Alpha sarcoglycan is required for FGF-dependent myogenic progenitor cell proliferation in vitro and in vivo

Sarcoglycans are dystrophin-associated proteins (DAPs) that form a membrane complex necessary for the physiological function of skeletal and cardiac muscles (Henry and Campbell, 1996; Lapidos et al., 2004). Mutations in genes encoding sarcoglycans produce muscular dystrophy and cardiomyopathy in animal models and humans (Vainzof et al., 1996). Mice null for α-sarcoglycan (Sgca-null mice) (Duclos et al., 1998), a model for human limb girdle type 2D muscular dystrophy (LGMD2D), develop a more severe myopathy than mdx mice (Lefaucheur et al., 1995; McArdle et al., 1995), the standard model for Duchenne muscular dystrophy (DMD). In humans however, DMD is more severe than LGMD2D (McNally et al., 1994) and the reason for this discrepancy is unknown. Similarly, why the mdx mouse develops a relatively mild myopathy compared with human DMD is unknown. However, it has been proposed that there is a more robust self-renewal of myogenic progenitor cells (MPCs) in murine DMD as compared with the human disease (Blau et al., 1983). This has been demonstrated in telomerase knockout mdx mice, which exhibit premature exhaustion of the MPC pool and a more severe disease phenotype (Sacco et al., 2010). These mice more closely represent the human DMD phenotype, highlighting the role of MPC self-renewal in DMD severity.

MPCs are a population of muscle progenitors that are able to exit quiescence, proliferate and differentiate into new muscle fibers in response to muscle damage. They also play a key role in muscle regeneration (Alameddine et al., 1989; Le Grand and Rudnicki, 2007; Mauro, 1961; Shinin et al., 2006; Zammit et al., 2004; Zammit et al., 2006). We observed that MPCs from Sgca-null mice have a reduced proliferative ability in comparison to MPCs from wild-type or mdx dystrophic mice. This is due to a molecular interaction between SGCA and fibroblast growth factor receptor 1 (FGFR1) during the proliferative phase. We propose that SGCA modulates FGFR1 activity via the formation of a complex, and that the absence of SGCA destabilizes FGFR1, leading to the proliferative impairment of MPCs. This phenomenon, together with the enhanced apoptosis previously reported (Allikian et al., 2004; Anastasi et al., 2004; Yoshida et al., 1998), might explain the severity of this form of muscular dystrophy in mice and the success of previous cell transplantation experiments (Sampaolesi et al., 2003).

A monoclonal antibody (mAb) against SGCA was purchased from Novocastra (Newcastle, UK) and Kevin Campbell (Iowa University, USA) kindly provided an SGCA polyclonal antibody. Biotinylated rat mAb anti-SM/C2.6 was kindly provided by So-ichiro Fukada (Osaka University, Japan). Mouse monoclonal myogenin, MF20 and PAX7 antibodies were purchased from the Developmental Studies Hybridoma Bank. Mouse mAb against β-tubulin, rabbit polyclonal antibody (pAb) against MYOD and MYF5 were from Santa Cruz (San Diego, CA, USA). Mouse mAb anti-FGFR1 (#05-149, clone 19B2), rabbit pAb anti-FGFR4 and pHH3 antibodies were purchased from Chemicon (Millipore, Billerica, MA, USA). Phospho-p44/42 MAPK (Thr202/Tyr204) rabbit mAb and p44/42 MAPK rabbit mAb were from Cell Signaling (Danvers, MA, USA). mAbs anti-β-galactosidase, V5 and GFP were purchased from Invitrogen (Carlsbad, CA, USA). Rabbit mAb anti-CycD, mouse mAb anti-GAPDH and rabbit pAb anti-syndecan 1 were provided by Abcam (Cambridge, MA, USA). Alexa Fluor secondary antibodies were purchased from Molecular Probes (Invitrogen).

Apoptosis was also evaluated by immunohistochemistry using a TUNEL assay (BioVision). Cytokine treatments were prepared by adding 5 ng/ml recombinant murine HGF (eBiosciences, San Diego, CA, USA) or bFGF (Peprotech, Rocky Hill, NJ, USA) to cell culture media. Human bFGF (5 ng/ml) and human recombinant HGF (5 ng/ml) were from Invitrogen. YP-740 was from Biotrend Chemikalien (Köln, Germany) and phorbol 12-myristate 13-acetate (PMA) was from Sigma-Aldrich (St Louis, MO, USA). Nuclei were always counterstained with DAPI (Invitrogen).

Plasmids were prepared using standard procedures. All plasmid preparations were obtained using a GenElute HP Endotoxin-Free Plasmid Maxiprep Kit (Sigma-Aldrich) and quantified using a Nanodrop ND-1000 spectrophotometer. pCMV-SPORT6 plasmids carrying murine Fgfr1 or Sgca cDNAs were purchased from Open Biosystems (Thermo Scientific, Huntsville, AL, USA).

