The peculiar clinical features and the pathogenic mechanism related to calpain-3 deficiency (impaired sarcomere remodelling) suggest that the ubiquitin-proteasome degradation pathway may have a crucial role in Limb Girdle Muscular Dystrophy 2A (LGMD2A). We therefore investigated muscle atrophy and the role of the ubiquitin-proteasome and lysosomal-autophagic degradation pathways.
We selected 25 adult male LGMD2A patients (and seven controls), classified them using clinical severity score, analysed muscle fibre size by morphometry and protein and/or transcriptional expression levels of the most important atrophy- and autophagy-related genes (MuRF1,atrogin1,LC3,p62,Bnip3).
Muscle fibre size was significantly lower in LGMD2A than in controls and it was significantly correlated with patients' clinical disability score recorded at the time of biopsy, suggesting that functional and structural muscle impairment are dependent. The large majority of atrophic fibres originate from a mechanism different from regeneration, as assessed by neonatal myosin immunolabelling. As compared with controls, LGMD2A muscles have higher MuRF1 (but not atrogin1) protein and MuRF1 gene expression levels, and MuRF1 protein levels significantly correlated with both muscle fibre size and clinical disability score. LGMD2A muscles have slightly increased levels of LC3-II and p62 proteins and a significant up-regulation of p62 and Bnip3 gene expression.
In LGMD2A muscles the activation of the atrophy programme appeared to depend mainly upon induction of the ubiquitin-proteasome system and, to a lesser extent, the autophagic-lysosomal degradation pathway.
Limb Girdle Muscular Dystrophy 2A (LGMD2A) is a progressive muscle disease caused by the deficiency of the muscle-specific protease calpain-3, which leads to an impairment of a physiological and dynamic process called sarcomere remodelling [1, 2]. Sarcomere remodelling consists of a continuous turnover of sarcomeric proteins, where the synthesis of novel proteins is balanced by the degradation of misfolded proteins, by activation of stress-signalling pathways. Muscle atrophy resulting from chronic disease rather than disuse may arise from primary disease of muscle. All atrophic conditions share an imbalance between dynamic anabolic and catabolic reactions, where the increased myofibrillar protein breakdown exceeds the protein synthesis. Muscle atrophy is an active process controlled by specific signalling pathways and transcriptional programmes involving many atrophy-promoting genes called atrogenes [3, 4]. Three proteolytic systems are involved in the maintenance of the sarcomeric function: the cytosolic calpain system, the ubiquitin-proteasome system (UPS) and the autophagic-lysosomal pathway. Although they are all activated in muscle atrophy, the key player in the degradation of myofibrillar proteins is the UPS (Figure 1). Proteasomes are multiprotein complexes that predominantly degrade short-lived, misfolded proteins. Degradation is initiated by labelling of the targeted proteins with ubiquitin molecules, with the coordinated action of three classes of enzymes (including E3 ubiquitin-protein ligases). As the ubiquitin ligase is the rate-limiting enzyme in the ubiquitination reaction, an increase in its expression is sufficient to enhance the UPS protein degradation.
All the conditions of muscle atrophy so far studied have shown an induction of two atrogenes, the muscle-specific ubiquitin-ligase Muscle Atrophy F box (MAFbx or Atrogin1) and the Muscle-specific RING Finger protein 1 (MuRF1), which lead to increased protein degradation through the UPS [3-6]. Knockout animals lacking either MuRF1 or atrogin1 show a reduced rate of muscle atrophy after denervation , confirming that these ligases are necessary for the atrophy programme. Atrogin1 and MuRF1 are considered the master genes and the best markers for muscle atrophy in humans, as they have been found to be significantly up-regulated in different conditions of atrophy even before significant loss of muscle mass can be detected. Furthermore, MuRF1 has been shown to be up-regulated in multiple models of muscle atrophy, where it promotes the degradation of key sarcomeric proteins, including titin, myosin, nebulin, troponins, myotilin, titin-cap [7, 8].
