S100B protein has been shown to exert anti-myogenic and mitogenic effects in myoblast cultures through inhibition of the myogenic p38 MAPK and activation of the mitogenic ERK1/2. However, the receptor mediating these effects had not been identified. Here, we show that S100B increases and/or stabilizes the binding of basic fibroblast growth factor (bFGF) to bFGF receptor 1 (FGFR1) by interacting with bFGF, thereby enhancing FGFR1 activation and the mitogenic and anti-myogenic effects of FGFR1. S100B also binds to its canonical receptor RAGE (receptor for advanced glycation end-products), a multi-ligand receptor previously shown to transduce a pro-myogenic signal when activated by HMGB1, and recruits RAGE into a RAGE–S100B–bFGF–FGFR1 complex. However, when bound to S100B–bFGF–FGFR1, RAGE can no longer stimulate myogenic differentiation, whereas in the absence of either bFGF or FGFR1, binding of S100B to RAGE results in stimulation of RAGE anti-mitogenic and promyogenic signaling. An S100B–bFGF–FGFR1 complex also forms in Rage−/− myoblasts, leading to enhanced proliferation and reduced differentiation, which points to a dispensability of RAGE for the inhibitory effects of S100B on myoblasts under the present experimental conditions. These results reveal a new S100B-interacting protein – bFGF – in the extracellular milieu and suggest that S100B stimulates myoblast proliferation and inhibits myogenic differentiation by activating FGFR1 in a bFGF-dependent manner.
Myogenesis is a multistep process during which cells committed to the myoblast lineage proliferate, migrate to body sites of skeletal muscle formation, cease to proliferate and eventually fuse to form myotubes, the precursors of myofibers (Chargé and Rudnicki, 2004; Buckingham, 2006). Similarly, regeneration of injured skeletal muscles requires the activation of muscle satellite cells (i.e. quiescent cells located beneath the basal lamina of myofibers), followed by their migration, proliferation and differentiation into fusion-competent myocytes (Brack et al., 2007; Chargé and Rudnicki, 2004; Kuang et al., 2008). Myoblast proliferation is a crucial step of the myogenic process; diminished proliferation severely affects myogenesis owing to insufficient cell density for subsequent myotube formation. By contrast, excess proliferation interferes with myogenesis because proliferation and differentiation are mutually exclusive. A relatively large number of extracellular factors take part in the regulation of myoblast proliferation and differentiation, such as insulin and insulin-like growth factors, basic fibroblast growth factor (bFGF), hepatocyte growth factor (HGF), myostatin, transforming growth factor-β, members of the Wnt family and several cytokines, with some factors stimulating and others inhibiting, myogenesis (Brack et al., 2007; Chargé and Rudnicki, 2004; Guasconi and Puri, 2009; Kuang et al., 2008).
S100B, a Ca2+-binding protein of the EF-hand type (Donato et al., 2009) that is expressed in mature myofibers (Arcuri et al., 2002), has been shown to modulate myoblast differentiation by reducing the activation of p38 mitogen-activated protein kinase (MAPK) and inhibiting the expression of myogenin (Sorci et al., 2003), a muscle-specific transcription factor essential for myogenic differentiation. It is known that activation of p38 MAPK is indispensable for myoblast differentiation in vitro and in vivo (Bennet and Tonks, 1997; de Angelis et al., 2005; Guasconi and Puri, 2009; Lluis et al., 2006; Perdiguero et al., 2007). S100B can affect myogenesis at subnanomolar–nanomolar concentrations, levels that are similar to the S100B levels measured in serum during development (Portela et al., 2002) and those found in crushed muscle extract (CME) (see below). In addition, S100B stimulates the proliferation of myoblasts in differentiation medium (DM) and reduces their apoptosis (Riuzzi et al., 2006b).
However, although RAGE (receptor for advanced glycation end-products) (Bierhaus et al., 2005; Schmidt et al., 2001) is a recognized S100B receptor, transduces S100B regulatory effects in several cell types (Donato, 2007; Donato et al., 2009; Hofmann et al., 1999; Leclerc et al., 2009) and is expressed in myoblasts (Sorci et al., 2003), RAGE was observed not to transduce the effects of S100B in myoblasts (Sorci et al., 2003). By contrast, RAGE, activated by its ligand, high-mobility group protein box 1 (HMGB1), has been shown to transduce a pro-myogenic, anti-proliferative and pro-apoptotic signal in myoblasts (Riuzzi et al., 2006a; Riuzzi et al., 2007; Sorci et al., 2004).
We show here that: (i) S100B is found in CME and contributes to CME-induced myoblast proliferation; (ii) the effects of S100B on C2C12 and primary myoblasts in DM vary depending upon the duration of treatment of myoblasts with the protein; (iii) clearance of S100B results in enhanced myoblast differentiation as a consequence of an S100B-induced increase in the cell number; (iv) S100B interacts with bFGF and increases and/or stabilizes the binding of bFGF to FGFR1, thereby enhancing FGFR1 signaling; and (v) that the S100B–bFGF–FGFR1 adduct recruits RAGE into a RAGE–S100B–bFGF–FGFR1 complex, resulting in the blockade of RAGE signaling. In addition, in Rage−/− myoblasts, S100B engages FGFR1 through prior binding to bFGF, thereby stimulating proliferation and inhibiting differentiation. We propose that S100B participates in the myogenic process by activating FGFR1 in a bFGF-dependent manner.
