuscularly by using adenovirus-associated viral vectors (AAV). After 6 months, treated muscles showed a significant atrophy even though muscle fibres were not EC uncoupled and were still capable to contract when stimulated. We also show that decreasing α1S induced redistribution of the neuronal nitric oxide synthase (nNOS), activation of FoxO3A, increased expression of key autophagy-related genes and initiated autophagosome formation. Finally, confocal investigations allowed to identify which subcellular compartment displayed the decrease of α1S after exon skipping treatment. We found that atrophy and ultrastructural disorganization of muscle fibres correlated with α1S disappearance from the sarcolemma while α1S remained present in T-tubules and triads. All together, these results suggest that the minor sarcolemmal fraction of α1S has a role in a signalling pathway controlling muscle maintenance. At that level, α1S might act as a sensor of muscle activity in charge of controlling muscle mass and morphogenesis.

Results

Knockdown of the α1S subunit

The α1S subunit is encoded by the Cacna1s gene, made of 46 exons and located, in mouse, on chromosome 1 (Drouet et al, 1993). Exon 16 encodes an amino acid sequence forming a cytoplasmic loop (II–III loop) connecting trans-membrane repeats II and III of the α1S subunit, which interacts directly with RyR1 and is crucial for skeletal muscle EC coupling (Tanabe et al, 1990; Casarotto et al, 2006; Cui et al, 2009). To knock down the α1S subunit we have designed U7 snRNA chimaeras interfering specifically with its pre-mRNA splicing. We have chosen to target exon splicing enhancers (ESE) and the acceptor splice site (SA) of exon 16 (U7-ESE, U7-SA). A U7-Ctrl, not affecting α1S splicing, was used as mock control (Figure 1A). Engineered U7 genes were cloned into AAV2 vector backbones, which were type 1 pseudotyped for intramuscular delivery: AAV2/1 (U7-ESE), AAV2/1 (U7-SA), AAV2/1 (U7-Ctrl). To improve skipping of exon 16, both AAV2/1 (U7-ESE) and AAV2/1 (U7-SA) vectors were co-injected into tibialis anterior muscles (TA^ΔDHPR). Contralateral muscles were used as controls (TA^Ctrl) and received mock AAV2/1 (U7-Ctrl). To show that these genomes were still active and produced the interfering U7 transcripts at 6 months post injection, total RNA from TA^ΔDHPR muscles was extracted, reverse transcribed into DNA and PCR amplified using a forward primer in the antisense sequence (ESE) and a reverse primer in the U7 stem loop. Results confirmed that vector expression was persistent up to 6 months post injection (Supplementary Figure S1).

Figure 1.

Generation of a local knockdown of α1S subunit. (A) Schema of exonic chaining of pre-mRNA, exons 15, 16 and 17 are represented by blue boxes. Sequences of exon acceptor site SA (in blue) and exon splicing enhancer ESE (in black) of exon 16 were targeted by antisense sequences introduced in U7smOPT cassettes. (B) qRT–PCR analysis using primers overlapping exons 15 and 16 permitted to quantify the non-skipped α1S form: 18±6% (^**P⩽0.001, n=19) and 10±7% (^**P⩽0.001 n=19) (black bars) of the total α1S mRNA (grey bars) at 2 and 6 months post injection in TA of AAV1-(U7-ESE) and AAV1-(U7-SA):ΔDHPR or AAV1-(U7-Ctrl): c as a control, respectively. (C) Six months post injection, lysates from TA^ΔDHPR (ΔDHPR) and TA^Ctrl (c) were immuno-blotted for α1S or α-actin for four mice. Graph depicts mean±s.e.m. of relative expression of α1S subunit determined by densitrometry and nomalized to the α-actin expression for each muscle. Results were expressed in protein levels of α1S subunit in TA^ΔDHPR normalized to TA^Ctrl for each mice, ^**P⩽0.001, n=4. (D) Longitudinal cryo-sections from TA^ΔDHPR (ΔDHPR) and TA^Ctrl (c) were stained with anti-α1S subunit (red) and anti-laminin (green) antibodies, nuclei were visualized by Dapi (blue) and imaged by confocal microscopy. Bars represent 20 μm.