Recent studies have shown that the lysosomal-autophagic pathway also plays a critical role in the control of muscle mass [9, 10]. Autophagy (macroautophagy) is a highly conserved process that degrades long-lived, aggregation-prone misfolded proteins by their sequestration within double-membrane vesicles (autophagosomes) and degradation by fusion with lysosomes . Aggregates of ubiquitinated proteins are selectively degraded by delivery to autophagosomes, through p62/sequestosome-1 protein , which binds ubiquitin and microtubule-associated protein Light Chain-3 (LC3) protein. The level of the lipidated form of LC3 (LC3-II isoform), which is generated during autophagosome formation, reflects autophagosome number and is used as a marker of autophagy. Endogenous LC3 can also be visualized by fluorescence microscopy either as a diffuse cytoplasmic pool or as ‘punctate’ structures. However, the simple quantification of autophagosomes is insufficient for an overall estimation of autophagic activity. Indeed, the autophagic flux can be monitored by the levels of other autophagy substrates, such as p62 protein .
UPS and autophagic-lysosomal systems communicate by their common regulator, transcription factor Forkhead box class O (FoxO3) [14-16], which induces atrophy by promoting atrogin1 and MuRF1 gene expression (Figure 1), and promotes autophagy by controlling the transcription of autophagy-related genes, including LC3, Beclin1, GABARAP11 and Bnip3 . Autophagy is up-regulated when cells need to generate intracellular nutrients and energy (for example, starvation), undergo structural remodelling or need to rid themselves from damaging cytoplasmic components (for example, protein aggregates).
UPS and autophagy activities occur at low basal levels in all cells to perform homeostatic functions, such as the control of protein quality and quantity, but when either the proteasomal or the lysosomal degradation is impaired, or when the amount of material to be degraded exceeds the normal capacity, ubiquitinated proteins aggregate and form autophagic vacuoles. A dysfunction in the UPS and/or autophagic pathways has been implicated in many pathological conditions, but recently their role have been recognized also in some muscle diseases, including inclusion body myositis, Danon disease and Glycogenosis type II [17-19].
The present study was undertaken in order to characterize muscle fibres atrophy in LGMD2A muscle and to investigate the potential role of the UPS and the lysosomal-autophagic degradation pathways in regulating muscle fibre size.
Selection criteria of patients and muscle biopsies
The Neuromuscular Telethon BioBank (which contains about 8500 diagnostic muscle biopsy specimens from patients affected with neuromuscular disorders, collected after written consent has been obtained) was surveyed for samples from patients with LGMD2A in whom the diagnosis was proven by the identification of mutations in the calpain-3 gene . In order to avoid the problem of small muscle fibre size occurring both in children's and in females' muscles , and of smaller fibre size typically observed in distal as compared with proximal muscles, we selected for this study only those samples that have been obtained from proximal limb muscles (usually vastus lateralis, or in few cases biceps and triceps brachii) of adult male patients. Such a restrictive selection criteria reduced the number of available samples, but produced more consistent results. We therefore examined the muscles from 25 adult male patients with LGMD2A (age at biopsy 14–59 years, mean 29 ± 13), and seven age-matched men used as normal controls (age at biopsy 18–41 years, mean 26 ± 9).
Clinical assessment of functional muscle involvement
To establish whether muscle fibre size was related with patients' functional muscle involvement recorded at the time at muscle biopsy, we used a ‘clinical disability score’, which corresponds to the widely used, conventional 10-graded scale of Gardner-Medwin & Walton : grade 0 = hyperCKemia; grade 1 = normal gait, unable to run freely, myalgia; grade 2 = waddling gait, fatigability; grade 3 = overt muscle weakness, difficulty climbing stairs; grade 4 = difficulty rise from floor, Gowers' sign; grade 5 = unable to rise from floor; grade 6 = unable to climb stairs; grade 7 = unable to rise from a chair; grade 8 = unable to walk unassisted. As this functional scale looks primarily at lower limb functions, it appears particularly suitable to correlate the functional involvement of lower limbs with the fibre size measured in the vastus lateralis muscle.