S100B is present in CME and contributes to CME-induced primary myoblast proliferation
S100B is expressed in skeletal muscle fibers (Arcuri et al., 2002), which means that there is the potential for its release in the case of muscle injury. Indeed, CME obtained as described previously (Chen and Quinn, 1992), contained approximately fourfold the amount of S100B present in control uncrushed muscle fibers (i.e. ~1 nM S100B compared with ~0.25 nM S100B) (Fig. 1A). In addition, CME stimulated primary mouse myoblast proliferation, but it was less able to do so when pretreated with an S100B-neutralizing antibody (Fig. 1B), suggesting that the CME mitogenic effect was due in part to released S100B and that, in the case of muscle injury, released S100B could play a regulatory role in the regeneration process. To obtain more detailed information about this latter issue, we next addressed whether the ability of S100B to stimulate myoblast proliferation might be beneficial to the myogenic process and what the membrane receptor transducing the effects of S100B on myoblasts might be.
S100B reversibly inhibits C2C12 and primary myoblast differentiation
When S100B was administered at the time of transfer of C2C12 mouse myoblasts from GM into DM and the cultures were left undisturbed for up to 3 days (i.e. without renovation of the medium), S100B inhibited differentiation (supplementary material Fig. S1A–D), stimulated proliferation (supplementary material Fig. S1E), activated the mitogenic extracellular signal-regulated kinases (ERK)1/2 and inhibited p38 MAPK (supplementary material Fig. S1F). Similar results were obtained using primary myoblasts (supplementary material Fig. S1A–C).
However, when myoblasts were left undisturbed for 6 days, no differences were seen between S100B-treated and untreated cultures in terms of the levels of the late myogenic differentiation marker myosin heavy chain (MyHC) (Fig. 2A), and numbers and size of myotubes (Fig. 2B). Compared with that in controls, myoblasts treated with a single dose of S100B showed reduced myogenin and MyHC levels and enhanced MyoD levels, along with a lower fusion index, at days 1 and 3, but they had a similar MyoD, myogenin and MyHC level and fusion index at day 6 (supplementary material Fig. S2). Because the levels of S100B in culture medium from S100B-treated and untreated myoblasts decreased substantially after day 3 (Fig. 2C), we concluded that S100B inhibited myoblast differentiation and fusion as long as it was present in sufficient amounts and that reduction of the S100B levels in culture medium after day 3 (Fig. 2C), coupled to the S100B-induced increase in the myoblast number (supplementary material Fig. S1E), might result in enhanced differentiation at day 6 (Fig. 2A,B; supplementary material Fig. S2). Indeed, administration of a second dose of S100B at day 4 reduced the recovery effects from the first dose (Fig. 2D, compare lanes 3 and 4 with lanes 5 and 6, respectively) and daily administration of S100B for 6 days (to keep the extracellular S100B concentration relatively high; Fig. 2E, third panel from top) caused a dramatic inhibition of myogenic differentiation (Fig. 2E) and stimulation of proliferation (data not shown).
S100B effects on myoblast differentiation vary depending on the duration of treatment
On the basis of the results in Fig. 2A,B, we reasoned that the S100B-induced increase in the myoblast number might favor myotube formation once S100B has been cleared. Treatment of myoblasts with S100B for 24 hours followed by S100B washout resulted in inhibition of differentiation, as measured at the end of the next 24 hours (Fig. 3A). In this same condition, S100B-treated myoblasts showed more activated ERK1/2, less activated p38 MAPK and a reduced MyHC level compared with that in controls (Fig. 3B), as well as reduced levels of M-cadherin (Fig. 3B), which is known to be required for myoblast fusion (Charrasse et al., 2007; Zeschnigk et al., 1995). However, at 2 and 5 days after washout, S100B-treated myoblasts showed enhanced MyHC levels compared with those in controls (Fig. 3C), suggesting that the S100B-induced increase in the myoblast number during the first 24 hours of treatment (Riuzzi et al., 2006b) (see supplementary material Fig. S1E) favored myogenic differentiation after S100B clearance. The 48-hour latency of the effects of S100B washout on MyHC levels in S100B-treated myoblasts is probably due to the large number of myoblasts completing the cell cycle during that time interval, although we cannot exclude other reasons. Thus, the duration of treatment with S100B strongly influences the effects of S100B on myoblast differentiation: a long-term treatment causes stimulation of proliferation and inhibition of differentiation, whereas a short-term treatment resulted in stimulation of differentiation, as measured at 2 or 5 days after S100B washout, owing to increased myoblast number.