Total RNA from MPCs was extracted using the PureLink RNA Mini Kit (Invitrogen) and genomic DNA was removed using the TURBO DNA-Free Kit (Ambion). Muscle extracts were homogenized in Trizol (Invitrogen) before proceeding to column purification. RNA was quantified and then reverse transcribed using the Superscript III Reverse Transcriptase Kit (Invitrogen). For PCR analysis we used Platinum Taq DNA polymerase (Invitrogen). For quantitative (q) PCR analysis, Fast SYBR Green Master Mix (Invitrogen) was diluted with cDNA and primers. RT-PCR and qPCR were performed using the following primers (Fw, forward; Rv, reverse; 5′ to 3′):

• Sgca, Fw TCTGTCACTCACCGGGCGG and Rv GTGACCTCAATGACCTGGAG;

• Gapdh, Fw TTCACCACCATGGAGAAGGC and Rv GGCATGGACTGTGGTCATGA;

• Fgfr1, Fw ACAAACCCAACCGTGTGACCAAAG and Rv TTGGAGGCGTACTCCACAATGACA; and

• Fgfr4, Fw AGCCTGGTGATGGAAAGTGTGGTA and Rv ACACCTTGCAGAGTAGCTCCACAT.

MPCs were prepared as described (Cossu et al., 1980). Briefly, CD1 mice were sacrificed by cervical dislocation and the hindlimb muscles were removed. Muscle fragments were digested with 2% collagenase II (Invitrogen) for 60 minutes at 37°C. Digested cells were discarded and fragments were incubated again with 0.05% trypsin for 15 minutes at 37°C with gentle agitation. After the incubation, isolated cells were collected and fragments were incubated again until the whole tissue was digested. Isolated cells were pooled, centrifuged and resuspended in DMEM supplemented with 20% pre-screened fetal bovine serum (FBS), 5% horse serum (HS), 5% chicken embryo extract (CEE), 1% gentamycin, and then plated onto collagen-coated dishes at 10⁴ cells/cm². Cells were pre-plated on plastic dishes for 3 hours in order to reduce contamination by non-myogenic cells.

For FACS-based sorting methods, cells from tissue extracts were counted and incubated with biotin-conjugated anti-SM/C2.6, anti-CD45-FITC and anti-CD31-PE for 1 hour at room temperature (RT). A second incubation step with streptavidin-APC was performed for 1 hour at RT. SM/C2.6⁺/CD45^–/CD31^– MPCs were sorted by flow cytometry (BD FACSDiva option).

Healthy human myoblasts were prepared from the skeletal muscle tissue of a 30-year-old patient undergoing post-traumatic surgery (all patients signed a specific informed consent approved by the institutional Ethics Committee at San Raffaele Hospital, Milan, Italy) and were maintained in culture as previously described (Dellavalle et al., 2007). Additionally, cells were FACS screened for the myoblast marker CD56 (NCAM1) (Miltenyi Biotec, 130-090-755) to confirm their phenotype. LGMD2D myoblasts were obtained from a biopsy sample (32-year-old patient) kindly provided by the Telethon Genetic BioBank Network, specifically by Dr Maurizio Moggio (Director, Bank of DNA, Cell Lines and Nerve-Muscle-Cardiac Tissues, Ospedale Maggiore Policlinico Mangiagalli e Regina Elena, Milan, Italy).

Mouse myogenic cell line C2C12, human myoblasts and the 293T cell line were maintained in DMEM with 10% FBS. C2C12 and MPCs were induced to differentiate into myotubes by replacing 10% FBS with 2% HS. Differentiation was completed in 5 days. For MAPK stimulation experiments, proliferating Sgca-null MPCs were stimulated for 30 minutes at 37°C with YP-740 (10 μM) or PMA (100 ng/ml).

Extensor digitorum longus (EDL) and soleus muscles from adult C57BL10 mice were isolated and digested in 0.008% collagenase type I (Sigma-Aldrich) for 20 minutes. Bundles of fibers protruding from the muscles were then collected using a wide-mouth glass Pasteur pipette. Following serial dilutions, single fibers were plated in 24-well plates coated with Matrigel (BD Biosciences). At day 1, plating medium was switched to proliferation medium (DMEM with 20% FBS, 1% CEE and 1% Pen/Strep) to allow proliferation of myogenic precursors. At day 5, myogenic precursors were fixed and stained for PAX7 and MYF5.