Muscle fibre morphometry
Muscle morphometric study was conducted on cross cryosections stained for haematoxylin-eosin because it allowed by simple microscope inspection to exclude blood vessels and nuclear clumps from the calculation. We digitalized five to seven non-overlapping, randomly selected microscope fields (Zeiss Axioskop, Gottingen, Germany) and outlined the borders of 200–500 fibres for each muscle (ImageJ software, v.1.34, National Institutes of Health, Bethesda, MD, USA) in order to measure and calculate the following morphometric parameters:
fibre diameter, defined as the shortest dimension bisecting the fibre in a plane through the fibre centre , which is used in order to avoid the distortion that occurs when fibres are sectioned obliquely (adult male subjects have values ranging from 40 to 80 μm) ; fibre cross-sectional area, which measure is fully reliable only when the section is perfectly orientated in the transverse plane (assuming a circular shape, the cross-sectional areas of normal fibres should range between 1256 and 5024 μm2); coefficient of fibre size variability, defined as the diameter standard deviation × 1000/mean diameter, which is used to evaluate the variability of fibre size (adult male subjects have values ranging from 0–250 units) ; fibre atrophy factor, which has been developed to give different importance to fibres with mild or severe change in size and to detect atrophy that may not be otherwise apparent by simply calculating the average fibre diameter. Among the abnormally small fibres, few fibres in the range of 30–40 μm would have less significance than the same number of fibres in the range of 10–20 μm. This is taken into account by multiplying the number of fibres with a diameter between 30 and 40 μm by 1, those with a diameter between 20 and 30 by 2, those from 10 to 20 μm by 3, and those less than 10 μm by 4. These products are added together and divided by the total number of fibres to put the result on a proportional basis. The resulting number is then multiplied by 1000 to obtain the atrophy factor. Vastus lateralis, biceps and triceps muscles from adult, normal male subjects have atrophy factor values ranging from 0–150 .
Muscle cross cryosections, 6–8 μm thick, were processed by indirect immunofluorescence to investigate the extent of muscle regeneration, using a monoclonal antibody against neonatal myosin (MHCn, Novocastra Lab., Newcastle upon Tyne, UK; dilution 1:100) . Sections were incubated for 1 h at room temperature with primary antibody, and the specific labelling was developed using cyanine-3-conjugated immunoglobulins (Caltag Medsystems, Buckingham, UK; dilution 1:100). The extent of regeneration was expressed as a percentage of fibres showing positive labelling for neonatal myosin on total fibres. Rare nuclear clumps (neonatal-myosin positive) have been excluded from the calculation.
A double immunofluorescence labelling was used to investigate autophagy features, using antibodies against p62, as a marker of protein aggregates (Gp62C, Progen Biotechnik, Heidelberg, Germany; dilution 1:200), and LC3, as a marker of autophagosomes (2775, Cell Signaling Technology, Danvers, MA, USA; dilution 1:100). For this purpose, sections were fixed with 4% paraformaldehyde, treated with 0.1% Triton, incubated in blocking solution (0.5% bovine serum albumin, 10% horse serum in PBS) for 20 min, and then incubated overnight at 4°C with primary antibodies . After incubation with appropriate secondary fluorescent-conjugated antibodies (Alexa-Fluor, Invitrogen, Paisley, UK), slides were mounted using Vectashild medium (Vector, Burlingame, CA, USA) with nuclear DAPI (4′,6-diamidino-2-phenylindole) stain and examined under a fluorescence microscope (DM5000B, Leica Microsystems, Wetzlar, Germany).
Western blot analysis
In a subset of 19 LGMD2A cases, which were selected on the basis of there being sufficient muscle tissue available, we conducted immunoblot analysis using antibodies against atrogin1 (AP2041, ECM Biosciences, Versailles, KY, USA; dilution 1:300), MuRF1 (MP3401, ECM Biosciences; dilution 1:500), p62 (Gp62C, Progen; dilution 1:1000) and LC3 (L7543, Sigma Chemicals, St. Louis, MO, USA; dilution 1:2000). Briefly, muscle sections were lysed in a buffer containing 50 mM Tris, pH 7.5, 150 mM NaCl, 10 mM MgCl2, 0.5 mM Dithiothreitol, 1 mM EDTA, 10% glycerol, 2% SDS, 1% Triton X-100 and protease inhibitors (Complete Protease Inhibitor Cocktail, Roche, Milan, Italy). The samples were immunoblotted as previously described  and the reaction was visualized with the chemiluminescent substrate. When necessary, the membranes were stripped and reprobed using a stripping buffer consisting of 25 mM glycine-HCl, pH = 2 and 1% SDS. The quantity of proteins in each sample was determined by densitometry, normalized to the amount of muscle tissue loaded, as determined by myosin quantity in the post-transfer Coomassie blue-stained gel, and expressed as a percentage of control mean. LC3 protein amount was expressed as the ratio between LC3-II and LC3-I bands.