Skipping exon 16 would allow abortion of α1S synthesis by provoking a shift of its mRNA reading frame creating a premature stop codon in position 33 of exon 17 (Figure 1A). Skipping efficiency was assessed by real-time quantitative RT–PCR (qRT–PCR) on total mRNA (Supplementary Table S1). Such analysis showed that levels of the unskipped α1S represented 18±6% (P⩽0.001, n=19) and 10±7% (P⩽0.001 n=19) of the total α1S mRNAs at 2 and 6 months post injection, respectively (Figure 1B). Western blotting revealed a significant decrease of α1S. Six months post injection, α1S levels in TA^ΔDHPR were about 56±13% (P⩽0.001, n=5) compared with the contralateral TA^Ctrl (Figure 1C). It is unlikely that skipped mRNAs could still be translated into a truncated protein. Indeed, such a truncated protein should display a theoretical molecular weight of about 80 kDa. Western blots using polyclonal antibodies made against the N-terminal protein never detected such a band (Supplementary Figure S2). Levels of α1S were also assessed by immuno-fluorescence staining on TA^ΔDHPR muscle sections. At 6 months, the majority of fibres showed a decrease in staining intensity (Figure 1D).

Atrophy, fibrosis and reduced maximal force after α1S knockdown

A significant atrophy was observed in injected TA^ΔDHPR. Six months after injection, TA^ΔDHPR showed a loss of 41±3% of muscle mass compared to contralateral TA^Ctrl (P⩽0.001, n=8) (Figure 2A; Table 1). Morphometric analysis showed that there was no change in the number of fibres per muscle. Atrophy was due to a decrease of the size of muscle fibres (Table 1) as shown by the distribution of cross-section diameters, which was shifted towards lower values (Figure 2B). Average fibre diameter was 24±2 μm in injected TA^ΔDHPR compared to 43±4 μm in contralateral TA^Ctrl (Table 1; Figure 2C). This atrophy was not accompanied by a modification of muscle fibre phenotype. Typically, the TA was almost entirely composed of fast-twitch fibres, expressing type 2b and 2x myosin heavy chains (MHC) isoforms and a very small number of muscle fibres expressing the MHC-2a isoform (Raffaello et al, 2006). No significant modifications were observed in the number of fibres expressing MHC-2x and -2b showing that the major muscle phenotype was maintained in treated muscle compared to contralateral (Table 1). However, there was a slight decrease in the number of fibres expressing MHC-2a in TA^ΔDHPR compared to TA^Ctrl (Table 1). Still, TA^ΔDHPR muscles showed striking morphological abnormalities without evidence of necrosis or core formation. There was no sign of infiltration by scavenger cells, which is a hallmark of necrosis, nor positive immuno-labelling for neonatal and embryonic MHCs (data not shown), a more sensitive marker of regeneration given that this isoform is expressed during muscle development and particularly the formation of new fibres from satellite cells (Ecob-Prince et al, 1986; Mouly et al, 1993). In treated muscles (TA^ΔDHPR), subsequent atrophy was associated with fibrosis (Figure 2C) as shown by morphometric analysis using Sirius red for staining collagen deposition (Figure 2D). Injected TA^ΔDHPR muscles displayed a four-fold increase in fibrosis compared with the contralateral muscles (Figure 2E). Muscle function was assessed by the measurement of in situ muscle force in response to nerve stimulation, as described earlier (Vignaud et al, 2005a). Compared with values measured in the contralateral limb, absolute and specific maximal isometric tetanic force measured in injected TA^ΔDHPR were shown to be reduced by 62% (40±11 g compared to 106±20 g, P⩽0.001, n=8) and 34% (1.1±0.3 g/mg compared to 1.6±0.2 g/mg, P⩽0.05, n=8), respectively (Table 1).

Figure 2.

Consequence of α1S subunit knockdown. (A) Six months post injection, TAs from eight mice were dissected and weighed. (B) The internal diameters (shortest diameter) from all fibres throughout the total muscle section were recorded and analysed. Muscles from five different animals were examined. The bar graph presents mean±s.e.m. of the number of myofibres by fibre diameter class for TA^ΔDHPR (black) and TA^Ctrl (grey). (C) Transversal sections of TA^ΔDHPR (ΔDHPR) and TA^Ctrl (c) were stained with haematoxylin and eosin, bars represent 100 μm. (D) To quantify fibrosis, transversal sections of total muscle were stained with Red Sirius and quantified using Histolab Software (marked in blue), data were normalized with total surface of each muscle (orange line surrounding the sections), bars represent 500 μm. (E) Quantification is presented in bar graph and showed 4.1±0.1 fold increase of fibrosis in TA^ΔDHPR (ΔDHPR) compared to TA^Ctrl (c) (^**P<0.001, n=4).

Table 1.

