Endocrinology and Metabolism

Ambient glucose levels qualify the potency of insulin myogenic actions by regulating SIRT1 and FoxO3a in C2C12 myocytes

Taku Nedachi, Akito Kadotani, Miyako Ariga, Hideki Katagiri, Makoto Kanzaki

Abstract

Nutrition availability is one of the major environmental signals influencing cell fate, such as proliferation, differentiation, and apoptosis, often functioning in concert with other humoral factors, including insulin. Herein, we show that low-serum-induced differentiation of C2C12 myocytes is significantly hampered under low glucose (LG; 5 mM) compared with high glucose (HG; 22.5 mM) conditions, concurrently with nuclear accumulation of SIRT1, an NAD+-dependent deacetylase, and FoxO3a, both of which are implicated in the negative regulation of myogenesis. Intriguingly, insulin appears to exert opposite actions, depending on glucose availability, with regard to the regulation of SIRT1 and FoxO3a abundance, which apparently contributes to modulating the potency of insulin's myogenic action. Namely, insulin exerts a potent myogenic effect in the presence of sufficient glucose, whereas insulin is unable to exert its myogenic action under LG conditions, since insulin evokes massive upregulation of both SIRT1 and FoxO3a in the absence of sufficient ambient glucose. In addition, the hampered differentiation state under LG is significantly restored by sirtinol, a SIRT1 inhibitor, whereas insulin abolished this sirtinol-dependent restoration, indicating that insulin can function as a negative as well as a positive myogenic factor depending on glucose availability. Taken together, our data reveal the importance of ambient glucose levels in the regulation of myogenesis and also in the determination of insulin's myogenic potency, which is achieved, at least in part, through regulation of the cellular contents and localization of SIRT1 and FoxO3a in differentiating C2C12 myocytes.

  • forkhead box O
  • differentiation

skeletal muscle cells have provided a useful model for exploring the molecular mechanisms involved in cellular differentiation (13), and insulin and insulin-like growth factors (IGFs) have been implicated in the process of myogenesis by activating the IRS-PI 3-kinase signaling pathway (7, 22, 25, 53) that also serves as a pivotal intracellular signal for exerting metabolic actions in mature skeletal muscle (9). However, despite our general understanding of the effects of ambient glucose levels on insulin responsiveness with regard to metabolic actions in skeletal muscle cells (37), the possible interrelationship between glucose and insulin acting on myogenesis remains to be clarified.

Skeletal muscle differentiation is a well-organized process governed by muscle-specific transcription factors belonging to the MyoD family, such as MyoD and myogenin (42), and the myocyte enhancer factor-2 (MEF2) family, such as MEF2A and MEF2C (35). In addition to these muscle-specific transcription factors, positively regulating myogenesis, the forkhead box O (FoxO) class of transcription factors, ubiquitously expressed in various cell types, has been shown to negatively regulate myogenesis (27). Insulin/IGF-induced repression of FoxO transcription factors, resulting from their nuclear exclusion in response to Akt-mediated phosphorylation, has been implicated in a key aspect of insulin/IGF actions not only for stimulating myogenesis but also for preventing muscle atrophy (21, 46).

Myogenesis is also directly influenced by the acetylation status of histones and nonhistone proteins including MyoD and MEF2, and class I and II histone deacetylases (HDACs) have been shown to regulate muscle gene expression by inhibiting MyoD and MEF2 factors (31, 33, 43). Recently, silent information regulator-2 (Sir2), a class III deacetylase originally characterized as controlling the life spans of animals in response to nutritional availability, was also identified to serve as a key regulator for myogenesis (14) via overexpression of SIRT1, the mammalian ortholog for Sir2, in C2C12 myoblasts by strongly inhibiting differentiation into myotubes, whereas suppression of SIRT1 expression by RNA interference enhanced myogenesis (14).

Given the unique property of SIRT1 that the cofactor nicotinamide adenine dinucleotide (NAD+) drives deacetylation activity, SIRT1 has been thought to serve as an energy and/or oxidation sensor, being directly involved in the nutritional regulation of gene transcription events in various tissues including skeletal muscle (38, 40, 45). Intriguingly, recent studies have also demonstrated that SIRT1 controls cellular functions by deacetylating FoxO transcription factors in response to various stimuli including nutritional availability (4, 36, 38). Thus, myogenesis is likely to be regulated cooperatively by SIRT1 serving as a sensor of the nutritional environment in concert with FoxOs serving as an insulin/IGF sensor in various situations in which glucose and insulin levels are fluctuating. However, no data are available on the potential interplay between ambient glucose levels and insulin in the regulations of SIRT1 and FoxOs, or on the regulation of myogenesis.

To gain insight into these issues, we investigated the effects of ambient glucose levels on differentiation of C2C12 myocytes and found the potency of insulin's myogenic action to be remarkably affected by extracellular glucose levels and that insulin exerts its maximum myogenic effect only in the presence of a relatively high level of glucose, whereas its potency is significantly compromised under low glucose (LG) conditions, a state in which massive upregulations of SIRT1 and FoxO3a are induced by insulin treatment. Thus, these findings reveal an important interplay between ambient glucose and insulin favoring alterations in the cellular contents of SIRT1 and FoxO3a, both of which are tightly coupled to the regulation of myogenesis.

MATERIALS AND METHODS

Materials.