RNA isolation, cDNA amplification and real-time PCR (qPCR)
A transcriptional study was conducted in the same subset of 19 LGMD2A cases. Total RNA was isolated from these muscles using Trizol reagent and treated with DNase I (Invitrogen, Paisley, UK) following the manufacturer's protocol. The yield and the purity of the extracted RNA were determined using Nanodrop spectrophotometer (Thermo Scientific, Epsom Surrey, UK). Complementary DNA, generated with iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA, USA), was analysed by real-time PCR using a MJ Mini Opticon Thermal Cycler (Bio-Rad) and the IQ SYBR green supermix (Bio-Rad).
Results were normalized to glyceraldehyde 3-phosphate dehydrogenase gene (GAPDH, Fw TGCACCACCAACTG CTTAGC, Rev GGCATGGACTGTGGTCATGAG), using the 2−ΔΔCT calculation method. Four different genes were investigated using the following primers sequences: atrogin1 (RefSeq NM_058229, Fw GCAGCTGAACAACATTCAGATCAC, Rev CAGCCTCTGCATGATGTTC AGT), MuRF1 (RefSeq NM_032588, Fw CCTGAGAGCCATTGACTTTGG, Rev CTTCCCTTCTGTG GACTCTTCCT), Bnip3 (RefSeq NM_ 004052, Fw GTCAAGTCGGCCGGAAAATA, Rev TTCATGACG CTCGTGTTCCT), p62 (RefSeq NM_003900, Fw GCTTCCAGGCGCACTACC, Rev CATCCTTCAC GTAGGACATGG). Values were expressed as fold change in gene expression after normalization to the control muscle. Fold change values more or less than 1.5 were considered significant.
Values were expressed as mean ± standard deviation. Differences of parametric variables between patient and control groups were assessed using the analysis of Variance by Student's unpaired t-test. Correlation between two parametric variables was assessed by the Linear Regression Analysis (r coefficient). Correlation between a parametric variable and the clinical disability score (a nonparametric, nonlinear variable) was assessed by the Spearman's Rank Correlation Analysis (ρ coefficient). Statistical significance was set at P < 0.05.
In our series of 25 LGMD2A patients, the age at muscle biopsy ranged from 14 to 59 years (mean 29.4 ± 13), the age at onset of symptoms ranged from 3 to 54 years (mean 19.7 ± 12), and the duration of symptoms (time elapsed from onset to biopsy) ranged from 1 to 39 years (mean 9.7 ± 10). At the time at biopsy, patients presented a large variability of clinical severity: indeed, muscle impairment, graded according to the clinical disability score, ranged from grade = 0 (asymptomatic hyperCKemia) in one patient to grade = 8 (unable to walk) in three patients.
Reduced muscle fibre size in LGMD2A
To asses whether LGMD2A muscles have fibres of reduced size (referred as ‘atrophic fibres’), we calculated four different morphometric parameters (diameter, cross-sectional area, coefficient size variability, atrophy factor). As compared with controls, LGMD2A muscles had lower values of fibre diameter (50.8 ± 4.4 vs. 41.0 ± 10.6 μm, respectively, P < 0.001), cross-sectional area (3302 ± 568 vs. 2533 ± 1228 μm2, respectively), and significantly higher levels of fibre atrophy factor (119 ± 89 vs. 1079 ± 715 units, respectively, P < 0.001) and coefficient of size variability (162 ± 19 vs. 404 ± 132 units, respectively, P < 0.001), demonstrating that are the majority of muscle fibres is atrophic in LGMD2A (Figure 2).