S100B activates FGFR1 but not RAGE in myoblasts
Because the effects exerted by S100B in myoblasts were reminiscent of those of bFGF and HGF in several aspects, we explored the possibility that S100B activated FGFR1 and/or the HGF receptor Met. Treatment of myoblasts with an FGFR1-neutralizing antibody resulted in a reduced basal ERK1/2 phosphorylation, an unchanged basal p38 MAPK phosphorylation, an S100B-dependent inhibition of ERK1/2 phosphorylation (Fig. 4A) and a stimulation of p38 MAPK phosphorylation (Fig. 4A) and myoblast differentiation (Fig. 4B,C), with an inability of S100B to stimulate proliferation (Fig. 4E). In accordance with these findings, knockdown of FGFR1 in myoblasts by RNA interference resulted in enhanced differentiation and reduced proliferation, and the presence of S100B amplified these effects (Fig. 4F). We could not analyze the effects of S100B in the presence of a Met-neutralizing antibody because commercial antibodies to Met are not neutralizing (however, see below). Because RAGE is a recognized S100B receptor in several cell types (Donato, 2007; Leclerc et al., 2009) and is expressed in myoblasts (Sorci et al., 2003), we also analyzed the effects of S100B on ERK1/2 and p38 MAPK in the presence of a RAGE-neutralizing antibody. In myoblasts treated with the anti-RAGE antibody, control cells showed reduced p38 MAPK phosphorylation and enhanced ERK1/2 phosphorylation (Fig. 4A), in line with the ability of RAGE to activate p38 MAPK and inactivate ERK1/2 in myoblasts in DM (Riuzzi et al., 2006a; Sorci et al., 2004). Because, in this condition, the effects of S100B on ERK1/2 and p38 MAPK were essentially the same as in native conditions (Fig. 4A), and S100B reduced myogenin and MyHC levels (Fig. 3B) and myotube formation (Fig. 3D), we conclude that RAGE does not mediate the effects of S100B in myoblasts. In the presence of an FGFR1-neutralizing plus a RAGE-neutralizing antibody, ERK1/2 and p38 MAPK phosphorylation levels were low and the same in the absence or presence of added S100B (Fig. 4A). Thus, the ability of S100B to activate ERK1/2, stimulate proliferation, inactivate p38 MAPK and inhibit myogenesis relies on functional FGFR1. However, upon neutralization of FGFR1 or reduction of its abundance in myoblasts, S100B inhibited ERK1/2, stimulated p38 MAPK, and increased MyHC expression and myotube formation (Fig. 4A,C–F), suggesting that in the absence of functional FGFR1, S100B activates the pro-myogenic RAGE. Thus, short-term treatment with S100B results in the activation of FGFR1 but not RAGE in myoblasts; however, blocking FGFR1 means S100B is able to activate the pro-myogenic RAGE.
bFGF, FGFR1 and RAGE co-immunoprecipitate with S100B in myoblasts
Because treatment with S100B resulted in activation of FGFR1 but not RAGE in myoblasts, and blocking FGFR1 meant S100B was able to activate the promyogenic RAGE, we asked whether S100B was able to interact with FGFR1 and/or RAGE. RAGE, but not FGFR1 or Met, was detected in the fraction of polypeptides eluted from an S100B–Sepharose column previously equilibrated with the Triton X-100 extract of myoblast plasma membranes in the presence of Ca2+ (Fig. 5A). Given the relatively high amount of the Sepharose-bound S100B used, competition between RAGE, FGFR1 and Met for binding to S100B could be excluded. RAGE binding to S100B–Sepharose was dependent on Ca2+, as RAGE was eluted by EGTA (Fig. 5A); no RAGE, FGFR1 or Met bound to S100B–Sepharose when Ca2+ was omitted (data not shown). However, these results seemed to be contradictory to the RAGE-independency and FGFR1-dependency of the S100B effects (Fig. 4). Thus, we performed co-immunoprecipitation analyses to identify the receptor(s) engaged by S100B in myoblasts.
RAGE and FGFR1, but not Met, co-immunoprecipitated with S100B, and S100B co-immunoprecipitated with RAGE and FGFR1 (Fig. 5B). Thus, S100B bound to RAGE without activating it and activated FGFR1 without binding to it. Hence, we investigated possible mechanisms whereby S100B could engage FGFR1 without binding to it and bind to RAGE without activating it in myoblasts. Because at 24 hours after the switch to DM and in the absence of additions, culture medium contained bFGF (Fig. 5C), we hypothesized that bFGF in the culture medium mediated the S100B interaction with FGFR1 and the S100B-dependent stimulation of FGFR1 signaling. Indeed, we found that bFGF did co-immunoprecipitate with S100B (Fig. 5D); however, pre-treatment of cultures with a bFGF-neutralizing antibody negated bFGF and FGFR1, but not RAGE, co-immunoprecipitation with S100B (Fig. 5D). In addition, S100B, bFGF and RAGE co-immunoprecipitated with FGFR1 (Fig. 5B,E); FGFR1, RAGE and S100B co-immunoprecipitated with bFGF (Fig. 5F); and S100B, FGFR1 and bFGF co-immunoprecipitated with RAGE (Fig. 5B,G). These results suggest that, first, S100B activates FGFR1 through prior binding to bFGF, thereby stimulating bFGF–FGFR1 mitogenic and anti-myogenic activity; second, that S100B also binds to RAGE, but, owing to the simultaneous S100B binding to bFGF–FGFR1, no S100B-dependent activation of RAGE occurs; and, third, when no S100B binding to bFGF–FGFR1 occurs (Fig. 5D), S100B can activate RAGE (Fig. 4A–C). Other membrane molecules such as M-cadherin, N-cadherin and N-CAM, which have been shown to play a role in myoblast terminal differentiation and fusion (Charrasse et al., 2002; Charrasse et al., 2007; Cifuentes-Diaz et al., 1993; Gavard et al., 2004; George-Weinstein et al., 1997; Péault et al., 2007; Zeschnigk et al., 1995), were not found in S100B immunoprecipitates (data not shown).