The Western blot detection kit (West super femto detection reagents) was obtained from Pierce Biotechnology (Rockford, IL). Dulbecco's modified Eagle's medium (DMEM), penicillin-streptomycin and trypsin-EDTA were purchased from Sigma Chemicals (St. Louis, MO). Cell culture equipment was from BD Biosciences (San Jose, CA). Calf serum (CS) and fetal bovine serum (FBS) were obtained from BioWest (Nuaille, France). Immobilon-P and anti-SIRT1 antibody were from Millipore (Bedford, MA). Anti-myosin heavy chain (MHC; MF20) and anti-myogenin (F5D) antibodies were obtained from Iowa Hybridoma Bank (University of Iowa, Iowa City, IA). Anti-phospho-S6 (Ser235/236), anti-Akt, anti-phospho-Akt (Ser473), and anti-phospho Akt (Thr308) antibodies were purchased from Cell Signaling Technology (Danvers, MA). Anti-β-actin antibody was obtained from Sigma Chemicals. Unless otherwise noted, all chemicals were of the purest grade available from Sigma Chemicals.

Cell culture.

Mouse skeletal muscle cell line C2C12 myoblasts (54) were maintained in DMEM supplemented with 10% FBS, 30 μg/ml penicillin, and 100 μg/ml streptomycin (growth medium) at 37°C under a 5% CO2 atmosphere. For biochemical study, the cells were grown on 4-well plates (Nalgen Nunc International, Rochester, NY) at a density of 1 × 105 cells/well in 5 ml of growth medium or on 6-well plates (BD Biosciences) at a density of 3 × 104 cells/well in 3 ml of growth medium. Three days after plating, cells had reached ∼80–90% confluence (day 0). Differentiation was then induced by switching the growth medium to DMEM supplemented with 2% CS, 30 μg/ml penicillin, and 100 μg/ml streptomycin (differentiation medium). The differentiation medium was changed every 24 h. For the immunofluorescent staining study, cells were grown on 22-mm glass coverslips (C022221; Matsunami, Osaka, Japan) in 6-well plates.

Immunofluorescent studies.

C2C12 myoblasts were cultured on coverslips placed on 6-well plates. After differentiation, the cells were stimulated with 100 nM insulin for 60 min. Then, the cells were fixed with 2% paraformaldehyde in PBS (without Ca2+ and Mg2+), followed by immunocytochemistry using anti-SIRT1 antibody (Millipore), and anti-mouse IgG antibody conjugated with Alexa 555 or Alexa 594 (Invitrogen, Carlsbad, CA). Images were monitored and analyzed using Olympus Fluoview FV1000 confocal microscopy and the associated application program, ASW v. 1.3 (Olympus, Tokyo, Japan).

Nuclear extract preparation.

Nuclear extract preparation was performed as follows. Briefly, the cells were washed three times with PBS (-) and resuspended in buffer A (10 mM HEPES-OH, pH 7.9, 1.5 mM MgCl2, 10 mM KCl). After a 20-min incubation on ice, the cells were destroyed with a vortex mixer (maximum speed), and the pellets were then collected. The pellets were resuspended in 50 μl of buffer C (HEPES-OH, pH 7.9, 420 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 25% glycerol) and then frozen (−80°C) and thawed twice. The supernatants were collected as nuclear extracts, and the protein concentration was measured and then stored at −80°C until Western blot analysis.

Immunoprecipitation.

The cell lysates were prepared using Triton X-100-NP40 lysis buffer (50 mM Tris-Cl, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 1% NP-40) and the protein concentrations of each sample were measured using a bicinchoninic acid assay (BCA) protein assay kit (Pierce). Five hundred micrograms of protein were mixed with 2 μg of anti-SIRT1 polyclonal antibody. The mixtures were incubated at 4°C for 3 h and continuously incubated in the presence of protein A-Sepharose. The immunoprecipitates were washed with Triton X-100-NP-40 lysis buffer three times. The adsorbed proteins were eluted with 1× Laemmli's buffer, boiled, and subjected to Western blot analysis.

Western blot analysis.

The expression and phosphorylation of each protein were analyzed by Western blot analysis. In brief, the harvested cell lysates were subjected to 5% or 12% SDS-polyacrylamide gel electrophoresis (1:30 bis:acrylamide). Proteins were transferred to a PVDF membrane (Immobilon-P, Millipore), and the membranes were then blocked for 2 h at room temperature with 5% nonfat dry milk in Tris-buffered saline (TBS) containing 0.1% Tween-20. Immunostaining to detect each protein was achieved with a 1-h incubation with a 1:1,000 dilution of anti-SIRT1 antibody, anti-myosin heavy chain antibody, and anti-myogenin antibody. Specific totals or phosphoproteins were visualized after subsequent incubation with a 1:5,000 dilution of anti-mouse or rabbit IgG conjugated to horseradish peroxidase and a SuperSignal chemiluminescence detection procedure (Pierce Biotechnology). Protein concentrations were determined using a BCA assay. Three independent experiments were performed for each condition. Coomassie blue staining was also performed to assess the efficiency of protein transfer.

Real-time PCR.

Fluorescence real-time PCR analysis was performed using a Light Cycler instrument and SYBR Green detection kit according to the manufacturer's instructions (Roche Diagnostics, Indianapolis, IN). PCR primers for measuring each of the secreted factors were as follows: for SIRT1, 5′-GAT CCT TCA GTG TCA TGG TT-3′ and 5′-GAA GAC AAT CTC TGG CTT CA-3′; for FoxO3a, 5′-TGC CTT GTC AAA TTC TGT C-3′ and 5′-TGC ACT AGC TGA ATA CAG TGA G-3′; for GAPDH, 5′-GGA GAA ACC TGC CAA GTA TGA-3′ and 5′-GCA TCG AAG GTG GAA GAG T-3′.