Morphometric and functional characteristic of TA^ΔDHPR

	TACtrl	TAΔDHPR
Mass (mg)	65±8	39±3**
Total number of fibres	3405±161	3249±443
Mean of CSA (mm)	43±4	25±2**
MHC composition (%)
MHC-1	0.09±0.03	0.03±0.03
MHC-1/2a	0.00	0.02±0.01
MHC-2a	1.5±0.7	0.3±0.06*
MHC-2a/2x	2.7±0.4	1.08±0.50
MHC-2x	38.90±1.60	33.50±3.50
MHC-2x/2b	16.40±3.90	16.50±1.30
MHC-2b	40.40±3.50	47.80±4.80
Absolute MTI force (g)	106±20	40±11**
Specific MTI force (g/mg)	1.6±0.2	1.1±0.3*
Investigations 6 months post injection of AAV1 (U7-ESE) and AAV1 (U7-SA) into the TA (TAΔDHPR), or AAV1 (U7-Ctrl) (TACtrl) were carried out; n=8 for mass, total number of fibres, mean of cross-section area (CSA), absolute and maximal tetanic isometric (MTI) force investigations; and n=3 for MHC composition analyses. Values are mean±s.e.m. and were significatively different for TAΔDHPR from TACtrl, **P⩽0.001; *P⩽0.05.

Altogether, these data show that α1S knockdown in normal muscle resulted in important morphological changes involving atrophy and fibrosis. These alterations were not caused by muscle paralysis although fibre dismantling ultimately affected muscle function.

α1S knockdown induces nNOS redistribution, FoxO3A nuclear translocation and upregulation of autophagy-related gene expression

It has been proposed that nNOS, a molecular partner of the dystrophin–glycoprotein complex, has a role in the mechanism of atrophy in denervation-induced and hindlimb unloading models (Suzuki et al, 2007). In TA^ΔDHPR muscles, levels of nNOS mRNA were found to be reduced by 41±1% (P⩽0.001, n=9) (Figure 3A), a significant decrease that was confirmed at the protein level by western blot analysis (Supplementary Figure S3A). Immuno-histochemical analysis revealed that nNOS staining shifted from the sarcolemma to the cytoplasm (Figure 3B), a redistribution that was not due to the disruption of the dystrophin–glycoprotein complex, which was not affected in TA^ΔDHPR muscles (see stainings for dystrophin and associated α, β and γ sarcoglycans in Supplementary Figure S4). We evaluated NOS activity by using NADPH-d histochemical analysis (Chen et al, 2008). As expected, NOS activity was redistributed towards the cytoplasm in TA^ΔDHPR atrophic fibres (Supplementary Figure S3B).

Figure 3.

nNOS membrane dissociation, FoxO3A accumulation in myonuclei and upregulation of autophagy-related genes as consequences of α1S subunit loss. (A) mRNA from TA^ΔDHPR and TA^Ctrl tissues were extracted and nNOS expression was quantified by qRT–PCR. Results are expressed as mean±s.e.m., ^**P<0.001, n=4. (B) Transversal cryo-sections of TA^ΔDHPR (ΔDHPR) and TA^Ctrl (c) were stained with anti-nNOS (red), anti-laminin (green) antibodies, nuclei with Dapi (blue) and imaged by confocal microscopy. Bars represent 20 μm. (C) The tissues from extracts were analysed by western blot with FoxO3a-P antibody and normalized with α-actin antibody. Graph depicts mean±s.e.m. of relative expression of FoxO3a-P determined by densitometry and nomalized to the α-actin expression for each muscle, ^*P⩽0.005, n=3. (D) Longitudinal cryo-sections of TA^ΔDHPR (ΔDHPR) and TA^Ctrl (c) were stained with anti-FoxO3a (red), anti-laminin (green) antibodies, nuclei with Dapi (blue) and imaged by confocal microscopy. Bars represent 20 μm. (E) mRNA from TA^ΔDHPR and TA^Ctrl tissues were extracted and regulation of autophagy genes expression was followed by qRT–PCR. Bnip, CathepsinL, LC3 and PI3KIII expression (noted in red) were significantly increased in TA^ΔDHPR compared with the controlateral TA^Ctrl, ^**P<0.001, n=4.

It has been shown in tail suspension experimental models that nNOS redistribution induced muscle atrophy through regulation of FoxO3A (Suzuki et al, 2007), a pivotal transcription factor involved in muscle mass regulation (Mammucari et al, 2007). Phosphorylated forms of FoxO3A are found in the cytoplasmic compartment, whereas dephosphorylated forms translocate to the nucleus to activate transcription of target genes (Cao et al, 2005). Western blot analysis of TA^ΔDHPR muscles using an antibody recognizing the phosphorylated form showed that FoxO3A was mainly dephosphorylated (Figure 3C). Concurrently, immuno-staining of TA^ΔDHPR muscle sections with an antibody recognizing the non-phosphorylated form of FoxO3A showed its accumulation in myonuclei (Figure 3D), thus confirming its translocation after activation.