Cell pellets or tissue extracts were suspended in lysis buffer (RIPA buffer, pH 7.5; Sigma-Aldrich) supplemented with a Protease Inhibitor Cocktail (Roche). The whole-cell lysate or tissue extract was then centrifuged at 17,000 g for 10 minutes at 4°C. Protein lysates were boiled in SDS sample buffer for 5 minutes, subjected to 10% SDS-PAGE under reducing conditions and transferred onto Immobilon-P membranes (Millipore) using a semi-dry blotting apparatus (BioRad). The membranes were blocked with TBS-T buffer (Tris-buffered saline with 0.1% Tween 20) containing 5% skimmed milk for 30 minutes at RT and then incubated in TBS-T containing 5% skimmed milk with primary antibodies. The membranes were washed with TBS-T and incubated with horseradish peroxidase (HRP)-conjugated secondary antibody. Protein bands were detected using an enhanced chemiluminescence substrate (Amersham Biosciences) and a Fuji scanner equipped with a CCD camera.

Transfection experiments were carried out using Lipofectamine 2000 (Invitrogen). The day before transfection, cells were seeded according to the manufacturer's protocol. On day 0, DNA-Lipofectamine complexes were incubated for 20 minutes at RT, added directly to the cell lines and left overnight. The following day, transfection medium was removed and replaced with proliferation medium.

Active caspase 9 (Casp9) levels were measured using a mouse Caspase 9 ELISA Kit (#CSB-E08864m, Hoezel Diagnostika, Germany). Wild-type and Sgca-null MPCs cultured for 3 days in low-serum media with or without bFGF and HGF were processed to obtain cell lysates. Protein extracts were diluted to obtain equal starting concentrations and incubated in 96-well plates pretreated with anti-active Casp9 antibody. The total amount of active Casp9 was quantified using a Casp9-biotin/streptavidin-HRP conjugating system and measured using an ELISA reader at 450 nm.

Surface protein biotinylation was performed using the Cell Surface Protein Isolation Kit (#89881, Pierce). Briefly, four T75 flasks containing ∼1×10⁶ wild-type or Sgca-null MPCs were used as starting material and treated for 2 hours at 4°C with sulfo-NHS-SS-biotin. Biotin-labeled surface proteins were incubated for 2 hours with NeutrAvidin Agarose resin columns. Finally, surface proteins were eluted and subject to western blot analysis.

Fifty microliters of Dynabeads Protein G (#10007D, Invitrogen) were incubated for 1 hour at RT with 1.5 mg rabbit polyclonal anti-SGCA antibody to allow antibody coupling. Protein extracts (1-4 mg) were incubated overnight at 4°C with the antibody-coupled magnetic beads. As a negative control, parallel incubations with an irrelevant antibody were performed. The immune complexes were washed five times, eluted in 20 μl elution buffer and then incubated with 20 μl 2× Laemmli buffer at 70°C for 5 minutes. Immunoprecipitated complexes were detected with FGFR1 and SGCA antibodies.

MPCs were transduced as described (Sampaolesi et al., 2003) with pLenti6-V5/Sgca or with lentiviral vector pLenti-CMV/NLS-nlacZ. pGIPZ-shFgfr1 and pGIPZ-shScramble vectors were purchased from Open Biosystems. Following infection, transduced cells were selected in media containing puromycin (5 μg/ml).

Primary cultures of MPCs were washed with PBS and fixed with 4% paraformaldehyde (PFA) for 10 minutes at 4°C. Cells were stained with primary antibodies overnight at 4°C and then incubated with species-specific secondary antibodies for 40 minutes at RT. For tissue histology, serial muscle sections were fixed with 4% PFA, permeabilized with 0.2% Triton X-100, rinsed with 1% BSA in PBS and then immunostained with β-galactosidase antibody.

Phase contrast images were taken with a Nikon ECLIPSE TS-100 microscope. Immunofluorescence staining analysis was performed using a Nikon ECLIPSE Ti-S inverted microscope. Image editing was performed using ImagePro software (Media Cybernetics). Confocal images were taken with a Leica TCS SPE.

Values are expressed as mean ± s.e.m. When two groups were compared, an unpaired Student's t-test was used. P<0.05 was considered statistically significant. When three groups were compared, and a two-way ANOVA was used, P<0.05 was also considered statistically significant (Bonferroni post-hoc test). The statistical significance of the differences between percentage values was assessed using a Kruskal-Wallis one-way analysis of variance by ranks test. α represents the significance in each independent case.

We analyzed the clonogenic ability of MPCs isolated from the soleus and tibialis anterior muscles of wild-type (wt), Sgca-null and mdx mice (Fig. 1A). The total number of MPC clones per muscle was increased for Sgca-null mice, but the total number of cells per single clone was dramatically reduced (Fig. 1B,C). We hypothesized that the increased number of MPC clones could be due to continuous regeneration attempts in dystrophic mice that might augment the percentage of activated MPCs. This feature is indeed common to the mdx mouse (Fukada et al., 2010). However, the increased clone number does not compensate for the smaller clone size as the overall number of MPCs obtained per single isolation was strongly reduced in Sgca-null compared with both wt and mdx mice (see Fig. S1A in the supplementary material). However, mdx-isolated MPCs had approximately the same number of cells per clone as wt.