Glucose concentration assay.

Glucose concentrations in the cultured media were measured using a determiner GLE kit (Kyowa Medex, Tokyo, Japan).

Statistical analysis.

Statistical analysis was performed using one-way ANOVA followed by Tukey's multiple comparison test or Student's paired t-test for independent samples. Data are expressed as means ± SE unless otherwise specified.

RESULTS

Extracellular glucose influences low-serum-induced C2C12 differentiation.

To characterize the effects of extracellular glucose levels on myogenesis of C2C12 cells, we first examined whether the glucose concentration in the low-serum differentiation medium (DMEM + 2% CS) affects C2C12 differentiation. Under the LG (5 mM glucose) conditions, the process of myogenesis was obviously delayed and the number of well-developed myotubes was decreased compared with high glucose (22.5 mM glucose; HG) conditions on day 4 of differentiation (Fig. 1A). We quantified differentiation status by counting the number of myotubes defined as multinuclear myotubes that contained more than 5 nuclei (Fig. 1B), as we previously reported (37). The effect of the extracellular glucose concentration on myogenesis was confirmed by Western blot analysis of differentiation marker proteins using anti-skeletal muscle type MHC and anti-myogenin antibodies, as not only were their expressions detected later, but their amounts were also lower than in C2C12 cells differentiated under LG conditions (Fig. 1C).

Fig. 1.

Low-serum-induced C2C12 myoblast differentiation was affected by extracellular glucose levels. A and B: C2C12 myoblasts were cultured in low glucose (LG)-DMEM (5 mM glucose) + 10% FBS and then switched to differentiation medium [DMEM + 2% calf serum (CS); D-medium] containing 5 mM (LG) or 22.5 mM high glucose (HG) glucose (day 0). Cells were continuously cultured with D-medium changes every 24 h. A: at day 4 of differentiation under LG (a) or HG (b) conditions, myotube formations were observed under a microscope. B: relative numbers of myotubes, defined as multinuclear myotubes containing more than 5 nuclei, were determined. Statistical analysis was performed using paired t-test. #P < 0.05 (n = 5). C: on the indicated day, cell lysates were prepared, and the same amounts of proteins were subjected to Western blotting using anti-myosin, anti-myogenin, and anti-β-actin antibodies. Each experiment was repeated 3 times and representative results are shown.

SIRT1 predominantly localizes in the nucleus under LG conditions but is excluded under HG conditions in C2C12 myotubes.

To understand the mechanisms by which extracellular glucose alters the process of C2C12 differentiation, we initially focused on SIRT1, an NAD-dependent protein deacetylase with enzymatic activity sensitive to changes in cellular energy levels (for a review, see Ref. 2), since a recent study showed direct involvement of SIRT1 in myogenesis (14). Consistent with previous reports showing that Sir2 and its mammalian homolog SIRT1 are localized in the nucleus (19, 34), immunofluorescent analysis using anti-SIRT1 antibody demonstrated that, when C2C12 cells were differentiated under LG conditions, predominant localization of SIRT1 was observed in the nucleus (Fig. 2A, a) similar to what was observed in undifferentiated myoblasts (data not shown). In contrast, when C2C12 myoblasts were differentiated under HG conditions, the number of SIRT1-positive nuclei was remarkably reduced (Fig. 2A, c, see arrowheads), as confirmed by Western blotting analysis of nuclear proteins extracted under each culture condition (data not shown). Counterstaining was performed with DAPI (Fig. 2A, b and d). Immunofluorescent analysis demonstrated that the nuclear exclusion of SIRT1 was not acutely induced by HG administration in either myoblasts (data not shown) or LG-differentiated myotubes (Fig. 2B). However, we found that total SIRT1 protein was significantly reduced when C2C12 cells were differentiated under HG conditions (Fig. 2C, bottom, HG-SIRT1) compared with those under LG conditions (top, LG-SIRT1). The HG-dependent reduction of SIRT1 was obvious from day 2 of differentiation (Fig. 2C, bottom, HG-SIRT1, lane3).

Fig. 2.

Cellular abundance of SIRT1 and its localization are controlled by extracellular glucose levels during myogenesis. A: C2C12 myoblasts were differentiated into myotubes under LG conditions (5 mM glucose in DMEM + 2% CS) or HG conditions (22.5 mM glucose in DMEM + 2% CS) for 6 days. Myotubes were then fixed and immunostained using anti-SIRT1 antibody (a, c) or anti-IgG as a negative control (e). DAPI staining was performed at the same time to confirm the position of the nucleus (b, d, f). B: C2C12 myoblasts were differentiated into myotubes under LG conditions for 6 days. Myotubes were then incubated under LG or HG conditions for 2 h. Cells were fixed and immunostained using anti-SIRT1 antibody and DAPI. C: C2C12 myoblasts were differentiated into myotubes under LG or HG conditions for the indicated days. Cell lysates were prepared as described in materials and methods, and the same protein samples were subjected to Western blotting using anti-SIRT1 antibody. D: C2C12 myoblasts were differentiated into myotubes under LG conditions for 6 days. Next, media were switched to LG or HG conditions, and myotubes were then further incubated for the indicated times. Cell lysates were prepared as described in materials and methods, and the same protein amounts were subjected to Western blotting using anti-SIRT1 antibody. E: densitometric analysis of D. Statistical analysis was performed using 1-way ANOVA, as described in materials and methods [###P < 0.001 vs. control (0 h), HG, n = 3]. F and G: C2C12 myoblasts were differentiated into myotubes under HG conditions for 6 days. Next, media were switched to LG conditions, and myotubes were then further incubated for the indicated time. F: cell lysates were subjected to Western blotting using anti-SIRT1 antibody. G: total RNAs were purified and subjected to real-time PCR analysis to measure SIRT1 mRNA amounts as described in materials and methods. Statistical analysis was performed using paired t-test (#P < 0.05, n = 3). All experiments were performed at least 3 times, and similar results were obtained.