It is known that the ubiquitin-proteasomal pathway is markedly increased in atrophying muscle because of transcriptional activation of ubiquitin, several proteasomal subunit genes and the two muscle-specific ubiquitin ligases atrogin-1/MAFbx and MuRF1 (Cao et al, 2005). In TA^ΔDHPR muscle extracts, qRT–PCR revealed that atrogin-1/MAFbx and MuRF1 were not over-expressed at 6 month post injection (Figure 3E). Moreover, atrogin-1/MAFbx and MuRF1 were not even transiently upregulated before this time point (expression of these genes was measured 1, 2, and 4 months post injection) suggesting that the ubiquitin-proteasome pathway was not involved in TA^ΔDHPR induced atrophy. We then inquired whether the induction of atrophy was accompanied by an upregulation of autophagy-related genes, LC3, Gabarapl1, Atg4b and Atg12; the genes involved in regulating autophagy Vps34 (a class III PI3K), Bnip and finally the lysosomal proteinase cathepsinL gene (Kelekar, 2008). By using qRT–PCR, we showed that Bnip3, LC3, PI3KIII and CathepsinL expression was significantly increased in TA^ΔDHPR compared to contralateral TA^Ctrl (Figure 3E).

These results indicate that knockdown of α1S levels induced atrophy through the autophagy pathway that was most probably triggered by nNOS redistribution, FoxO3A translocation to myonuclei and activation of autophagy-related gene expression.

Fall of α1S induces autophagosome formation and alters morphology of the sarcoplasmic reticulum

Autophagic activity and presence of autophagic vacuoles were revealed at the ultrastructural level by using electron microscopy (EM). Characteristic double-membrane-limited structures corresponding to autophagosomes were observed in atrophic TA^ΔDHPR muscle fibres (Figure 4A). Autolysosome structures, characterized by highly electron dense signals and corresponding to the fusion of autophagosomes and lysosomes, were frequently observed. The occurrence of autophagosomes was also assessed on tissue sections and isolated fibres from TA^ΔDHPR muscles by using LC3 immuno-staining, a specific marker for autophagosome formation (Mizushima et al, 2004; Mizushima and Kuma, 2008). We found that LC3 staining on TA^ΔDHPR transverse sections labelled small vacuolar structures surrounded by dystrophin (Figure 4B). We then carried out double staining on isolated fibres by using antibodies against LC3 and the polyubiquitin-binding protein p62 (also known as sequestosome SQSTM1), which serves as an adaptor molecule mediating the degradation of polyubiquitinated proteins by the autophagic pathway (Raben et al, 2008). Confocal analysis showed widespread rising of colocalized LC3 and p62 staining all along ΔDHPR atrophied fibres, whereas control fibres did not displayed such a phenomenon (Figure 4C).

Figure 4.

Formation of autophagosomes in TA^ΔDHPR. (A) Ultrathin sections were imaged by electron microscopy, M, mitochondria; C, collagen fibres; T, T-tubule. Arrows show double-membrane vesicules called auphagosomes. (B) Transversal cryo-sections were stained with anti-LC3b (red), anti-dystrophin (green) antibodies, nuclei with Dapi (blue) and imaged by confocal microscopy. Upper panel: TA^Ctrl (Ctrl) and lower panel: TA^ΔDHPR (ΔDHPR). Bars represent 20 μm. (C) Myofibres isolated from control (Ctrl) or 6 months post-injected FDB (ΔDHPR) muscles were processed for immuno-fluorescent labelling for P62 (green) and LC3b (red) and imaged by confocal microscopy. Scale bars, 10 μm.

EM analysis of TA^ΔDHPR muscles also revealed that sarcomeres were smaller and less organized than controls, although Z and M lines were still properly aligned. Triads were correctly located and displayed electron dense structures enclosing the EC coupling machinery, but most of the terminal SR cisternae were unusually dilated and the longitudinal SR displayed unusual circumvolutions and invaginations (Figure 5A). Confocal analysis by using antibodies against SERCA, a marker of longitudinal SR, and RyR1, a marker of triads, confirmed this SR disorganization at the level of whole isolated fibres (Figure 5B). Western blot analysis showed that levels of RyR1 were only slightly reduced in TA^ΔDHPR muscles (Figure 6A). Triadin, a protein associated to the EC complex, was also not significantly affected (data not shown). Confocal double stainings showed that α1S and RyR1 were still co-localized in atrophic TA^ΔDHPR fibres (Figure 6B), a result in agreement with EM observations and the fact that atrophic TA^ΔDHPR fibres were still EC coupled.

Figure 5.