The number of nuclei per myotube was consistently reduced in Sgca-null clones compared with both wt and mdx clones (Fig. 1D). After 5 days in differentiation medium, anti-myosin heavy chain (MyHC) immunostaining revealed large MyHC⁺ myotubes in both wt and mdx clones but only a few, thin myotubes in Sgca-null clones (Fig. 1E), as a consequence of the reduced MPC number.

The contamination rate of non-myogenic cells in our MPC clonal cultures was extremely low (wt, 2.75±0.65; Sgca null, 3.45±0.97; percentage of non-myogenic cells ± s.e.m.) and comparable to a FACS-based isolation method (wt, 1.23±0.33; Sgca null, 2.867±0.96), excluding the possibility that the biological differences observed could be attributed to a significant presence of non-myogenic cells (see Fig. S1B-F in the supplementary material).

The proliferation defect of Sgca-null myogenic precursors was further confirmed by single fiber analysis (Fig. 1F-L). At early time points, Sgca-null and mdx mice showed a small increase in the total number of PAX7⁺ satellite cells per fiber compared with the wt counterpart, confirming the results obtained from clonal MPCs (Fig. 1F-H; statistical analysis not shown). After 5 days in proliferation medium, the total number of satellite cell-derived MPCs was dramatically decreased for Sgca-null-derived myofibers compared with wt and mdx MPCs (Fig. 1I-L). However, the proportion of cells in the different compartments (PAX7⁺/MYF5^–, PAX7⁺/MYF5⁺ and PAX7^–/MYF5⁺) did not appear to be drastically altered (Fig. 1I, red, green and yellow bars) when compared with wt and mdx counterparts.

Sgca is known to be expressed together with other dystrophin-glycoprotein complex (DGC)-associated components in differentiated myofibers, but not during proliferation, at least in the C2C12 myogenic cell line in which it has been investigated (Liu and Engvall, 1999). We examined Sgca expression during the proliferation and differentiation of wt MPCs. As expected, C2C12 cells did not express SGCA during proliferation (Fig. 2A, lane pC2C12), whereas expression was observed in proliferating MPCs at an early stage and this increased during differentiation (Fig. 2B and see Fig. S2A in the supplementary material). Differentiation endpoints corresponded to the highest expression level of SGCA (Fig. 2A,B), when the protein assumes its structural role as a member of the DGC complex (Lapidos et al., 2004). Interestingly, we observed that some as yet undifferentiated wt MPCs expressed SGCA before myogenin (Fig. 2C), while many cells co-expressed both proteins (Fig. 2C′′, yellow arrows). Moreover, several PAX7⁺ wt MPCs, corresponding to self-renewing cells, were found to be positive for SGCA expression (Fig. 2D′′, yellow arrows); for quantification of the PAX7⁺/SGCA⁺ pool (54.67±2.34%) from freshly isolated wt MPCs, see Fig. S2B in the supplementary material.

Cytokine treatment is known to enhance MPC proliferation (Allen and Boxhorn, 1989). We tested the effect of two potent mitogenic cytokines – hepatocyte growth factor (HGF) and basic fibroblast growth factor (bFGF; also known as FGF2) – on wt and dystrophic MPCs. As already reported, bFGF stimulates robust wt and mdx MPC proliferation in vitro (Fig. 3A, gray lines) (DiMario and Strohman, 1988; Johnson and Allen, 1995; Smith and Schofield, 1994). Moreover, in wt MPCs, cyclin D (CycD) expression is maintained due to the mitogenic effect of bFGF (Fig. 3B). By contrast, Sgca-null MPCs appeared to be insensitive to bFGF treatment and their proliferation curve was essentially the same with or without bFGF (Fig. 3A, compare black and gray lines). Indeed, CycD expression was barely detectable in bFGF-treated Sgca-null MPCs (Fig. 3B). By contrast, HGF treatment induced proliferation in wt MPCs and in both dystrophic MPCs, but no substantial differences existed at any of the time points analyzed (Fig. 3A, dashed lines). Furthermore, HGF-induced CycD expression was comparable in wt and Sgca-null MPCs (Fig. 3B).