Myogenesis is a highly organized and regulated sequence of multiple processes orchestrated by a wide variety of functional proteins, including SIRT1, requiring a relatively long time, and all of these processes could be affected by ambient glucose levels. In an attempt to dissect the effects of glucose on SIRT1 regulation and myogenesis, we conducted an experiment to specify the time required for glucose to exert its effect on SIRT1 suppression. Thus, C2C12 myotubes differentiated under LG conditions (days 5–6) were transferred to HG or the same LG medium, and time course changes in SIRT1 contents were analyzed by Western blotting (Fig. 2, D and E). The time course experiment revealed that 24-h exposure to HG is sufficient for reducing the cellular SIRT1 content and its nuclear localization (data not shown) in LG-differentiated C2C12 myotubes. In addition, we performed the converse experiments; that is, C2C12 myotubes were cultured under HG conditions and then switched to LG for the indicated time to confirm that amounts of SIRT1 were reversibly controlled by ambient glucose levels (Fig. 2F). We also measured SIRT1 mRNA levels under these conditions and found that SIRT1 mRNA was significantly induced by switching to the LG condition (P < 0.05, n = 4; Fig. 2G). Thus, these results suggest that SIRT1 protein induction is regulated, at least in part, by its mRNA levels.

FoxO3a localizes in predominantly the nucleus under LG conditions but is excluded under HG conditions in C2C12 myotubes.

We next examined whether FoxO3a associates with the spatiotemporal changes in SIRT1 localization in response to alterations in glucose availability, since recent reports revealed functional and physical interactions between SIRT1 and FoxO transcription factors in response to various stimuli including oxidative stress (4, 36) and nutritional circumstances (38). Similar to the pattern observed in SIRT1 subcellular localization, FoxO3a was predominantly detected in the nuclei of myotubes when the cells were differentiated under LG conditions (Fig. 3A, a), whereas no nuclear localization of FoxO3 was observed in those differentiated under HG conditions (Fig. 3A, c). Again, HG administration failed to induce acute redistribution of FoxO3a (Fig. 3B) but resulted in a remarkable reduction of FoxO3a when the LG-differentiated myotubes (days 5 and 6) were exposed to HG for an additional 24 h (Fig. 3, D and E). Furthermore, we confirmed that FoxO3a protein was reversibly controlled by extracellular glucose (Fig. 3F) and also that FoxO3a mRNA levels were regulated by extracellular glucose (P < 0.001, n = 4; Fig. 3G). Thus, FoxO3a displayed spatiotemporal regulation similar to that observed in SIRT1 in response to altered extracellular glucose levels. However, the changes in cellular FoxO3a content during myogenesis were obviously different from those of SIRT1. As shown in Fig. 3C, little FoxO3a expression was observed in undifferentiated myoblasts (day 0), but its expression gradually increased upon differentiation only under LG conditions. This differentiation-dependent increase in the cellular content of FoxO3a was abolished when the cells were differentiated under HG conditions.

Fig. 3.

Cellular abundance of FoxO3a and its localization are controlled by extracellular glucose levels during myogenesis. A: C2C12 myoblasts were differentiated into myotubes under LG conditions (5 mM glucose in DMEM + 2% CS) or HG conditions (22.5 mM glucose in DMEM + 2% CS) for 6 days. Myotubes were then fixed and immunostained using anti-FoxO3a antibody (a, c) or anti-IgG as a negative control (e). DAPI staining was performed at the same time to confirm the position of the nucleus (b, d, f). B: C2C12 myoblasts were differentiated into myotubes under LG conditions for 6 days. Myotubes were then incubated under LG or HG conditions for 2 h. Cells were fixed and immunostained using anti-FoxO3a antibody and DAPI. C: C2C12 myoblasts were differentiated into myotubes under LG or HG conditions for the indicated days. Cell lysates were prepared as described in materials and methods, and the same protein amounts were subjected to Western blotting analysis using anti-FoxO3a antibody. D: C2C12 myoblasts were differentiated into myotubes under LG conditions for 6 days. Next, media were switched to LG or HG conditions, and myotubes were then further incubated for the indicated time. Cell lysates were subjected to Western blotting analysis using anti-FoxO3a antibody. E: densitometric analysis of D. Statistical analysis was performed using 1-way ANOVA followed by Tukey's posttest [##P < 0.01, ###P < 0.001, n = 3) vs. control (0 h, HG)]. F and G: C2C12 myoblasts were differentiated into myotubes under HG conditions for 6 days. Next, media were switched to LG conditions, and myotubes were then further incubated for the indicated time. F: cell lysates were subjected to Western blotting using anti-FoxO3a antibody. G: total RNA was purified and subjected to real-time PCR analysis to measure FoxO3a mRNA amounts as described in materials and methods. Statistical analysis was performed using paired t-test (###P < 0.001, n = 3). All experiments were performed at least 3 times, and similar results were obtained.