S100B enhances bFGF binding to FGFR1, and the S100B–bFGF complex cross-links RAGE and FGFR1
In myoblasts pretreated with the anti-S100B antibody, virtually no S100B, bFGF or FGFR1 co-immunoprecipitated with RAGE (Fig. 5G). This result, and the observation that, in myoblasts pretreated with anti-bFGF antibody, RAGE but not FGFR1 co-immunoprecipitated with S100B (Fig. 5D), indicates that the presence of both S100B and bFGF is required for FGFR1 and RAGE to co-immunoprecipitate and that the S100B–bFGF complex can crosslink RAGE and FGFR1. In addition, we found that larger amounts of bFGF co-immunoprecipitated with FGFR1 when bFGF was added to myoblasts, compared with that in basal conditions, and even higher amounts did so when myoblasts were treated with bFGF plus S100B, compared with that in basal conditions and internal controls (Fig. 5H). Moreover, a higher amount of FGFR1 co-immunoprecipitated with S100B in myoblasts treated with 1 nM bFGF plus 1 nM S100B compared with that upon treatment with S100B alone (Fig. 5I). These results suggest that the S100B–bFGF–FGFR1 complex formation is dose-dependent with respect to bFGF and S100B, and that S100B enhances and/or stabilizes bFGF binding to FGFR1. Accordingly, in the presence of a bFGF-neutralizing antibody, S100B was no longer able to stimulate myoblast proliferation (Fig. 5J) and S100B actually stimulated myoblast differentiation (Fig. 5J), which is similar to the results with myoblasts treated with S100B in the presence of an FGFR1-neutralizing antibody (Fig. 4C–E) and with myoblasts in which FGFR1 was knocked down by RNA interference (Fig. 4F).
By utilizing an in situ proximity ligation assay (PLA) (Söderberg et al., 2006), we detected immunocomplex dot formation indicative of a close association between FGFR1 and RAGE in myoblasts that had been treated with S100B, but not in myoblasts that had been pretreated with either an S100B-neutralizing or a bFGF-neutralizing antibody (Fig. 6A). In addition, dot formation was observed when the test was performed using anti-S100B and anti-bFGF antibodies, pointing to close association of the two ligands, with absence of immunocomplex dot formation in cultures pretreated with a bFGF-neutralizing antibody (Fig. 6B). Moreover, close association between S100B and FGFR1 was detected in the presence, but not in the absence, of bFGF (Fig. 6C). Finally, close association of S100B with RAGE was seen irrespective of the absence or presence of bFGF (Fig. 6D). Collectively, these results suggest that S100B promotes the formation of a RAGE–S100B–bFGF–FGFR1 complex, thereby enhancing the stimulatory effects of FGFR1 on myoblast proliferation, and its inhibitory effects on differentiation, while simultaneously blocking the pro-myogenic signaling of RAGE (see Discussion).
S100B enhances FGFR1 phosphorylation and inhibits myogenic differentiation independently of its mitogenic effect
After 30 minutes, myoblasts treated with S100B showed higher phosphorylation levels of FGFR1 (an index of FGFR1 activation) compared with those in untreated cultures and cultures treated with bFGF (Fig. 7A). In addition, after 120 minutes, myoblasts treated with S100B showed augmented FGFR1 phosphorylation levels compared with those in the control, although less intensely than at 30 minutes and less intensely than that caused by added bFGF (Fig. 7A). Remarkably, no S100B-dependent increase in FGFR1 phosphorylation levels was observed in myoblasts pretreated with the anti-bFGF antibody (Fig. 7B). In addition, higher levels of phosphorylated tyrosine residues were detected in FGFR1 immunoprecipitates from cultures treated with S100B compared with those from control cultures (Fig. 7C). These results support the conclusion that S100B enhances FGFR1 signaling by binding to FGFR1-bound bFGF.
Transfection of myoblasts with the proliferation inhibitor p21WAF1 resulted in reduced proliferation, as shown by measuring the levels of proliferation stimulator cyclin D1, compared with that of mock-transfected cells, irrespective of the absence or presence of S100B (Fig. 7D). However, S100B was still able to inhibit myogenic differentiation in this condition, as evidenced by the similar and lower levels of MyHC in S100B-treated and untreated p21WAF1-expressing myoblasts compared with those in controls (Fig. 7D). This suggests that the anti-myogenic effect of S100B is unrelated to its ability to stimulate myoblast proliferation and relies on the ability of S100B to inhibit p38 MAPK independently of its ability to activate ERK1/2.
S100B activates FGFR1 in wild-type and Rage−/− myoblasts
We found that RAGE, bFGF and FGFR1, but not Met, co-immunoprecipitated with S100B in wild-type myoblasts, and FGFR1 and bFGF, but not Met, co-immunoprecipitated with S100B in Rage−/− myoblasts (Fig. 8A). S100B also stimulated the proliferation and inhibited the differentiation of wild-type and Rage−/− myoblasts through activation of ERK1/2 and inhibition of p38 MAPK (Fig. 8B–D), and treatment of Rage−/− myoblasts with anti-FGFR1 antibody negated the effects of S100B (Fig. 8C). Notably, control Rage−/− myoblasts proliferated faster (Fig. 8B) and showed lower myogenin and MyHC levels in DM than those of wild-type myoblasts (Fig. 8C), in agreement with the pro-myogenic and anti-proliferative activities of RAGE signaling in myoblasts (Riuzzi et al., 2006a; Sorci et al., 2004).