To provide a further link between the absence of Sgca expression and bFGF signaling, loss-of-function experiments were performed after FGFR1 knockdown in wt MPC clones using a short hairpin (sh) RNA strategy. Silencing was very effective, as demonstrated by RT-PCR and western blot (WB) analysis of shFGFR1-infected wt MPCs (see Fig. S2C,D in the supplementary material). When treated with bFGF, shFGFR1 wt clones proliferated at a slower rate than scramble-infected control cells (Fig. 3C). The proliferation curve of the shFGFR1 wt clones matched that of the bFGF-treated Sgca-null clones (compare gray lines in Fig. 3A and 3C) and expression of CycD (Fig. 3E, lane wt + shFGFR1) and phosphorylated (p) ERK1/2 (also known as MAPK3/1) (see Fig. S3C in the supplementary material) was barely detectable.

Using YP-740 (Derossi et al., 1998) and PMA (Weiss et al., 1994) stimulation, we determined that downstream activation of the bFGF pathway in Sgca-null MPCs restores the proliferative rate to levels comparable to wt MPCs, as demonstrated by growth curves (Fig. 3D), CycD (Fig. 3E, lanes Sgca null + YP-740 and Sgca null + PMA) and pErk1/2 (see Fig. S3C in the supplementary material) expression analysis. A defective bFGF response is also present in humans, with patient-derived bFGF-treated LGMD2D myoblasts showing impaired proliferation compared with their healthy counterparts (Fig. 3F, gray lines). However, HGF treatment led to an increase in the total cell number of both healthy and LGMD2D-derived myoblasts (Fig. 3F, dashed lines). CycD expression analysis of human myoblasts revealed higher levels in healthy myoblasts treated with both bFGF and HGF, whereas LGMD2D-derived myoblasts responded to HGF stimulation only (Fig. 3G). Importantly, WB analysis of pERK1/2 confirmed the impairment of proliferation in Sgca-null myogenic cells during bFGF stimulation (see Fig. S3A,B,D in the supplementary material). Additionally, a strong decrease (by 75%) in phosphohistone H3 (pHH3)/MYF5 immunoreactivity in Sgca-null MPCs further reinforces the idea that a bFGF-associated impairment in proliferation is present (see Fig. S4A-C in the supplementary material).

It has been reported that the absence of SGCA can alter integrin-dependent adhesion and motility pathways, causing enhanced apoptosis (Allikian et al., 2004; Anastasi et al., 2004; Yoshida et al., 1998). To test this possibility, active caspase 9 (Casp9) levels were measured in wt and Sgca-null MPCs cultured both with and without bFGF or HGF (see Fig. S4D in the supplementary material). Only a modest increase (15%) in Casp9 levels was observed in Sgca-null MPCs cultured in the presence of bFGF, making it unlikely that enhanced apoptosis is the main cause for the reduced number of cells observed in Sgca-null MPCs. Altogether, these data confirm that the impairment of proliferation in Sgca-null MPCs is due to a defective bFGF-FGF receptor axis.

The data reported above suggest a possible link between Sgca and bFGF-driven MPC proliferation. To prove this hypothesis, we characterized the expression levels of FGFR1 and FGFR4, which are involved during the myogenic program in vitro and in vivo (Kastner et al., 2000), in wt and Sgca-null MPCs. Interestingly, we found that Sgca-null MPCs upregulated both Fgfr1 and Fgfr4 at the mRNA level in proliferating medium, as compared with their wt counterparts, possibly as a compensatory mechanism (Fig. 4A). These data are consistent with gene expression profiling studies on tibialis anterior muscle from Sgca-null mice (Biressi et al., 2007), which show marked upregulation of Fgfr1 mRNA (see Fig. S5A in the supplementary material).

FGFR1 is normally upregulated during the early differentiation stage of C2C12 cells (see Fig. S5B in the supplementary material). WB analysis revealed similar levels of FGFR1 expression in wt and Sgca-null MPCs (Fig. 4B). Similarly, wt and Sgca-null MPCs did not show any substantial differences in the expression levels or cellular localization of FGFR4 (see Fig. S5C,D in the supplementary material). These results provide evidence that the proliferative defect of Sgca-null MPCs is not a secondary effect of an enhanced FGFR4-dependent spontaneous differentiation.

During proliferation, FGFR1 localized on the cell membrane together with SGCA only in wt MPCs (Fig. 4D,D′), whereas in Sgca-null MPCs FGFR1 expression was mainly confined to the nucleus (Fig. 4C,C′) (Kilkenny and Hill, 1996; Myers et al., 2003; Peng et al., 2002). For this reason, we analyzed the subcellular distribution of FGFR1 and SGCA in wt MPCs. Immunohistochemical staining revealed partial colocalization of FGFR1 and SGCA in small, rounded cells, which are likely to be proliferating or early differentiating MPCs (Fig. 4E′, white arrow). Confocal analysis of MPCs at day 1 of differentiation confirmed partial colocalization of the two proteins at the plasma membrane (Fig. 4F′, white arrows). At late stages of differentiation, when myotubes are formed, FGFR1 no longer colocalized with SGCA and was mainly restricted to the nucleus (Fig. 4G′), as previously reported (Myers et al., 2003). To support the claim that FGFR1 and SGCA colocalize on the cell membrane, we assessed the amount of FGFR1 on the cell membrane in wt and Sgca-null MPCs by surface protein biotinylation analysis (Fig. 4H). Again, FGFR1 localized on the cell surface only in wt and not in Sgca-null MPCs, providing direct evidence that SGCA is required for proper localization of the receptor. Therefore, FGFR1 mislocalization cannot be attributed to the altered expression of heparan sulfate proteoglycans (HSPGs) (Filla et al., 1998; Steinfeld et al., 1996), as demonstrated by WB analysis for syndecan 1 in wt and Sgca-null MPCs (Fig. 4B). Taken together, these results suggest that an interaction exists between bFGF signaling and SGCA during MPC proliferation but not after differentiation.