Taken together, these data demonstrate that, although acute HG treatment fails to induce subcellular redistributions of SIRT1 and FoxO3a, with chronic HG treatment (24 h) there is an obvious nuclear exclusion of these proteins concomitant with the significant reductions in their cellular contents in differentiating C2C12 cells. These data also suggest the compromised myogenesis under LG conditions to be attributable to nuclear accumulation of relatively high levels of SIRT1 and FoxO3a in C2C12 cells. In addition, our data indicate that 24-h exposure to HG is sufficient to produce obvious reductions in both SIRT1 and FoxO3a proteins in C2C12 myotubes.

Extracellular glucose levels modify the stimulatory effect of insulin on myogenesis by altering cellular contents of SIRT1 and FoxO3a.

Effects of HG appeared to be elicited within 24 h as assessed by the obvious reductions of both SIRT1 and FoxO3a proteins in the LG-differentiated C2C12 myotubes (Figs. 2 and 3). We therefore took advantage of this phenomenon to explore the possible interplay between insulin and ambient glucose during the regulation of myogenesis. Namely, the LG-differentiated C2C12 myotubes (days 5 and 6) were transferred to LG or HG medium in the absence or presence of the indicated insulin concentration and cultured for an additional 24 h, and the cellular contents of SIRT1 and FoxO3a were then analyzed by Western blotting (Fig. 4). In the presence of LG, insulin significantly increased cellular contents of both SIRT1 (Fig. 4A, top, lanes 1–6) and FoxO3a (Fig. 4A, middle, lanes 1–6) in a dose-dependent manner (Fig. 4, B and C, open symbols) although the amount of β-actin as a loading control was not changed (Fig. 4A, bottom). In the presence of insulin under LG conditions for 24 h, SIRT1 predominantly displayed nuclear localization, whereas increased FoxO3a displayed both cytoplasmic and nuclear localization (data not shown). In sharp contrast, insulin treatment tended to decrease SIRT1 protein in the presence of HG (Fig. 4A, top, lanes 7–12, and Fig. 4B, •). Moreover, insulin completely failed to increase FoxO3a contents in the presence of HG (Fig. 4A, bottom, lanes 7–12; Fig. 4C, ▴). In the presence of insulin under HG conditions for 24 h, both of these proteins were barely detectable by immunofluorescent staining (data not shown).

Fig. 4.

Effects of insulin on SIRT1 and FoxO3a abundances depend on extracellular glucose levels. A: C2C12 myoblasts were differentiated into myotubes under LG conditions for 6 days. Then, media were switched to HG (22.5 mM glucose in DMEM + 2% CS) or control LG (5 mM glucose in DMEM + 2% CS) in the presence or absence of the indicated concentration of insulin for 24 h. Amounts of SIRT1, FoxO3a, and β-actin were monitored by Western blotting. B: densitometric analysis of SIRT1 abundance during time course experiments. Statistical analysis was performed using 1-way ANOVA followed by Tukey's posttest [*P < 0.05, **P < 0.01 (n = 3) vs. LG control (LG, 0 nM insulin); #P < 0.05, ##P < 0.01 (n = 3) vs. HG control (HG, 0 nM insulin)]. C: densitometric analysis of FoxO3a abundance during time course experiments. Statistical analysis was performed using 1-way ANOVA followed by Tukey's posttest [LG: *P < 0.05, **P < 0.01, ***P < 0.001 (n = 3) vs. LG control (LG, 0 nM insulin)].

To address the possibility that these changes in SIRT1 and FoxO3a proteins induced by the combined actions of insulin and glucose contribute to regulating the differentiation status of C2C12 myotubes, the amount of MHC as a myogenic differentiation marker under each condition was also analyzed by Western blotting (Fig. 5). Consistent with the results depicted in Fig. 1, the stimulatory effect of HG on MHC expression was also detected even after incubation for just 24 h with HG (Fig. 5A, lane 1 vs. lane 6). In the presence of HG, insulin displayed a potent myogenic-stimulating action, and MHC amounts were significantly increased in response to 100 nM insulin (P < 0.05, n = 3; Fig. 5A, lanes 6–10) under conditions in which the SIRT1 and FoxO3a contents were marginal (Fig. 4). In contrast, the myogenic stimulating potency of insulin was apparently limited (Fig. 5A, lanes 1–5) under LG conditions in which massive amounts of SIRT1 and FoxO3a were expressed (Fig. 4). The synergistic action of insulin and HG on the upregulation of MHC was clearly demonstrated by a quantitative densitometric analysis (Fig. 5B).

Fig. 5.

Insulin myogenic potency is altered by extracellular glucose levels. A: C2C12 myoblasts were differentiated into myotubes under LG conditions for 6 days. Then, media were switched to HG (22.5 mM glucose in DMEM + 2% CS) or control LG (5 mM glucose in DMEM + 2% CS) in the presence or absence of indicated concentration of insulin for 24 h. Amounts of myosin heavy chain (MHC) were monitored by Western blotting. B: densitometric analysis of A. Statistical analysis was performed using 1-way ANOVA followed by Tukey's posttest [HG: #P < 0.05 (n = 3) vs. control (HG, 0 nM insulin)].