Impact of the α1S subunit knockdown in maintenance of muscle ultrastructure. (A) Ultrathin sections of TA^ΔDHPR (ΔDHPR) and TA^Ctrl (c) were imaged by electron microscopy. Cis, terminal cisternae; T, tubule; SR, sarcoplasmic reticulum. Bars: 500 nm. (B) Myofibres isolated from Ctrl or ΔDHPR FDB muscles 6 months post injection were processed for immuno-fluorescent labelling for RyR1 (green), SERCA (red), Dapi (blue) and imaged by confocal microscopy. Scale bars, 10 μm.

Figure 6.

α1S subunit knockdown and RyR1 expression. (A) Lysates from TA^ΔDHPR (ΔDHPR) and TA^Ctrl (c) were immuno-blotted for RyR1 or α-actin. (n=4). Graph depicts mean±s.e.m. of relative expression of RyR1 determined by densitometry and nomalized to the α-actin expression for each muscle. Results were expressed in protein levels of RyR1 in TA^ΔDHPR normalized to TA^Ctrl for each mice, ^*P⩽0.005, n=4. (B) Longitudinal cryo-sections of TA^ΔDHPR (ΔDHPR) and TA^Ctrl (c) were stained with anti-α1S subunit (red), anti-RyR1 (green) antibodies and imaged by confocal microscopy. Bars=20 μm.

Atrophy in ΔDHPR muscle fibres correlates with the disappearance of the sarcolemmal fraction of α1S

As some α1S expression remained preserved in exon skipping-treated muscles undergoing atrophy, we wondered in which compartment α1S was really decreased. To address this point, we have investigated more deeply α1S localization in treated muscles, particularly when atrophy was clearly established. This study was carried out by 3D confocal analysis on isolated flexor digitorum brevis (FDB) muscle fibres (Figure 7; Supplementary Movies S5A and S5B). In control fibres, α1S was chiefly found in the inner compartment encompassing T-tubules and triads. However, 3D-projection analysis revealed that a minor fraction of α1S was present on the sarcolemma as already shown by using EM techniques and suggested by subcellular fractionation (Jorgensen et al, 1989; Munoz et al, 1995; Vassilopoulos et al, 2009). In exon skipping-treated fibres undergoing atrophy, α1S was still present in the inner compartments and displayed a quite preserved striated distribution. The major effect concerned the sarcolemmal α1S fraction, which was missing in treated fibres.

Figure 7.

Localization of α 1S subunit. Whole skeletal muscle fibres enzymatically isolated from Ctrl (A–D) or ΔDHPR (E–H) FDB muscles 6 months post injection were processed for immuno-fluorescent labelling for α1 S subunit (red) and laminin (green). Scale bars, 10 μm; Arrows indicate α1 S subunit expression on sarcolemma; antibody labelling was visualized by serial confocal microscopy and represented as movies of the optical sections (Supplementary Movies S5A and S5B). (D, H) Present projections of confocal Z-series (step between each frame is 1 μm) along XZ and YZ planes as indicated.

Discussion

α1S knockdown in skeletal muscle induces atrophy

In skeletal muscle, it is well known that α1S has a dual function: it forms the L-type Ca²⁺ channel in T-tubules and is the voltage sensor of EC coupling at the level of triads. Muscular dysgenesis (mdg/mdg), a null mutation on the Cacna1s gene encoding α1S, has been described in the mouse (Pai, 1965; Platzer and Gluecksohn-Waelsch, 1972; Chaudhari, 1992). In mdg/mdg, skeletal muscles are paralysed because of a complete lack of EC coupling and muscle fibres never fully differentiate: and show a poor index of fusion, reduced basal lamina, sarcomere disorganization and triad misfolding (Bournaud and Mallart, 1987; Rieger et al, 1987; Tanabe et al, 1988). As animals died at birth, it was impossible to discriminate additional roles for this protein, particularly in adult healthy muscle, and subsequent studies mainly focused on the role of α1S during skeletal muscle development (Tanaka et al, 2000). Here, by using an exon skipping approach (Goyenvalle et al, 2004), we could, for the first time, achieve loco-regional long-lasting downregulation of α1S in a normal well established adult muscle without impairing the quality of life of treated animals. This strategy allowed silencing α1S mRNAs by about 80% at 2 months after intramuscular injection of AAV2/1 vectors harbouring U7 chimaeras designed to disrupt its reading frame. Levels of U7 chimaeras were stable over time, thereby maintaining and improving α1S silencing. Surprisingly, no significant decay in α1S protein was observed until 6 months of treatment (its expression was tested by western blot at 2 and 4 months post injection) suggesting that α1S has a natural turnover roughly between 2 and 4 months. At 6 months, treated fibres displayed about a 56% loss of α1S protein. Interestingly, this fall did not abrogate either the absolute or the specific tetanic force, which was only slightly decreased, indicating that EC coupling was still effective. Despite these results, the foremost surprising finding was that treated fibres underwent massive atrophy. The fact that α1S knockdown was far from being complete confirmed that subsequent effects were not trivial consequences of a major defect affecting primary functions of muscle fibres such as action potential genesis, EC coupling, contraction or even innervation.