S100B upregulates FGFR1, but not RAGE, expression in myoblasts
S100B dose-dependently increased FGFR1, but not RAGE levels, in myoblasts when added at a concentration of 1–10 nM in DM (Fig. 9A), and this effect was FGFR1-dependent because it was not observed in cultures treated with the anti-FGFR1 antibody (Fig. 9B). Thus, the higher abundance of FGFR1 in myoblast cultures treated with S100B might promote the interaction of the S100B–bFGF complex with FGFR1, thus resulting in the stimulation of proliferation and the inhibition of differentiation. However, S100B upregulated RAGE in myoblasts pretreated with anti-FGFR1 antibody (Fig. 9B).
S100B has been identified as a potential regulator of myoblast differentiation, promoting proliferation and survival and inhibiting differentiation (Riuzzi et al., 2006b; Sorci et al., 2003). Here, we addressed the questions of whether S100B might participate in myogenesis by virtue of its ability to stimulate myoblast proliferation and what cell surface molecule transduces the effects of S100B in myoblasts. We have shown that clearance of S100B results in enhanced myogenic differentiation owing to an S100B-induced increase in cell number; S100B interacts with FGFR1-bound bFGF, thereby stabilizing bFGF–FGFR1 interactions, and the FGFR1–bFGF–S100B adduct recruits RAGE into a FGFR1–bFGF–S100B–RAGE complex in which FGFR1 mitogenic and anti-myogenic signaling is enhanced and RAGE pro-myogenic signaling is blocked. However, neutralizing either bFGF or FGFR1 or reducing FGFR1 abundance by RNA interference causes S100B to stimulate the pro-myogenic activity of RAGE. In addition, we found that S100B binds directly to RAGE and bFGF, but not to FGFR1, and in Rage−/− myoblasts S100B stimulates the mitogenic and anti-myogenic activity of FGFR1 through prior binding to bFGF. Importantly, S100B is found in CME at nanomolar concentrations, pointing to release of the protein from injured muscle tissue, and neutralization of the CME S100B results in reduced myoblast proliferation. Of note, serum levels of S100B increase following intense physical exercise (Dietrich et al., 2003; Hasselblatt et al., 2004), raising the possibility that S100B leaks from injured myofibers in vivo; conceivably, the S100B concentration at the site of myofiber injury is significantly higher than in serum, thus approaching the subnanomolar–nanomolar amounts shown to exert mitogenic and anti-myogenic effects on myoblasts.
Short-term treatment of myoblasts with S100B resulted in stimulation of ERK1/2 and inhibition of p38 MAPK, with a resultant increase in the myoblast number and an inhibition of differentiation during the S100B treatment and for the first 24 hours after S100B washout; however, at 2 and 5 days after washout, S100B-treated cultures showed enhanced MyHC levels, thus supporting the possibility that the S100B-induced increase in the cell number during the initial period of treatment of myoblasts (supplementary material Fig. S1E; Fig. 1A; supplementary material Fig. S2; Fig. 4E; Fig. 5J; Fig. 8B) translates into an enhanced myotube formation after S100B clearance. Consistently, an inhibition of differentiation was observed in myoblast cultures that received S100B at the time of the transfer from GM to DM and were left undisturbed for 3 days; however, when left undisturbed for 6 days, S100B-treated and untreated cultures showed the same (and high) extent of myoblast differentiation (Fig. 2A,B; supplementary material Fig. S2). Under these conditions, the level of S100B in the culture medium was higher in S100B-treated myoblasts than in the control myoblasts until day 3, declining in S100B-treated and untreated cultures by day 6 (Fig. 2C) – probably owing to degradation of the protein. Indeed, myoblast cell lines do not secrete S100B and any S100B in the culture medium of S100B-untreated myoblasts derives from the serum used to cultivate them (Sorci et al., 2003) (our unpublished data). Thus, the combination of S100B-induced expansion of myoblasts during the first 3 days and the reduction of S100B in the culture medium thereafter might cause S100B-treated myoblasts to form rapidly myotubes, thus reaching the same extent of differentiation as that of myoblasts never exposed to exogenous S100B.
S100B-dependent stimulation of myoblast proliferation and inhibition of differentiation required the presence of bFGF and functional FGFR1, irrespective of the presence or absence of RAGE, and relied on S100B–bFGF complex formation and activation of FGFR1. Indeed, S100B enhanced bFGF binding to FGFR1 (Fig. 5H) and increased FGFR1 phosphorylation (activation) levels compared with those in controls (Fig. 7A,C); this latter event did not occur in myoblasts pretreated with a bFGF-neutralizing antibody (Fig. 7B). S100B also induced the formation of a complex between RAGE and bFGF–FGFR1, as studied by co-immunoprecipitation assays and in situ PLAs, with no complex formation occurring in myoblast cultures pretreated with an S100B- or a bFGF-neutralizing antibody (Figs 4, 5 and 6). Moreover, a 24-hour treatment of myoblasts with S100B resulted in augmented FGFR1 expression levels with no effects on RAGE expression levels (Fig. 9A). However, in the absence of either bFGF or FGFR1, S100B activated the pro-myogenic signaling of RAGE (Fig. 4A–C,F; Fig. 5J). Because the ability of S100B to enhance FGFR1 mitogenic and anti-myogenic signaling required the presence of bFGF (Fig. 5J), it is tempting to speculate that the co-release of bFGF and S100B from injured myofibers is important for the activation and/or expansion of activated satellite cells. That injured myofibers release bFGF is a well-established notion (Chen et al., 1994), and we show here that S100B is also released by injured muscle tissue (Fig. 1). We also propose that S100B switches from mitogenic and anti-myogenic (mediated through bFGF–FGFR1) signaling to pro-myogenic (through RAGE engagement) in the case of a decrease in the local bFGF concentration.