Muscle fibre size correlates with clinical functional disability in patients
We assessed whether muscle fibre size correlates with patients' clinical functional impairment at the time at biopsy. We found that the clinical disability score was significantly correlated with all myofibre size parameters (diameter: ρ = −0.66, P < 0.001; cross area: ρ = −0.63, P < 0.001; atrophy factor: ρ = 0.78, P < 0.001), indicating that functional and morphological muscle impairment are dependent. Conversely, both the age at onset and the duration of symptoms were unrelated to myofibre size.
Muscle fibre atrophy and regeneration
To better understand the reason why LGMD2A is characterized by muscle fibre atrophy, we considered the role of possible physiopathological factors. Regeneration process, which is known to vary considerably depending on the different rate of fibre degeneration and necrosis, appeared as a potential cause of fibre atrophy in LGMD2A. We therefore evaluated the contribution of muscle regeneration in LGMD2A using neonatal myosin as an immunohistochemical marker (Figure 3). As expected, regenerating fibres were indeed significantly smaller than normal fibres (cross area: 543 ± 848 vs. 3302 ± 568 μm2, respectively, P < 0.001; diameter: 16.3 ± 10.6 vs. 50.8 ± 4.4 μm, respectively, P < 0.001), but they accounted, in average, for only 2.8% of total fibres. These data indicate that the poor regeneration level we found in LGMD2A is not sufficient to counteract the loss of muscle fibres, which has been observed at various levels in this disorder. Given that the proportion of abnormally small-sized fibres (diameter <40 μm) was significantly higher in LGMD2A than in controls (48% vs. 6.5% of total, respectively, P < 0.01), we concluded that the majority of fibres were truly atrophic, demonstrating that a process different from regeneration may explain the overall muscle fibre atrophy observed in LGMD2A.
Denervation is not a feature of muscular dystrophies, nor is it characteristic of LGMD2A; the variability of muscle fibre type composition has been minimized by the selection of proximal muscles, which have similar fibre type composition.
Muscle fibre atrophy and UPS
Following the demonstration of morphological fibre atrophy, we determined whether the molecular programme of muscle atrophy was activated in LGMD2A muscles, and therefore conducted a combined protein and gene expression study, involving two of the most important biomolecular markers of the UPS degradation pathway, atrogin1 and MuRF1.
LGMD2A muscles showed nearly normal levels of atrogin1 protein (77 ± 26% of control mean) and atrogin1 gene expression (1.3 ± 0.8-fold increase), and these were unrelated to both fibre size parameters and patients' clinical disability score. Conversely, as compared with controls, LGMD2A muscle showed higher level of MuRF1 protein (146 ± 64% of control mean; P < 0.05) (Figure 4), which was also significantly correlated with both fibre size parameters (cross area: r = −0.52, P < 0.05; diameter: r = −0.59, P < 0.01; atrophy factor: r = 0.59, P < 0.01) and patients' clinical disability score (ρ = 0.72, P < 0.001). Furthermore, LGMD2A muscles showed significantly higher MuRF1 gene expression levels (2.9 ± 2.3-fold increase). Overall, these results indicate that LGMD2A muscles showed an up-regulation of MuRF1, which resulted in the activation of the UPS degradation pathway.
Muscle fibre atrophy and autophagic-lysosomal pathway
The potential role of autophagic-lysosomal pathway in causing muscle fibre atrophy was investigated by protein and gene expression analysis of some biomolecular markers of autophagy, that is, LC3, p62 and Bnip3.
Immunofluorescence analysis revealed a cytoplasmic ‘punctate’ reaction for p62 (protein aggregates), which was present only in a few atrophic fibres of some patients, and scattered LC3-positive reaction in the same or in other fibres (Figure 4). Immunoblot analysis showed that, as compared with controls, LGMD2A muscles have slightly increased levels of LC3-II (162 ± 74% of control mean) and p62 proteins (128 ± 76% of control mean) (Figure 4), whereas qPCR analysis showed a significant up-regulation of p62 (2.1 ± 1.7-fold increase) and Bnip3 (3 ± 2.2-fold increase) gene expression.
Muscle fibre atrophy and the molecular pathways underlying this process have not yet been investigated in LGMD2A. Some lines of evidence, based both on the peculiar clinical features (that is, the disease is characterized by atrophy and proximal muscle group involvement) and on the specific pathogenetic mechanism (that is, impaired sarcomere remodelling), suggested that the UPS degradation pathway may have a crucial role in LGMD2A.