It is known that atrophy caused by experimental denervation induces modifications of fibre MHC phenotype in both fast and slow type muscles. In denerveted rat soleus and TA, both slow and fast fibres undergo a rapid atrophy (Dedkov et al, 2003), which is associated with a marked transformation towards a fast phenotype (Huey and Bodine, 1998; Loughna and Morgan, 1999). In our model, ΔDHPR-induced atrophy was not associated with a variation of the total number of fibres. Rather, morphometric analysis indicated that all fibres shifted towards a smaller diameter and no modification of MHC-2x and -2b distribution was observed.

ΔDHPR atrophy is mediated by autophagy

Hindlimb unloading is widely used for studying muscle atrophy because of muscle inactivity while innervation is preserved (De-Doncker et al, 2005; Suzuki et al, 2007). In this model, redistribution of nNOS has been described to mediate muscle atrophy through regulation of FoxO3A (Suzuki et al, 2007). In our model of ΔDHPR-induced atrophy, nNOS was also redistributed to the cytoplasm and activated FoxO3A was found translocated into muscle nuclei, suggesting a similar atrophic pathway. Activation of FoxO3A has been shown to be essential for fibre atrophy on denervation and fasting (Sandri et al, 2004, 2006) through the ubiquitin/proteasome pathway by upregulating two crucial ubiquitin ligase genes encoding atrogin-1 and MuRF1 (Sandri et al, 2004). FoxO3A can also control atrophy independently of the ubiquitin-proteasomal pathway through the autophagic/lysosomal pathway. Indeed, it has been shown that constitutive over-expression of active FoxO3A stimulated the autophagic process selectively (Zhao et al, 2008). Here, we did not observe atrogin-1 or MuRF1 over-expression. Therefore, it was unlikely that α1S-knockdown muscle atrophy involved the canonical ubiquitin-proteasomal pathway. Rather, it involved the autophagic pathway as confirmed by the upregulation of several autophagy-related genes such as Bnip3, LC3, PI3KIII and CathepsinL (Zhao et al, 2007). Moreover, LC3 protein was observed in large vacuolar structures expressing dystrophin, a hallmark of a select group of autophagic vacuolar myopathies including Danon disease and X-linked myopathy with excessive autophagy (Malicdan et al, 2008). The precise relevance of these structures and the mechanism by which they are formed remain to be clarified. Autophagy is the engulfment of cytosol and organelles by double-membrane vesicles termed autophagosomes. Autophagosome formation is known to occur near the endoplasmic reticulum (ER). In live imaging experiments, a vesicular punctuate compartment is in dynamic equilibrium with the ER and could provide a membrane platform for accumulation of autophagosomal proteins, expansion of autophagosomal membranes and emergence of fully formed autophagosomes (Xie and Klionsky, 2007). At the ultrastructural level, ΔDHPR fibres displayed many double-membrane-limited structures characteristic of autophagosomes in addition to disorganization of the SR, which displayed abnormal circumvolutions and invaginations. This disorganization could suggest that the SR might also be a donor for membranes required to form autophagosomes.

Underlying mechanism for muscle atrophy induced by α1S knockdown

The α1S subunit has a dual function depending on its localization. It is the voltage sensor for EC coupling in triads whereas it forms the L-type Ca²⁺ channel along T-tubules (Romey et al, 1988). The biological significance of the L-type Ca²⁺ channel is still unknown, although Bannister et al (2009) recently proposed it could contribute to calcium entry involved in the maintenance of myoplasmic calcium levels during either sustained depolarization or during trains of repetitive brief stimuli. It has also been reported in neurons that L-type Ca²⁺ channels could modulate transcriptional activity through EC coupling (Wheeler et al, 2008). It is likely that this class of channels could have a similar role in skeletal muscle by mediating the effect of muscle electrical activity on gene expression. Our data strongly suggest that a minor fraction of α1S located on the sarcolemma could have this role and be involved in muscle differentiation.