In the high-Ca2+ and non-reducing conditions found extracellularly, S100B forms octamers and higher-order oligomers (Dattilo et al., 2007; Ma et al., 2007; Ostendorp et al., 2007), and S100B has been shown to cause RAGE oligomerization, through binding to RAGE VC1 domains (Dattilo et al., 2007), and/or stabilization of RAGE oligomers, which is required for RAGE signaling (Xie et al., 2008). On the basis of the present results, we propose that S100B octamers might crosslink RAGE and FGFR1-bound bFGF, with a resultant lack of RAGE oligomerization and/or stabilization of RAGE oligomers, and hence prevent RAGE activating p38 MAPK and enhance FGF–FGFR1 signaling. In this scenario, S100B might serve the function of reducing the activity of the pro-myogenic and anti-mitogenic RAGE (Riuzzi et al., 2006a; Sorci et al., 2004). By stimulating bFGF–FGFR1 signaling and blocking RAGE, S100B would contribute substantially to myoblast expansion. In addition, and at variance with the notion that S100B is a RAGE agonist in several cell types (Donato et al., 2009), our data suggest that, when bound to bFGF–FGFR1, S100B behaves as a competitive RAGE antagonist by blocking access of the RAGE activator HMGB1 (Bianchi et al., 2007; Rauvala and Rouhiainen, 2007) to the RAGE V domain (Hori et al., 1995; Rauvala and Rouhiainen, 2007) in myoblasts. Future studies should elucidate the details of the S100B–bFGF interaction. At present, our results showing that S100B can interact with FGFR1-bound bFGF on the myoblast surface suggest that the S100B-binding site on bFGF differs from both that recognizing FGFR1 and that recognizing heparan sulfate, a cofactor that is found on myoblast plasma membranes and is required for bFGF-dependent activation of FGFR1 (Olwin and Rapraeger, 1992).
Notably, it has been shown that it is only at high doses that HMGB1 can stimulate the differentiation of myoblasts in the presence of S100B (Sorci et al., 2004). In retrospect, that finding can be explained by the present observation that S100B can bind to RAGE without activating it; thus, it is only at relatively high doses that HMGB1 can displace S100B from RAGE. In this scenario, it is not only that FGFR1–bFGF-bound S100B cannot activate RAGE, but that it actually blocks RAGE signaling by impeding access of HMGB1 to the receptor.
S100B independently activates ERK1/2 and inhibits p38 MAPK, thereby stimulating proliferation and inhibiting differentiation (Riuzzi et al., 2006b). Our present results suggest that this is partly due to the formation of a RAGE–S100B–bFGF–FGFR1 complex, resulting in the blockage of RAGE promyogenic and anti-mitogenic signaling, and partly due to S100B-dependent enhancement of intrinsic properties of bFGF–FGFR1 signaling (because S100B stimulates proliferation and inhibits differentiation of Rage−/− myoblasts through enhancement of bFGF–FGR1 signaling; Fig. 8). Moreover, S100B upregulates FGFR1, but not RAGE, expression levels (Fig. 9A), which might contribute to sustained myoblast proliferation. Finally, S100B-dependent inhibition of myogenic differentiation in mitotically arrested (i.e. p21WAF1-expressing) myoblasts suggests that the effects of S100B on proliferation and differentiation are distinct from one another.
In conclusion, we have shown, for what we believe to be the first time, that in myoblast cultures extracellular S100B binds to bFGF, thereby enhancing the ability of bFGF to activate FGFR1-mediated mitogenic and anti-myogenic signaling. The concomitant S100B-dependent blockage of the pro-myogenic activity of RAGE (either through the lack of RAGE oligomer formation and/or stabilization, or through negation of the access of HMGB1 to RAGE) might contribute by preventing precocious myoblast differentiation; conversely, clearance of S100B might interrupt myoblast proliferation and allow the HMGB1-stimulated pro-myogenic signaling of RAGE. We propose that, by enhancing bFGF–FGFR1 signaling and blocking RAGE signaling, S100B plays a role in the process of activation of satellite cells and/or expansion of activated satellite cells in the course of skeletal muscle regeneration; the finding that CME contains S100B and that neutralization of S100B reduces CME-stimulated proliferation of primary myoblasts supports this possibility.
Materials and Methods
Expression and purification of S100B
Production of CME, measurement of S100B in CME and effect of CME on myoblast proliferation
Adult C57BL/6 mice were killed by cervical dislocation, and leg muscles were isolated and processed to obtain CME (Chen and Quinn, 1992). CME and the bathing liquid from uncrushed muscles were subjected to sandwich ELISA to measure S100B content. Mouse primary myoblasts were cultivated in GM in the absence or presence of untreated CME, CME supplemented with an S100B-neutralizing antibody or CME previously absorbed to an anti-S100B antibody, as described previously (Sorci et al., 2003), and subjected to a BrdU incorporation assay.