As isolated proteasomes are unable to degrade intact myofibrils , the atrophy process may depend on calpain-3 in the initial stage and on the UPS in the later stages . Gene expression profiling in LGMD2A showed an overexpression of UPS-related genes [22, 23]. While the expression of atrogin1 and MuRF1 was not increased in mouse models of LGMD2A, FoxO1 was strongly up-regulated, and induced muscle atrophy in calpain-3 deficient mice .
A first conclusion from our morphological study is that LGMD2A muscle fibres are significantly atrophic, and that the degree of structural fibre atrophy matched that of functional muscle impairment recorded at clinical level. Although conceptually expected, our results corroborate earlier correlations between clinical and muscle imaging findings [25, 26] and clearly indicate that structural and functional muscle atrophy are mutually dependent.
To interpret the determinant of fibre atrophy, we showed that in LGMD2A the large majority of atrophic fibres are not regenerating, suggesting that this rather results from a different mechanism, possibly leading to defective fibre growth and protein turnover [2, 27].
As the UPS is the major pathway responsible for protein degradation in muscle, we investigated the expression of some atrogenes, and we found a differential up-regulation of MuRF1 and atrogin1. This variable expression pattern might be attributed to distinct regulatory mechanisms: MuRF1 seems to be associated with muscle proteolysis, while atrogin1 seems to be related to down-regulation of protein synthesis . Interestingly, we found a significant increase of MuRF1 (but not atrogin1) protein levels in LGMD2A, which correlated with muscle fibre size and clinical functional muscle impairment, indicating that MuRF1 is a reliable marker of muscle atrophy.
Although our study investigated only few of the molecular players acting in the atrophy process, the results indicate an activation of the UPS pathway leading to an induction of the muscle atrophy programme in LGMD2A. This conclusion agrees with the earlier suggestion that UPS activation might follow the impaired sarcomere remodelling and the increased turnover of myofibrillar proteins, which result from the primary defect of calpain-3. We excluded the possibility that the UPS activity may be enhanced by regeneration in LGMD2A, as regeneration occurs at very low levels in this disorder [20, 29], and the primary impairment involves sarcomere remodelling.
When considering the signalling pathway that might have induced MuRF1 overexpression in LGMD2A, we attributed the main role to the FoxO-Akt pathway. Indeed, even if the up-regulation of the NF-κB transcription factor [30, 31] and other signalling pathways has been reported to be able to up-regulate MuRF1 expression, we believe that this concurrence is unlikely in LGMD2A, as activation of NF-κB depends on calpain-3 . However, the finding of activated NF-κB in LGMD2A muscle cells  suggests that only a subset of NF-κB-dependent genes is sensitive to the deficiency of calpain-3.
A second conclusion from our study is that LGMD2A muscles showed an up-regulation of the autophagy-related genes p62 and Bnip3, but the scarcity of p62-positive cytoplasmic aggregates and the slightly increased LC3-II/LC3-I protein ratio and p62 protein, would suggest that the autophagic-lysosomal degradation pathway is involved to a lesser degree than the UPS pathway.
It is evident that any intervention aimed at reducing muscle atrophy could improve the course of the disease. Several factors have been described as reducing the otherwise elevated expression of MuRF1 and atrogin1 and therefore as potentially counteracting muscle atrophy: muscle exercise training [34, 35], treatment with Bortezomib and Fenofibrate [36, 37], branched-chain amino acids , l-carnitine  and long-chain ω-3 fatty acid .
The present study highlights the importance of accurate investigations in determining the pathways that regulate muscle atrophy in muscular dystrophies, as this issue has potentially important consequences for the future development of therapeutic strategies.
This work was supported by research grants from the Association Française contre les Myopathies (No. 13859 to M.F., No. 14199 to A.C.N., No. 14999 and 15696 to C.A.) and the Comitato Telethon Fondazione Onlus (No. GTB07001 to C.A.).
Dr Fanin designed the study, analysed and interpreted the data, performed statistical analysis and drafted the manuscript. Dr Nascimbeni analysed and interpreted the data, and revised the manuscript for intellectual content. Dr Angelini revised the manuscript for intellectual content.