Obviously, substantial knockdown of any important protein for muscle function would cause muscle atrophy. Our results advocate for the involvement of α1S in a pathway controlling muscle mass through a specific mechanism. Indeed, although denervated muscles also undergo an atrophic process, they never show upregulation of autophagy-related genes and autophagosome formation (Supplementary Figure S6). In addition, the mechanism involved in atrophy after α1S downregulation cannot be explained by a simple loss of muscle contractility as treated muscles are still EC coupled. Moreover, quantitative defect of RyR1 has been reported in families presenting a recessive form of congenital myopathies with cores (Monnier et al, 2008). In some patients, the measurement of protein expression showed a decrease of about 50% of the RyR1 while these patients did not display atrophy of skeletal muscle. These data strongly suggest that a fall of any important protein for muscle function would not systematically result in the same physiological and molecular consequences.

Hence, our results support a new role for the sarcolemmal subpopulation of α1S, which might act as a voltage sensor of a still unknown molecular machinery located on the sarcolemma, in the close vicinity of muscle nuclei, in charge of controlling transcriptional activity of genes involved in the muscle anabolism/catabolism balance.

Materials and methods

Antibodies

For western blot, primary antibodies used were α1S subunit, phosphorylated FoxO3A and nNOS (Abcam); for immuno-staining α1S subunit (Chemicon), FoxO3 and LC3b (Cell Signaling), laminin and α-actin (Sigma-Aldrich), P62 (Progen), SERCA and nNOs (Abcam). For MHC isoform analysis primary antibodies were MHC-1 (hybridoma#BA-D5, Deutsche Sammlung von Mikroorganismen und Zellkulturen DSMZ), MHC-2a (hybridoma#SC-71, DSMZ), MHC-2x (hybridoma#6H1, Developmental Studies Hybridoma Bank) and MHC-2b (hybridoma#BF-F3, DSMZ). RyR1 antibody was a gift from Dr I Marty, secondary antibodies were from Invitrogen and Jackson ImmunoResearch.

α1S subunit targeting vector construction

Antisense sequences (Supplementary Table S1) were introduced into the U7smOPT gene as described earlier (Goyenvalle et al, 2004): U7-c was a non-functional construction used as negative control. U7-SA, U7-ESE and U7-c sequences were introduced separately or at the XbaI site of the pSMD2 AAV2 vector. AAV2/1 pseudotyped vectors were prepared by transfection in 293 cells as described earlier (Riviere et al, 2006) and vector particles were purified on cesium chloride gradients from cell lysates obtained 72 h after transfection and tittered by qPCR. Titres for U7-SA and U7-ESE were 2.4 × 10¹³ and 3.0 × 10¹³ vector genomes (vg) ml⁻¹, respectively.

In vivo gene transfer

All procedures were in accordance with the institutional and European agreement for human treatment of animals. Experiments were performed on adult C57/Black6 mice. Anaesthesia was achieved with a mix of 100 mg/kg ketamine and 10 mg/kg xylasine. Two intramuscular injections in 24 h of a mixture (50 μl/TA) containing AAV (U7-SA) and AAV (U7-ESE) were carried out in TA of the right hindlimb; the contralateral muscles were injected in the same procedure with vector AAV (U7-Ctrl) and used as control.

Gene expression analysis

Total RNA was prepared from 600 μm of TA using Quiagen total RNA isolation kit. Complementary DNA generated with Invitrogen Superscript II Plus reverse transcriptase (Invitrogen) was analysed by real-time qPCR performed on Opticon2 (Bio-Rad) using iTaq SyberGreen Supermix with ROX (Bio-Rad). We therefore quantified in all samples transcripts of the PO gene encoding human acidic ribosomal phosphoprotein ubiquitly expressed as the endogenous RNA control and each sample was normalized on the basis of its PO content. Primers used are listed in Supplementary Table S1.

Force production

Muscle performance was evaluated by the measurement of in situ muscle contraction in response to nerve stimulation, as described earlier (Vignaud et al, 2005a, 2005b). Animals were anaesthetised (i.p., pentobarbital sodium, 50 mg × kg⁻¹). The distal tendon of the TAs was attached to an isometric transducer (Harvard Bioscience) using a silk ligature. All data provided by the isometric transducer were recorded and analysed on a microcomputer. The sciatic nerves were distally stimulated by a bipolar silver electrode. Responses to tetanic stimulation (pulse frequency 50–143 Hz) were successively recorded. Absolute maximal tetanic forces were determined. Muscle masses were measured to calculate specific tensions. After contractile measurements, the animals were killed with an overdose of pentobarbital.

Morphometric analysis

For muscle analysis, tissues were sectioned at 10 μm on a cryostat (Leica CM3050S). The internal diameters (shortest diameter) of all fibres in the total transversal muscle section were recorded and analysed with Elix software (MDS Analytical Technologies).