We investigated a possible interaction between SGCA and FGFR1 by immunoprecipitation and WB analysis in non-myogenic and myogenic cells. Human 293T cells expressing Sgca and Fgfr1 under the control of a CMV promoter were subjected to immunoprecipitation analysis. Proteins were resolved from cell lysates using SGCA antibodies as bait and then immunoblotted with FGFR1 antibodies (Fig. 4I). Co-immunoprecipitation experiments on double-transfected 293T cells revealed that SGCA effectively binds FGFR1 (Fig. 4I). The absence of SGCA signal in immunoprecipitates of single-transfected 293T cells (Fig. 4I) highlight the specificity of the interaction. SGCA polyclonal antibodies were also able to co-immunoprecipitate SGCA-FGFR1 complexes in wt skeletal muscle and MPC cultures (lanes 1d, 3d and wt in Fig. 4J), confirming the above results. Moreover, cardiotoxin-induced muscle damage and regeneration augmented the amount of SGCA-FGFR1 complexes detected (lane wt + ctx in Fig. 4J). Thus, SGCA might regulate MPC proliferation via the formation of a complex with FGFR1, rather than playing a structural role as it does after differentiation.

To address the possible role of Sgca in regulating the self-renewal and myogenic potential of MPCs, wt and dystrophic MPCs were transplanted into an Sgca-null recipient. We isolated GFP⁺ MPCs from wt and Sgca-null mice backcrossed with human phosphoglycerate kinase (hPGK)-GFP transgenic animals. Before injection, MPC purity was evaluated by double staining for GFP and MYOD (Fig. 5A-B′) and found to be ∼70% for both Sgca-null and wt cells (Fig. 5C). The extent of engraftment was evaluated by measuring GFP mRNA expression in the injected muscles at different time points. Direct comparison of GFP expression from wt and Sgca-null MPCs revealed a profound impairment of engraftment of dystrophic MPCs (Fig. 5D). Wt MPC transplantation resulted in a modest but progressive increase in GFP expression. Conversely, the GFP expression level was initially similar in mice injected with dystrophic MPCs but decreased rapidly to become barely detectable 6 days after injection.

To test whether the engraftment impairment of dystrophic MPCs was due to poor survival, we analyzed apoptosis of donor cells in vivo. MPCs isolated from wt and Sgca-null mice were transduced with pLenti-CMV/NLS-nlacZ, a lentiviral vector that expresses a nuclear β-galactosidase (β-gal) reporter. Following lentiviral infection, β-gal⁺ MPCs were transplanted into the tibialis anterior of juvenile Sgca-null mice and apoptotic donor cells were identified by β-gal⁺/TUNEL⁺ nuclei. An increase of apoptotic events (more than 2.5-fold higher after 3 days, both in resident and donor cells) was observed in dystrophic mice, reflecting a survival defect in resident Sgca-null MPCs (Fig. 5E,). Therefore, at day 3 after injection, the apoptotic rate of transplanted cells (the total number of cells that were β-gal⁺/TUNEL⁺) was 3-fold higher in dystrophic MPCs than in their wt counterparts (Fig. 5F). At day 10 after injection, when apoptotic events were almost completed, there was no longer any significant difference in the apoptotic rate between wt and dystrophic MPCs (Fig. 5E,F). These data show that SGCA deficiency interferes functionally with the biological properties of MPCs, which become more susceptible to apoptosis and fail to properly colonize skeletal muscle.

To understand to what extent proliferation defects and increased cell mortality can be attributed to Sgca deficiency, genetic correction of dystrophic MPCs was performed. We used a lentiviral vector carrying the murine coding sequence of Sgca under the control of a CMV promoter and fused it in frame with a synthetic V5 epitope to further discriminate protein localization. We obtained 95% transduction efficiency after one round of infection at a multiplicity of infection of 10 (data not shown). Genetic correction led to an increase in the number of cells per Sgca-null MPC clone as compared with non-corrected MPCs (Fig. 6A-C). Moreover, when the myogenic potential of corrected MPCs was analyzed under differentiating conditions, cells formed large multinucleated myotubes (Fig. 6D,D′) with the number of nuclei per myotube comparable to that in wt (Fig. 6E). In addition to the rescue of proliferation, genetic correction restores the SGCA-FGFR1 interaction, as pull-down experiments in corrected Sgca-null MPCs detected the presence of FGFR1 in protein extracts precipitated with an anti-SGCA antibody (Fig. 6F).