Glucose-dependent opposing effects of insulin on SIRT1 and FoxO3a are achieved through PI 3-kinase and mTOR activities.

The myogenic action of insulin/IGFs has been shown to be mediated through the activation of PI 3-kinase and Akt signaling pathways (7, 23, 25, 53). In addition, several lines of evidence have demonstrated mTOR involvement in myogenesis (39, 47). To test whether the activities of PI 3-kinase and mTOR are involved in the regulation of SIRT1 and FoxO3a contents in response to insulin under LG or HG conditions, we used LY-294002 and rapamycin, potent inhibitors of PI 3-kinase and mTOR, respectively. Consistent with previous reports (8, 26, 41), the myogenic action of insulin as assessed by MHC expression, clearly detectable only in the presence of HG, was completely abolished by these compounds (Fig. 6, A and C, bottom, lanes 4–6). Importantly, these compounds also inhibited the insulin-mediated suppression of SIRT1 while increasing its cellular content during 24-h HG exposure (Fig. 6, A and C, top, lanes 4–6). Interestingly, these compounds were also effective under LG conditions and completely abolished insulin's action (Fig. 6, A and C, middle, lanes 1–3), although insulin exerted completely opposing effects on the regulation of SIRT1 and FoxO3a abundances (Fig. 6, A and B, top and middle, lane 2 vs. lane 5). We also monitored insulin-dependent phosphorylations of Akt and S6 in the presence of LY-294002 and rapamycin (Fig. 6, B and D). As expected, LY-294002 abolished insulin-dependent phosphorylation of Akt (Ser473 and Thr308); on the other hand, rapamycin tended to increase insulin-dependent Akt phosphorylation (Fig. 6, B and D, panels 1 and 2) as previously reported (52). Phosphorylation of S6 was high under basal conditions but was slightly induced by insulin, and either LY-294002 or rapamycin had a negative effect on this phosphorylation (Fig. 6, B and D, panel 4).

Fig. 6.

Dissecting the interplay between glucose and insulin effects on SIRT1 and FoxO3a expressions and their involvements in myogenesis. A and C: C2C12 myoblasts were differentiated into myotubes under LG conditions for 6 days. Media were then switched to HG or LG with or without 100 nM insulin in the presence or absence of 10 μM LY-294002 (A) or 50 nM rapamycin (C) for 24 h. Amounts of MHC, SIRT1, and FoxO3a were monitored by Western blotting. B and D: C2C12 myoblasts were differentiated into myotubes under LG conditions for 6 days. Then, myotubes were treated with 100 nM insulin for 5 min in the presence or absence of 10 μM LY-294002 (B) or 50 nM rapamycin (D) under LG or HG conditions. Phosphorylations of Akt (Ser473 and Thr308) and S6 as well as total Akt and β-actin were analyzed by Western blotting. E: C2C12 myoblasts were differentiated into myotubes under LG conditions for 6 days. Cells were then cultured for 24 h in the presence of indicated concentrations of sirtinol under LG conditions (lanes 1–4) or LG + 100 nM insulin (lanes 5–8). Amounts of MHC, SIRT1, and FoxO3a were monitored by Western blotting. All experiments were repeated at least 3 times, and representative results are shown.

Reduction of SIRT1 activity is not sufficient to restore C2C12 differentiation status in the presence of insulin.

Finally, we attempted to restore the poor differentiation seen under LG conditions by exposure to sirtinol, an inhibitor of SIRT1 (20, 32). Consistent with a previous study (14), 24-h exposure to sirtinol under LG conditions restored MHC expression to levels comparable to those observed in HG-exposed C2C12 myotubes (Fig. 6E, panel 3, lanes 1–4). However, sirtinol failed to restore MHC expression (Fig. 6E, panel 3, lanes 5–8) in the presence of insulin, a condition in which FoxO3a is highly expressed (Fig. 6E, panel 2, lanes 5–8).

Changes in extracellular glucose levels during the course of overnight incubation.

To precisely document the importance of ambient glucose levels in the phenomena described above, extracellular glucose concentrations were monitored during the course of overnight incubation of differentiated C2C12 myotubes. As shown in Fig. 7, extracellular glucose levels gradually decreased but remained at high levels (∼13 mM) even after an 18-h incubation when differentiated C2C12 myotubes were cultured in HG-DMEM, whereas glucose in the medium was almost completely exhausted when the cells were cultured in LG-DMEM for 18 h.

Fig. 7.

Glucose consumption by C2C12 myotubes. C2C12 myoblasts were differentiated into myotubes for 6 days. Media were then switched to fresh medium containing LG (•) or HG (○), and glucose concentrations were measured at the indicated time.