Cell culture conditions and transfections
C2C12 myoblasts seeded at a density of ~443103 cells per cm2, were cultured for 24 hours in Dulbecco's modified Eagle medium (DMEM; Invitrogen) supplemented with 20% fetal bovine serum (FBS; Invitrogen), 100 units/ml penicillin and 100 μg/ml streptomycin (growth medium; GM), under a humidified 5% CO2 atmosphere at 37°C, and were then shifted into DMEM containing 0.5% horse serum (DM) to induce myoblast differentiation. C2C12 and primary myoblasts were cultivated in DM in the absence or presence of S100B, plus or minus other additions as indicated in figure legends. Myoblasts in GM were transfected for 6 hours with muscle creatine kinase (MCK)-luc or myogenin-luc reporter genes or with empty vector (Sorci et al., 2003), and they were then transferred into DM plus or minus S100B. After 24 hours, cells were harvested to measure luciferase activity. C2C12 myoblasts were transiently transfected with p21WAF1 expression vector (pCEP-WAF1-AS) or empty vector in GM using jetPEI (Polyplus Transfections) as recommended by the manufacturer. After 24 hours the cultures were transferred into DM for 48 hours, washed, lysed and probed by western blotting using an anti-cyclin D1 antibody (1:1000; Cell Signaling Technology) or anti-MyHC antibody (1:500; Novocastra).
Isolation of primary myoblasts
C57BL/6 mice were obtained from Charles River. Rage−/− mice, generated as described previously (Liliensiek et al., 2004), were obtained from Angelika Bierhaus (University of Heidelberg, Germany). Primary myoblasts were isolated from 3-day-old wild-type C57BL/6 or Rage−/− pups, cultivated as described previously (Neville et al., 1997), and characterized by immunofluorescence using a polyclonal anti-Met antibody (1:50; Santa Cruz Biotechnology) after fixation with cold methanol for 7 minutes at −20°C. Greater than 95% cells were Met-positive. Approval for use of animals was obtained by the Ethics Committee of the Perugia University and the Ministero della Salute, Italy.
Cell proliferation was measured by either fluorescence-activated cell sorting (FACS) (Riuzzi et al., 2006b) or with a bromodeoxyuridine (BrdU) incorporation assay. BrdU was added to cultures 2 hours before fixation with cold methanol at −20°C and processing for immunofluorescence using a monoclonal anti-BrdU antibody (1:50; Santa Cruz Biotechnology). BrdU-positive and total cells were counted.
Myoblasts were lysed and cell lysates were subjected to western blotting as described previously (Sorci et al., 2003). The following antibodies were used: monoclonal anti-MyHC antibody (1:1000; Novocastra), monoclonal anti-M-cadherin antibody (1:1000; Santa Cruz Biotechnology), monoclonal anti-α-tubulin antibody (1:10,000; Sigma), polyclonal anti-phosphorylated (Thr180 and Tyr182) p38 MAPK antibody (1:1000; Cell Signaling Technology), polyclonal anti-p38 MAPK antibody (1:2000; Cell Signaling Technology), polyclonal anti-phosphorylated (Thr202 and Tyr204) ERK1/2 antibody (1:2000; Cell Signaling Technology), polyclonal anti-ERK1/2 antibody (1:20,000; Sigma), monoclonal anti-MyoD antibody (1:500; Santa Cruz Biotechnology), monoclonal anti-myogenin antibody (1:1000, Santa Cruz Biotechnology), anti-S100B antibody (1:1000; Epitomics), polyclonal anti-RAGE antibody (1:1000; Santa Cruz Biotechnology), monoclonal anti-FGFR1 antibody (1:1000; Chemicon), polyclonal anti-bFGF antibody (1:1000; Santa Cruz Biotechnology), polyclonal anti-Met antibody (1:500; Santa Cruz Biotechnology), monoclonal anti-cyclin D1 antibody (1:1000; Cell Signaling Technology) and monoclonal anti-phosphorylated-tyrosine antibody (PY20) (1:1000; sc-508, Santa Cruz Biotechnology). Analyses of the levels of S100B and bFGF in culture medium were performed as described previously (Riuzzi et al., 2007). Briefly, culture medium was collected and precipitated using trichloroacetic acid. Pellets were resuspended in Laemmli buffer and subjected to western blotting using anti-S100B or anti-bFGF antibodies. The immune reaction was developed by enhanced chemiluminescence (ECL) (SuperSignal West Femto Maximum or SuperSignal West Pico, both from Pierce).
Immunocytochemistry and immunofluorescence
MyHC was detected by immunocytochemistry, as described previously (Sorci et al., 2003). Immunofluorescence analyses were performed, as described previously (Tubaro et al., 2010), using a polyclonal anti-myogenin antibody (1:20; Santa Cruz Biotechnology) and a monoclonal anti-MyoD antibody (1:20; Santa Cruz Biotechnology).
C2C12 myoblasts cells in GM were transfected with control siRNA-A (sc-37007, Santa Cruz Biotechnology) or Flg siRNA (m) (sc-29317, Santa Cruz Biotechnology) using Interferin Polyplus transfection (Euroclone) according to the manufacturer's instructions. After 48 hours the cells were switched to DM, with or without S100B, and cultivated under these conditions for a further 24 hours. The cells were lysed and cell lysates were subjected to western blotting using anti-FGFR1, anti-myogenin or anti-cyclin D1 antibodies.