Corrected Sgca-null GFP⁺ MPCs were injected into dystrophic mice to further investigate their capacity for functional rescue. In vivo engraftment of Sgca-null MPCs was restored, as demonstrated by high levels of GFP expression in injected muscles at all time points analyzed (Fig. 6G).

Altogether, these data support the hypothesis that lack of Sgca expression affects MPC proliferation as well as cell survival by interfering with bFGF signaling.

Satellite cells are quiescent mononucleated cells, located between the basal lamina and sarcolemma of adult skeletal muscle (Gnocchi et al., 2009; Mauro, 1961; Zammit et al., 2006). They contribute to postnatal growth by supplying myonuclei to the growing muscle fibers and become quiescent at the end of postnatal growth (Le Grand and Rudnicki, 2007). However, upon injury to the tissue, satellite cells become activated and proliferate as MPCs. MPCs function to repair or replace damaged fibers (Cassano et al., 2009; Collins et al., 2005) and then part of the cell population returns to quiescence in order to replenish the satellite cell pool (Zammit et al., 2004). Therefore, satellite cells are considered the unipotent adult stem cells of skeletal muscle (Zammit et al., 2006).

Satellite cells are the most relevant to muscle repair because of their abundance, location and robust myogenic commitment (Cossu and Biressi, 2005). However, the development of cell therapies based upon satellite cell transplantation has been hampered by poor engraftment, poor ability to migrate inside the tissue and an inability to cross the endothelial barrier when injected into the vasculature (Cossu and Sampaolesi, 2007; Sampaolesi et al., 2003). In previous work, we observed that systemic delivery of mesoderm progenitors, termed mesoangioblasts, rescued the dystrophic phenotype of Sgca-null mice. The success of this approach was mainly due to widespread distribution of donor cells through the capillary network. As in other cases of successful engraftment (Aiuti et al., 2002), we questioned whether mesoangioblast transplantation into Sgca-null mice had been favored as a result of poor competition with resident satellite cells. In other cases of cell transplantation in mdx dystrophic mice (Montanaro et al., 2003; Torrente et al., 2004), although engraftment and dystrophin reconstitution have been documented, the extent of recovery was less successful than for the Sgca-null mice. This might depend upon the better engraftment potential of mesoangioblasts in comparison with side population (SP) or AC133 cells, or, alternatively, be influenced by a different microenvironment in the two types of dystrophic mice. Differences in transplantation efficiency might concern the level of inflammatory cell infiltration, the deposition of fibrous tissue, as well as the biological features of resident myogenic progenitors. Therefore, we focused on the satellite cells of Sgca-null mice, assuming that a defect in these cells might result in reduced competition with donor cells, thus explaining the extensive engraftment and phenotypic rescue elicited by mesoangioblast transplantation in the Sgca-null mouse.

In this work, we have studied the role of SGCA in satellite cell proliferation and differentiation. In the late 1980s, DiMario and Strohman reported that when satellite cells cultured from mdx and control mouse hindlimb muscles are stimulated with bFGF, they proliferate and fuse to form muscle fibers within 4 to 5 days (DiMario and Strohman, 1988). We found that Sgca-null satellite cells show reduced proliferation in comparison with their wt or mdx counterparts owing to reduced sensitivity to bFGF, a major mitogen for these cells. The proliferative defect of Sgca-null myogenic precursors was confirmed by both clonogenic assays and by single fiber analysis. Importantly, this effect was specific for bFGF, as proliferation in response to HGF, another mitogen for satellite cells (Gal-Levi et al., 1998), was unaltered.

In order to elicit any effect on proliferating MPCs, SGCA should be expressed at detectable levels during proliferation, when dystrophin and most of the other associated proteins are not yet expressed. Indeed, we observed expression of SGCA at the early stages of satellite cell activation, when we detected PAX7⁺/SGCA⁺ MPCs. At these stages, SGCA colocalizes with FGFR1 and a lack of SGCA expression in MPCs alters the localization of the receptor, which leads to intracellular accumulation of FGFR1. Although modest amounts of SGCA may be sufficient to help FGFR1 localization on the cell membrane, higher levels are needed for SGCA to play its structural role as a DGC member. This fact might well explain the upregulation of SGCA expression during differentiation. Our co-immunoprecipitation studies confirmed the formation of a complex between SGCA and FGFR1 in MPCs, but also confirmed complex formation in a non-myogenic cell system in which both proteins were overexpressed. Thus, under normal conditions, SGCA positively modulates FGF signaling by forming stable complexes with FGFR1.