Dissociated fibre cultures

Myofibres were isolated from the dissected FDB muscle of 6 months post-injection mice by digestion with collagenase 1a (SIGMA) and mechanical dissociation (Kaisto et al, 1999). Isolated fibres were cultured (37°C, 5% CO₂) in Dulbecco's Modified Eagle medium with high glucose, 1% horse serum (GIBCO) and penicillin/streptomycin on coverslips coated with Matrigel (Becton Dickinson Labware, Bedford, MA). Myofibres were used for experiments after overnight culture.

Histology and immuno-fluorescence analysis

Muscle sections were stained with haematoxylin and eosin or processed for immuno-histochemistry. Fibrosis was labelled by Red Sirius (3%) then dehydrated with ethanol and cleared in xylene as already described (Dooley et al, 2003). Analyses of labelled zone were carried out on Histolab Software (MDS Analytical Technologies). For immuno-staining procedures, 10 μm sections were cut on a Tissue-Tek II cryostat and fixed on glass slides. Slides were rehydrated in PBS, fixed with 4% PAF for 10 min or not (LC3b and dystrophin), blocked in 4% BSA (Sigma-Aldrich) for 1 h, incubated with primary antibodies for overnight at 4°C, washed in PBS, incubated for 1 h with secondary antibodies, thoroughly washed in PBS, incubated with DAPI for nuclear staining for 5 min and mounted in Fluoromount (Southern Biotech) and finally imaged by confocal laser scanning microscope (LEICA SPE DM2500), the brightness/contrast adjustements done with Photoshop CS version 9.0.2 (Adobe) in control and test samples were identical. For myofibres isolated from FDB, immuno-labelling was performed directly in the 24-well culture plate. Cells were fixed (15 min, 4% paraformaldehyde in PBS, RT) washed (2 × , 5 min, PBS), permeabilized (10 min. PBS with 0.5% Triton X-100), and immuno-labelled using the same procedure described for muscle sections. Myofibres were analysed by confocal laser scanning microscopy using a LEICA SP2 operating system using an ApoPlan × 63, 1.40 NA oil lens. DAPI, Alexa-488, Alexa-568 or Cy3 fluorescence were sequentially excited using lasers with wavelengths of 405 nm (DAPI), 488 nm (Alexa-488) and 561 nm (Alexa-568, Cy3). Z-series from the top to the bottom of fibres were sequentially collected for each channel with a step of 1 μm between each frame. For MHC analysis, frozen unfixed 10 μm sections were blocked 1 h in PBS plus 1% BSA, 1% sheep serum, 0.01% Triton X-100 and 0.001% sodium azide. Sections were then incubated overnight with primary antibodies against laminin and MHC isoforms. After washes in PBS, sections were incubated 1 h with secondary antibodies. After washes in PBS, slides were finally mounted in Fluoromont (Southern Biotech). Morphometric analyses were made on two serial sections of whole 6 months post-injetion TA muscles. Images were captured using a digital camera (Hamamatsu ORCA-AG) attached to a motorized fluorescence microscope (Zeiss AxioImager.Z1), and morphometric analyses were made using the software MetaMorph 7.5 (Molecular Devices).

Western blotting

For α1S subunit and RyR1, muscle tissues were lysed in a buffer containing 200 mM sucrose, 20 mM Hepes pH 7.4, 0.4 mM CaCl₂ with a mix of protease inhibitors (Roche). Samples were quickly micro centrifuged. Laemmli buffer was added and samples were denatured at RT. Soluble proteins were resolved by SDS–PAGE (3–8% Invitrogen) and transferred to nitrocellulose membranes. For FoxO3A and nNOS, muscles were lysed in RIPA buffer containing 150 mM Nacl, 50 mM Hepes, 1% NP-40, 0.5% sodium deoxylate, 0.1% SDS, 5 mM EDTA with 1 mM PMSF and a mix of protease inhibitors (Roche). Soluble proteins were resolved by SDS–PAGE and western blotting. Western blots were revealed with enhanced chemiluminescence (Thermo Scientific), films were scanned and band signal intensities were determined using Image Processing and Analysis in Java (version 1.37v).

Ultrastructure analysis

For EM, muscles were fixed with 2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4, for 2 h. After post fixation for 1 h at 4°C in 2% osmium tetroxide and acetone dehydration, the samples were embedded in polybed 812 resin. Ultrathin sections were contrasted with uranyl acetate/methanol (7%) and lead citrate and observed with a Philipps CM120 electron microscope.

Statistical analysis

Statistical analysis of differences between mouse cohorts was performed using the Student's t-test. Excel (Microsoft) was used for calculations, and results were expressed as mean±s.e.m.

Supplementary Material

Supplementary Table S1

Supplementary Figures S1–S4 and S6

Supplementary Movie S5A

Supplementary Movie S5B

Supplementary Movie Legends

Review Process File