DISCUSSION

In the present study, we demonstrated that low-serum-induced differentiation of C2C12 cells is significantly influenced by extracellular glucose levels (Fig. 1), concurrently with glucose-dependent alterations in the amounts and subcellular localizations of SIRT1 and FoxO3a (Figs. 2 and 3) both of which have recently been implicated as negative regulators of myogenesis (14) (21). Consistent with many previous reports studying myogenesis by using mostly IGFs (6, 7, 12), insulin also exerts a myogenic action in a manner dependent on PI 3-kinase and mTOR activities, as assessed by MHC expression (Fig. 6); however, we cannot rule out the possibility that insulin activates IGF-I receptors since a relatively high concentration of insulin was required to stimulate MHC expression within 24 h (Fig. 5). Surprisingly, however, we found the potency of insulin's myogenic action to also be remarkably affected by extracellular glucose levels and that insulin exerts its potent myogenic effect only in the presence of relatively high levels of glucose, whereas its potency is significantly compromised in the absence of sufficient glucose (Fig. 5), perhaps due to massive increases in SIRT1 and FoxO3a, serving as negative regulators of this process, which are induced by insulin treatment under LG conditions (Figs. 4 and 6). Taken together, these results reveal an important interplay between glucose availability and insulin in the regulation of myogenesis, which is achieved, at least partially, through alterations in the cellular contents and nuclear abundances of SIRT1 and FoxO3a. Our data also document opposing effects of insulin, depending on glucose availability and thereby on the regulation of SIRT1 and FoxO3a amounts in differentiating C2C12 myotubes (Fig. 6A). Since insulin/IGFs often display opposite biological effects, e.g., proliferation and differentiation, depending on conditions and circumstances (8, 11, 28, 30), our data presented herein provide a conceptual framework for understanding the mode of insulin actions (and presumably those of IGFs) by providing evidence that insulin's myogenic action is profoundly influenced by glucose availability mediated through regulation of SIRT1 and FoxO3a, both of which have been shown to be directly involved in the determination of cellular fates, including proliferation, differentiation, and senescence, in various cell types (1, 4, 5, 44, 50, 51).

Effects of ambient glucose levels on regulation of SIRT1 and FoxOs during myogenesis.

A recent report revealed an important SIRT1 role in myogenesis by showing that overexpression of SIRT1 inhibited myogenesis, whereas either siRNA-mediated suppression of SIRT1 or sirtinol enhanced it by altering the acetylation states of MyoD and the histone acetylase p300/CBP (14). Although the SIRT1 expression level was shown to be slightly decreased upon differentiation of C2C12, the authors focused primarily on the importance of the regulation of its enzymatic activity rather than the expression level of SIRT1. In the present study, we found that SIRT1 expression is influenced by glucose levels and that SIRT1 abundance is significantly reduced in C2C12 myotubes differentiated under HG conditions, a culture condition obviously potentiating myogenesis (Fig. 1). In addition, 24 h exposure of differentiating C2C12 myotubes to HG (days 5–6) was also sufficient to decrease SIRT1 abundance (Fig. 2C), which apparently contributed to the potentiation of myogenesis as assessed by MHC expression levels (Fig. 5). Although acute redistribution was not observed with either glucose or insulin in the present study (Fig. 2B), recent reports revealed the existence of a nucleocytoplasmic shuttling mechanism for SIRT1 (24, 50). Thus, our data strongly suggest that SIRT1 abundance and its localization, being sensitively influenced by ambient glucose levels, are also directly involved in the regulation of myogenesis as a prerequisite for regulating its enzymatic activity at suitable site(s). Consistent with this idea, a recent report demonstrated SIRT1 transcription to be regulated by metabolic states via the HIC1:CtBP corepressor complex (55). These changes in SIRT1 abundance in skeletal muscle cells may contribute not only to myogenesis but also to metabolic adaptations during glucose deprivation, such as regulation of mitochondrial gene expression and fatty acid utilization (17, 29).

Similar to what was observed in SIRT1 regulation, we found decreases in both the cellular and the nuclear abundance of FoxO3a to also be coupled to the potentiation of myogenesis, which can be induced under HG conditions (Figs. 1 and 3). These findings are consistent with a previous report showing that short interfering RNA-mediated suppression of FoxOs enhanced myogenesis, whereas its overexpression inhibited differentiation (21). Although the importance of nucleocytoplasmic shuttling of FoxOs by posttranslational modifications, such as the interplay between the Akt-mediated phosphorylation and the SIRT1-mediated deacetylation in response to growth factors and oxidative stress, is well established (3, 4), acute redistribution of FoxO3a was not detected with changes in the ambient glucose level (Fig. 3B), whereas its nuclear exclusion was rapidly induced by insulin, as previously reported (49). Since C2C12 myocytes also express other FoxO family transcription factors, including FoxO1 and FoxO4, in addition to FoxO3a (21), our data at this stage cannot address the magnitude of the FoxO3a contribution to myogenic inhibitory action. However, our present data support previous reports showing that both SIRT1 and FoxO3a serve as negative myogenic regulators (14, 21) and also provide evidence supporting the important participation of these proteins in the process of myogenesis achieved through the regulation of their amounts and subcellular localizations. As discussed below, the physiological importance of glucose-dependent alterations in cellular SIRT1 and FoxO3a abundance in myogenesis further underscores our striking finding that insulin exerts distinct effects regulating the abundances of these proteins depending on glucose availability, which is tightly coupled with myogenic differentiation status (Fig. 4 and Fig. 5).

Interplay between ambient glucose and insulin in regulation of SIRT1 and FoxO3a and its involvement in myogenesis.