FGFR1 is present in both myogenic and non-myogenic cells in adult skeletal muscle (Kastner et al., 2000) and muscle development and homeostasis are severely affected when FGFR1 is absent (Marics et al., 2002). Previous studies put forth the hypothesis that FGFR4 expression patterns correlate with the growth-differentiation transition of satellite cells, whereas FGFR1 is mainly involved in maintaining the proliferative state of MPCs (Itoh et al., 1996; Marics et al., 2002; Olwin et al., 1994; Weinstein et al., 1998). The observation that the expression levels and cellular localization of FGFR4 were unaffected in Sgca-null as compared with wt MPCs rules out the possibility that the proliferation impairment in Sgca-null MPCs is a consequence of enhanced FGFR4-dependent differentiation. Impaired proliferation was reverted by transduction with lentiviral vectors carrying the Sgca gene, confirming a direct role of Sgca in MPC proliferation impairment.

MPCs isolated from Sgca-null mice failed to properly colonize the regenerating muscle, whereas wt cells engrafted efficiently and survived in vivo. Interestingly, an association of SGCA with integrins has been reported (Allikian et al., 2004; Anastasi et al., 2004; Yoshida et al., 1998), and thus the absence of SGCA might also contribute to impaired migration through inadequate integrin-dependent adhesion and motility. This might also explain, at in least in part, the enhanced apoptosis observed in Sgca-null MPCs in vitro and in vivo.

The regulation of the cell cycle and cell death are often coupled and a fine balance between these processes is required for many biological processes, such as tissue development and homeostasis and regeneration (Gil-Gomez et al., 1998). Here, we present evidence for a synergistic effect between defective bFGF signaling and enhanced apoptosis resulting in a functional defect of Sgca-null MPCs.

Two main features distinguish muscular dystrophy in Sgca-null versus mdx mice: the early onset of the disease and the severity of the pathology. In mdx mice, the pathology is mild after a degeneration-regeneration crisis at 3-4 weeks of age (McArdle et al., 1995), whereas the persistence of degeneration in Sgca-null mice results in a more severe pathology.

In our initial attempt at cell therapy for muscular dystrophy, we selected Sgca-null mice for the reasons outlined above. The results we obtained were better than those obtained with mdx, but we did not know at that time that the success in engraftment was due, at least in part, to the proliferation impairment of Sgca-null satellite cells. In line with this observation, the phenotype of mdx mice is dramatically worsened if proliferation of satellite cells is hampered by deletion of the telomerase gene (Sacco et al., 2010) or when the endogenous progenitors are exhausted with long-term serial and intense exercise, with consequent enhanced mesoangioblast engraftment (Tedesco et al., 2011). Interestingly, the same proliferation defect is observed in human MPCs from LGMD2D patients, whose dystrophy, however, remains milder than DMD. This could be explained by the fact that human cells show a reduced proliferative potency compared with those of mice, and that other factors, such as the residual expression of dystrophin (Klinge et al., 2008; Vainzof et al., 2000), might act at the level of fiber stability rather than in the regeneration process. The phenotypic difference observed between LGMD2D and Sgca-null mice, together with the large variability in the severity of the human disease, might also reflect the mutation type of the SGCA gene together with species-specific compensatory and regulatory mechanisms (Kobuke et al., 2008).

SGCA is crucial for postnatal myogenesis. It forms a complex with dystrophin in differentiated muscle and participates in bFGF and integrin signaling as part of a program that regulates satellite cell proliferation and survival. bFGF stimulates proliferation and prevents the precocious differentiation of satellite cell-derived MPCs (Lapidos et al., 2004; Sheehan and Allen, 1999). The SGCA-FGFR1 interaction is necessary for this proliferative signaling. When the myogenic pattern is upregulated, as satellite cell progeny differentiate, the SGCA-FGFR1 interaction ceases and the SGCA acquires its definitive role in the membrane dystrophin complex (Fig. 7).

We thank Gianpaolo Papaccio for critical discussion of the manuscript; K. P. Campbell for providing Sgca-null mice; Laura Perani, Rossana Tonlorenzi and Rudi Micheletti for technical assistance; Christina Vochten for professional secretarial service; Paolo Luban for a kind donation; Dr Shea Carter for manuscript editing; and Dr Maurizio Moggio (Bank of DNA, Cell Lines and Nerve-Muscle-Cardiac Tissues/Telethon Genetic BioBank Network, Ospedale Maggiore Policlinico Mangiagalli e Regina Elena, Milan, Italy) for providing the LGMD2D sample.