C2C12 myoblasts (107 cells) were subjected to differential centrifugation (Sanchez-Heras et al., 2006) to obtain a fraction enriched in plasma membranes. These were centrifuged and resuspended in buffer containing 50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% Triton X-100, 1 mM EGTA, 10 mM NaF and 1 mM sodium orthovanadate in the presence of a mixture of protease inhibitors (Roche Applied Science). The Triton X-100 extract was adjusted to ~1 mM free CaCl2 and loaded at room temperature onto an S100B–Sepharose column (1.532.5 cm, 3.8 mg S100B per ml) equilibrated with buffer containing 50 mM Tris-HCl pH 7.4 150 mM NaCl, 1% Triton X-100 and 1 mM CaCl2. The column was extensively washed with the equilibration buffer and S100B-bound proteins were eluted with buffer containing 2 mM EGTA, instead of CaCl2, followed by buffer containing 1 M NaCl. Individual eluates were concentrated by trichloroacetic acid precipitation and further processed for western blotting, as described previously (Riuzzi et al., 2006a), for detection of RAGE, FGFR1 and Met.
C2C12 and primary myoblasts, cultivated for 30 minutes in the absence of additions or in the presence of S100B, bFGF (Peprotech) or both, were lysed in buffer containing 50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% Triton X-100, 1 mM CaCl2, 10 mM NaF and 1 mM sodium orthovanadate in the presence of a mixture of protease inhibitors (Roche Applied Science). Solubilized proteins were subjected to immunoprecipitation using a polyclonal anti-S100B antibody (SWant), a polyclonal anti-RAGE antibody (Santa Cruz Biotechnology), a monoclonal anti-FGFR1 antibody (Chemicon), a polyclonal anti-bFGF antibody (Santa Cruz Biotechnology) (all at 2 μg per mg of total protein) or non-immune IgG. The immunoprecipitates were probed with the following antibodies: polyclonal anti-S100B antibody (1:1000; Epitomics), polyclonal anti-RAGE antibody (1:1000; Santa Cruz Biotechnology), monoclonal anti-FGFR1 antibody (1:1000; Chemicon), polyclonal anti-bFGF antibody (1:1000; Santa Cruz Biotechnology), polyclonal anti-Met antibody (1:500; Santa Cruz Biotechnology) and monoclonal anti-phosphorylated tyrosine antibody (PY20) (1:1000; Santa Cruz Biotechnology).
Neutralization of ligands or receptors
Myoblasts in DM were treated for 2 hours with polyclonal anti-RAGE antibody (10 μg/ml; Santa Cruz Biotechnology), monoclonal anti-FGFR1 antibody (2 μg/ml; Chemicon) or both and cultivated for 24 hours in the absence of additions or in the presence of S100B. In experiments on the effects of neutralization of S100B or bFGF in the culture medium, polyclonal anti-S100B antibody (50 μg/ml; SWant) or monoclonal anti-bFGF antibody (5 μg/ml; clone bFM-1, Millipore) was added to myoblasts in DM for 2 hours, followed by treatments as described in figure legends. Control samples received non-immune IgG (50 μg/ml).
In situ proximity ligation assay
To investigate whether S100B interacts with bFGF, FGFR1 and/or RAGE, we also used an in situ proximity ligation assay (PLA), which allows the visualization of subcellular localization and protein–protein interactions in situ (Söderberg et al., 2006). Myoblasts in DM were pretreated with polyclonal anti-S100B antibody (50 μg/ml; SWant), monoclonal anti-bFGF antibody (5 μg/ml; clone bFM-1, Millipore) or non-immune IgG (50 μg/ml) for 2 hours before incubation with 1 nM S100B or vehicle for 30 minutes. Cells were then fixed, as described previously (Sorci et al., 2003), and treated with a mixture of anti-RAGE and anti-FGFR1 antibodies, anti-S100B and anti-bFGF antibodies, anti-S100B and anti-FGFR1 antibodies or anti-S100B and anti-RAGE antibodies, before being subjected to PLA (OLINK Bioscience, Uppsala) according to the manufacturer's instructions. In control experiments, anti-RAGE antibody, anti-FGFR1 antibody, anti-S100B antibody or anti-bFGF antibody was omitted.
Each experiment was repeated at least three times. Representative experiments are shown unless stated otherwise. The data were subjected to analysis of variance (ANOVA) with SNK post-hoc analysis using a statistical software package (GraphPad Prism version 4.00). Where appropriate, data are represented as the means(±s.e.m.) for three determinations. Statistical significance was taken as P<0.05.
This work was supported by Ministero dell'Università e della Ricerca (PRIN 2004054293, 2007LNKSYS and 2007AWZTHH_004), Association Française contre les Myopathies (Project 12992), Associazione Italiana per la Ricerca sul Cancro (Project 6021) and Fondazione Cassa di Risparmio di Perugia (2004.0282.020, 2007.0218.020 and 2009.020.0021). We thank Eyal Bengal (Haifa, Israel) for providing the myogenin-luc construct, Pier Lorenzo Puri (La Jolla, CA) for providing the MCK-luc construct, Bert Vogelstein (Baltimore, MD) for providing the p21WAF1 expression vector and Angelika Bierhaus (Heildeberg, Germany) for providing the Rage−/− mice. The authors declare no conflict of interest.