In the present study, we found that the effects of ambient glucose levels can be exerted within 24 h, according to the cellular abundances of both SIRT1 and FoxO3a (Figs. 2 and 3), in differentiating C2C12 myotubes. Similarly, the drastic changes in SIRT1 and FoxO abundances in various tissues, including skeletal muscle, have been reported for in vivo experiments showing that starvation increases SIRT1 and FoxOs, which are resuppressed by refeeding within 24 h (15, 16, 38, 45). In this regard, we observed that the extracellular glucose concentration falls to less than 0.5 mM after a 24-h incubation when differentiating C2C12 cells are cultured in LG-DMEM (Fig. 7), even though DMEM containing 5 mM glucose (LG) is a conventional medium routinely utilized to maintain various cell types. This is probably because myotubes are postmitotic multinuclear cells that consume vast amounts of glucose as an energy source. Consequently, the cells cultured under LG conditions were perhaps experiencing an environment similar to the condition of glucose starvation, of varying degrees, during the 24-h incubation, even though the LG media were replaced daily, whereas when cells were cultured under HG conditions they were continually exposed to pathophysiologically high levels of glucose for 24 h (Fig. 7). Hence, our observations indicate that these gross culture environments, including the consequences of glucose consumption and/or exhaustion, not just the initial glucose concentration, apparently contribute to regulating SIRT1 and FoxO3a, which is in turn responsible for the modulation of myogenesis during the 24-h incubation.

One of the most intriguing observations is that insulin remarkably increases the cellular contents of both SIRT1 and FoxO3a under only LG conditions (Fig. 4A, lanes 1–6), whereas insulin completely fails to increase FoxO3a, instead decreasing the SIRT1 amount in the presence of HG (Fig. 4A, lanes 7–12). Moreover, our most striking finding is that insulin is unable to exert its myogenic action under LG conditions (Fig. 5A, lanes 1–5), whereas its potency is maximized under HG conditions, as assessed by MHC expression levels (Fig. 5A, lanes 6–10). Although it is well established that insulin and IGFs serve as potent myogenic stimulators (7, 12, 21), our present data reveal insulin's myogenic action to be significantly influenced by glucose availability. In addition, our data strongly suggest that the insulin-induced massive accumulations of these negative myogenic transcriptional regulators, SIRT1 and FoxO3a, provoked under LG conditions in differentiating C2C12 myotubes counteract the myogenic stimulatory potency that insulin intrinsically possesses.

The important functional interrelationships between SIRT1 and FoxOs have been established in various organisms (18), and SIRT1 and FoxOs, including FoxO3a, have been shown to physically interact with each other to regulate their functions in a wide array of cell types (4, 10, 36). Thus, although either SIRT1 or FoxO3a alone is reportedly able to interfere with the process of myogenesis (21), it is likely that SIRT1 and FoxO3a are cooperatively involved in this interference, properly responding to alterations of culture circumstances such as glucose availability and the presence of insulin. In an attempt to evaluate the contribution of the myogenic inhibitory actions of these proteins, we utilized sirtinol to eliminate the deacetylase activity of SIRT1 and found that, although the poor differentiation state under LG conditions is significantly restored within 24 h by sirtinol (Fig. 6C, bottom, lanes 1–4), as previously reported (14), the sirtinol-dependent restoration of increased MHC expression is completely abolished in the presence of insulin (lanes 5–8). These data not only further confirm the opposing actions of insulin, i.e., that insulin can serve as a negative, rather than a positive, myogenic factor when glucose availability is low, but also suggest that the reduction of SIRT1 enzymatic activity alone might be insufficient to overcome the poor differentiation status in the presence of insulin, a condition under which FoxO3a is remarkably increased (Fig. 6C, middle, lanes 5–8). Thus, the considerably augmented FoxO3a may still be functional to some extent even in the presence of insulin during a 24-h incubation, which perhaps contributes to interference with the promotion of myogenesis. Moreover, since both FoxO1 and FoxO3a have been shown to increase the expressions of the ubiquitin ligases MAFbx and MuRF1, responsible for muscle atrophy via increased protein degradation (46, 49), the increased FoxO3a may also participate in the stimulation of protein degradation governed by these FoxO-inducible ubiquitin ligases. In any case, together with previous studies showing that overexpression of FoxOs results in reduced muscle mass in transgenic mice (27) and also retards myogenesis (21), our present data suggest that the massively increased FoxO3a plays a pivotal role in exerting the inhibitory actions of insulin, at least under these experimental conditions.

Another interesting observation presented in this study is that the opposing actions of insulin depending on ambient glucose levels were both completely abolished by LY-294002 (Fig. 6A) or rapamycin (Fig. 6B). These results indicate crucial involvements of the PI 3-kinase and mTOR activities stimulated by insulin in exerting insulin actions on SIRT1 and FoxO3a depending on ambient glucose environments (Fig. 6), although the insulin-induced decrease in FoxO3a under HG conditions is not apparent due to its undetectable expression (Fig. 6, A and B, middle, lanes 4–6), as was the case with that observed in SIRT1 (Fig. 6, A and B, top, lanes 4–6). Recently, Southgate et al. (48) showed that the elevation of FoxO1 protein levels induced the nonphosphorylated form of 4EBP1, followed by reductions in Raptor and mTOR protein amounts. Together with our finding that the opposing actions of insulin on FoxO3a protein levels, which depend upon ambient glucose concentrations, are both abolished by rapamycin (Fig. 6), it is reasonable to speculate that insulin stimulates either negative or positive feedback loops on the mTOR-FoxO axis, depending on ambient glucose levels. Future work should be directed toward increasing our understanding of the mechanisms underlying the differential effects of insulin on SIRT1 and FoxOs, depending on glucose availability to solve the mystery of how insulin/IGFs exert diverse, and in some instances opposing, biological actions.

GRANTS

This work was supported by Special Coordination Funds for Promoting Science and Technology. This work was also supported in part by grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan and the New Energy and Industrial Technology Development Organization (NEDO).

Acknowledgments

We thank Fumie Wagatsuma and Natsumi Emoto for technical assistance.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract